[Debian-ha-commits] [pacemaker] annotated tag upstream/1.1.13 created (now e667fa1)
Christoph Berg
myon at debian.org
Wed Jan 6 20:15:28 UTC 2016
This is an automated email from the git hooks/post-receive script.
myon pushed a change to annotated tag upstream/1.1.13
in repository pacemaker.
at e667fa1 (tag)
tagging 21f40a38e3d33fbf8ec35226d6bacf1639582f5d (commit)
replaces Pacemaker-1.1.7
tagged by Ferenc Wágner
on Tue Sep 8 16:00:31 2015 +0200
- Log -----------------------------------------------------------------
Upstream version 1.1.13
Adam Spiers (2):
fix broken link to "Ordering Explained" PDF
fix node names to match small cluster transition graph
Adrián (2):
Wrong command parameters
pcs has the command move
Andreas Grüninger (2):
Fix: logging: Illumos/Solaris do not define some error codes
Refactor: cib: Illumos/Solaris don't have the type in dirent, stat() is more portable
Andrei Belov (1):
Fix: Use custom OCF_ROOT_DIR if requested
Andrew Beekhof (2920):
Merge pull request #54 from davidvossel/rhbz#805147
High: Core: Fix the logging of XML and conditional corefile creation by asserts
Medium: Tools: Gracefully handle status sections with duplicate node_state entries
Low: Build: Refresh spec file from fedora
Low: cib: Disabled diagnostics
Low: doc: Update the vast majority of Clusters from Scratch for Corosync-2.0/Fedora-17
High: RA: controld - use the correct dlm_controld when membership comes from corosync directly
High: crmd: Detect when the quorum API isn't configured for corosync 2.0
Low: PE: Formatting
Merge pull request #55 from davidvossel/bug-5045
High: Tools: crm_simulate - Correctly detect status sections with duplicate node_state entries (broken 6 days ago)
Low: PE: Allow single regression tests to be initiated in debug mode
Low: PE: Include descriptions for tests related to old OSDL bugs
Low: PE: Check for regressions in single-shot mode
Merge pull request #61 from davidvossel/as_xml_oneshot
Low: PE: Reduce the amount of special cases when calculating what actions are needed
Merge pull request #58 from davidvossel/bug-5051
Merge pull request #51 from davidvossel/bug-5025
Merge pull request #57 from davidvossel/bug-5053
Merge pull request #60 from davidvossel/bug-5047
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: Build: Don't build docs by default, unless 'all'is requested
Medium: heartbeat: Correctly observe const keyword in get_uname()
High: PE: Supply the correct number of arguments to tracing log
Merge pull request #63 from jnpkrn/patches
Merge pull request #66 from inouekazu/bugfix
Low: cts: Generate test configs without the shell
Merge pull request #67 from gao-yan/crmd-lrm-reconnect
Merge pull request #65 from davidvossel/bug-5058
Merge pull request #64 from jnpkrn/patches
High: Remove the shell as requested by its author/maintainer
Low: doc: Remove shell documentation now that the code is gone
Low: Reduce debug logging
Low: crmd: Move heartbeat-specific functionality to its own file
Low: cts: Support nested CibXML objects
Low: cts: Remove developmental hacks
Low: cts: Further refinement to config builder
Low: Corosync: Move deprecated corosync variants to their own file
Low: Core: Indicate where various node state information originated
High: corosync: Detect when the first corosync has not yet formed and handle it gracefully
Low: cts: Avoid blocking indefinitely when we do not have permission to read a logfile
Low: cts: Earlier and slightly more explicit detection of local/remote logging configuration
Low: cib: Additional detail in cib_process_diff() logs
Low: crmd: Move some debug logs to trace
Low: crmd: Attempt to detect endless election loops so we can debug them
Medium: mcp: Children only inherit the corosync setting of debug
Medium: mcp: Default to debug=off if no setting is found in corosync
Medium: crmd: Detect when peers have an incorrectly set process list
Medium: PE: Observe the promotion score of probed resources
Medium: Overhaul the way peers are determined to be sufficiently active
Medium: tools: crm_node - Test stack membership only when displaying partition members
Low: tools: Rename the source file for crm_node
Low: crmd: Rename the source file for heartbeat specific functionality
Low: tools: Adjust the default name for archived CTS tests
Low: extra: cluster-init include a stub for enabling debug-mode for the quorum API
Merge branch 'master' of github.com:ClusterLabs/pacemaker
High: corosync: Fix compilation against 1.4.x
Low: Build: Turn on --enable-fatal-warnings by default when building rpms
Low: cts: Remove redundant test tool (now that crm_report can extract tests too)
Low: crmd: Remove the file that got moved to crmd/heartbeat.c
Low: Fix compilation with fatal-warnings enabled
High: cman: Successful connection to CPG implies we're a valid cluster member (Backport from corosync.c)
High: cman: Correctly determine node state as active when connected to CPG
High: corosync: Determine node state for plugin and cman based cluster types
High: cman: Ignore the process list from the mcp and use CPG membership to decide if peers are active (backport from corosync.c)
Low: Build: Drop the -fno-default-inline compiler flag as it is 'valid for C++/ObjC++ but not for C'
Low: legacy: Fix compile warnings so that we can turn on -Werror by default
Low: cts: Add the second type of fencing log that we look for
Low: cts: Add stonith BadNew ignore patterns automatically when someone calls fencing_cleanup
Low: cts: Look for 'we got quorum' fencing during NearQuorumPoint
Low: cts: Lower the bar for deciding which log message means corosync has stopped
Low: Build: Explicitly include mainloop.h where necessary to avoid include loops
Low: Build: Abort compilation on the first failure when calling 'make core'
Low: Corosync: Don't connect to the mcp for corosync-2.0 based clusters
Merge pull request #73 from davidvossel/bug-5045
Merge pull request #70 from yuusuke/fix_crm_config.h
Merge pull request #69 from gao-yan/controld-ra
Medium: Core: Fix compilation when corosync is disabled
Low: Core: Tweak the way bit flag changes are logged
Low: PE: Improved tracing for determining who and when (un)set an action flag
Low: PE: No need to pre-filter the flags argument to update_actions()
High: PE: Bug cl#5059 - Use the correct action's status when calculating required actions for interleaved clones
Low: PE: Bug cl#5059 - Use the correct action's status when calculating required actions for interleaved clones (regression test)
Merge branch 'master' of github.com:beekhof/pacemaker
Medium: Core: Implement the glue to use libqb for IPC
Medium: Fencing: Convert the stonith daemon to libqb's IPC code
Medium: mcp: Convert to libqb's IPC code
Medium: crmd: Convert to libqb ipc
Medium: pengine: Convert to libqb's IPC code
Medium: attrd: Convert to libqb ipc
Medium: cib: Convert to libqb ipc
High: Core: Replace use of IPC from clplumbing with IPC from libqb
Low: mcp: Additional documentation for runtime configuration options
High: crmd: Increase the default message limit to 5Mb for PE transition graphs
High: PE: Default to shared memory for IPC connections
Low: Core: Ensure trace logging goes to either stderr or a file
Low: crmd: Improved error message
High: Core: Make logging with libqb mandatory
Low: crmd: Reduce some state change logs to LOG_INFO
Low: crmd: Bump a log message for CTS
Medium: Core: Modify the glib IO channel wrappers to expose file descriptor support
Medium: Add file desciptor based connections to mainloop using IO Channels
Low: Core: Move heartbeat specific XML helpers to their own, conditionally built, file
Low: ais: Remove dead file
Low: cib: Remove low-value call stats
Medium: ipc: Block by default when sending events/responses but allow lower priority events to be discarded with a log message
High: cib: Add the ability to use sockets and/or shared memory for IPC
Low: mcp: Switch to the GLib API for watching child processes
Low: crmd: Switch to the GLib API for watching child processes
Low: crmd: Remove dead code
Low: Core: Simplify uuid creation
Low: crmd: The wrong message got demoted to debug
Low: cts: Update for new IPC patterns
Low: Fencing: Simplify the logic for replying to the original caller
Low: crmd: Stop checking for errors from a CIB call that doesn't produce them
Low: Core: Reduce log severity when unsupported clusters are encountered
Low: Fencing: Use g_child_watch_add() for tracking agent operations
High: crm_report: Always include system logs (if we can find them)
Low: Fencing: Remove the 'local_first' option because the intended user wont be using stonith-ng
High: crm_report: Correctly locate logfiles in common directories
Low: Core: Additional uuid->uname tracing
Medium: ipc: Prevent multiple calls to qb_ipcc_disconnect()
Low: cts: Don't auto-cleanup stale IPC and corosync data files
Medium: cib: Do not leak resources in clients if the server is killed
Low: cts: Look for files that are indicative of problems
Low: cluster-clean: remove stale IPC files too
Low: crm_report: Suppress collection errors when corosync isn't running
Low: ipc: Additional tracing
Low: cib: Use g_child_watch_add() for writing the CIB to disk
Low: crm_verify: Remove redundant code
Low: Core: Remove dead code
Low: crm_mon: Consistently log errno failures
Low: Core: Obtain md5sum functionality from gnulib
Low: Build: Only perform heartbeat-specific library checks if heartbeat is supported
Medium: Fencing: Disconnect from the CIB before exit
High: cib: Correctly disconnect clients regardless of whether mainloop is used and/or active
High: Fencing: Correctly disconnect clients regardless of whether mainloop is used and/or active
Medium: Fencing: Clean up IPC connections and memory on exit
Low: corosync: Logging improvements
Low: Build: Provide defaults for cluster-glue defines
Low: cts: New corosync BadNews pattern
Low: corosync: Additional detail when logging CPG membership changes
High: cib: Clean up if we detect a failed connection
High: fencing: Clean up if we detect a failed connection
High: cib: Ensure references to disconnected channels are released so they can be destroyed
High: cib: Stop the IPC servers on clean disconnect
High: fencing: Stop the IPC servers on clean disconnect
Low: Core: Make the trigger struct private
High: attrd: Clean up IPC server on exit
High: Core: Allow mainloop triggers to be used for long running processes
Low: corosync: Downgrade a log message now that we know the code is working
Low: crmd: Downgrade a warning to debug
Merge pull request #76 from davidvossel/bug-5063
Merge pull request #75 from davidvossel/bug-5064
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge branch 'libqb-ipc'
Low: Core: Simplify logging initialization helpers for cli tools
High: Core: Support a blackbox for additional logging detail after crashes/errors
High: Core: SIGUSR1 now enables blackbox logging, SIGTRAP to write out
Medium: Core: Allow tracing to be turned on at runtime with SIGPROF
High: Legacy: Fix compilation when building for corosync < 2.0
Low: extra: cluster-init - include the ability to sync node times
Low: Convert the last _debug_N logging calls to traces and remove stale macros
High: pengine: Move API additions to avoid ABI incompatabilities
Merge pull request #77 from gao-yan/controld-ra
Low: Build: Move cluster.h out of the common subdirectory
Low: Build: Avoid using backslashes to continue _SOURCES lines in Makefiles so we can programatically determine the variable's contents
High: Build: Bump library versions
Low: Build: Add a tool for keeping the library versions correct
Low: Build: Calls to 'make changes' use the previous release as a starting point
Low: Build: Speed up coverity runs
Low: Build: Tweaks to the bumplib script
Low: Core: Remove cruft from the common library
Medium: mcp: Improved startup logging
Medium: crmd: Ensure the attrd IPC connection is cleaned up, even if attrd dies unexepectedly
Medium: cib: Allow cibadmin to create blank configurations of a specific validation type
Low: Fencing: Support the power_wait option in our fake fencing agents
Low: Fencing: Implement a sane monitor op for our fake fencing agents
Low: cts: Add support for testing fencing levels
Low: cts: Add two new startup patterns - because the actual logs might dissappear if the node is fenced
Low: cts: Scripts often implicitly add -- as the last argument, handle it gracefully
Medium: tools: Automatically find pacemakerd trace logs if present
High: crmd: Prevent crash when using incorrect log format string
Low: fencing: Provide more detail when encountering commands deom unexpected clients
Low: Core: Indicate when closing an active IPC connection
Low: fencing: Don't complain about nodename in dummy fence agents
Low: extra: cluster-init - install fence-agents too
Low: Core: Remove redundant 'extern' keyword from function definitions in public include files
Low: Replace our wrappers with standard memory allocators
Low: Replace our crm_free wrapper with standard calls to free()
Low: Build: Move the ad-hoc test file
High: Core: Correctly enable blockbox logging
Low: Build: Rename 'make scratch' to 'make dirty'
Low: extra: cluster-clean - remove stale blackbox files
High: Core: Log blackbox contents, if enabled, when an error occurs
Low: Core: Differentiate between user-requested blackbox dumps and ones due to an error
High: crmd: Prevent double-free in crmd_authorize_message()
Low: extra: cluster-clean - fix mass deletion of state and log files
Low: Core: Increase the blackbox size to 1Mb
Low: Replace strncpy with the more efficient memcpy
Low: Core: Replace use of strlcat() with memcpy()
High: Fencing: Only remove active timers
High: fencing: Prevent duplicate conversion of timeouts to milliseconds
Low: fencing: Remove duplicate xml dumps
Low: Core: Use the value of LOG_TRACE from libqb if available
Low: Core: Remove unnecessary trace logging from xml->text conversion
Low: Build: strlcat is no longer used
Low: cts: Use power_timeout instead of power_wait, the latter is applied many times per call
Low: Core: Avoid duplicate blackbox dumping on assertion failures
Low: fencing: Timeouts should normally originate in the server
Low: fencing: Full messages are already logged in the IPC/CPG dispatch functions
High: fencing: Results of complex operations should only go to the originator (not broadcast)
High: fencing: Instruct slaves what type of logic is required
Low: fencing: Additional logging
High: fencing: Dividing the timeout equally between the available devices
Low: Tools: iso8601 - Fix the output from the date tool
Medium: Core: Write the blackbox out to a persistant location
Low: fencing: Check that the context is correctly set before continuing
High: Core: Avoid assertion error when underflowing days of the month in iso8601 date code
Low: cib: Supply the correct number of arguments to terminate_cib()
Low: Core: Ensure blackbox recorders are cleaned up on exit
High: cib: Prevent memory corruption of blackbox structures after calls to fork()
High: fencing: Correctly distribute the timeout between configured devices
Medium: tools: crm_report - Extract blacbox data files too
Low: tools: crm_report - Tweak the CTS filename pattern
Low: cts: Bump the action timeout now that there are multiple stonith devices called in series
Low: cts: Pattern updates
High: mcp: Correctly retry the connection to corosync on failure
Low: Core: Avoid forcing a logfile if the blackbox is enabled
High: crmd: Correctly re-establish failed attrd connections
Merge pull request #80 from fghaas/pkgconfig
Merge pull request #79 from gao-yan/proc-flag
Merge pull request #74 from gao-yan/ra-fixing
Merge pull request #78 from davidvossel/mainloop_fd_race_condition
Merge pull request #71 from davidvossel/bug-5054
Merge branch 'bug-5056' of https://github.com/davidvossel/pacemaker
Merge branch 'build-fixes' of git://github.com/lge/pacemaker
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: Build: Disable _FORTIFY_SOURCE when building for gcov
High: mcp: Include missing headers
High: fencing: Only select peers with the required device(s)
High: fencing: Do not prematurely set the operation state (ignoring future query results)
Low: ipc: Fix and improve trace message
Merge branch 'new_lrmd' of https://github.com/davidvossel/pacemaker
Low: cib: Include the call ID in the ack
Low: Build: Clean up gcov files created during configure
High: lrmd: Prevent use-after-free on double disconnect
High: legacy: Ensure the plugin starts the new lrmd
Merge pull request #84 from davidvossel/bug-5069
High: fencing: Delegate complex self fencing requests - we wont be around to see it to completion
Low: Core: Prevent blackbox over-dumping
Medium: fencing: Take extra steps to avoid parallel invocations of fencing devices
Low: cts: Start pacemaker only if corosync started
Low: extra: cluster-init - Keep debug messages away out of syslog and only send to the logfile
High: fencing: Ensure all peers are notified of complex fencing op completion
High: Fencing: Only the request owner should broadcast results of complex fencing operations
High: plugin: Fix use-after-free in legacy find_corosync_variant()
Low: fencing: Backport fixes from fence_true to fence_false
Low: cts: Code for testing the advanced fencing logic
Low: Core: Support tagged logs
Low: Core: Support enabling logs based on tags
Low: Git: Update ignore file
Low: Build: Clean up changelog tempfile
Low: Fencing: Account for newlines when logging output in run_stonith_agent()
Medium: crmd: Handle more fenced DCs in more situations
Low: Core: Consolidate logfile handling
Low: mcp: Always start pacemakerd in the forground
Low: Bump some log messages to notice
Low: Logging tweaks
High: Core: Only LOG_NOTICE and higher are sent to syslog
Low: Core: Move logging code to its own file
Low: Core: Improve the output format of logged XML diffs
High: Core: Make sure logging callsites are correctly evaluated
Low: cts: Update patterns now that info is no longer logged to syslog
Merge pull request #3 from davidvossel/no_pending_error
Merge pull request #5 from davidvossel/fixes_for_broken_stuff
Low: crmd: Move a development log message to 'trace'
Merge pull request #88 from kskmori/cib-memleak
Low: Build: Update the specfile changelog
Low: tools: Update regression tests
Merge pull request #86 from gao-yan/acl-libqb
Merge pull request #89 from gao-yan/acl-libqb
High: Core: Repair log message causing segfault in crm_ipcs_send()
Merge pull request #85 from davidvossel/bug-5068
Merge branch 'cts_rsc_recovery_fix' of https://github.com/davidvossel/pacemaker
High: Tools: crm_shadow - Bug cl#5062 - Correctly set argv[0] when forking a shell process
Low: Remove dead code identified by coverity
High: PE: Fix use-after-free errors detected by coverity
High: Core: Fix use-after-free error in decodeNVpair() detected by coverity
Low: Core: Fix minor memory leak in logging initialization reported by coverity
High: Core: Fix memory leak run_stonith_agent() reported by coverity
High: PE: Fix use-after-free errors detected in check_action_definition() by coverity
Low: crmd: reload cannot be NULL at this point unless malloc failed
Low: lrmd: Fix typo spotted by coverity
High: crmd: Prevent use-of-NULL in cancel_op() reported by coverity
Low: Build: Fix output of core-* helpers
Low: Build: Clean up after coverity
Low: Fix some CONSTANT_EXPRESSION_RESULT warnings reported by coverity
High: PE: Fix use-of-NULL in master_score() reported by coverity
Merge branch 'master' of github.com:ClusterLabs/pacemaker
High: PE: Fix additional use-after-free introduced in cs: 6a0d5dc - Fix use-after-free errors detected by coverity
Medium: Purge some legacy utilities
Low: Build: Move the install location for lrmd_test to comply with packaging guidelines
Low: Build: Ignore dictionary warnings from rpmlint
Low: Build: Ignore rpmlint's complaints that OCF agents are in /usr/lib
Low: cts: Whitespace
Low: cts: Ignore failures by... FenceFail
Low: Use standard exit codes from sysexits.h
Low: Core: Only include ha_msg.h if we're building heartbeat support
Low: Core: Provide defaults for some defines that we used to inherit from heartbeat
Low: Move more heartbeat specific code into conditional blocks
Low: Explicitly include required headers
Low: Remove dead code
Low: lrmd: Use the simpler uuid generation utility
Low: Tools: Only build crm_uuid when supporting heartbeat
Low: Fencing: Support building without support for the LHA agents
Low: crmd: Use defines from the new lrmd implementation
Low: Build: Remove dead code
Low: Build: Consolidate heartbeat specific files
Low: mcp: Ensure core directories are created at startup
Low: Build: Make building packages against cluster-glue optional
Low: Build: Default to /not/ building documentation for the mock and rpm helper targets
Low: tools: cibadmin - Fix regression tests
Low: Add a todo list
Merge pull request #90 from davidvossel/lrmd_indention_fix
Low: Todo list updates
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: Remove the remaining uses of crm_malloc/crm_free
Low: Remove the remaining uses of crm_malloc/crm_free
Low: Build: Add missing header to automake
Low: services: Import upstart functionality from cluster-glue
Low: services: Massage the upstart functionality to be suitable for use in Pacemaker
High: lrmd: Add support for upstart services
Low: service: Convert upstart to the GDBus API
High: services: 'service' should expand to the correct standard instead of relying on the binary of the same name
Medium: services: Implement an alternate systemd bridge that uses the GDBus API
Medium: services: Use the GDBus API for talking to systemd and upstart
High: services: Support Asynchronous systemd/DBus actions
High: services: Support Asynchronous upstart/DBus actions
Medium: services: More reliably detect if upstart and/or systemd are supported on the local machine
upstart sync fix
systemd sync fix
Medium: lrmd: Add API call for listing supported standards
Low: Build: Add headers to the dist tarball
Low: Core: Support unaliased long options for CLI tools
Low: services: Automatically remap monitor->status when appropriate
Low: Tools: crm_resource - add support for querying the list of supported agents/providers/standards
High: services: Strip the path prefix from upstart job lists
crm_resource fix
Low: services: Support operations that can block forever
Medium: lrmd: Support systemd and upstart metadata queries from the lrmd API
Medium: tools: crm_resource - Support metadata querying
Merge branch 'lrmd-upstart'
Low: services: Remove dead code
Low: services: Prioritize LSB scripts if they exist when using the 'service' alias
High: services: systemd_unit_by_name - Call LoadUnit if the initial call to GetUnit fails
High: services: systemd_unit_exec - Pass the agent type to systemd, not the resource name
Low: lrmd: Additional debug messages
Low: lrmd: Indicate the command being run in the regression test
Low: Build: Install the lrmd regression test for buildbot to run
High: Build: Fix the install directory for lrmd regression tests
Medium: services: Produce an error if an unsupported resource standard is requested
Low: Build: Allow support for upstart and/or systemd to be disabled
Low: Build install the lrmd regression test as a script so automake doesn't strip the execuatble bit (which rpmlint then complains about)
Low: lrmd: Update regression test, systemd now knows about the failed state
Low: doc: Update todo list
High: services: upstart_job_exec - Pass the agent type to upstart, not the resource name
Merge pull request #93 from davidvossel/lrmd_err2txt
Low: Build: Install services.h
Low: doc: Sort the todo list by target release
Merge pull request #91 from davidvossel/lrmd_coverity_fixes
Merge pull request #94 from davidvossel/lrmd_regression_tests
Merge pull request #96 from davidvossel/systemd_upstart_cleanup
Low: Use the standard g_list_free_full() function instead of our own custom one
Low: doc: Indicate that the docs are now written for corosync 2.0
Medium: Create a compatability header for software written against older Pacemaker versions
High: Unified error codes based on errno.h
Low: Don't bother wrapping realloc
Low: Fix spelling of compatibility header file
Low: Remove useless test
Low: Deprecate the slist_destroy() macro
Low: Deprecate the xml_prop_iter() and xml_prop_name_iter() macros
Low: Deprecate the crm_strdup() macro
Low: Tools: Update the regression tests for the new error codes and text
Merge pull request #97 from t-matsuo/disable-systemd-patch
Merge pull request #98 from davidvossel/bug-5072
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: cib: Remove old declaration of cib_error2string()
Low: Doc: TODO list update
Medium: fencing: Detect if we can support legacy Linux-HA fencing agents at runtime using dlopen()
Low: Build: Clean up some legacy baggage
Low: Build: Move private function declarations to an internal header
Low: Build: Remove unneeded replace functions
Low: Build: Mark private functions as internal
Revert "Low: Build: Mark private functions as internal"
Merge pull request #99 from davidvossel/master_monitor_regression
High: fencing: Fix compilation when cluster-glue is not present
Medium: lrmd: Correctly locate the required test binaries
Medium: services: Gracefully handle failure to connect to the upstart/systemd daemon
Low: lrmd: Add option processing
High: services: Be more tolerant of failures to connect to systemd/upstart
High: fencing: Tolerate agents that report empty metadata in the admin tool
High: fencing: Gracefully handle metadata requests for unknown agents
High: fencing: Stop building fake metadata for broken agents
Low: lrmd: Simplify the code for listing agents for various classes
High: lrmd: Indicate when stonith agents are supported
Low: lrmd: Only run regression tests for standards supported by the host/build
Low: lrmd: Add metadata to the fake fencing agent
Medium: services: Obtain all agents if no provider is specified for resources_os_list_ocf_agents()
Medium: services: Obtain all agents if no class is specified for resources_list_agents()
Low: lrmd: Simplify the list_ocf_agents() logic now that the services API is smarter
Low: Git: Update .gitignore
High: Build: Install error.h
Low: Build: Leave the mock buildroot available for subsequent testing
Low: Build: Additional mock helper targets
Merge branch 'master' of github.com:ClusterLabs/pacemaker
High: Build: Fix the previous commit
Merge pull request #100 from gao-yan/build-fix
High: services: Translate the upstart jobs back into their real name
High: services: Produce valid metadata even when the systemd description is not available
High: services: Support metadata queries when upstart/systemd are not available
High: lrmd: Correctly co-ordinate the destruction of asynchronous resource operations
High: services: Indicate when the requested systemd service is not installed
Low: lrmd: New upstart tests and minor corrections to existing ones
Merge pull request #101 from davidvossel/update_cfs
Low: doc: Prevent asciidoc from thinking yum output is a blockdef listing
Low: Build: Spec file tweaks
Low: TODO list updates
Low: Build: Move some headers to more appropriate locations
Low: Build: Move internal cluster stack functions into a non-public header in the correct location
Low: Build: Move private fencing defines to an internal header
Low: Build: Move private attrd declarations and/or deprecated functions to better locations
Low: Build: Remove dead header
Low: crmd: Remove dead code
High: Build: Fix compilation after moving headers around
Low: Build: Move private PE functions out of public headers
Low: Build: Move private core functions out of public headers
Low: Build: Move private CIB functions out of public headers
High: Fencing: Re-expose fields necessary for notification
Merge pull request #103 from t-matsuo/fix-compile-error-on-rhel6
Merge pull request #104 from gao-yan/check-gio-support
Low: Core: Promote crm_strerror() to a full function
Low: Build: Fix header installation
Low: Tools: New command line tool for getting the text for error codes
High: Fencing: Convert notifications API to use a formal struct instead of XML
Medium: Tools: crm_mon - optionally watch for fencing operations and send snmp/smtp traps
High: crmd: Block after 10 failed fencing attempts for a node
High: Promote support for service and systemd to the 1.2 schema
High: Promote support for node starting scores to the 1.2 schema
High: Promote support for per-instance constraints to the 1.2 schema
High: Promote support for advanced fencing logic to the 1.2 schema
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: Core: Move glib compatibility code into portability.h
High: cib: Include required header
Low: Build: Update spec file
Merge pull request #110 from gao-yan/build-fix
Merge pull request #111 from jnpkrn/patches
Merge pull request #112 from davidvossel/stonith_admin_fixup
Medium: Legacy: Rename the pluing's send_cluster_msg() function to avoid comflicting with heartbeat
Low: doc: TODO list update
Low: cib: Move more heartbeat specific functionality into a conditional block
Low: Better names for node state #defines
Medium: crmd: Drop the 'ha' field from node state as it was inferred from the other fields
Low: PE: Variable renaming
Medium: PE: Write a saner version of determine_online_status_fencing()
High: PE: Bypass online/offline checking resource detection for ping/quorum nodes
Low: crmd: Move more code to the heartbeat specific area
High: crmd: Do not overwrite any configured node type (eg. quorum node)
High: mcp: Update the SYSV init script to put pacemakerd into the background
Low: PE: Drop support for status entries using the parent:child naming standard dating back to the heartbeat 2 days
Low: PE: Create the dot graphs based on the resource names from the status section
Low: doc: Fix the URL for novell/suse docs
Low: doc: TODO list update
Low: crmd: Centrally include th cib header file
Revert "High: Promote support for per-instance constraints to the 1.2 schema"
Low: Build: Correctly detect uuid functionality on OSX
Medium: Core: Some error codes do not exist on OSX
Low: PE: Add another dimension to the regression test output
Low: PE: Sort the node list in abreviated clone details
Merge pull request #114 from mkubecek/xmlstrip
Medium: PE: Simplify anonymous clone handling by discarding recorded instance numbers
Low: PE: Simplify anonymous clone handling by discarding recorded instance numbers (update regression tests)
Low: PE: Simplify the logic for resource searching
Low: PE: Simplify pe_find_resource()
High: PE: Do not append instance numbers to anonymous clones
Low: PE: Do not append instance numbers to anonymous clones (update regression tests)
Low: PE: Reformat regression test inputs
Low: doc: TODO update
Merge pull request #115 from davidvossel/stonith_api_cleanup
Merge pull request #113 from davidvossel/fix_fake_fencing_agents
Merge pull request #109 from davidvossel/blocked_demote_promote_log
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: PE: Update regression tests
Merge branch 'colocation_changes' of https://github.com/davidvossel/pacemaker
Medium: crmd: Fix compilation with cman enabled
Low: PE: Show failures in buildbot
Low: PE: Install regression test summary files for buildbot
Low: PE: Fix regression test installation
Low: PE: Stop rpmlint from moaning
Merge pull request #107 from kskmori/version-hash
High: cib: No longer expose the the variant_op API call
Merge pull request #118 from gao-yan/fencing-fix
Merge pull request #116 from jnpkrn/patches
Merge pull request #117 from davidvossel/stonith_regression_tests
High: corosync: Populate the cib nodelist from corosync if available
High: corosync: Use the CFG API and DNS to determine node names if not configured in corosync.conf
High: corosync: Obtain a node name from DNS if one was not already known
Low: corosync: Consolidate node name lookup functionality
High: corosync: Obtain a full list of configured nodes, including their names, when we connect to the quorum API
Low: corosync: Reuse an existing CMAP connection if one exists, delay connecting to CFG until required
High: corosync: Fix ipv4 detection in corosync_name_is_valid()
Low: corosync: Additional tracing
Medium: xml: Node type is now optional
Low: PE: Minor code cleanup
Merge pull request #120 from davidvossel/lrmd_cleanup
High: Overhaul the recording of node state in the CIB
High: PE: Correctly set the status for nodes not intended to run Pacemaker
Medium: PE: Correctly detect nodes that failed the join process
Low: PE: New fencing regression test
Low: PE: Restore the order-wrong-kind regression test
Low: doc: TODO list update
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: PE: Add another fencing testcase
High: crmd: Set the join state to down for nodes no longer running the crmd
High: PE: Improve detection of rejoining nodes (with testcase)
High: PE: Stonith resources are allowed to start even if their probes haven't completed on partially active nodes
Low: extra: Support the creation corosync configurations with nodelists
High: crmd: Another overhaul the way node state is recorded and updated in the CIB
High: Core: Do not loop forever if an IPC client fails to empty its inbound queue
Medium: doc: Choose a more appropriate (site-local) multicast address for people to copy from
Medium: Schema: Officially support A after ( B or C or D )
Low: Core: Improved IPC logging
Low: Core: Only log callsite details once per process
High: ipc: Allow clients to make some sort of progress if a synchronous operation times out
Medium: IPC: Handle another common error code from qb_ipcs_event_sendv()
High: ipc: Prevent use-of-NULL when destroying an non-existant IPC server
Low: PE: Clean up after regression tests
Low: Build: Ignore coverity directories
Low: extra: Allow cluster processes to be killed without erasing configs
Low: PE: Minor logging change
Medium: cib: Drop unused API parameter while we're breaking everything else
High: ipc: Prevent use-of-NULL when closing connections in the client
High: attrd: Differentiate between updating all our attributes and everybody updating all theirs too
Low: cts: Find more daemon errors
Low: Core: Document the location of system errors
Low: attrd: Better logging of failed updates
Medium: ipc: Do not try to send to disconnected peers
Low: ipc: Log the event queue length when sending to a client fails
High: crmd: Reset the expected node state to pending when the cluster returns there
Low: ipc: Logging updates
High: ipc: Expose mainloop priority when adding ipc/fd sources
High: ipc: Read up to 10 messages per mainloop invocation
High: Core: Correctly extract the version details from a diff
Low: PE: Improved logging of failed actions
Medium: PE: Tollerate node_state entries with only uname specified
Low: attrd: Improved logging
Medium: cib: Include and verify a digest when replacing the entire CIB
Low: Core: Improved tracing for digests and failed diff applications
High: crmd: Ensure node_state updates have a valid uuid set
Low: Cluster: Better logging when all processes are marked as offline
Low: Core: free_xml() is a superset of free_xml_from_parent(), deprecate the latter
High: ipc: Retry sending requests that fail with -EAGAIN
High: cib: Resolve memory issues reported by valgrind
Low: crmd: Minor tweak to error message
Low: doc: TODO list update
High: lrmd: Prevent use-after-free by the client if the connection dies
Medium: crmd: Only close the attrd connection if sending failed permanently
Low: attrd: Improved client logging of failures
High: ipc: Fix error handling
Low: attrd: Expose the raw return code when IPC fails in attrd_update_delegate()
Low: ipc: Leave it to callers of crm_ipc_send() to decide if failures are an error
Low: crmd: Improved logging
Low: fencing: Delay error reporting when constructing a notification event
Low: cts: Update some BadNews patterns to ignore
Low: PE: Update regression test, stonith can now almost always make progress
High: cluster: Fix compilation when plugin support is enabled
High: Build: Fix conditional compilation of plugin support
High: legacy: Fix compilation after change to mainloop_add_ipc_client() signature
High: crmd: Fix compilation when plugin support is enabled
High: Library version bumps
Low: Build: Make the tool for bumping shared library versions smarter
Merge pull request #125 from davidvossel/pengine_stderr_fix
Merge pull request #124 from davidvossel/test_coverage_improvements
High: cib: Move cib.xml out of the heartbeat tree for new installations
Low: Core: Improve the output of xml diffs
Low: Core: Send a short-form xml diff to the logs unless debugging
Low: PE: Sort nodes alphabetically so that they display consistently in CLI tools
Low: PE: Sort nodes alphabetically so that they display consistently in CLI tools (update regression test)
High: Tools: Teach crm_report how to find the new config and PE file locations
Low: Tools: crm_report - Only search for files inside directories that exist
Low: Build: Default to /var/lib/pacemaker for core files
Low: todo
Low: Build: Update spec file
Medium: Update defines containing the stonith and cib library versions
Low: Core: Correctly determine if we should log callsite trace information
Low: cts: Wait until the cluster is stable before checking for startup fencing actions
Low: todo
Medium: attrd: Disconnect clients manually so we can log them
Medium: crmd: Prevent use-of-NULL when startup fails
Low: PE: Improved logging when nodes need to be fenced because of active resources
Low; PE: Drop unnessary log message
Low: PE: Improved logging when a node has been shot but still contains the terminate flag
Low: Fencing: Some additional logging when complex self fencing cannot be forwarded
High: PE: Prevent use-of-NULL when tracing is enabled
High: ipc: Re-allow truely asynchronous requests by no longer requiring a Response for every Request
High: crmd: Prevent use-of-NULL when free'ing empty hashtables
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge branch 'rsc_tag_logs' of https://github.com/davidvossel/pacemaker
Merge pull request #119 from davidvossel/stonith_regression_tests
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge branch 'stonith_poweroff_vs_off' of https://github.com/davidvossel/pacemaker
Medium: Fencing: Backwards compatibility code for the pcmk_poweroff_action option when the default is 'off'
High: ipc: Prevent use-of-NULL after a connection is closed
Medium: crmd: Ensure failed attrd connections are always cleaned up
Low: cts: Downgrade a log message now that startup fencing detection is reliable
Medium: Core: Only log command line arguments to syslog when -V is specified
High: attrd: Have single-shot clients wait for an ack before disconnecting
Low: PE: Supress 'multi active' error for fencing devices on unclean nodes
High: Fencing: Specify the action to perform using action= instead of the older option=
Low: attrd: Additional detail in trace logs
High: Core: Ensure failed IPC connections are cleaned up
Fix for 3d91917: High: attrd: Have single-shot clients wait for an ack before disconnecting
Low: cib: Create a common function for updating a local cib copy from an update event
Low: cts: Relax the fencing test - look for one less pattern
Medium: crmd: Wait for the last resource update to complete before shutting down
High: Prevent use-after-free when disconnecting from cib/lrmd/stonithd
Low: Core: Downgrade an error to a warning, its normal for fd-based connections
Low: Core: Prevent mainloop message from being sent to stderr
Low: cts: Correctly determin the highest known IP for a subnet
Low: extra: cluster-init - sync timezone details as well
Low: Build: Use the new core file location in the spec file
Merge pull request #126 from davidvossel/merge_fence_cmds
Low: Create a tools to run all our various regression tests
Low: Support gcov testing from the source tree
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: PE: Remove dead code
Medium: Build: Fix installation of cli regression test files
Low: BasicSanity - fail if any subtest failed
Low: Build: gcov magic moved to coverage.sh
Merge pull request #129 from davidvossel/lrmd_stonith_regression_test_exitcode
Merge pull request #128 from davidvossel/external_fence_transition_abort
Low: Tweak the mock helpers
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: doc: Update AUTHORS list
Merge branch 'master' of github.com:beekhof/pacemaker
Low: test: Disable fencing regression tests for now - they dont run in a chroot
Low: buildbot: Always install valgrind and lcov
Low: todos
Low: todos
Low: test: fencing and lrmd tests need root
Low: Build: Install helpful packages into the buildroot
Low: test: sudo doesn't work in a buildbot environment
Medium: ipc: Do not wait when trying to read Events
Low: Fix coverage tests
Merge pull request #130 from davidvossel/crm_resource_list_stderr
Low: Build: Simplify the task of creating a chroot
Low: Test: Fix argument passing for lrmd and fencing tests
Low: cluster: Smarter retry logic for connecting to corosync
High: crmd: If we're getting messages from a node, then it must be up
Low: test: Fix coverage testing within a chroot
High: PE: Fix memory leaks found by valgrind
Low: Tools: Clean up logging at exit
High: PE: Fix failcount expiration
Low: PE: Fix failcount expiration (regression test)
Merge pull request #131 from davidvossel/stonith_deadlock
Low: PE: Use slightly different scores in the hope that 64 and 32 bit arches can agree on a common result
Low: Build: Fix generation of SRPMs for opensuse
Medium: cib: Treat changes to remote-{tls,clear}-port as config changes
Low: tools: Bug rhbz#816881 - Indicate in the xml output of crm_mon when the resource state is cached
Low: Tools: Remove annoying line of '=' chars from crm_mon output
Medium: Resolve some coverity errors
Low: Tools: Clean up lrmd API memory on exit
Medium: crmd: Discard defered fencing updates at the correct time
Medium: Cluster: Clean up the common connection function
Low: Cluster: Move some more unnecessary elements to a private header
High: mcp: Prevent use-after-free when running against corosync 1.x
Low: tools: Ensure memory is cleaned up on exit to avoid false positives in valgrind tests
Medium: Tools: Deprecated option -h for cibadmin requires an argument (prevents use-of-NULL)
Low: Build: libqb is required on opensuse too
Low: Build: Add a control file for the travis build service
Merge pull request #134 from Xarthisius/master
Medium: PE: Bug cl#5089 - Process colocation with clones before primitives and groups
Low: Fencing: Avoid using the stdout keyword as a struct member
Low: Remove dead code
Merge pull request #135 from gao-yan/crm_ticket
High: Core: Bug cl#5032 - Rewrite the iso8601 date handling code
Low: Core: Add regression tests for the iso8601 date handling code
Medium: Core: Cleanup the iso8601 API
Low: legacy: Remove dead code
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: Core: Fix memory leak in iso8601 time comparision
High: Core: Fix parsing of iso8601 time intervals
Low: lrmd: Fix LSB regression tests
Merge pull request #137 from gao-yan/fsa-actions-i_fail
Merge branch 'stonith_timeout_3' of https://github.com/davidvossel/pacemaker
Low: Build: Place coverage results in the current directory
High: Fencing: Once again record the node's uname when connecting to the cluster (prevents use-of-NULL)
Low: cts: Update help text
Low: services: Keep the compiler happy when upstart and systemd are not supported
High: ipc: Correctly detect old replies
Merge pull request #139 from davidvossel/ipc_send_improvements
Low: Core: All logs are sent to the Blackbox if enabled
Low: Core: Better ipc trace logging
Low: Core: Trial a 5Mb blackbox buffer
Low: crmd: Reduced logging format for incomplete transtions
Low: Core: Improved ipc trace logs
High: mcp: Do not shut down IPC until the last client exits
Low: Core: Simplify blackbox handling now that all messages go there if it is enabled
Low: cts: New stonith patterns
Low: tools: crm_report - tolerate log entries without dates when determining log ranges to extract
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: crmd: Do not try to invoke the PE if we're not in S_POLICY_ENGINE
Low: Core: Support dynamic blackbox trigger points
Medium: PE: Do not block stop operations for nodes that wish to shut down
Low: crmd: Further improved output format for aborted transitions
Low: PE: Include probe_complete more often to increase clarity when displaying the transition graph
Low: extra: Use a corosync config that more closely matches the default
Low: crmd: Swallow warnings for old elections
Medium: crmd: Aborted transitions are no longer an error condition
Low: cts: Look for more reliable corosync patterns
Merge branch 'stonith_bcast_origin_node' of https://github.com/davidvossel/pacemaker
Low: crmd: Downgrade a log message now that we know the logic works
High: crmd: Improved handling of node up/down events
Low: crmd: Exclude completed operations from the 'waiting' section of printed transition graphs
Low: cts: Look for terminated transitions
Merge branch 'master' of github.com:ClusterLabs/pacemaker
High: crmd: Prevent memory-leak when printing transition graphs
Medium: Core: Improved command line logging
Medium: fencing: Log the command-line of any (un)fence actions
Low: cts: Correctly handle nodes that are fenced during the ComponentFail test
High: Fencing: Prevent use-of-NULL when fencing is initiated by stonith_admin
High: Fencing: Delegation of complex fencing operations requires that we know the node name of all peers
Low: Fencing: Add whitespace to log message
Low: cts: More quickly determine which log method to use
Merge pull request #143 from gao-yan/load_stopped_loop
High: Fencing: Bug rhbz#853537 - Prevent use-of-NULL when the cib libraries are not available
Low: Logging tweaks
Medium: Fencing: Allow node suicide when fencing via stonith_admin
Low: cts: Use stonith_admin for StonithdTest
Merge branch 'stonith_service_lib_usage' of https://github.com/davidvossel/pacemaker
Low: Tools: Rename the crm_simulate source file
Medium: Tools: crm_report - Improved handling of malformed lines
Low: Fencing: Addition logging
Merge branch 'stonith_cts_fixes' of https://github.com/davidvossel/pacemaker
Low: Core: Tweak the blackbox location/filename
Low: cts: The output from nmap is already in increasing order, sort messes it up
Low: doc: CfS - the crm shell may not be present anymore
Low: doc: Draft changelog for 1.1.8
Low: doc: Reformat the changelog
Low: build: Fix the 'make changes' target
Low: Build: Package the blackbox directory in the rpm
Low: extra: Correctly remove old blackboxes
Low: tools: crm_report - send stderr from corosync-blackbox to the stored file too
Medium: fencing: Improved logging of fencing initiation
Low: Git: Ignore lcov artifacts
Low: fencing: Have fence_pcmk record failures
Medium: fencing: Do nothing if an equivalent request succeeded less than 'tolerance' seconds earlier
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Medium: fencing: Always send our node name when new nodes join - they won't know us even if we know them
Medium: fencing: Improve log comprehension by including the name+pid of the caller in more logs
Low: build: Update the changelog helpers
High: Core: Prevent use-after-free in mainloop_del_fd()
Low: PE: Minor cleanup
Medium: Core: Keep counting in write_last_sequence() when no series max is specified
Low: Change OpenAIS references to Corosync
Low: crmd: Move function to the only place it is used
Low: PE: Fix the output of failed resources
Merge branch 'lrmd_stop_recurring' of https://github.com/davidvossel/pacemaker
High: cman: Prevent SIGPIPE by not having the crmd write to the fence_override pipe after fencing completes
High: crmd: Wait for stonith-ng to reply before declaring stonith ops complete
High: corosync: Use unsigned nodeid's in the cib
High: PE: Correctly find failcounts for /stopped/ anonymous clones
Low: Minor logs changes
Merge pull request #149 from davidvossel/lrmd_only_notify_changes
Low: Core: Only log the command line once
High: Core: Ensure file descriptors and IPC channels attached to mainloop are correctly cleaned up on disconnection
High: Core: Ensure server-side IPC connections are cleaned up on client exit
Merge pull request #150 from davidvossel/send_by_nodeid
High: fencing: Merge top-level fencing requests
High: Fencing: Report timeout values for merged operations
Low: Fencing: Wait for killed operations
Low: Fencing: Some additional logging changes
Low: cts: Handle stonith_admin timeouts in the Stonithd test
Low: cts: Some extra BadNews patterns to watch for
Low: lrmd: Display lrmd logs in verbose regression test output
Low: Fencing: Simplify the generated corosync.conf file for regression tests
Low: Fencing: Better pattern matching in regression tests
Low: Fencing: Display the unmodified operation timeout
Medium: Fencing: Target is not required for query operations, search for the parent of @action instead
Low: Fencing: Increase the timeouts for some regression test commands
Merge pull request #6 from davidvossel/stonith_regression_fixes
Low: PE: Remove stopped resource status entry from test
High: PE: Correctly unpack anonymous clones
Low: PE: Additional tracing
High: PE: Correctly find action definitions for anonymous clones
Merge branch 'master' of github.com:ClusterLabs/pacemaker
High: Core: Don't attempt to use empty hashtables
High: crmd: Ensure crm_get_peer() is always passed a valid uname by join_query_callback()
Merge branch 'scheduling-priority' of https://github.com/gao-yan/pacemaker
Merge pull request #151 from davidvossel/cfs_pcs_update
Low: doc: Fix/improve various aspects of Clusters from Scratch
Low: doc: Document the way we use asciidoc
Low: Build: Add fence-agents to the chroot for fencing tests
Low: doc: Tweak CfS layout
Low: Build: Update version to 1.1.8
Low: doc: Update the changelog
High: cib: Expose the correct #define for external users
Merge pull request #157 from davidvossel/stonith_merge_tests
Merge branch 'master' of github.com:ClusterLabs/pacemaker
High: Cluster: Allow cman and corosync 2.0 nodes to use a name other than uname()
Low: PE: Consoldate duplicated code
Low: Cluster: Minor code cleanup
Medium: mcp: Correctly disable syslog output when requested
Merge pull request #159 from davidvossel/stonith_wait_query_results
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Medium: fencing: Clean up after operations that fail during unpack
Low: Core: Warn about node names containing a capital letter
Medium: Improved compatibility with older distros
Low: cts: Add corosync to the ComponentFail test
Low: doc: Update the todo list
Merge branch 'master' of github.com:ClusterLabs/pacemaker
High: Move g_strcmp0() compatibility function to be with other 2.16 compat code
High: mcp: Only define HA_DEBUGLOG to avoid agent calls to ocf_log printing everything twice
Merge branch 'stonith_cib_device_cache' of https://github.com/davidvossel/pacemaker
Low: PE: Set a better task id for stonith actions
High: PE: Support a 'requires' resource meta-attribute for controlling whether it needs quorum, fencing or nothing
Medium: PE: Improved summary of resources being moved
Medium: PE: Always require fencing if the resource is configured that way
High: PE: Support resources that require unfencing before start
Low: PE: Log when resources require fencing but fencing is disabled
Low: Core: Obtain more information when someone tries to parse an empty string as xml
Low: Build: Backport some specfile updates suggested by rpmdiff
Low: doc: Convert more parts of Pacemaker Explained to asciidoc
Low: doc: Complete the conversion of Pacemaker Explained to asciidoc
Low: doc: Correctly indicated when sourcecode listings are examples and other formatting updates
Low: doc: Use C comments for commands because they're a nicer color than fluro magenta
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #164 from jnewland/fix-crm_uuid
Low: doc: Improved text for cross-references
Low: doc: Add details on systemd, upstart and service resource types
Low: doc: It is preferred to set 'requires' on the resource instead of the start operation
Low: doc: order.kind replaces order.score
Low: doc: Overhaul the P-E index
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: doc: Small tweaks to the layout
Low: doc: Tweak the build process
Low: cluster: More accurately log nodes with no active processes
Medium: PE: Resource.requires also applies to promote actions
High: legacy: Relax the 'active' check for plugin based clusters to avoid false negatives
High: PE: Remove special cases for fencing resources
Low: PE: Remove special cases for fencing resources (regression test updates)
Low: Update .gitignore
Low: PE: Inline comments for recovery policies
Low: Build: Build both pcs and crm versions of the documentation for clusterlabs.org
Low: doc: Fix unterminate conditional asciidoc blocks
High: legacy: Skip peer process check if the process list is empty in crm_is_corosync_peer_active()
High: Heartbeat: Remove incorrect assert during cluster connect
Low: PE: Fix regression tests
Low: doc: Fix compilation with older versions of asciidoc
Merge pull request #168 from davidvossel/stonith_inherit_pending_op
Low: doc: Refer to the original shell as crmsh to avoid confusion
Low: Build: De-deuplicate the Requires for -devel
Medium: mcp: At startup, ensure important directories exist or are created
Merge pull request #7 from grueni/upstream
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge branch 'master' of github.com:beekhof/pacemaker
High: cib: Fix compilation by restoring const qualifier
Low: doc: Copy in the content from the now defunct http://www.29west.com/docs/THPM/multicast-address-assignment.html
Low: doc: Fix table formatting
Low: doc: Fix the google analytics link
Low: doc: Tweak build process
Low: Build: Add basic doxygen support
Low: fencing: Update the description of fence_pcmk
High: PE: Bug cl#5101 - Ensure stop order is preserved for partially active groups
Low: Build: Fix anchor for STONITH chapter
Medium: Fencing: Implement a dummy function that uses -lpils so that smartass linkers don't drop the reference
Low: doc: Fix escaped brackets in xml examples
Medium: crmd: Bug cl#5109 - Do not record pending cancel operations in the CIB
Low: doc: Ensure pcs and crmsh variants get built in serial
Medium: tools: Fix handling of cibadmin --node
High: IPC: Bug cl#5110 - Prevent 100% CPU usage when looking for synchronous replies
Merge branch 'stonith_retry_until_timeout' of https://github.com/davidvossel/pacemaker
Low: cts: Disable NearQuorumPoint when testing CMAN
High: Fencing: Supply the correct number of arguments for new error messages
Merge branch 'stonith_retry_until_timeout' of https://github.com/davidvossel/pacemaker
High: PE: Fix clone_zero() and clone_strip() for single character resource names
High: PE: Correctly unpack active anonymous clones
Low: PE: Reduce needless copying
Low: doc: Update for pcs_passwd -> passwd/PAM transition
Low: doc: Include an Italian translation of Pacemaker Explained
Low: doc: Remove duplicate images from the ro-RO translation
Low: doc: Update translation templates
Low: cts: More patterns to ignore when corosync is killed
Low: Build: Support the public coverity instance
Low: crmd: Correctly log when quorum is lost
Low: cib: Include the digest when displaying xml diffs
High: Core: Correctly process XML diff's involving element removal
Low: doc: Updated completed items in todo list
Low: Core: Update our configuration header with new defines
Low: Build: Create pkg-config files for our libraries
Merge pull request #170 from phmarek/master
Low: Fencing: Provide metadata for our two helper agents
High: cib: Reduce duplication and ensure all diffs contain an md5 digest
High: Core: Prevent ordering changes when applying xml diffs
High: Replace the use of the insecure mktemp(3) with mkstemp(3)
Medium: cib: Bug cov#739457, cov#739458 - Time of check time of use
Medium: Cluster: Bug cov#739454 - Copy into fixed size buffer
Medium: Upstart: Bug cov#739433, cov#739432 - Dereference after null check
Low: cts: Install the cts helper to a location that doesn't get erased on reboot
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Medium: Core: Non-logging tracepoints shouldn't be enabled by the blackbox alone
Medium: Ensure shared memory for blackboxes are removed from disk before exit
Merge pull request #171 from davidvossel/bug-5111
Low: Core: Bug cov#739430, cov#739431 - Dereference after null check
Low: IPC: Bump the timeout sent to qb_ipcc_recv() for synchronous replies
Low: extras: perl-TimeDate is needed for CTS
Low: Core: Bug cov#739444 - Resource leak
Low: Core: Bug cov#739451 - Insecure temporary file (claimed)
Low: Core: Big cov#739460 - Time of check time of use in crm_lock_pidfile()
Low: Fencing: Bug cov#739425 - Logically dead logging code
Low: Fencing: Bug cov#739426 - Correctly pass back failures to exec agents by internal_stonith_action_execute()
Low: Fencing: Bug cov#739428, cov#739429, cov#739447 - Dereference before null check
High: Fencing: cov#739453 - String not null terminated
Low: Core: Bug cov#739707 - Dereference after null check
Low: cib: Bug cov#739564 - Resource leak
Low: crmd: Bug cov#739459 - Suppress 'time of check time of use' defect
Medium: lrmd: Fix removal of old lrmd instances
Merge pull request #172 from davidvossel/mcp_pid_reattach
High: Remove functionality for terminating old lrmd and pengine instances
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: mcp: Tweak log message
Low: systemd: Differentiate between start and stop timeouts
Low: mcp: Speed up shutdown in the presence of detatched child daemons
Low: tools: Improved documentation for the --op-fail and --op-inject options
High: Fencing: Fix extraction of agent output
Medium: Fencing: Suppress an error during normal shutdown
Low: Tools: Tweak the crm_simulate examples
High: Core: Default to uname(2) for corosync clusters if no DNS or /etc/hosts entry is found
Low: cts: Work around a bug in fence_xvm
Medium: Allow node removal by name
Medium: Tools: crm_node - Delete cib entries when removing nodes
Medium: Fencing: Suppress (correctly) an error during normal shutdown
Medium: Tools: Support crm_node --remove on Heartbeat and CMAN
Low: doc: Update documentation for crm_node --remove
Merge pull request #177 from gao-yan/ptest-fix-build
Merge pull request #176 from gao-yan/obsolete-lrm-resource
Low: Tools: Allow cibadmin --xpath queries to format the results as XPath addresses
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: cts: New patterns to ignore in ComponentFail
High: Cluster: Strip domains from node names by default
High: Cluster: Detect node name collisions in corosync
Low: PE: Fix regression test input - invalid xml
Merge pull request #181 from jnpkrn/patches-xml
Merge pull request #180 from jnpkrn/patches
Low: doc: Remove dead file
Medium: Corosync: Use nodelist.$id.name verbatim if supplied
High: Corosync: Stop using reverse DNS lookups to try and infer node names
High: Cluster: Do not strip the domain from node names calculated from uname(2)
Low: doc: Document where Pacemaker gets the node name from
High: legacy: Reduce confusion by aborting when we find an active corosync cluster with no CMAN/pcmk plugin loaded
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: doc: Tweak node name text
High: Fencing: Fix passing of parameters from CMAN containing '='
Medium: Fencing: Make pcmk_fencing_max_retries action-specific and lower the default to 2
Revert "High: IPC: Bug cl#5110 - Prevent 100% CPU usage when looking for synchronous replies"
Low: doc: Add missing images
High: extra: Correctly look for globally_unique parameter in o2cb
Low: Core: Use strcmp() directly in XML related code - 15% performance boost comparing CIBs
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge branch 'clone-from-template' of https://github.com/gao-yan/pacemaker
Low: Define a common utility for printing strings with newlines
Low: cts: Support openstack fencing
High: cib: Do not pass local variables to mainloop_add_fd() when creating remote connections
Low: Use common functions for daemon option handling and help text
Low: extra: Make cluster-init a bit more flexible
Low: cts: Helper script for developers
Low: extra: cluster-init - Support corosync udpu
Low: cts: Fix the helper script to work from its new location
Medium: crmd: Don't rely on $HA_cluster_type when pushing the cluster type into the cib
Low: cts: Always use 'service' to start/stop the cluster
Low: cts: Make it easier to run cts from an uninitialized source tree
Low: cts: Drop unused variable
High: Core: Supply the correct number of arguments for the format string
Low: Remove duplicate log initialization
Low: Core: Don't complain if the requested signal handler is already installed
Low: doc: Guide for configuring a cluster of openstack instances
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: tools: crm_report -S should replace any previously supplied values
Low: corosync: Correctly log node IDs as unsigned ints
Low: extra: Default to $cluster_name for cluster-helper
Low: cts: Correctly detect the location of cts
High: attrd: Correctly handle deletion of non-existant attributes
Low: fencing: Log output consistently
Low: cts: Add a shorter alias for --test-ip-base
Low: cts: Make sure the fake fencing agents are executable
Medium: mcp: Clean up after partially active cluster instances with no pacemakerd process
Low: cts: Have 'local-init' set up crm_report too
Low: tools: Remove dead code
Low: cts: Smarter detection of node fencing due to quorum
Low: fencing: Make our dummary agents executable
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: cts: Fix fenced peer detection
Low: cts: Grab local copies of helpers utilities
Low: extra: Fix comparision when installing fencing files
Low: cts: Speed up fencing searches
High: Fencing: Do not try to copy NULL metadata strings from stonith plugins
High: corosync: Avoid errors when closing failed connections
Low: cts: Template for rhev-m based fencing
Low: cts: Wait for fenced nodes to return before continuing
Medium: Fencing: Have legacy agents log consistently with the rest of pacemaker
Low: cts: Simplify the startup fencing detection
Low: cts: Always save CTS logs to $HOME
High: mcp: chmod the correct directory instead of /var/run
Low: Build: New dependancies
Low: cts: Additional debug for fencing_cleanup()
High: mcp: Fix SYS-V shutdown
Low: cts: Re-enable startup fencing detection
Medium: Legacy: Attempt to trace the source of SIGBUS
High: PE: Bug rhbz#880249 - Ensure orphan masters are demoted before being stopped
High: PE: Bug rhbz#880249 - Teach the PE how to recover masters into primitives
Low: PE: Bug rhbz#880249 - Teach the PE how to recover masters into primitives (regression test)
Low: PE: Minor improvement to the transition summary for masters that restarted in a different role
High: cman: Ignore qdisk 'nodes'
Low: cts: Support -g for consistency
Merge branch 'master' of github.com:ClusterLabs/pacemaker
High: cman: Ensure reliable shutdown
Low: Avoid exceeding the blackbox line character limit
High: Fencing: Correctly terminate when all device options have been exhausted
High: legacy: Re-enable logging from the pacemaker plugin
Low: Build: Fix libqb dependancy
Low: PE: Fix regression test
High: Fencing: Record delegated self-fencing operations in case they fail
Low: Bump library versions for changes since 1.1.8
High: cman: Avoid delaying shutdown - confirm is a standard function
Low: cts: StatusCmd requires a node name
High: cman: Do not wait for unmanaged resources during shutdown
High: Cluster: Preserve corosync membership state when matching node name/id entries
Medium: cman: Backport the corosync 2.x version of pcmk_cpg_membership()
Low: Build: refine the list of public library headers
Low: cts: Central point to enable logwatcher/rsh tracing
Low: cts: Speed up log watching
High: tools: crm_report - Fix node list detection
High: corosync: Ensure peer state is preserved when matching names to nodeids
High: crmd: Prevent election storms caused by getrusage() values being too close
Low: fencing: Additional code clarity
High: Core: Prevent use-of_NULL in IPC code
High: cib: Avoid use-after-free by correctly support cib_no_children for non-xpath queries
High: crmd: Avoid filling the ipc queue with data we don't need
Low: crmd: Remove some trace logging
Low: crmd: Move the election storm logic
Medium: ipc: Halve the number of retries before giving up sending to a slow client
Low: crmd: Simplify the logic in process_graph_event()
Merge pull request #202 from xiali/master
Low: PE: Update regression tests for new feature version
High: PE: Ensure previous migrations are closed out before attempting another one
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Medium: crmd: Only look for election storms if we won the election
Medium: Fencing: Supply the correct number of arguments for INFO log
Medium: Core: Write out a blackbox for non-fatal asserts
Low: crmd: Reset the election count after logging an existing storm
Medium: mcp: Reduce the interval for find dead children in detached parent mode
Low: cts: Another pattern to look for
Low: doc: Update TODO list
Low: crmd: Do not count no-votes towards an election storm
Low: crmd: Obtain more information on two potential defects
Low: tools: crm_report - Ignore timestamps of remote files
Low: Core: Allow the blackbox to be dumped once per second, a lot can happen in that time
High: tools: Have crm_resource generate a valid transition key when sending resource commands to the crmd
High: tools: Fix compilation
Merge branch 'probe-timeout' of git://github.com/gao-yan/pacemaker
Merge pull request #203 from gao-yan/on-fail-stop-log
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Low: Core: Improved logging
High: Core: Do not attempt to call free() in a signal handler
Low: extra: cluster-init - repair defaults logic
Medium: crmd: Improved logging on lrmd disconnect
Low: PE: Log formatting
Medium: crmd: Always store the data when stalling the FSA
High: crmd: Only call set_slave() if we were the DC to avoid spamming the the cib
Low: Core: Only increment the blackbox count when writing to disk
High: crmd: Have cib operation timeouts scale with node count
Low: cib: Use register_callback() directly in clients
Low: crmd: Slightly improved logging
Medium: crmd: Don't go into a tight loop while waiting for something to happen
Merge branch 'master' of github.com:ClusterLabs/pacemaker
High: crmd: Store all current actions when stalling the FSA
Low: cib: Attempt to track down a pause that happens in or before cib_process_diff()
High: Fencing: Prevent use-after-free in debug logging
Low: cib: Logging updates
Medium: cib: Improve performance by only validating every 20th slave update
Low: Trace logging updates
High: cib: Performance improvements for non-DC nodes
Low: Core: Obtain more data when an IPC request times out
Log: ipc: Correctly log disconnected sessions after calling qb_ipcs_response_sendv()
Log: cib: Log when forwarding messages
Log: Require some xml logging to be enabled explicitly, even for the blackbox
Log: cib: Find out why some commands take a long time
Refactor: Core: A faster and more consistant digest function
Log: More xml that needs to be explicitly enabled
Fix: crmd: Prevent use-after-free of fsa_message_queue during exit
Fix: pacemakerd: Ensure libqb blackbox files are cleaned up on exit
Fix: attrd_updater: Ensure libqb blackbox files are cleaned up on exit
Doc: Update todo list
Refactor: Safe version of g_hash_table_size() that checks for NULL first
Feature: Drop per-user core directories
Fix: Correctly stall the FSA when waiting for additional inputs
Refactor: IPC: Create a common wrapper for client connections
Log: Remove trace messgae
Log: Better trace logging for resources active at shutdown
Test: Fix ignore patterns for the Reattach test
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Refactor: Clean up the internal send/recv API for remote connections
Low: Additional tracing for remote connections
Test: lrmd: Clean up daemon log on exit
Feature: IPC: Use queues to prevent slow clients from blocking the server
Feature: Use shared memory for IPC by default
Feature: Reliably detect when an IPC message size exceeds the connection's maximum
Feature: Compress messages that exceed the configured IPC message limit
Refactor: Use our custom xml-to-string function for performance
Test: Reduce scratch.c to essentials
Test: fencing: Update fence_false metadata
Fix: ipc: Ensure synchronous replies are decompressed as necessary too
Merge pull request #210 from gao-yan/RA-SysInfo
Refactor: ipc: Allow messages to be constructed once and sent multiple times
Refactor: cib: Construct each notification once, not once per client
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: Core: Implement true escaping for the five predefined XML entities
Doc: Clusters_from_Scratch: Fix pacemaker version
Merge pull request #216 from gao-yan/acl-compilation
Merge pull request #211 from gao-yan/fencing-registration
Merge pull request #195 from gao-yan/container
Log: crm_verify: Correctly disable the 'Use -V for more details' message
Style: crm_resource - combine formatting strings
Refactor: Remove unused elements of clone and group private structures
Feature: Check for master/slave resources that are not OCF agents
Test: Only use OCF agents for master/slave tests
Merge pull request #221 from gao-yan/nagios
Remove dead file
Merge pull request #222 from inouekazu/set_env_variables
Merge pull request #224 from gao-yan/process-ticket-constraints
Log: Update debug message with more details
Test: Allow xmlNodeDump() to be easily re-enabled further debugging
Refactor: Core: Switch to a 'double' memory re-allocation stratgey for xml buffers
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: Remove extra newline from cibadmin output
Log: crmd: Improved message when resources are active at shutdown
Refactor: Clean up ipc includes
Test: ComponentFail: Ignore resources active during shutdown
Fix: cman: Force fenced to terminate on shutdown
Log: fence_pcmk: Supply logger with a tag to use
Fix: Correctly clean up blackbox data on cli tool exit
Fix: crm_report: Ensure policy engine logs are found
Fix: TLS: Remove references to long since deprecated types
Fix: PE: Fix memory leak in dead code
Fix: Services: Newer versions of glib no longer need g_type_init() to be called
Log: fencing: Remove duplicate trace output
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: Fencing: Advertise the pcmk_[action]_timeout options for stonith agents
Fix: ipc: Prevent segfault in crm_client_get_by_id()
Fix: fencing: Correctly record completed but previously unknown fencing operations
Style: Prepare to run indent
Style: Reformat C-code with indent for consistency
Fix: pengine: Avoid leaking memory when cleaning up failcounts and using containers
Fix: cib: Prevent error message by correctly handling peer replies
Test: Clean up the primary configuration file
Fix: Bug cl#5135 - Improved detection of the active cluster type
Build: Do not run indent on the gnu extras
Test: ComponentFail: Determine the correct patterns to look for when killing corosync
Fix: corosync: Correctly detect corosync 2.0 clusters even if we don't have permission to access it
Fix: fencing: Attempt to provide more information that just 'generic error' for failed actions
Fix: Date/time: Bug cl#5118 - Correctly convert seconds-since-epoch to the current time
Fix: fencing: Unique return code when agents do not support the requested operation
Test: Support for collecting load stats while cts is running
Merge pull request #229 from l-mb/master
Fix: xml: Remove over-use of memory cleanup function
Log: Trace log for return code on exit
Log: crmd: Additional tracing of 'age' during elections
Log: A_LOG actions to reach the file logs
Fix: corosync: Clean up CMAP connections after querying node name
Fix: cts: Don't crash if stats are not enabled
Refactor: crmd: Rearrange the election logic
Log: Reduce non-essential logging
Fix: pengine: Re-log the name of unchanged input files
Fix: systemd: Gracefully handle unexpected DBus return types
Fix: core: Provide a sane default for the compression target in crm_compress_string()
Feature: corosync: Use queues to avoid blocking when sending CPG messages
Feature: ipc: Support compressed messages from clients
Feature: crmd: Enable A_DC_JOIN_OFFER_ONE
Refactor: crmd: Consolidate the pacemaker level membership negotiation logic
Fix: crmd: Prevent timeouts when performing pacemaker level membership negotiation
Fix: ipc: Prevent assertion failure by specifying the correct condition
Fix: ipc: Use 50k as the default compression threshold
Refactor: cluster-init: Support aliases for rhel*
Test: cts: Reduce timeouts for fencing agents
Test: cts: Additional patterns to ignore when killing corosync
Fix: crmd: Detect and recover when we are evicted from CPG
Merge pull request #231 from haibinshi/ra_controld
Log: crmd: Indicate the current join state if do_dc_join_finalize() fails
Fix: crmd: Process join ack's in S_INTEGRATION
Log: xml: Drop unuseful trace log
Refactor: crmd: Sort switch cases by value
Log: crmd: Reduce severity of message now that join_offer_one is enabled
Fix: crmd: Differentiate between a node that is up and coming up in peer_update_callback()
Test: cts: Look for fencing operations even if node startup failed
Test: cts: Wait longer for the server in fence_rhevm
Test: cts: A less strict regex for BadNews in ComponentFailure tests
Feature: fencing: Ability to identify fencing operations with a tag
Test: cts: Look for any node being fenced during CMAN startup
Log: corosync: Better description for CS_OK in ais_error2text()
Fix: crmd: Improved continue/wait logic in do_dc_join_finalize()
Fix: fencing: Do not wait for the query timeout if all replies have arrived
Fix: Invoke destroy functions if we are evicted from the CPG group
Log: PE: Display expanded details for cloned groups when necessary
Refactor: Fencing: Move comment to a more accurate location
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: crmd: Correctly determin if cluster disconnection was abnormal
Refactor: PE: Remove unnecessary code duplication
Fix: PE: Any location constraint for the slave role applies to all roles
Log: crmd: Improved logging when chosing a node's configuration to sync everywhere
Fix: Fencing: Do not merge new fencing requests with stale ones from dead nodes
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: Bug rhbz#913093 - Use crm_node instead of uname
Fix: rgmanager: Bug rhbz#911461 - Remove redundant usage of xmlCleanupParser()
Fix: crm_mon: Do not try to render a blank cib and indicate the previous output is now stale
Merge pull request #237 from gao-yan/crmd-relay-messages
Fix: crm_mon: Look for the correct error codes when applying configuration updates
Fix: fencing: Fix non-comparison when sorting devices by priority
Fix: fencing: Update priority metadata to reflect what the code actually does
Fix: fencing: Maintain an in-memory copy of the cib so that we have access to the full device definitions
Feature: fencing: Allow target-role to disable devices completely
Feature: fencing: Allow location constraints to disable devices for a specific node
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Refactor: PE: Clean up native_print
Fix: PE: Don't unconditionally move curser to the line start for ncurses
Fix: crmd: Correctly detect foreign fencing events
Log: Reduce severity for internal (recoverable) conditions
Doc: Pacemaker Explained: Remove shell-specific details from Stonith chapter
Doc: Pacemaker Explained: Document the differences between stonith and normal resources
Doc: Pacemaker Explained: Document fencing topologies
Doc: Pacemaker Explained: Expand on the implications of instantiating fencing devices on-the-fly
Feature: extra: Script for producing SNMP traps for crm_mon events
Build: doc: Remove duplicate descriptive text from the publican build files
Doc: Pacemaker Explained: Improved text describing when fencing completes
Doc: Pacemaker Explained: Import the options for fencing devices from the stonithd metadata
Log: Pacemaker Explained: Tweak some text around fencing topologies
Fix: PE: Bug cl#5140 - Allow set members to be stopped when the subseqent set has require-all=false
Refactor: PE: Remove dead code
Fix: PE: Bug cl#5143 - Prevent shuffling of anonymous master/slave instances
Test: PE: Bug cl#5143 - Prevent shuffling of anonymous master/slave instances
Refactor: Use more specific error codes than -EREMOTEIO
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: OSX: clock_gettime() is not available on Darwin
Fix: OSX: Fix compilation when no stacks are supported
Fix: lrmd: Ensure portability.h is included before lrmd_private.h
Merge pull request #245 from gao-yan/build-acl
Merge pull request #246 from gao-yan/ipc-memory-leak
Merge pull request #247 from JamesGuthrie/fix-crm_report_autoconf
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: cts: Repair the local-init option now that report_common is generated
Merge pull request #236 from gao-yan/route-ais-message
Log: crmd: Improved logging of cib API call failures
Feature: loggin: Bypass over-dumping logic for CRIT messages
Log: crmd: Try and trace the source of a disconnection error
Test: abi-check: Upload to the correct location
Merge branch 'master' of github.com:beekhof/pacemaker
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Log: crmd: Improved logging of pacemaker level membership changes
Test: fencing: Have agents log consistently with the fencing python library
Log: fencing: Include human readable text corresponding to the numeric error code
Refactor: crmd: Avoid g_hash_table_foreach() for logging tasks
Build: Tweak the description for the remote subpackage
Merge pull request #252 from inouekazu/add_logpriority_option
Fix: crmd: Indicate that we expect lrmd connections to being closing
Test: cts: Patterns to ignore in ComponentFail tests
Log: crmd: Additional trace logging
Fix: crmd: Bug cl#5144 - Do not updated the expected status of failed nodes
Build: Separate Feature from Fixes in the ChangeLog
Log: Prepare the 1.1.9 changelog
Build: bumplibs.sh: Ignore whitespace changes
Build: Bump library versions as appropriate
Build: bumplibs.sh: Keep library version defines up to date
Build: Move pacemaker_remoted to sbin and add a SYS-V init script
Fix: fencing: Don't delay execution if all topology queries have arrived
Fix: corosync: Reduce excessive delays when resending CPG messages
Fix: mcp: Incorrect processname in SYS-V initscript metadata
Test: cluster-clean: Kill pacemaker_remote processes too
Build: lrmd: Add man page for pacemaker_remoted
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Doc: mcp: Update the text for PCMK_logpriority
Merge pull request #253 from yuusuke/fix_facility
Fix: fencing: Only initiate topology based fencing if all topology queries have arrived AND we are still in the query phase
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Update changelogs and version strings
Merge pull request #256 from yuusuke/fix_memory_leak
Merge pull request #254 from gao-yan/doc-nagios
Merge pull request #255 from yuusuke/fix_pipe_close
Merge pull request #259 from gao-yan/nagios-schema
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Refactor: Fencing: Re-use the policy engine libraries for filtering devices
Log: mcp: Include the feature set in the output for --features
Fix: lrmd: Handle --verbose consistently with the rest of pacemaker
Fix: logging: Better checks when determining if file based logging will work
Fix: ipc: Restore the ability for members of the haclient group to connect to the cluster
Build: Show help text before building man pages
Merge pull request #261 from yuusuke/fix_memory_leak
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Refactor: Build: Update the libqb version checks
Test: Update the valgrind suppressions file
Merge pull request #264 from gao-yan/crm-node-remove
Refactor: Remove all use of g_child_watch_add() due to its dependance on threads
Refactor: lrmd: Clean up the LSB metadata template
Fix: crmd: Suppress secondary errors when no metadata is found
Log: Fix output format for logged xml
Test: Kill pacemaker processes running under valgrind too
Fix: Pass errors from lsb metadata generation back to the caller
Test: lrmd: Update search pattern
Log: pe: Reduce info logging
Fix: fencing: Correctly filter devices when no nodes are configured yet
Fix: xml: Prevent use-after-free when not processing all xpath query results
Refactor: Prevent use-of-NULL when processing xpath results
Fix: crmd: Prevent use-after-free when the blackbox is enabled
Fix: fencing: Fail the operation once all peers have been exhausted
Fix: Skip WNOHANG when waiting after sending SIGKILL to child processes
Log: PE: Turn off some errors when doing a location check for stonith-ng
Fix: crmd: Clean up memory at exit
Fix: fencing: Clean up memory at exit
Test: Ignore another non-recurring libqb memleak
Merge pull request #267 from davidvossel/require-all-docs
Fix: xml: Prevent use-after-free in cib_process_xpath()
Fix: lrmd: Avoid memory leak in resources_action_create()
Fix: fencing: Avoid memory leak in can_fence_host_with_device()
Refactor: fencing: Some additional debugging for parse_host_line()
Log: fencing: Additional trace logging
Fix: fencing: Correctly unpack device parameters before using them
Fix: fencing: Prevent invalid read in parse_host_list()
Fix: fencing: Prevent memory leak when registering devices
Log: cib: Indicate why the blackbox is being dumped by cib_process_request()
Fix: fencing: Fix memory leaks during query phase
Test: Fix 'cts local-init'
Test: Clean up after valgrind runs
Refactor: fencing: More appropriate error codes
Log: cib: Additional tracing
Fix: xml: Prevent lockups by setting a more reliable buffer allocation strategy
Refactor: cib: Create replies with existing values
Refactor: Switch to malloc() for invocations that do not require initialized memory
Fix: PE: Do not use functions from the cib library during unpack
Log: Additional trace logging of digest calculations
Fix: Prevent use-of-NULL when reading CIB_shadow from the environment
Test: cts: Additional patterns to ignore when corosync is killed
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge remote-tracking branch 'origin/pr/263'
Log: xml: Correctly indent xml elements
Fix: Ensure custom error codes are less than 256
Fix: cib: Use unique error codes for write_cib_contents()
Test: tools: Update regression tests for new error codes
Merge pull request #268 from gao-yan/node-cache
Test: tools: Update regression tests for new error codes
Fix: Allow programs in the haclient group to use CRM_CORE_DIR
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #270 from realloc/master
Feature: mcp: Prevent the cluster from trying to use cman even when it is installed
Log: crmd: Record trace data when election loops are detected
Log: pengine: Drop message severity
Log: crmd: More accurate log for A_RECOVER
Log: mcp: Update help text
Fix: crmd: Only performa dry run for deletions if built with ACL support
Fix: ipc: Allow unpriviliged clients to clean up after server failures
Feature: crm_resource: Allow individual resources to be reprobed
Fix: crmd: Indicate completion of re-probe to callers
Fix: crmd: Indicate completion of refresh to callers
Refactor: crm_resource: Make --refresh and --reprobe an alias for --cleanup [--resource name]
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #275 from seinocluster2/fix_fd_leak
Merge pull request #276 from inouekazu/add_reset_st_fail_count
Build: mcp: Fill in configure variables for Upstart job
Merge pull request #277 from gao-yan/cibadmin-query-quiet
Fix: cman: Do not unconditionally start cman if it is already running
Merge pull request #274 from gao-yan/old-failure
Test: tools: Ignore whitespace changes in regression tests
Fix: fencing: Prevent empty call-id during notification processing
Fix: crmd: Ensure we return to a stable state if there have been too many fencing failures
Doc: crmd: Additional comments describing the various timers used
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: ipc: Repair compilation when ACLs are turned on
Merge pull request #282 from t-matsuo/quoting-variables
Fix: PE: Implement the rest of get_timet_now() and rename to get_effective_time
Test: PE: Tell valgrind to ignore glib type initialization
Fix: crm_resource: Allow --cleanup without a resource name
Fix: cman: Skip cman_pre_stop in the init script if fenced is not running
Doc: Fix the product so Pacemaker_Remote gets uploaded to the correct location
Build: pacemaker-cts subpackage depends on pacemaker-libs
Fix: crm_report: Correctly collect logs when 'uname -n' reports fully qualified names
Fix: crm_report: Suppress irrelevant error message
Test: Make our dummy LSB script more realistic
Doc: Pacemaker_Explained: Document the stop-all-resources option
Fix: init: Unless specified otherwise, assume cman is in use if cluster.conf exists
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #284 from t-matsuo/improve-ping-ra
Fix: crm_mon: Ensure stale pid files are updated when a new process is started
Fix: crm_mon: Check if a process can be daemonized before forking so the parent can report an error
Build: spec: Correctly call chkconfig for pacemaker-remote
Fix: Core: Ensure the last field in transition keys is 36 characters
Fix: ipc: Reference the correct type when inspecting IPC messages
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Refactor: crmd: Simplify the logic for resetting node status in the CIB after fencing events
Bug: cf#5153 - Correctly display clone failcounts in crm_mon
Bug: cl#5154 - Do not expire failures when on-fail=block is present
Test: cl#5154 - Do not expire failures when on-fail=block is present
Fix: PE: Ensure per-node resource parameters are used during probes
Test: PE: Ensure per-node resource parameters are used during probes
Spec: Always stop pacemaker before uninstalling/upgrading
Test: PE: Ensure per-node resource parameters are used during probes - update old tests
Build: Helper for installing dependancies with yum
Test: cts: Add an option for copying /etc/hosts during setup
Doc: Update the openstack quickstart
Fix: Detect child processes that terminate before our SIGCHLD handler is installed
Fix: Check for and replace non-printing characters with their octal equivalent while exporting xml text
Fix: xml: Preserve newlines, tabs and line feeds in some form
Build: Update top-level helpers
Build: Update library versions for 1.1.10
Feature: cluster-init: Support inaddr_any for udpu clusters
Test: PE: Supress additional glib memory leaks
Doc: Updates to the openstack quickstart
Merge pull request #288 from gao-yan/pending-delete
Merge pull request #287 from satoyoshi/stonith_action_log
Refactor: Use GLIB_CHECK_VERSION when deciding what compatability code to include
Fix: xml: Restore the ability to embed comments in the cib
Fix: xml: Log comments as well
Build: Display suggested configure arguments for rpm based systems
Refactor: Common defines for resource operation timing fields
Fix: crmd: Store last-run and last-rc-change for fail operations
Test: crm_simulate: Ensure fake actions have values for last-run and last-rc-change
Fix: crmd: Ensure pending and lost actions have values for last-run and last-rc-change
Fix: PE: Correctly handle resources that recover before we operate on them
Test: PE: Correctly handle resources that recover before we operate on them
Log: Indicate which container resource a node belongs to
Refactor: Use a common valgrind suppressions file
Feature: PE: Display a list of nodes on which stopped anonymous clones are not active instead of meaningless clone IDs
Feature: PE: Suppress meaningless IDs when displaying anonymous clone status
Merge pull request #290 from t-matsuo/exclude-rc
Fix: PE: Log the correct number of stopped anonymous clones
Bug cl#5133 - PE: Correctly observe on-fail=block for failed demote operation
Test: PE: Fix operation digests
Log: crm_simulate: Indicate which resource is having its failcount reset
Log: PE: Add resource tags to some log points
Refactor: TE: Use #define instead of raw strings
Fix: PE: Re-initiate _active_ recurring monitors that previously failed but have timed out
Test: PE: Remove noise from the failcount testcase
Test: PE: Additional testcase for expired failcounts
Refactor: Remove unused function
Refactor: consolidate code for writing xml to disk
Fix: Conditionally perform cib-specific modifications when writing xml to disk
Test: PE: Ignore blank lines in regression test output
Log: crmd: Attempt to display the node on which changes occurred
Fix: lrmd: Don't complain when heartbeat invokes us with -r
Merge pull request #292 from inouekazu/always_block
Create a cheat sheet for users switching between pcs and crmsh
Doc: Cheatsheet whitespace
Fix: PE: Resolve memory when displaying anonymous clone status
Test: More valgrind suppresions for glib
Bug cl#5151 - Ensure node names are consistently compared without case
Fix: crmd: Insert async failures so that they appear in the correct order
Fix: Fencing: CID#1023874 - Dereference before null check
Fix: Core: CID#1023875 - Dereference before null check
Fix: Core: CID#1023870 - Resource leak
Fix: Core: CID#1023868 - Dereference null return value
Build: Find all bugs for changelog
Fix: Core: CID#1023867 - Dereference null return value
Fix: Core: CID#1023866 - Dereference null return value
Fix: cib: CID#1023865 - Dereference null return value
Fix: Fencing: CID#1023864 - Dereference null return value
Fix: cib: CID#1023863 - Dereference null return value
Fix: Core: CID#1023861 - Dereference after null check
Fix: Core: CID#1023859 - Dereference after null check
Fix: cib: CID#1023858 - Explicit null dereferenced
Fix: cib: CID#1023857 - Dereference after null check
Fix: crm_simulate: CID#1023856 - Logically dead code
Fix: crm_simulate: CID#1023854 - Logically dead code
Fix: Core: CID#1023853 - Unchecked return value
Fix: crmd: CID#1023852 - Unchecked return value
Fix: crmd: CID#1023850 - Side effect in assertion
Fix: crm_resource: CID#739442 - Dereference null return value
Fix: Core: CID#1023861 - Dereference after null check
Fix: cib: CID#739562 - Improper use of negative value
Fix: Core: CID#1025092 - Dereference before null check
Fix: cib: CID#1023862 - Improper use of negative value
Fix: lrmd: CID#1023855 - Logically dead code
Fix: lrmd: CID#1023851 - Truncated stdio return value
Fix: cib: CID#739562 - Improper use of negative value
Fix: cib: CID#1025093 - Resource leak
Doc: Pacemaker Explained: Document missing cluster options
Fix: lrmd: CID#1023872 - Resource leak
Fix: lrmd: CID#1023871 - Resource leak
Fix: lrmd: CID#1023869 - Resource leak
Fix: Fencing: CID#739461 - Double close
Fix: Core: CID#739448 - Dereference before null check
Fix: Fencing: CID#739453 - String not null terminated
Fix: cib: CID#739437 Dereference null return value
Fix: crmd: CID#739438 - Dereference null return value
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: Fencing: Restore the ability to manually confirm that fencing completed
Fix: cman: Our daemons have no need to connect to pacemakerd in a cman based cluster
Fix: cib: Compress backups when possible
Log: cluster: additional debug
Test: cts: Additional patterns to look for when corosync is killed
Test: cts: Find a way to display the output of asynchronous commands
Fix: crm_simulate: Memory leak
Test: cts: Add a hook to bring up the cluster and then leave it
Fix: Fencing: Do not send duplicate replies for manual confirmation operations
Fix: Fencing: Correctly mark manual confirmations as complete
Fix: Fencing: Correctly broadcast manual fencing ACKs
Log: Fencing: Indicate who initiated fencing actions, not just the node name
Test: cts: Simplify the cman start command
Test: cts: Wait for in-progress fencing operations to complete before continuing
Fix: cib: Archived versions are never compressed
Log: crmd: Indicate why feature_update_callback() raised an error
Fix: ping: Ensure OCF_RESKEY_failure_score has a value
Build: Update library versions for 1.1.10-rc3
Build: Use the spec file that came with the specified version
Build: Reduce duplication
Build: Make a generic version of yumdep
Merge pull request #295 from t-matsuo/without_doc
Merge pull request #296 from inouekazu/fix_typo
Fix: crm_resource: Do not crash when calling --force-start on an invalid resource
Merge pull request #294 from gao-yan/unmanaged-stop
Fix: crmd: Do not get stuck in S_POLICY_ENGINE if a node we couldn't fence returns
Test: cman: Ignore additional valgrind errors
Fix: crmd: Clean up more memory before exit
Fix: Core: Correctly unreference GSource inputs
Fix: Clean up internal mainloop structures at exit
Log: init: Remove unnecessary detail from shutdown message
Refactor: crmd: Allocate a mainloop before doing anything to help valgrind
Fix: Free additional memory at exit
Fix: Fencing: Ensure API object is consistently free'd
Test: More glib suppressions
Test: cts: Fix the 'boot' scenario
Refactor: ais: Remove unused code
Fix: crmd: Clean up more memory before exit
Merge pull request #297 from gao-yan/stop-blocked
Fix: crmd: Ensure timers are stopped at exit
Log: cluster: Additional tracing at shutdown
Test: Ingore memory related to option parsing
Fix: Ensure crm_system_name is cleaned up at exit for valgrind
Fix: Ensure anything we free at shutdown is set to NULL too
Fix: crmd: Use standard error codes on exit
Fix: heartbeat: Failed crmd child processes should fail fast with crm_exit()
Refactor: crmd: Allow mainloop to terminate naturally on graceful shutdown
Fix: Core: Tolerate NULL in mainloop_destroy_trigger()
Fix: Correctly calculate crm_system_name
Refactor: mainloop: Saner ordering of reference counting events in mainloop_destroy_trigger()
Fix: crmd: Prevent use-after-free now that mainloop terminates naturally
Test: Remove unnecessary valgrind supressions now that mainloop terminates naturally
Merge pull request #298 from dubrsl/dubrsl
Test: Add additional valgrind supressions
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Build: Update library versions
Feature: cib: Allow multiple children to be removed from the object root (eg. constraints)
Feature: crm_resource: Implement --ban for moving resources away from nodes and --clear (replaces --unmove)
Test: tools: Additional tests for --ban, --move and --clear
Refactor: PE: Use common defines instead of raw strings for field names
Refactor: crm_resource: Make --move without --host a true alias for --ban without --host
Feature: crm_resource: Implement the --master option for the --move,--ban,--clear commands
Fix: crm_attribute: Return the actual error code rather than the negative of it
Fix: crm_attribute: Send details on duplicate values to stdout
Test: tools: Overhaul regression tests
Fix: Convert all exit codes to positive errno values
Fix: attrd: Correctly detect successful invocation of attrd_update()
Fix: Repair handling of attrd_update_delegate() return codes
Log: crmd: Display error details in crmd_exit()
Test: cts: Update error patterns
Test: cts: Correctly set peer status when startup fencing is detected
Fix: crmd: Ensure all membership operations can complete while trying to cancel a transition
Fix: crmd: Prevent use-of-NULL during emergency shutdown
Feature: cman: Support non-blocking CPG calls as we do for corosync-2
Fix: Core: Ensure the blackbox is saved on abnormal program termination
Fix: crmd: crmd_exit() is re-entrant, prevent it from draining mainloop multiple times
Fix: crmd: Prevent the IPC server from being free'd multiple times
Refactor: mainloop: Make mainloop_gio_destroy() tollerant of being called re-entrantly
Refactor: crmd: Further improve clean shutdowns
Log: crmd: Bump a non-recurring log message needed by CTS
Test: cts: Suppress blackbox messages from ComponentFail tests
Merge pull request #304 from t-matsuo/improve-regression
Test: cts: Display the process list when stale files are detected
Test: cts: Look for completion of the entire fencing operation before continuing
Feature: crm_error: Add the ability to list and print error symbols
Log: cluster: Additional detail when flushing CPG messages
Log: crmd: Additional detail when quorum updates fail
Refactor: crmd: Do the bare minimum when a graceful shutdown is not required
Log: mcp: Log error text when a child exits
Refactor: crmd: Use hashtable iterators instead of loop functions
Fix: corosync: Detect the loss of members for which we only know the nodeid
Fix: crmd: Update the status section with details of nodes for which we only know the nodeid
Test: cts: Update patterns
Fix: Correctly fail in attrd_update_delegate() when no connection to attrd is available
Fix: Allow uuid retrieval for nodes with no known name
Refactor: Remove unnecssary uuid and uname caches
Refactor: Cluster: Remove the pcmk_uname and pcmk_nodeid global variables
Test: cts: Update attrd exit pattern
Fix: Ensure local uuid is defined
Feature: crm_resource: Support OCF tracing when using --force-(check|start|stop)
Fix: crm_resource: Correctly implement --master
Fix: crm_resource: Gracefully fail when --force-* is attempted for stonith resources
Test: fencing: 'quiet' is no longer supported, remove it from fence_(true|false)
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: corosync: Do not pretend we know the state of nodes we've never seen
Test: cts: Fix patterns - the ()'s were not interpretted literally
Merge pull request #285 from inouekazu/update_spec
Fix: crmd: Initiate node shutdown if another node claims to have successfully fenced us
Fix: crmd: Everyone who gets a fencing notification should mark the node as down
Fix: corosync: Nodes that can persist in sending CPG messages must be alive afterall
Test: crm_simulate: Fix memory leaks reported by valgrind
Merge pull request #306 from yuusuke/fix_stonithd
Fix: corosync: Fall back to uname for local nodes
Fix: crm_resource: Avoid whitespace in transition keys when cleaning up and failing resources
Merge pull request #307 from yuusuke/crm_mon_customize
Fix: crm_simulate: Support systemd and upstart actions
Fix: corosync: Suppress harmless error message for non-crmd processes
Fix: systemd: Ensure we get shut down correctly by systemd
Test: cluster-init: Use the cluster_name variable instead of CTS_GROUP
Log: crm_report: Sort installed packages
Log: systemd: Adjust some messages for systemd resources
Doc: Indicate how to turn on synchronous log file writes
Fix: PE: If fencing is unavailable or disabled, block further recovery for resources that fail to stop
Log: PE: Inhibit error log if there is no quorum
Log: fencing: Tweak trace message
Feature: PE: Allow active nodes in our current membership to be fenced without quorum
Test: PE: Allow active nodes in our current membership to be fenced without quorum
Log: lrmd: Differentiate between stderr and stdout from agents
Log: lrmd: Indicate when no output was supplied
Test: Ignore bash leaks in cli regression tests
Test: lrmd: Drop trace logging from the verbose output
Test: lrmd: Attempt to track down the source of transient errors
Test: tools: Have valgrind generate suppressions for any leak found
Merge pull request #309 from davidvossel/fencing_retries
Merge pull request #308 from gao-yan/xml-diff
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: lrmd: Default to the upstream location for resource agent scratch directory
Test: cts: Combine fence_true and fence_false and remove dependancy on fencing.py
Build: Fix dodgy date in changelog
Log: PE: Unclean nodes are logged as WARN not ERROR
Fix: membership: Correctly merge partial entries in the peer cache
Test: Prevent systemd from respawning processes we're trying to kill
Test: Fix the bash leak suppression
Fix: crm_report: Find logs in compressed files
Test: BasicSanity: Do lrmd tests last
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Feature: Turn off auto-respawning of systemd services when the cluster starts them
Fix: Use the official fedora user/group ID for haclient/hacluster
Test: cts: Fix fence_dummy and its usage
Refactor: corosync: Merge three sets of CPG related code
Fix: membership: Simplify crm_peer_init()
Fix: logging: If SIGTRAP is sent before tracing is turned on, turn it on
Fix: cman: Give cman time to start before starting the rest of pacemaker
Log: mcp: Improved startup logging
Build: clean up after try_extract_header_define()
Test: Prevent clock drift when setting up a cluster
Test: cts: Update log patterns
Log: crmd: Downgrade an error message since we can recover
Fix: systemd: Reload systemd after adding/removing override files for cluster services
Refactor: Clean up cluster header
Build: Turn off fortify source when building with profiling support
Build: Suggest good defaults for configure on rpm based systems
Build: Merge systemd detection logic
Build: streamline the spec file
Test: Tell rpmlint to ignore TZ differences
Build: SUSE changed their package groups by the looks of it
Build: Force rpath off to comply with rpmlint on suse
Fix: crmd: Ensure operations for cleaned up resources don't block recovery
Build: Fix rpm macros on suse
Doc: Fix revision numbers in Pacemaker_Remote
Build: Massage helper targets
Build: Further simplify the spec file
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Build: Fix coverage builds
Build: Fix libqb configure check
Fix: Tools: CID#1036760 Truncated stdio return value
Fix: Tools: CID#1036764 Unused pointer value, CID#1036762 Resource leak
Fix: Corosync: CIB#1036763 Dereference before null check
Fix: crmd: CID#1036761 Dereference null return value
Doc: Note to self in spec file
Build: Re-instate support for --without=cman in spec file
Merge pull request #286 from oallart/master
Merge pull request #300 from gao-yan/delete-old-state
Merge branch 'node-maintenance' of https://github.com/gao-yan/pacemaker
Merge pull request #314 from gao-yan/plugin-header
Merge pull request #315 from oallart/master
Refactor: crmd: Remove dead code
Fix: cluster: Correctly remove duplicate peer entries
Merge pull request #318 from l-mb/master
Merge pull request #316 from gao-yan/plugin-dispatch
Fix: Fencing: User-defined arguments always trump calculated ones
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Refactor: plugin: Rename some functions for consistency
Fix: PE: Do the right thing when admins specify the internal resource instead of the clone
Fix: PE: Detect constraints with inappropriate actions (ie. promote for a clone)
Fix: Bug cl#5157 - Allow migration in the absence of some colocation constraints
Log: PE: Fix migration debug logs
Refactor: PE: Remove dead code
Fix: Bug cl#5157 - Allow migration in the absence of some colocation constraints - prevent graph loops
Test: Bug cl#5157 - Allow migration in the absence of some colocation constraints
Test: cts: New BadNews patterns
Merge pull request #319 from l-mb/master
Log: Fix typo
Fix: Fencing: Force use of the calculated command for support ops
Fix: Fencing: Avoid hashtable lookups in NULL
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #320 from yuusuke/fix_memory_leak
Merge pull request #322 from l-mb/master
Test: PE: Fix regression tests after lm-b's output change
Refactor: crmd: All resource events need to include the host name that performed it
Doc: crmd: Update comment
Fix: Cluster: Future flexibility
Doc: Additional comments
Fix: Bug cl#5168 - Prevent clones from being bounced around the cluster due to location constraints
Test: Bug cl#5168 - Prevent clones from being bounced around the cluster due to location constraints
Fix: lrmd: Prevent use-of-NULL in client library
Doc: Document the correct default for on-fail
Doc: mcp: Provide details for having 'service pacemaker stop' take corosync down too
Build: mcp: Add missing noinst header
Fix: cluster: Detect and warn about node names with capitals
Refactor: mcp: Re-use existing logic for tracking corosync-2 peers
Fix: corosync: Ensure removed peers are erased from all caches
Test: cts: Have 'cts start' bring nodes online in parallel
Test: mainloop: Validate glib is using the correct file descriptor
Merge pull request #326 from gao-yan/remote-client-message
Fix: tools: Fix compilation error
Fix: crm_node: Return 0 if --remove passed
Fix: fencing: Correctly detect existing device entries when registering a new one
Merge pull request #328 from yuusuke/fix_memory_leak
Fix: PE: Do the right thing when admins specify the internal resource instead of the clone
Log: Drop the severity of logs for recoverable conditions
Doc: Drop references to 2.x in todo list
Test: cts: Normal start for the first node so that the config is generated
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Bug: cl#5170 - Correctly support on-fail=block for clones
Test: cl#5170 - Correctly support on-fail=block for clones
Doc: Typos and clarifications in Clusters from scrtach
Log: crm_mon: Suppress expected votes for non-plugin clusters
Fix: PE: Re-allow ordering constraints with fencing devices now that it is safe to do so
Fix: cluster: Correctly construct the header for compressed messages
Doc: Update todo list
Fix: crm_resource: Present an error if an admin tries to --force-start an already active resource
Fix: Core: Suppress '-' arguments when creating the getopt string
Fix: cpg: Correctly free sent messages
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: cib: The result is not valid when diffs fail to apply cleanly for CLI tools
Refactor: Fix spelling of resource
Log: fencing: Indicate correct unit for timeouts
Build: cman: Only build migration tools for targets that may use them
Test: cib: Ensure we set up hacluster's groups in stand-alone mode
Log: crmd: Additional detail when destroying operation caches
Bug cl#5161 - crmd: Prevent memory leak in operation cache
Refactor: crmd: Clean up handling of pending operations at shutdown and improve logging
Log: fencing: Increase verbosity of message symptomatic of common configuration mistake
Build: spec: Incorporate rpmlint feedback
Doc: Update todo list
Test: cts: Update patterns for corosync failure test
Fix: cib: Correctly read back archived configurations if the primary is corrupted
Fix: cman: Do not pretend we know the state of nodes we've never seen
Feature: crm_resource: Allow options to be set recursively
Test: fencing: We want all defined options to be available for use
Test: cts: Some additional patterns to ignore when corosync is killed
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Release: Update changelog and versions for 1.1.10
Fix: cpg: Ensure the CPG group is always null-terminated
Fix: Make coverity happy, check for NULL return from malloc()
Refactor: legacy: Remove dead code
Refactor: Convince coverity there is really no memory leak
Build: Update version, changelog in sample spec file
Merge pull request #332 from inouekazu/add_fail_fast
Build: Fix compilation on OSX
Build: Remove unused variable
Merge pull request #333 from inouekazu/fix_pacemaker_service
Fix: crmd: Prevent recurring monitors being cancelled due to notify operations
Test: cts: Correctly configure the ping resource
Merge branch 'master' of github.com:beekhof/pacemaker
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: PE: Do not restart resources that depend on unmanaged resources
test: PE: Do not restart resources that depend on unmanaged resources
Fix: crmd: Prevent crash by passing log arguments in the correct order
Fix: fencing: Put all fencing agent processes into their own process group
Fix: PE: Do not re-allocate clone instances that are blocked in the Stopped state
Fix: PE: Do not allow colocation with blocked clone instances
Test: PE: Do not allow colocation with blocked clone instances
Fix: systemd: Prevent glib assertion - only call g_error_free() with non-NULL arguments
Fix: logging: glib CRIT messages should not produce core files by default
Fix: Command-line tools should stop after an assertion failure
Fix: Gracefully handle -ECHILD from waitpid() in crm_abort()
Fix: Prevent use-of-NULL when ping resources do not define a host list
Fix: crm_mon: Correctly extract alternate name for pingd attributes
Build: Clean up after coverity runs
Fix: PE: Gracefully handle bad values for XML_ATTR_TRANSITION_MAGIC
Fix: crm_simulate: Do not complain about CRM_OP_REPROBE actions
Fix: crmd: Gracefully handle actions that cannot be initiated
Doc: controld: Update the description
Fix: PE: Correctly account for the location preferences of things colocated with a group
Test: PE: Correctly account for the location preferences of things colocated with a group
Log: crm_mon: Clean up the output for failed operations
Fix: cib: Correctly log short-form xml diffs
Fix: crmd: Correcty update the history cache when recurring ops change their return code
Log: PE: Better indicate when a resource has failed
Test: PE: Better indicate when a resource has failed
Merge pull request #336 from adrianlzt/patch-2
Merge pull request #335 from adrianlzt/patch-1
Refactor: Reduce duplication by merging the lrmd_exec_rc enum with ocf_exitcode
Fix: lrmd: Do not overwrite any existing operation status error
Feature: services: Detect missing agents and permission errors before forking
Fix: services: Prevent use-of-NULL in g_error_free
Test: lrmd: Drop the default verbosity for lrmd regression tests
Log: Correctly log 'crit' messages from glib
Fix: services: Preserve errno when handling exec failres least it be reset as a side-effect of other calls
Fix: services: Correctly invoke callbacks for aborted operations
Merge pull request #338 from boda2004/master
Bug cl#5171 - PE: Don't prevent clones from running due to dependant resources
Test cl#5171 - PE: Don't prevent clones from running due to dependant resources
Merge pull request #339 from davidvossel/pcmk_remote_cts
Test: PE: Update regression test scores
Refactor: PE: Break up unpack_rsc_op() into something that doesn't make your eyes bleed
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: cpg: Correctly set the group name length
Feature: Fencing: Support agents that need the host to be unfenced at startup
Fix: crm_report: Collect corosync quorum data
Merge branch 'master' of github.com:beekhof/pacemaker
Fix: services: Use the supplied error code instead of errno which may have changed
Test: cts: Template for using fence_scsi
Test: cts: Ship the cts helper instead of the older cluster_test
Test: cluster-init: Detect when long names might be needed to reach a host
Test: cts: Add a helper for interacting with cluster-init
Fix: mcp: Remove LSB hints that instruct chkconfig to start pacemaker at boot time
Refactor: logging: Use the built-in abs() function
Fix: Additional paranoia courtesy of a Veracode review
Doc: lrmd: Indicate which function recieves the proxied command
Refactor: mainloop: Common wrapper for timer functionality
Merge pull request #342 from t-matsuo/improve-crm_mon
Merge pull request #345 from t-matsuo/hide_headers
Refactor: cluster: Create a reusable module from the crmd election code
Log: cib: Additional startup logging
Fix: xml: Location constraints are allowed to specify a role
Bug rhbz#902407 - crm_resource: Handle --ban for master/slave resources as advertised
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Build: Additional CFLAGS when building for profiling
Refactor: ipc: Common function for disconnecting ipc clients
Internal: election: Track the election state and allow it to be queried
Refactor: attrd: Move to its own directory and create a stub for attrd-ng
Feature: attrd: Untested reimplementation of a truely atomic attrd
Internal: attrd: Elect one peer to do all cib writes and sync attributes to new peers
Internal: attrd: Hook up the cib for writing
Fix: attrd: Ensure there is always someone to write out attributes
Feature: attrd: Use the truely atomic version where CPG is used for cluster communication
Fix: attrd: Any variable called 'cluster' makes the daemon crash before reaching main()
Fix: attrd: Fix memory handling when supplying defaults for id and set
Fix: attrd: Ensure we can vote in elections and recognize when we've won
Fix: attrd: Remember new attribute values so we can write them out
Fix: attrd: Prevent use-of-NULL
Fix: attrd: Timeouts need to be converted to milliseconds for g_timeout_add()
Log: attrd: Improve some logging formats
Feature: cib: Allow values to be added/updated and removed in a single update
Fix: attrd: Supply the correct number of arguments to log message
Fix: attrd: Write all attributes out after an election
Fix: attrd: Drop all attributes for peers that left the cluster
Fix: attrd: Local values take priority during synchronization
Fix: attrd: Ensure peer state is kept up-to-date
Fix: attrd: Write out attributes with node UUIDs
Fix: attrd: Support SI units for attribute dampening
Log: attrd: Refine logging
Fix: attrd: Indicate if we shut down due to an error
Feature: crm_node: Display partition output in sorted order
Test: cts: logs dissappear if systemd-journald is not restarted too
Fix: attrd: Do not pass NULL keys to g_hash_table_remove()
Fix: crm_mon: Do not display duplicates of failed actions
Feature: crm_report: Record the output of the collector
Test: cts: Fix log message
Merge pull request #349 from gao-yan/xml-diff-comment
Log: crm_report: Improve help text
Fix: crm_report: Do not print garbage when collecting from the local node
Test: cts: More reliably obtain a sane value for IPBase
Fix: crmd: Reliably detect and act on reprobe operations from the policy engine
Fix: attrd: Only restart the timer if an attribute changed this time
Log: attrd: Additional trace messages
Log: cib: Reduce verbosity of 'cib_mixed_update' results
Fix: cib: Correctly check for archived configuration files
Fix: attrd: Reliably determine the 'id' to use when writing out attributes
Fix: attrd: Only write out attributes if we are elected
Feature: crm_report: Collect logs directly from journald if available
Test: cts: Add the ability to watch for logs with journalctl
Test: cluster-clean: Erase systemd journals too
Test: cts: Fix the DiskAudit
Fix: attrd: Ensure attributes get written out if we loose a peer between attrd_client_message() and attrd_peer_update()
Log: crmd: Improved logging
Fix: attrd: Defer calculation of set key names so that the correct host prefix is used
Fix: attrd: Prevent inflation of attribute dampen intervals
Test: cluster-init: Simplify the comamnd for synchronising dates
Test: cts: Update patterns for the new attrd
Log: cib: Additional logging for invalid diff updates
Fix: election: Do not update state when ACKs arrive
Log: attrd: Improved trace logging
Fix: fencing: Skip empty notifications from the cib
Fix: attrd: Infer when some updates will timeout
Fix: crm_shadow: Simplify prompt generation and impose a sane limit on length
Refactor: cib: Simplify set_standby
Fix: cib: Guard against overly long user specified strings
Refactor: cib: Avoid holding open the file we are renaming
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: decode_transition_key: Older systems do not support '%ms' in sscanf
Fix: attrd: Use the legacy version if no stack is supported
Refactor: cib: Drop logic for compatability with releases prior to 1.0
Merge pull request #353 from gao-yan/expected-quorum-votes
Test: cluster-init: Keep the port number to 4 digits for large clusters
Test: cts: Some defaults that make my life easier
Feature: ipc: Raise the default buffer size to 128k
Fix: crm_resource: Provide a meaningful error if --master is used for primitives and groups
Fix: services: Prevent use-of-NULL when executing service actions
Test: cluster-helper: New command for installing ssh keys
Fix: Fencing: Observe pcmk_host_list during automatic unfencing
Merge pull request #356 from l-mb/master
Log: crmd: Supply arguments in the correct order
Fix: cman: Do not start pacemaker if cman startup fails
Fix: cibadmin: Report errors during sign-off
Merge pull request #357 from ldzhong/master
Merge pull request #359 from l-mb/master
Fix: Bug rhbz#1011618 - Consistently use 'Slave' as the role for unpromoted master/slave resources
Fix: crm_report: Correctly redirect error message to /dev/null
Fix: crm_report: Correctly redirect quorumtool error messages
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Feature: PE: Allow location constraints to take a regex pattern to match against resource IDs
Feature: PE: Add a special attribute for distinguishing between real nodes and containers in constraint rules
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Test: lxc_autogen.sh: Improved scalability and repeatability
Fix: PE: Location constraints with role=Started should prevent masters from running at all
Test: PE: Location constraints with role=Started should prevent masters from running at all
Fix: crm_resource: Observe --master modifier for --move
Fix: iso8601: Prevent dates from jumping backwards a day in some timezones
Fix: crmd: Correctly update expected state when the previous DC shuts down
Test: tools: Update regression test output
Merge pull request #362 from gao-yan/services-action-sync
Fix: cman: Start clvmd and friends from the init script if enabled
Log: crm_mon: Correctly show the queued and exec times for an action
Fix: crm_resource: Wait for all replies when cleaning up resources
Log: membership: Indicate which node entry we're destroying
Fix: cli: Do not enabled blackbox for cli tools
Fix: fencing: Ignore changes to non-fencing resources
Test: cts: Split off the xml part of generating a config so that a pcs version can be added
Test: cts: Figure out how to have a help start in the background
Test: cts: Don't overwrite defaults
Test: cts: Cope with patterns from both attrd implementations
Test: cluster-clean: Correctly remove blackbox files
Test: cluster-init: Allow creation of a cluster using only a subset of the available nodes
Fix: mcp: Some distros complain when LSB scripts do not include Default-Start/Stop directives
Log: crmd: No need to relog the contents of every message when it gets routed
Log: xml: Fix typo
Fix: crm_mon: Reduce flickering in interactive mode
Fix: lrmd: Do not interrogate NULL replies from the server
Feature: crmd: Do not block on proxied calls from pacemaker_remoted
Fix: ipc: Suppress asserts for proxied connections
Refactor: ipc: Consolidate client/server flags and add version field to header
Feature: crmd: Track CPU usage on cluster nodes
Feature: remote: Properly version the remote connection protocol
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Log: remote: Improved logging of recv() errors
Disable prototype CPU usage tracking code
Merge pull request #372 from phmarek/master
Merge pull request #369 from gao-yan/services-reset-scheduler
Feature: crmd: Notify peers of our relative load
Fix: transition: spelling mistake
Feature: crmd: Use the load on our peers to know how many jobs to send them
Feature: crmd: Make the utilization limit configurable
Feature: crmd: Slow down recovery on nodes with IO load
Feature: crmd: Tweak the throttling code based on migration-limit and/or LRMD_MAX_CHILDREN
Feature: crmd: Make the per-node action limit directly configurable in the CIB
Fix: crmd: Simplify the logic for calculating the number of available cores
Log: crmd: Adjust throttle log messages
Fix: crmd: Simplify option processing for throttling options
Fix: crmd: Tweak the throttling threshold multipliers
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Revert "Low: mainloop: Don't allow client buf size to be determined by hardcoded value given to mainloop"
Fix: ipc: Use the higher of the configured buffer size or the default
Revert "Fix: crm_resource: Wait for a reply for each request to crmd"
Merge pull request #380 from lge/for-beekhof
Merge pull request #379 from gao-yan/crm_master-node
Fix: crmd: Do not send throttle updates to versions that don't support them
Fix: crmd: Don't start throttling until we're over the target threshold
Refactor: crmd: Fold the enforcement of migration-limit into the throttling code
Feature: crmd: Track the CIB's CPU usage
Refactor: crmd: Tell the DC our job limit instead of the number of cores
Fix: crmd: Honor load thresholds lower than the cib maximum
Test: PE: Ignore changes to crm_feature_set and batch-limit
Feature: crmd: Allow non-CIB based throttling to be disabled
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Refactor: crmd: A more reliable method of looking up migration options
Merge pull request #374 from dirkmueller/aarch64
Fix: remote: Handle endian changes between client and server and improve forward compatibility
Merge pull request #382 from SynetoNet/master
Fix: cpg: Only process one message at a time to allow other priority jobs to be performed
Fix: crm_resource: Wait for the correct number of replies when cleaning up resources
Test: cts: Teach cts about ipv6 addresses
Log: remote: Correctly log connections to IPv6 servers
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #385 from gao-yan/fix-casts
Merge pull request #383 from inouekazu/apply_score
Feature: crmd: Enable cluster-wide throttling when the cib heavily exceeds its target load
Fix: crmd: Prevent over-polling of throttle inputs
Fix: lrmd: Correctly calculate metadata for the 'service' class
Fix: controld: Use the correct variant of dlm_controld for corosync-2 clusters
Fix: attrd: Improve compatibility with older attrd versions
Fix: crmd: Correctly observe the configured batch-limit
Fix: attrd: Do something sane when older non-atomic versions send us messages
Fix: attrd: Memory leak
Fix: crmd: When a peer expectedly shuts down, record the new join and expected states into the cib
Fix: crmd: When the DC gracefully shuts down, record the new expected state into the cib
Fix: PE: Correctly handle demotion of grouped masters that are partially demoted
Test: PE: Correctly handle demotion of grouped masters that are partially demoted
Merge pull request #388 from phmarek/master
Merge pull request #389 from gao-yan/first-throttle-update
Log: PE: Supproess warning when probes return PCMK_OCF_NOT_INSTALLED in asymmetric clusters
Test: PE: Additional regression tests
Bug cl#5179 - Corosync: Attempt to retrieve a peer's node name if it is not already known
Fix: fencing: Wait until all possible replies are recieved before continuing with unverified devices
Fix: crmd: Improve throttling response on clusters of single core machines
Log: Node IDs are unsigned
Fix: election: Correctly hook up the timeout callback
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Test: buildbot needs psmisc for lrmd regression tests
Bug cl#5181 - corosync: Ensure node IDs are written to the CIB as unsigned integers
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Doc: Pacemaker_Explained: Fix formatting
Log: attrd: Quote attribute names
Fix: crmd: Avoid deleting the 'shutdown' attribute
Merge pull request #392 from yuusuke/fix_cts
Fix: crm_attribute: Do not swallow hostname lookup failures
xml: Track changes to xml trees as they happen
xml: Create a more robust diff format
Merge pull request #393 from inouekazu/master
xml: allow either patch format to be created and applied automatically
Fix: tools: Ensure cibadmin initializes xml functionality
Fix: Prevent use-after-free after calculating changes
Fix: xml: Log changes and create patches without side-effects
Fix: xml: Correctly construct create operations for v2 patchsets
Fix: xml: Indicate when the new diff format fails to apply
Fix: xml: Detect changes in global replace operations
Fix: xml: Correctly handle nodes that are created and removed in a single transaction
xml: Use the new patch application code path
Fix: xml: Preserve the attribute order in the patch for subsequent digest validation
Fix: xml: Return valid error codes from xml_apply_patchset_v1
Fix: xml: Ensure patches always have version details set
Fix: xml: Detect content that is added and removed in the same update
xml: Use an optimized xpath search function for applying patch sets
Merge pull request #398 from krig/bnc854060
Log: cib: Log the v2 patch format
Feature: cib: Send all r/w operations via the cluster connection and have all nodes process them
Fix: xml: Prevent use-of-NULL when applying v1-style patch sets
Fix: xml: Correctly detect subtree additions in v1-style patch sets
xml: Allow the source of xml library errors to be traced
Fix: xml: Remove optimization that could incorrectly _not_ mark parts of the tree as dirty
Log: cib: More trace information
Fix: cib: Simplify operation reply/notification handling
Fix: xml: Prevent use-after-free of ID values
Fix: xml: Only set epoch fields if there was a non-status change
Fix: xml: Handle NULL arguments when repairing diffs
Fix: cib: Some operations need v2 exceptions
Fix: xml: Correctly infer attribute changes after a replace operation
Log: cib: Additional v2-style patchset tracing
Fix: xml: Correctly flag created subtrees
Fix: xml: Finish implementing the new diff format so that attribute order is preserved
Fix: cib: v2 patchset test code cannot have side-effects
Fix: xml: Prevent use-after-free when applying v2-style patchsets
Fix: xml: Prevent use-of-NULL when inferring differences between two xml trees
Fix: xml: Preserve child ordering when applying v2 patch sets
Log: xml: Detect child ordering changes in xml_calculate_changes()
Test: cib: Calculate the v1 diff before testing v2 functionality (so the counters are updated)
Fix: xml: Empty patchsets are considered to have applied cleanly
Fix: xml: Introduce a 'move' primitive for v2 patch sets
Fix: cib: Auto-detect which patchset format to use
Fix: xml: Detect attribute ordering changes
Fix: xml: Mark children of created nodes as also created
Refactor: xml: New function for calculating patchset version details
Fix: crmd: Add support for parsing resource operations in the v2 patch format
Feature: cib: Enable the more efficient xml patchset format
Feature: cib: Implement zero-copy status update
Refactor: cib: Make patchset validation a run-time switch
Fix: crmd: Correctly handle create directives in v2 patchsets
Fix: crmd: Handle patchset directives that change the configuration top-level
Fix: crmd: Correctly handle modify patchset directives
Fix: crmd: Correctly search xpath patterns in v2 patchsets
Refactor: crmd: Reduce code duplication
Fix: crmd: Support failcount processing with the new patch format
Log: xml: Additional debug
Log: cib: Additional comments
Fix: xml: Correctly compare the current feature set when determining which patch format to use
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: crm_report: Suppress logging errors after the target directory has been compressed
Log: cib: Additional change logging
Log: crmd: Additional trace logging
Fix: crmd: Correctly find resource operations in 'lrm' and 'lrm_resources' update blocks
Fix: xml: Prevent use-after-free, version may be free'd as part of applying the v1 patchset
Fix: xml: Prevent use-after-free when calculating v1 patchsets
Merge pull request #406 from davidvossel/ipc_cleanup_fixes
Merge pull request #405 from yuusuke/fix_pacemakerd_logging
Merge pull request #397 from xrobau/patch-2
Merge pull request #407 from inouekazu/default_node-action-limit
Merge pull request #403 from gao-yan/crm_ticket-atomic-op
Merge pull request #399 from gao-yan/location-sets-templates
Merge pull request #408 from davidvossel/fix_action_cancel_crash
Fix: upstart: Allow comilation with glib versions older than 2.28
Fix: Make coverity happy
Fix: crmd: Prevent memory leak in error paths
Fix: Prevent potential use-of-NULL
Fix: stonith_admin: Ensure pointers passed to sscanf() are properly initialized
Doc: Clusters From Scratch: Fix image sizes in PDF output
Log: PE: Downgrade message showing detected masters
Revert "High: cman: Prevent SIGPIPE by not having the crmd write to the fence_override pipe after fencing completes"
Fix: crmd: Prevent SIGPIPE when notifying CMAN about fencing operations
Fix: crmd: Repair memory leak re-introduced when old patch was reverted
Merge pull request #410 from davidvossel/stonith_timeout_fix
Fix: upstart: Allow comilation with glib versions older than 2.28
Fix: Prevent potential use-of-NULL
Fix: stonith_admin: Ensure pointers passed to sscanf() are properly initialized
Merge pull request #412 from gao-yan/lrmd-children-reset-nice
Merge pull request #413 from gao-yan/fencing-update-device
Merge pull request #414 from gao-yan/services-reset-scheduling
Merge pull request #417 from krig/grep-text
Merge pull request #428 from krig/doc-update
Fix: Bug rhbz#1057697 - Use native DBus library for systemd support to avoid problematic use of threads
Fix: Bug rhbz#1057697 - Use native DBus library for systemd async support to avoid problematic use of threads
Refactor: systemd: Simplify dbus API usage
Build: Add dbus as an rpm dependancy
Fix: Portability: Use basic types for DBus compatability struct
Fix: Bug rhbz#1057697 - Use native DBus library for upstart support to avoid problematic use of threads
Refactor: dbus: Use native function for adding arguments to messages
Fix: cib: Use the correctly typed enum value during initialization
Fix: logging: clang does not like the definition of QB_LOG_INIT_DATA()
Refactor: ipc: Convince clang that our casting is intended
Refactor: Core: Remove dead code
Refactor: Standardize on __FUNCTION__ as clang's __PRETTY_ variant is too verbose
Test: tools: Set up override for local schema file locations
Test: lrmd: Allow regression tests to run from the source tree
Test: lrmd: Fix upstart metadata test and install required OCF agents
Test: Get travis integration working
Test: Travis: Add coverity support to travis integration
Test: Travis: Import the coverity script so we can fix/improve it
Test: Travis: Customize and use our own coverity/travis script
Test: Travis: Use sudo when testing the lrmd with travis
Test: Travis: Handle coverity as a new compiler
Test: Travis: Ensure state directories exist
Test: travis: Make config.h available on build failures
Test: lrmd: Kill uninstalled daemons by the correct name
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Refactor: Services: Move common code into pcmk_dbus_connect
Fix: Bug rhbz#1057697 - Use native DBus library for systemd support to avoid problematic use of threads
Fix: Bug rhbz#1057697 - Use native DBus library for systemd async support to avoid problematic use of threads
Refactor: systemd: Simplify dbus API usage
Build: Add dbus as an rpm dependancy
Fix: Portability: Use basic types for DBus compatability struct
Fix: Bug rhbz#1057697 - Use native DBus library for upstart support to avoid problematic use of threads
Refactor: dbus: Use native function for adding arguments to messages
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: cib: Move children to the correct offset when applying 'move' directives of v2-style diffs
Refactor: election: If an error occurrs in crm_uptime(), compare as {0,0}
Refactor: xml: Allow both patch formats to include and validate an xml digest
Fix: cib: Validate clients are correctly applying v2-style diffs at most once a minute
Refactor: cib: Remove redundant paranoia functionality
Fix: cib: Validate that everyone still sees the same configuration once all updates have completed
Fix: xml: Do not create a patchset at all if the target is unchanged
Refactor: xml: Consolidate patchset logging and apply functionality
Fix: xml: Create entries at the correct offset
Fix: xml: Handle logging of empty patchsets
Log: cib: Include node name in the operation record for addition context
Fix: xml: Correctly detect deletions when testing if the target changed
Fix: cib: Stop ignoring replies to sync and replace operations
Fix: xml: Correctly unset the document dirty flag when accepting changes
Fix: cib: Do not ignore replace operations resulting from cib_sync requests
Fix: xml: Correctly detect attributes removed during a replace operation
Test: cts: Ensure we correctly handle configuration merging
Test: cts: Ensure the cib directory is erased before the node _first_ starts
Test: cts: Some extra patterns of interest
Merge branch 'cib-performance'
Test: cts: Allow use of the IPaddr2 resource to be turned off in cases where it is too problematic
Test: cts: Prefer the local directory for test code if available
Merge pull request #418 from krig/crm_resource
Merge pull request #427 from gao-yan/manual-ack-unknown-node
Merge pull request #432 from inouekazu/update_attributes
Merge pull request #419 from gao-yan/crm_mon-brief
Merge pull request #404 from gao-yan/display-pending-ops
Merge pull request #436 from gao-yan/pending-task-leak
Log: PE: Additional trace logging
Test: PE: Ensure stop actions are viable when reporting what will happen to a resource
Fix: Bug rhbz#1036631 - Stop members of cloned groups when dependancies are stopped
Test: Bug rhbz#1036631 - Stop members of cloned groups when dependancies are stopped
Fix: PE: Default sequential to TRUE for resource sets for consistency with colocation sets
Fix: crmd: Do not overwrite existing node state when fencing completes
Log: crmd: Extra detail when a peer leaves CPG
Fix: crm_mon: Allow the file created by --as-html to be world readable
Fix: mcp: Tell systemd not to respawn us if we exit with rc=100
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Test: cts: Extar BadNews pattern to ignore
Fix: xml: Prevent use-of-NULL, create _private for comment nodes as well
Fix: cib: Erasing the cib is an admin action, bump the admin_epoch instead
Test: tools: Update CLI tests for new CIB code
Test: lrmd: Bump the lrmd test timeouts to avoid transient travis failures
Fix: crm_mon: Ensure resource attributes have been unpacked before displaying connectivity data
Bug cl#5196 - PE: Check values after expanding templates
Merge pull request #438 from inouekazu/attrd_log_level
Merge pull request #439 from gao-yan/rsc-maintenance-mode
Fix: PE: Ensure instance numbers are preserved for cloned templates
Merge pull request #437 from gao-yan/anti-colocation-order
Fix: crmd: Ignore 'move' operations to the cib, we only care about content changes
Fix: Correctly detect lrm refreshes
Fix: xml: Expose the function for calculating more useful xpath paths
Log: crmd: Improved logging of transition cancellations
Log: crmd: Improved logging of transition cancellations
Doc: Update crmsh/pcs quickref
Doc: pcs-crmsh-quick-ref: Fix resource stop/start on pcs
Log: crmd: Simplify the logging of completed lrm operations
Fix: crmd: Correctly recognise operation cancellations we initiated
Test: cts: Update the resource operation patterns
Feature: crm_report: Allow ssh alternatives to be used
Fix: crm_report: Additional dlm detail
Log: Reduce log severity in xml helper functions
Log: cib: Simplify logging of updates
Test: PE: Correctly create cancel operations
Fix: fencing: Teach stonith-ng about finding device and topology changes in the v2 patch format
Feature: cib: Default ACL support to on
Feature: cib: Check ACLs prior to making the update instead of parsing the diff afterwards
Refactor: cib: Simplify ACL handling for read-only cib queries
Refactor: cib: Remove dead code
Fix: cib: ACL improvements
Merge pull request #440 from gao-yan/install-compatibility.h
Refactor: xml: Add ACL flags for xpath selectors when tracking is enabled
Refactor: xml: Convert all ACLs to xpaths during unpack
Fix: xml: Support the creation of filtered copies based on ACLs
Fix: cib: Use the new acl filtering functionality
Fix: xml: Allow new objects to be created provided they subseqently match a configured ACL
Fix: xml: Correctly initialize and find v1-style patch version numbers
Fix: cib: Switch over completely to the new ACL code
Log: Fencing: Send details of stonith_api_time() and stonith_api_kick() to syslog
Fix: Fencing: Pass the correct options when looking up the history by node name
Feature: Fencing: Add the ability to call stonith_api_time() from stonith_admin
Merge pull request #441 from chrissie-c/master
Merge pull request #444 from davidvossel/stonith_histroy_fix
Merge pull request #442 from gao-yan/cibadmin-replace-reply
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: crm_report: Only gather dlm data if dlm_controld is running
Merge pull request #446 from davidvossel/pengine_fast
Fix: xml: Prevent memory leak in xml _private field
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #443 from yuusuke/fix_memory_leak
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: PE: Memory leak when updating cancel operations
Doc: crm_resource: Fix wikipedia reference
Fix: systemd: Prevent use-of-NULL when determining if an agent exists
Fix: Fencing: Remove shadow definition and use of variable 'progress'
Merge pull request #448 from gao-yan/xml-tracking-changes-segfault
Fix: xml: Memory leak when free'ing xml documents
Fix: xml: Convert misbehaving #define into a more easily understood inline function
Fix: cib: Include a request with file-mode cib actions for the acl code to interrogate
Refactor: cibadmin: Purge legacy command options deprecated since 2008 to make the short-forms available for new options
Fix: cibadmin: Remove dangerous commands that exposed unnecessary implementation internal details
Fix: cibadmin: Allow priviliged clients to perform tasks as unpriviliged users
Refactor: cib: Implement CIB_file-mode ACLs for testing
Fix: xml: Unpack and apply ACLs when tracking changes
Fix: cib: Turn ALCs off so we can update version counters once the op is complete
Fix: cib: Respect the CIB_user environment variable for ACLs
Fix: xml: Set flags to indicate when updates were rejected by an ACL
Refactor: xml: Clean up the API for ACLs
Fix: cib: Memory leaks caused by ACLs returning filtered copies
Fix: xml: Always allow new scaffolding - node with no attributes or only an id field
Test: Write some tests for ACLs
Fix: xml: Only test write permissions for attributes that are being changed
Test: BasicSanity: cli tests need to be run as root for acl testing
Test: tools: Always run additional cib queries as root
Fix: xml: Correctly observe ACLs when creating filtered copies of xml documents
Fix: crm_resource: Gracefully handle -EACCESS when querying the cib
Refactor: Stop using get_cib_copy() as it doesn't expose -EACCES errors
Merge branch 'acl-playground'
Fix: cib: Do not try to use an empty value for CIB_shadow
Fix: xml: Correctly enforce ACLs after a replace operation
Fix: xml: Mark the parents of a partial replace as dirty
Fix: cib: Prevent use-after-free when cleaning up after query operations
Fix: xml: Ignore DTD nodes
Log: crmd: Supply string arguments in the expected order
Fix: cib: Ingore patch failures if we already have their contents
Test: RemoteBaremetal: Update resource operation patterns
Log: cib: Improved logging of ping requests and replies
Log: cib: Reduce redundant logging
Refactor: crm_diff: Mirror how the cib operates for better testing
Fix: xml: Correctly calculate the adjusted offset of xml elements
Fix: xml: Better move detection logic for xml nodes
Test: cts: Functions for querying and updating PCMK_trace_functions entries
Log: xml: Improved logging of changes and patchsets
Log: cib: Better call logging
Log: xml: Tweak the format when logging v2 patchsets
Fix: xml: Ensure xml_calculate_changes() doesn't modify the attribute ordering as a side-effect
Test: cts: Clean out corosync state files and old logs
Fix: acl: Read access must be explicitly granted
Test: tools: Fix regression test output
Refactor: crmd: Avoid logging NULL
Merge pull request #450 from inouekazu/increase-loglevel
Merge pull request #453 from gao-yan/on-disk-cib-corrupt
Log: Fencing: Reduce severity of some noisy messages
Merge pull request #449 from inouekazu/decrease-loglevel
Test: BasicSanity: Quote arguments to su
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #455 from gao-yan/on-disk-cib-corrupt-2
Merge pull request #458 from chrissie-c/master
Test: cts: Refactor things a bit
Fix: xml: Create the correct diff when only part of a document is changed
Fix: xml: Find the minimal set of changes when part of a document is replaced
Fix: crmd: Prevent memory leak when accepting a new DC
Test: tools: Install expexted regression test output
Fix: crm_diff: Allow the generation of xml patchsets without digests
Fix: xml: Correctly mark the parent's of deleted attributes as dirty
Fix: xml: Do not log deleted attributes
Test: cts: Isolate the patterns required for tests
Merge pull request #461 from inouekazu/kill_process_group
Test: cts: Continue to decouple the various pieces of CTS
Fix: crmd: Fenced nodes that return prior to an election do not need to have their status section reset
Merge pull request #460 from gao-yan/location-sets-schema
Fix: Prevent use-of-NULL spotted by coverity
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: Prevent use-of-NULL spotted by coverity
Merge pull request #464 from inouekazu/proper_log
Fix: controld: Verify killall is available on resource start
Fix: acl: Authenticate pacemaker-remote requests with the node name as the client
Merge pull request #466 from oallart/master
Merge pull request #468 from oallart/master
Merge pull request #469 from krig/pcs-crmsh-updates
Merge pull request #471 from AnchorCat/shadow-reset
Test: tools: See if we can make the regression tests pass
Refactor: xml: Rename pacemaker-1.1 to the more clear pacemaker-next
Fix: xml: If validate-with is missing, we find the most recent schema that accepts it and go from there
Fix: crm_report: extract additional DLM lockspace data
Fix: xml: Allow first-acl-match-wins semantics to be supported if desired
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: xml: Add missing xslt file
Refactor: xml: Split up the RelaxNG schema files so they're easier to version individually
Feature: xml: Add the ability to have lightweight schema revisions
Refactor: xml: Generate the 1.0 schema too
Doc: xml: More information about schema versioning
Refactor: xml: Work out more schema details automatically
Fix: xml: best-match.sh helper needs to be included in make dist
Refactor: xml: Autogenerate versions.rng based on RNG_versions in the makefile
Refactor: xml: Replace CRM_DTD_VERSION with a function that returns the relevant known_schemas entry
Clone the latest resources schema in preparation for changes
Feature: xml: Support resources that require unfencing
Doc: xml: More details on schema versioning
Build: xml: Autodetect the available schema versions
Fix: xml: Return an actual string from xml_latest_schema() as advertised
Build: xml: tell make to build versions.rng
Fix: xml: Build a valid schema for versions.rng
Test: tools: Ingore changes to validate-with in regression tests
Refactor: PE: Indicate the type of fencing in the fencing operation key
Merge pull request #477 from yuusuke/fix_crm_mon
Fix: PE: Potential memory leak
Refactor: PE: Fix comments and formatting for pe_ordering enum
Merge pull request #472 from oallart/master
Refactor: PE: Ensure actions like shutdown and fence have a unique key
Test: PE: Include stonith operation in .dot graph regression test results
Test: PE: Include stonith operation in the summary results and handle unfencing
Refactor: PE: Store non-resource actions in a hashtable where we can easily (and cheaply) find them again
Test: PE: Regression tests for future unfencing logic
Merge pull request #484 from krig/crm_report-iso8601
Feature: PE: Automatically re-unfence a node if the fencing device definition changes
Test: Fix: ACLs are not important for PE regression tests
Fix: PE: Fencing devices default to only requiring quorum in order to start
Fix: PE: Delay unfencing until after we know the state of all resources that require unfencing
Log: acl: The message about enforcing ACLs doesn't need to be so aggressive
Fix: crmd: Correctly react to successful unfencing operations
Merge pull request #485 from HideoYamauchi/master
Merge pull request #478 from gao-yan/attrd-logging
Merge pull request #473 from jnpkrn/integer-overflow
Merge pull request #479 from gao-yan/corosync-logging
Merge pull request #476 from krig/crm_attribute_node
Merge branch 'master' of github.com:ClusterLabs/pacemaker
xml: Duplicate the nodes subschema ready for changes
xml: Restore type=remote for nodes
PE: Finish purging the experimental 'domain' construct - its unnecessary
Build: xml: Improve the command for showing the differences from and in the current schema
Merge pull request #486 from inouekazu/rsc_casesensitive
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Test: PE: Remove the domain test case too
Build: bc is required for schema generation
Fix: xml: Add missing schema file
Test: PE: node.type=remote is only available in the 1.3 schema
Test: PE: Update new regression test with unique keys for probe actions
Fix: acl: Ensure known users are not able to read the cib by default
Fix: cib: Determine the best value of validate-with if one is not supplied
Fix: cib: All schema upgrade requests must go via the DC
Refactor: xml: Construct the table of supported schemas
Test: cts: Fix the 'boot' scenario
Fix: xml: Correctly determine the lowest allowed schema version
Build: xml: INFINITY could overflow for 16-bit ints - explicitly cast it
Fix: acl: Prevent use-of-NULL
Merge pull request #425 from gao-yan/nvpair-id-ref
Merge pull request #420 from inouekazu/get_failcount_by_prefix
Merge pull request #493 from davidvossel/unfencing_fixes
Fix: lrmd: Provide stderr output from agents if available, otherwise fall back to stdout
Log: xml: Include patch format when logging digest failures
Log: xml: Reduce LOG_NOTICE noise
Fix: crmd: Report unsuccessful unfencing operations
Log: Fencing: Additional trace information
Fix: fencing: Automatically switch from 'list' to 'status' to 'static-list' if those actions are not advertised in the metadata
Fix: fencing: Filter self-fencing at the peers to allow unfencing to work correctly
Fix: PE: Ensure unfencing occurs before fencing devices are (re-)probed
Feature: logging: daemons always get a log file, unless explicitly set to configured 'none'
Refactor: logging: Simplify processing of logging directives
Fix: lrmd: Expose logging variables expected by OCF agents
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: Fencing: Correctly record which peer performed the fencing operation
Fix: crmd: Do not erase the status section for unfenced nodes
Fix: crmd: Actions can only be executed if their pre-requisits completed successfully
Log: crmd: device isn't hooked up yet, don't bother logging it
Fix: PE: Ensure unfencing only happens once, even if the transition is interrupted
Fix: PE: Do not unfence nodes that are offline, unclean or shutting down
Fix: xml: Resolve memory leak when logging xml blobs
Test: fence_dummy does not support the list command
Merge pull request #497 from inouekazu/fix_condition
Fix: PE: Unfencing is based on device probes, there is no need to unfence when normal resources are found active
Test: PE: Unfencing is based on device probes, there is no need to unfence when normal resources are found active
Merge pull request #496 from xiali/master
Merge pull request #495 from gao-yan/ticket-state-change
Merge pull request #463 from gao-yan/tags
Fix: crm_node: Only remove the named resource from the cib
Fix: PE: Do not initiate fencing for unclean nodes when fencing is disabled
Fix: PE: Memory leak
Merge pull request #475 from gao-yan/site-specific-params
Merge pull request #499 from gao-yan/expire-non-blocked-failure
Merge pull request #507 from jnpkrn/fencing-literals
Fix: fencing: Advertise support for reboot/on/off in the metadata for legacy agents
Build: Use the Fedora approved method of obtaining Source0 URLs from github
Test: cib: Trust CIB_user when operating directly on a file
Build: mock does not need to be run as root
Build: RHEL6 doesn't support the autosetup directive
Fix: cib: Resolve memory leaks in query paths
Fix: xml: Prevent use-after-free when comparing two xml documents
Doc: fencing: Spelling mistake
Refactor: Remove unnecessary create_cib_fragment function
Refactor: cib: Populate version and schema details in calls to createEmptyCib()
Feature: tools: Allow crm_shadow to create older configuration versions
Test: tools: Refactor the acl tests to allow running for multiple versions
Test: tools: Simplify the acl tests
Test: tools: Reduce output noise from acl tests
Build: Prepare for ACL schema changes
Feature: Changes to the ACL schema to support nodes and unix groups
Test: xml: Implement regression tests for new ACL syntax
Fix: xml: Check all available schemas when doing upgrades
Test: tools: Update regression tests
Build: Trick analyzers into understanding how CRM_ASSERT() actually works
Build: Move the version number to a separate file
Build: Teach make how to be even less verbose
Build: Create a common location for shared makefile functionality
Build: The use of INCLUDES is deprecated in favor of AM_CPPFLAGS
Build: Stop using the deprecated form of AM_INIT_AUTOMAKE
Build: doc: Make the default build output less verbose
Build: tools: Fix man page generation
Build: Fix the clang target
Fix: Resolve all clang analyser defects
Refactor: crmd: Remove dead code
Refactor: crm_shadow: Make use of standard patchset creation and logging functionality
Merge pull request #510 from gao-yan/cluster-site-specific-params
Merge pull request #509 from inouekazu/fix_reread
Fix: xml: Clean up schema list at exit
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #508 from gao-yan/xml-patchset-v1-fixes
Fix: iso8601: Different logic is needed when logging and calculating durations
Test: iso8601: Improved logging of durations
Fix: PE: Correctly handle origin offsets in the future
Test: PE: Correctly handle origin offsets in the future
Merge pull request #511 from gao-yan/xml-patchset-v1-fix
Fix: iso8601: Fix memory leak in duration calculation
Fix: crmd: Use correctly signed inputs to sscanf()
Refactor: crmd: Silence cppcheck error
Merge with devel
Merge branch '1.1' of github.com:ClusterLabs/pacemaker into 1.1
Merge pull request #514 from fuhu/master
Merge pull request #513 from gao-yan/default-cluster-name
Merge pull request #512 from yuusuke/fix_startup-fencing
Build: Allow upstream building for i386
Doc: Pacemaker_Explained: Fix options typo
Merge pull request #515 from inouekazu/1.1
Bug cl#5184 - Ensure pending probes that ultimately fail are correctly updated
Build: add a cppcheck helper
Bug rhbz#1097457 - Prevent invalid transition when resource are ordered to start after the container they're started in
Test: PE: Prevent invalid transition when resource are ordered to start after the container they're started in
Bug cl#5212 - Do not promote instances when quorum is lots and no-quorum-policy=freeze
Test: PE: Do not promote instances when quorum is lots and no-quorum-policy=freeze
Merge branch '1.1' of github.com:ClusterLabs/pacemaker into 1.1
Test: cts: allow use of the the 2.0 schema
Bug rhbz#1097457 - Limit the scope of the previous fix and include a helpful comment
Merge pull request #517 from davidvossel/1.1.12-testing
Merge pull request #520 from davidvossel/1.1.12-testing2
Fix: systemd: Try to handle dbus actions that complete prior to configuring a callback
Merge pull request #521 from gao-yan/unmanaged-block-restart
Merge branch '1.1' of github.com:ClusterLabs/pacemaker into 1.1
Bug cl#5213 - Ensure role colocation with -INFINITY is enforced
Test: PE: Ensure role colocation with -INFINITY is enforced
Bug cl#5213 - Limit the scope of the previous commit to the masters role
Log: cib: Reduce noise from query operations
Fix: attrd: Ensure attribute dampening is always observed
Merge branch 'master' into 1.1
Merge pull request #523 from b1-systems/master
Merge pull request #516 from jnpkrn/build-speedup
Merge pull request #524 from b1-systems/typo_fix
Merge pull request #522 from gao-yan/pengine-segfault
Merge branch 'master' into cts-playground
Fix: crmd: Avoid spamming the cib by triggering a transition only once per non-status change
Merge pull request #525 from berendt/pep8
Log: cib: Drop the message serverity now that the cause is well understood
Merge pull request #527 from b1-systems/master
Merge pull request #529 from yuusuke/fix_stonithd
Merge pull request #530 from davidvossel/cts-lxc-updates
Log: services: Some extra detail about which pid is being killed
Fix: services: Remove file descriptors from mainloop as soon as we have drained them
Fix: services: Remove cancelled recurring ops from internal lists as early as possible
Fix: mainloop: Better handle the killing of processes in the act of exiting
Merge pull request #531 from davidvossel/cts-lxc-updates
Fix: Fencing: Cache metadata lookups to avoid repeated blocking during device registration
Test: cts: Search logs in parallel to speed up large clusters
Merge branch 'master' into cts-playground
Test: cts: Fix qarsh support and non-IPaddr resources
Merge pull request #534 from gao-yan/remove-temp-shadow
Test: cts: Make partition detection order independant
Merge pull request #532 from gao-yan/load-stopped-migrate
Merge pull request #535 from gao-yan/cib-sync-request-assert
Merge pull request #536 from davidvossel/cts-lxc-updates
Fix: cib: Use a better option name for enabling the legacy cib diff code
Test: cts: Provide more instructions when setting up cts
Bug cl#5222 - cib: Repair rolling update capability
Merge pull request #537 from AnchorCat/mainloop
Fix: crm_report: Gracefully handle rediculously large logfiles
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Build: fix the cts file list
Merge pull request #539 from gao-yan/search-node-uname
Merge pull request #538 from inouekazu/1.1
Merge pull request #541 from yuusuke/fix_log
Merge pull request #540 from krig/readme-shells
Merge pull request #542 from gao-yan/xml-diff-check
Bug cl#5222 - Enable legacy mode whenever a broadcast update is detected
Bug cl#5219 - PE: Allow unrelated resources with a common colocation target to remain promoted
Bug rhbz#1054307 - cname pattern match should be more restrictive in init script
Merge branch '1.1'
Test: tools: Remove line numbers from ACL test output
Merge pull request #543 from gao-yan/cib_perform_op-assert
Fix: crm_report: Correctly detect large files
Merge pull request #544 from davidvossel/cts-lxc-updates
Merge pull request #547 from krig/crm_diff-filter
Merge pull request #546 from xiali/mybranch
Merge pull request #549 from ldzhong/1.1
Merge pull request #548 from gao-yan/ocf-log-setenv
Merge pull request #550 from gao-yan/remove-unknown-conflicting-node
Test: cts: sort partition details
Test: cts: Stop Popen from blocking
Merge branch '1.1'
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: lrmd: Handle systemd reporting 'done' before a resource is actually stopped
Fix: PE: Correctly observe requires=nothing
Fix: mcp: Let daemons know when they have been respawned by the mcp
Fix: xml: Prevent possible array underflow
Fix: ipc: Buffer size cannot be negative
Fix: PE: Potential use-of-NULL
Fix: PE: Earlier checks for NULL
Fix: tools: Non-recurring memory leak
Fix: PE: Potential use-of-NULL
Fix: crmd: Only write fence_averride if open() returns a positive file descriptor
Fix: xml: Ensure realloc didn't fail
PE: Fix type
Fix: PE: Correctly detect empty tags
Merge pull request #552 from xiali/1.1
Merge pull request #554 from fuhu/master
Merge pull request #555 from gao-yan/logrotate
Doc: Update some into images
Merge pull request #553 from gao-yan/remove-unknown-conflicting-node
Doc: Pacemaker Explained: Include information on using master/slave resource in sets
Doc: Pacemaker Explained: Improve the readability of the section on resource sets
Build: Use the 'correct' dependancy for packages using pkgconfig
Doc: Clusters from Scratch: Update the intro and installation chapters
Doc: Pacemaker Explained: Update the introduction
Doc: Clusters from Scratch: Fix the Installation chapter
Test: cts: Fix the --choose option
Fix: lrmd: Handle systemd reporting 'done' before a resource is actually stopped
Refactor: PE: Always use pe_fence_node to mark nodes as unclean
Log: PE: Correctly warn when resources require fencing but fencing is disabled
Fix: crm_verify: Perform a full set of calculations whenever the status section is present
Fix: crm_simulate: Remove resource states and transient attributes when a node is marked as down
Merge pull request #558 from davidvossel/master
Merge pull request #559 from liedekef/patch-1
Test: lrmd: Update the systemd agent to test long running actions
Fix: systemd: Ensure the retry logic is also exercised for the 'service' class
Fix: dbus: Return 'null' if we cannot connect via dbus
Fix: Upstart: Do not attempt upstart jobs if we cannot connect to dbus
Test: Install psmisc for killall
Build: Fix rpm builds where the pkg-config version for corosync is busted
Fix: Logging: Bootstrap daemon logging before processing arguments but configure it afterwards
Log: Default to emitting CRIT messages only until logging is fully configured
Fix: Tools: Non-daemons shouldn't abort just because xml parsing failed
Fix: logrotate: maxsize is not supported by older versions
Merge pull request #560 from yuusuke/fix_log
Test: cts: Don't overwrite the pattern set name
Test: cts: Turn off log audits when file logging is selected
Log: Do not unconditionally send crm_perror output to stderr
Log: Turn off the default log file if a real one is specified
Test: cts: Put a limit on the number of results returned by log watcher
Test: cli: Update regression test output
Test: cts: Adjust patterns so they can be used when looking at normal log files (not syslog)
Test: cts: Correctly read all logs emitted prior to the timeout
Test: cts: Reduce noise
Test: cts: One less use of ClusterManager as a variable array
Merge branch '1.1'
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: crmd: Disable debug logging that results in significant overhead
Build: Update library versions
Build: Update changelog and version numbers
Merge branch '1.1'
Fix: lrmd: Log with the correct personality
Merge branch '1.1'
Test: Fix travis builds... maybe
Test: Fix travis builds... maybe
Bug rhbz#1067544 - Tools: Correctly handle --ban, --move and --locate for master/slave groups
Fix: mcp: Teach pacemakerd to ignore SIGHUP
Fix: mcp: Ignore SIGHUP as early as possible
Fix: crm_mon: Repair notification delivery when the v2 patch format is in use
Build: xml: Prepare for changes to constraint sets
Fix: Fencing: Force 'stonith-ng' as the system name
Merge pull request #563 from gao-yan/crm_diff-no-version
Fix: stonith-ng: Reset mainloop source IDs after removing them
Doc: minor formatting issue
Feature: PE: Allow natural ordering of colocation sets
Test: PE: Allow natural ordering of colocation sets
Test: tools: Update regression tests
Merge pull request #565 from davidvossel/master
Feature: mcp: Include basic watchdog functionality
Feature: mcp: Two new error codes that result in machine reset or power off
Fetaure: mcp: Do not start the cluster if the configured watchdog cannot be opened
Feature: Fencing: Advertise the watchdog device for fencing operations
Merge pull request #566 from davidvossel/sigpipe-dfl
Refactor: mcp: Clean up some variable names
Fix: mcp: Shut down the cluster if sysrq/watchdog resets fail
Feature: crmd: Ensure a watchdog device is in use if stonith-watchdog-timeout is configured
Build: Exclude some compiler flags on solaris
Feature: crmd: If configured, trigger the watchdog immediately if we loose quorum and no-quorum-policy=suicide
Fix: mcp: Default to a shorter watchdog interval
Feature: Fencing: Allow the cluster to recover resources if the watchdog is in use
Fix: mcp: Do not initiate node reboot if crashdump is requested
Fix: mcp: Ensure the watchdog is closed on normal exit
Fix: mcp: Delay reopening the watchdog if old child processes are detected
Fix: solaris: Correctly replace the linux swab functions
Fix: solaris: Allow compilation when dirent.d_type is not available
Fix: mcp: Allow forced closure of the watchdog device
Fix: mcp: We don't need to bump the IO priority just for the watchdog
Refactor: mcp: Make watchdog debug flags a little more flexible
Fix: mcp: Re-enable logging in case we don't die when the watchdog expires
Merge pull request #568 from gao-yan/crm_resource-no-attr
Fix: solaris: Disable throttling since /proc doesn't exist
Merge pull request #569 from davidvossel/testfixes
Merge pull request #570 from davidvossel/cts-fixes
Feature: mcp: Drop direct watchdog integration
Build: Ensure the 'bool' type is available
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Test: cts: Terminate journald pattern searches after the timeout
Merge pull request #572 from davidvossel/services-refactor
Test: cts: Update child death patterns
Feature: Support machine panics for some kinds of errors (via sbd if available)
Feature: mcp: Allow orphan children to initiate node panic via SIGQUIT
Test: cts: Fix fencing detection during node startup
Test: cts: Update child daemon exit patterns
Fix: cib: Do not update on disk backups if we're just querying them
Merge pull request #574 from inouekazu/fix_job
Fix: corosync: Avoid unnecessary repeated CMAP API calls
Fix: crm_resource: Ensure fail-counts are purged when we redetect the state of all resources
Fix: watchdog: sync() may never return
Merge pull request #573 from inouekazu/set_env
Merge pull request #575 from gao-yan/conditionally-detect-conflicting-peer
Log: PE: Don't report an error when an uncompletable migration is detected
Merge pull request #576 from xiali/master
Test: cts: Correctly handle threading and journalctl commands that don't return the new cursor
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #571 from kjperry/spelling
Test: cts: Some patterns cannot be removed after fencing
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Test: cts: Do not bounce between complete and in-progress for fencing ops
Test: cts: Fix remote file watching
Fix: fencing: Reject actions for devices that haven't been explicitly registered yet
Refactor: Use the correct define for origin
Refactor: attrd: Decide which kind to build in configure
Test: cts: Additional patterns when stonith-ng is killed
Feature: Allow fail-counts to be removed en-mass when the new attrd is in operation
Merge pull request #578 from davidvossel/master
Merge pull request #579 from davidvossel/master
Merge pull request #577 from davidvossel/docker-cts
Fix: crm_attribute: Correctly update config options when -o crm_config is specified
Fix: systemd: Perform actions asynchronously
Refactor: dbus: Mention the symbolic name that -1 is equivalent to
Fix: crmd: Ensure throttle_mode() does something on Linux
Merge pull request #580 from davidvossel/master
Merge pull request #581 from davidvossel/master
Fix: dbus: Notice when dbus connections get disabled
Merge pull request #586 from fuhu/master
Merge pull request #583 from davidvossel/master
Merge pull request #585 from HideoYamauchi/upstart_crm_mon
Fix: systemd: Perform monitor operations without blocking
Fix: upstart: Perform more parts of monitor operations without blocking
Test: cts: Indicate which packages virsh needs for lxc_autogen
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: systemd: Ensure we don't call dbus_message_unref() with NULL
Fix: upstart: Correct ddfa810 - Perform more parts of monitor operations without blocking
Fix: upstart: Monitor actions always return 'ok' or 'not running'
Fix: cib: Ensure upgrade operations from a non-DC get an acknowledgement
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #591 from lge/for-beekhof
Merge pull request #590 from krig/ping-fixes
Merge pull request #584 from HideoYamauchi/change_log
Log: cib: Improved tracing of callbacks
Merge pull request #592 from gao-yan/xml-patch-no-versions
Merge pull request #593 from davidvossel/master
Fix: lrmd: Prevent glib assert triggered by timers being removed from mainloop more than once
Build: Simplify the process of building docs
Feature: mcp: Turn on sbd integration if pacemakerd finds it running
Fix: cts: Ensure we can install the log watcher when running as non-root
Merge pull request #598 from HideoYamauchi/changeLoglevel
Fix: Fencing: Ensure the hostlist parameter is set for watchdog agents
Fix: Fencing: Correctly wait for self-fencing to occur when the watchdog is in use
Fix: PE: Watchdog integration is sufficient for fencing
Fix: Resolve some potential issues spotted by clang and coverity
Fix: Fencing: If configured, wait stonith-watchdog-timer seconds for self-fencing to complete
Merge pull request #600 from davidvossel/master
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: sbd: Attempt to locate sbd via its pid file
Fix: stonith-ng: Correctly observe the watchdog completion timeout
Fix: Prefer to inherit the watchdog timeout from SBD
Merge pull request #601 from jnpkrn/fix-xml.c-memleak
Build: Prepare for additions to the constraint schema
Feature: Officially support the resource-discovery attribute for location constraints
Test: Tools: Update regression test
Merge pull request #602 from davidvossel/master
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Log: PE: Reduce 'no quorum' logging severity
Doc: Document the record-pending option for actions
Fix: dbus: Remove double-free introduced while trying to make coverity shut up
Fix: crmd: Correctly obtain SBD_WATCHDOG_TIMEOUT from the environment
Fix: crmd: Prevent use-of-NULL during reprobe
Fix: dbus: Fail gracefully if malloc fails
Fix: service: Prevent potential use-of-NULL in metadata lookups
Merge pull request #595 from gao-yan/master-asymmetrical-order
Fix: watchdog: Allow startup without sbd
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: Perform systemd reloads asynchronously
Fix: systemd: Ensure failed monitor operations always return
Test: Fix regression test when run from installed packages
Fix: dbus: Ensure both the read and write queues get dispatched
Test: lrmd: Ensure the lsb script is executable
Fix: crm_failcount: Better error reporting when no resource is specified
Fix: systemd: Convince clang that 'op' is not NULL
Feature: crm_resource: Implement an intelligent restart capability
Test: cts: sbd patterns to ignore
Merge pull request #605 from davidvossel/master
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: crm_resource: Calculate how long to wait for a restart based on the resource timeouts
Fix: crm_resource: Implement --timeout for resource restart operations
Fix: crm_resource: Allow resource restart operations to be node specific
Fix: crm_resource: Repair regression test outputs
Fix: attrd: Clean out the node cache when requested by the admin
Fix: attrd: Simplify how node deletions happen
Fix: Membership: Detect and resolve nodes that change their ID
Merge pull request #608 from davidvossel/master
Merge pull request #607 from gao-yan/services-action-conditional-cleanup
Fix: cib: Avoid nodeid conflicts we don't care about
Fix: fencing: Allow nodes to be purged from the member cache
Doc: CfS: Publican doesnt like something about this part of the chinese translation
Fix: crm_node: Correctly remove nodes from the CIB by nodeid
Merge pull request #613 from gao-yan/xml-diff-unset-attr
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Log: xml: Reduce severity of xslt upgrade messages
Fix: crm_resource: Memory leak in error paths
Fix: services: Prevent potential use-of-NULL
Build: Use a wrapper for realloc() which calls abort when we run out of memory
Fix: crm_resource: Clean up memory in --restart error paths
Fix: crm_resource: Ensure --restart sets/clears meta attributes
Fix: crm_resource: Include group members when calculating the next timeout
Log: xml: Report an error if someone creates an unnamed xml node
Fix: systemd: Correctly handle long running stop actions
Low: lrmd: Additional logging for long running systemd commands
Test: cts: Include a systemd resource where possible
Fix: crm_resource: Prevent use-after-free
Fix: xml: Prevent use of NULL
Merge pull request #617 from ultrabug/master
Merge pull request #622 from HideoYamauchi/patch-for-pacemakerd
Fix: cib: Ensure file-based backends treat '-o section' consistently with the native backend
Fix: xml: Ensure ACLs are reapplied before calculating what a replace operation changed
Test: xml: Ensure ACLs are reapplied before calculating what a replace operation changed
Fix: mcp: Avoid requiring 'export' for SYS-V sysconfig options
Build: clang: Ensure we scan everything, not just what changed
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #624 from davidvossel/require-all-feature
Merge pull request #616 from gao-yan/cib-diff-v1-prop-ordering
Merge pull request #626 from gao-yan/build-ncurses
Merge pull request #628 from davidvossel/bz1117341-pseudo-stop
Merge pull request #615 from gao-yan/cib-diff-v1-with-digest
Fix: acl: Correctly implement the 'reference' acl directive
Fix: PE: Do not aggregate children in a pending state into the started/stopped/etc lists
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Feature: attrd_updater: Allow attributes to be set for other nodes
Bug rhbz#1181824 - Ensure the DC can be reliably fenced
Doc: crmd: Note to future self regarding non-obvious fencing implications
Fix: cli: Ensure subsequent unknown long options are correctly detected
Fix: crm_resource: Use-after-free when restarting a resource
Merge pull request #631 from Ledest/master
Merge pull request #632 from kgaillot/tools
Merge pull request #633 from gao-yan/watchdog-option
Merge pull request #634 from kgaillot/doc
Update coding_guidelines.txt
Merge pull request #642 from kgaillot/sig
Merge pull request #637 from lge/for-beekhof
Merge pull request #645 from gao-yan/crm_mon-notify_node
Merge pull request #644 from kgaillot/sig
Merge pull request #635 from kgaillot/doc
Fix: service: Additional pointer safety
Fix: service: Leave synchronous calls to be free'd by the caller
Fix: attrd: refresh should force an immediate write-out of all attributes
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: extra: Use crm_simulate for the showscores script
Merge pull request #649 from gao-yan/build-non-heartbeat
Merge pull request #651 from gao-yan/not-reschedule-unneeded-monitor
Merge pull request #639 from gao-yan/fencing-random-delay
Merge pull request #650 from cvubrugier/cross-compilation-fixes
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Refactor: Avoid using g_strdup and friends because they don't use malloc
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #646 from kgaillot/attrd
Merge pull request #654 from kgaillot/cib
Merge pull request #647 from kgaillot/doc
Merge pull request #652 from gao-yan/reschedule-specific-monitor
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #658 from kgaillot/doc
Merge pull request #659 from kgaillot/doc
Merge pull request #660 from kgaillot/wait
Merge pull request #667 from kgaillot/doc
Merge pull request #655 from kgaillot/attrd
Merge pull request #664 from kgaillot/wait
Feature: PE: Support non-actionable degraded mode for OCF
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Merge pull request #666 from kgaillot/lsb
Merge pull request #665 from gao-yan/reset-stonith-failcount
Fix: PE: Do not record duplicate copies of the failed actions
Log: Record all signals we receive
Fix: cib: Correctly set up signal handlers
Fix: systemd: Tell systemd not to take DBus down from underneath us
Fix: systemd: Kindly ask dbus NOT to kill the process if the dbus connection fails
Fix: mcp: Allow a configurable delay when debugging shutdown issues
Fix: systemd: Trick systemd into not stopping our services before us during shutdown
Merge pull request #668 from kgaillot/mon
Merge pull request #669 from davidvossel/pcmk-docker-bff
Fix: crmd: Wait for all pending operations to complete before poking the policy engine
Fix: crmd: Cached copies of transient attributes cease to be valid once a node leaves the membership
Fix: crmd: All peers need to track node shutdown requests
Log: services: Be more generous when trace logging the RA output
Merge branch 'master' of github.com:ClusterLabs/pacemaker
Fix: cib: Correctly track node state
Fix: stonith-ng: No reply is needed for CRM_OP_RM_NODE_CACHE
Fix: stonith-ng: Correctly track node state
Fix: pengine: Fix colocation with unmanaged resources
Merge pull request #672 from gao-yan/crm_resource-anonymous-children
Build: Turn on travis builds for the 1.1 branch
Merge branch '1.1' of github.com:ClusterLabs/pacemaker into 1.1
Fix: acl: Do not delay evaluation of added nodes in some situations
Merge branch '1.1' of github.com:ClusterLabs/pacemaker into 1.1
Fix: PE: Turn off legacy unpacking of resource options into the meta hashtable
Fix: crm_report: More patterns for finding logs on non-DC nodes
Log: cib: cib_file_read_and_verify doesn't need to be this verbose
Merge pull request #676 from kgaillot/timeout-1.1
Merge pull request #674 from kgaillot/fixes-1.1
Build: Make the bumplibs script smarter
Build: Bump the library versions as appropriate
Merge pull request #684 from HideoYamauchi/1.1
Merge pull request #683 from kgaillot/fixes-1.1
Merge pull request #681 from gao-yan/display-resource-unclean
Fix: PE: Correctly compare feature set to determine how to unpack meta attributes
Merge branch '1.1' of github.com:ClusterLabs/pacemaker into 1.1
Test: PE: Change the feature set, saves recalculating all the digests
Doc: fencing: Clarify that stonith-timeout only applies to reboot/on/off actions
Agents: Fix the HealthSMART metadata
Fix: lrmd: Hint to child processes that using sd_notify is not required
Merge branch '1.1' of github.com:ClusterLabs/pacemaker into 1.1
Low: spec: move report.common and report.collector to cli package
Merge pull request #686 from HideoYamauchi/1.1
Fix: HealthSMART: OCF_RESKEY_drives could contain multiple values
Doc: Work-around until pcs pulls in psmisc automatically
Fix: corosync: Bug cl#5232 - Somewhat gracefully handle nodes with invalid UUIDs
Merge branch '1.1' of github.com:ClusterLabs/pacemaker into 1.1
Fix: fence_legacy: Avoid passing 'port' to cluster-glue agents
Merge pull request #687 from gao-yan/fence_legacy_nodename
Merge pull request #697 from gao-yan/crm_mon-packaging
Build: Attempt to fix travis builds
Merge branch '1.1' of github.com:ClusterLabs/pacemaker into 1.1
Merge pull request #700 from gao-yan/acl-xsl
Merge pull request #685 from kgaillot/fixes-1.1
Merge pull request #694 from kgaillot/bz1200849
Merge pull request #701 from gao-yan/rolling-upgrade-fence
Merge branch 'colocated-utilization' of https://github.com/gao-yan/pacemaker into 1.1
Build: Another travis update
Merge pull request #707 from kgaillot/onfail
Merge pull request #704 from gao-yan/sort_rsc_process_order
Merge branch '1.1' of github.com:ClusterLabs/pacemaker into 1.1
Build: Attempt to repair srpm generation for buildbot
Build: Generate a temporary changelog each time
Build: Prevent rpm packaging conflicts
Build: Ensure all spec files contain a changelog
Tools: Repair expected output for ACLs
Build: CRM_STATE_DIR is created dynamically when needed
Build: Update rpmlint suppresssions
Fix: Fencing: Gracefully handle invalid metadata from agents
Merge pull request #709 from gao-yan/cib-legacy-mode
Merge pull request #712 from gao-yan/legacy-mode-digest
Merge branch '1.1' of github.com:ClusterLabs/pacemaker into 1.1
Merge pull request #715 from gao-yan/patchset-v1-delete-attribute
Merge pull request #713 from gao-yan/potential-use-of-null
Merge pull request #717 from gao-yan/utilization-check-allowed-nodes
Merge pull request #716 from gao-yan/cib-update-attributes
Merge pull request #710 from HideoYamauchi/1.1
Merge branch '1.1' of github.com:ClusterLabs/pacemaker into 1.1
Fix: Move new struct members to the end for backwards compatibility
Build: Update library versions
Build: Update changelog for 1.1.13
Christian Berendt (1):
fixed PEP8 issues E303, E225 and E302 in CTS
Christine Caulfield (5):
Test: Make sure that fence_dummy has executable bit set
Doc: Clusters from scratch: Update for the final pcs syntax
crm_mon: Add -R switch to show clone details
crm_mon: add brackets around conditionals and remove tabs
Low: crm_mon: Add failures to --as-xml output
Christophe Vu-Brugier (2):
Build: Do not add /usr/include to the include path on Linux
Build: Do not add ${prefix}/include/heartbeat to the include path when cross-compiling
David Vossel (845):
Low: pengine: cl#5025 - Automatically clear failures when resource configuration changes.
Low: pengine: cl#5025 - Regression tests for failure cleanup on rsc config change
Low: pengine: cl#5025 Regression tests for automatic failure cleanup on restart due to rsc config change
Low: Build: rhbz#805147 - Use default value for HB_DAEMON_DIR define when clusterglue does not provide one.
Low: pengine: cl#5045 - Better handling of interrupted migrations.
Low: pengine: cl#5045 - Changes to pengine tests to reflect the new ability to perform partial migrations.
High: crmd: cl#5051 - Fixes file leak in pe ipc connection initialization.
High: crmd: cl#5053 - Fixes fail-count not being updated properly.
Low: cts: cl#5047 - Suppress cts badnews output related to fencing when fencing is expected.
Low: tools: rhbz#754216 - Make crm_mon --as-xml option default to --one-shot mode
Medium: pengine: cl#5058 - Fixes issue with on-fail option on operation being ignored.
Low: pengine: cl#5058 - Regression tests for on-fail operation fix
Low: pengine: cl#5056 - Do not allow order dependencies between stonith and non-stonith resources.
Low: pengine: cl#5056 - Regression tests for stonith order constraints
Medium: pengine: cl#5054 - Ensure demote occurs before stop of multistate resource in Master role.
Low: pengine: cl#5054 - Regression tests for demote before stop action for rsc in Master role.
Low: pengine: cl#5045 - Fix log results when partial migration takes place.
High: IPC: Check for failures when creating the server
Low: crmd: cl#5064 - Loss of cpg in process list should be reflected in node status.
Low: crmd: cl#5063 - Properly remove recurring status from cib when cancelled.
High: core: Fixes race condition in mainloop add/delete fd wrapper functions.
High: services: Introducing the services library.
High: lrmd: Introducing the new LRMD daemon and client library
High: lrmd: Introducing lrmd regression tests.
High: crmd: Enable use of new lrmd daemon and client library in crmd.
High: pengine: cl#5069 - Honor 'on-fail=ignore' even when operation is disabled.
Low: pengine: cl#5069 - Regression tests for on-fail=ignore
High: crmd: cl#5068 - Fixes crm_node -R option so it works with corosync 2.0
Low: crmd: Removes unnecessary error message from lrmd event processing.
Low: lrmd: Removes remaining linkage to cluster glue lrm .so
Low: Fencing: Fixes regression in recent change to run_stonith_agent().
High: fencing: If multiple fencing devices are present, don't stop with the first one if it fails to fence the node.
High: cib: cl#5026 - Synced cib updates should not return until the cpg broadcast is complete.
Medium: cts: Remove cts dependency on lrmadmin
Low: lrmd: Indention fixes for lrmd, tabs to spaces.
Low: lrmd: Fixes issues reported by coverity
Low: lrmd: Indention fixes for lrmd, tabs to spaces.
Low: lrmd: Lrmd api error code to string function
Low: lrmd: Adds ability to run lrmd test outside of source tree.
Medium: services: Fixes memory leaks and freeing uninitialized memory errors
High: lrmd: cl#5072 - Keep notify ops out of CIB.
High: pengine: cl#5072 - Fixes monitor op stopping after rsc promotion.
High: pengine: cl#5072 - Fixes pengine regression test failures
Low: pengine: cl#5072 - Updates remaining failing pengine regression test for recent changes
Low: pengine: cl#5072 - New regression test to ensure master monitor op starts
High: pengine: cl#5072 - Fixes pengine regression test failures
Low: pengine: cl#5072 - Updates remaining failing pengine regression test for recent changes
Low: pengine: cl#5072 - New regression test to ensure master monitor op starts
High: pengine: Fixes use of colocation scores other than +/- INFINITY
Low: pengine: Updates regression test scores for correct floating point calculation
Low: pengine: New regression test targeted at colocation floating point calculations
High: pengine: cl#5070 - Allow influence of promotion score when multistate rsc is left hand of colocation
Low: pengine: cl#5070 Updates a regression test to reflect colocation promotion score influences
Low: pengine: cl#5070 - New regression tests for multistate rsc colocation changes
Low: pengine: cl#5075 - Indicate in log output when demote/promote is blocked.
High: fencing: Fixes stonith_admin -o and -e options.
Low: fencing: Removes mixture of tabs and spaces in stonith_admin code.
Low: fencing: Fencing command.c and remote.c remove mixture of tabs and spaces
Low: fencing: Removes mixture of tabs and spaces from the python agent fence_true
Low: fencing: Fixes python dictionary usage error in fence_true
Low: fencing: Support fake output for 'list' action in fence_true
Low: fencing: Fixes fence_false indention, hashtable usage, and 'list' action operation
High: fencing: Fixes use of stonith client 'list' and 'status' agent commands
Low: fencing: Removes mixture of tabs and spaces from fencing header files.
High: fencing: Replaces stonith client api call() functionality with a targeted functions.
Low: fencing: Removes mixture of tabs and spaces from fencing/main.c
Low: fencing: Fixes mixture of tabs and spaces in fencing/test.c
Low: fencing: Adds '-t' option to stonith-test to sanity test client api using fence_true/fence_false
Low: fencing: Modifies stonith-test output so it won't conflict with fencing regression.py
High: fencing: Introducing focused regression test suite for stonith component.
High: fencing: Addition of stonithd -c option for standalone with cpg membership.
Low: fencing: Adds stonith cpg and standalone tests to fencing regression tests
Low: doc: Update TODO list to reflect recent changes with stonith testing.
Medium: fencing: Spread fencing timeout across all fencing devices in use.
Low: fencing: Fencing regression tests for fencing timeout distribution across devices.
Low: lrmd: removes unused code from lrmd regression test
Low: lrmd: Removes unused code from lrmd/test.c
High: fencing: cl#5073 - Add 'off' as an valid value for stonith-action option.
Low: fencing: Use terminate() instead of kill() in regression test cleanup.
Low: lrmd: Use terminate() instead of kill() in lrmd regression.py cleanup function.
Low: doc: Update TODO list reflect changes to recent test coverage
Low: lrmd: Add test coverage for lrmd disconnect and delete client api function.
Low: lrmd: New regression test covering duplicate rsc registrations
Low: lrmd: Test coverage for only sending notifications to lrmd client originating an action.
Low: lrmd: Test coverage for systemd and upstart metadata
Low: fencing: Add test coverage for removing topology levels
Low: doc: Remove test coverage TODO for clone_rsc_colocation_lh(), that function should never get called
High: pengine: rsc id trace log filter using PCMK_trace_tags variable
High: logging: Fixes filtering trace output by tag.
Low: pengine: Fixes mock pengine test failure, install *.stderr files to test directory.
High: fencing: rhbz#801355 - Merge fence requests for identical operations already in progress.
High: fencing: rhbz#801355 Report fencing operations external of pacemaker to cib
Low: fencing: Fixes memory leak in stonith notify event callback
High: fencing: Fixes usage of synced fence operation required stonith_admin and fence_pcmk resource.
Low: fencing: Updates stonith regression test log pattern so it passes
High: lrmd: cl#5090 - Do not block stonith monitor actions
High: lrmd: cl#5092 Fixes timeout value used when monitoring stonith resources
High: fencing: rhbz#801355 - Abort transition on DC when external fencing operation is detected
Low: lrmd: Adds -1 exit code for regression test on failure.
Low: fencing: Adds -1 exit code for regression test on failure.
Low: tools: Suppress stderr output on success
High: fencing: Return cached dynamic target list for busy devices.
High: fencing: rhbz#801355 - Fixes stonith op result cpg broadcast.
Merge pull request #132 from davidvossel/stonith_bcast_fix
High: fencing: cl#5093 Stonith per device timeout option
Low: fencing: Regression tests for stonith per device timeouts
Low: lrmd: Regression test for lrmd stonith connection failure
Low: lrmd: Increase timeout period for start/stop commands in regression test.
Low: fencing: Prevent stonith client crash due to registration of non-existent agent
Low: lrmd: Regression test for Stonith start timeout
Low: fencing: Autogen corosync.conf for stonith regression test
Low: fencing: Regression test options for running only tests not requiring corosync.
High: lrmd: cl#5094 - Immediately report monitor errors for all stonith devices when lrmd's stonith connection fails.
Low: fencing: Regression test updates for cl#5094
High: fencing: Fixes regression in stonith remote operations from recent cluster connection changes
Merge pull request #138 from davidvossel/stonith_cluster_uname_fix
Low: ipc: crm_ipc_send() cleanup
High: ipc: Stop waiting for ipc reply on server disconnect.
High: fencing: cl#5092 - Always timeout stonith operations if timeout period expires.
Low: fencing: Add stonith op return code string representation to log output.
Low: fencing: Remove xmlNode pointers from stonith callback function
Low: doc: Update todo list
Merge pull request #140 from davidvossel/stonith_callback_cleanup
High: fencing: Always broadcast the origin node with fencing results.
Merge pull request #142 from davidvossel/stonith_bcast_origin_node
Merge pull request #141 from davidvossel/stonith_signal_handling
Low: fencing: stonith_agent_run() code refactoring
Low: fencing: Move async_command struct out of header file
High: fencing: Guarantee non-blocking when fetching stonith metadata
High: fencing: Only allow node owning the fencing action to generate remote fencing operations to other nodes.
Low: fencing: Update fence_true to be able to perform reboot operation
Low: fencing: Add simple regression test for stonith reboot operation.
Low: fencing: Fixes memory leak in stonith event notifications
Low: fencing: Convert stonith_admin to use async fencing commands
Low: fencing: Have stonith_admin watch fencing notifications
Low: fencing: Add action to stonith notification events
Low: cts: Suppress an error in Stonithd test.
Low: lrmd: Fixes memory leak in lrmd client api
Merge pull request #146 from davidvossel/lrmd_mem_leak
High: lrmd: Cancel of recurring ops is now implied by rsc stop action.
Low: lrmd: Regression tests for stop implies cancel
Medium: lrmd: The lrmd op call id generation now happens on the lrmd daemon instead of the client.
Medium: lrmd: Drop recurring operations in lrmd when crmd disconnects.
Low: fencing: Stonith log message improvements
Low: lrmd: Flag for notify recurring operation changes only
Low: lrmd: Regression tests for notify recurring operation changes only
Medium: lrmd: Move logic to ignore recurring operations without changes from crmd to lrmd
Merge pull request #148 from davidvossel/stonith_log_improvements
Low: cpg: Added ability to send to specific node in cpg group by node id
Low: doc: Update TODO list
Medium: fencing: Prevent processing of timed out late fencing result
Low: fencing: Reply properly to fence query results initiated from stonith api.
Merge pull request #153 from davidvossel/cfs_pcs_update
Merge pull request #154 from davidvossel/cfs_pcs_update
Merge pull request #155 from davidvossel/cfs_pcs_update
Low: fencing: Stonith regression test now only starts/stops corosync once for cpg tests.
Low: fencing: Regression tests targeting merging of remote fencing operations
Merge pull request #156 from davidvossel/stonith_merge_tests
Low: fencing: remote_op_done refactoring
Low: fencing: Move merging of top level fencing operations into separate function
Low: fencing: Note about remote op state enum requiring values to be in order
Low: fencing: Remove access of remote_op_list in fencing/commands.c
Low: fencing: Remove overloaded use of remote op timers that signify things other than a timer id
Low: fencing: Function to find alternate_host when self fencing request is detected
Low: fencing: Give stonith request and replies separate code paths
Merge pull request #158 from davidvossel/stonith_refactoring_cleanup
High: fencing: Do not start fencing until entire device topology is found or query results timeout.
Low: fencing: Regression tests for incomplete fencing topologies
Low: fencing: documentation of remote op struct
Low: fencing: Separate confusing usage of the term 'remote' as both op id and remote peer
Low: fencing: remote_op_done documentation
Merge pull request #160 from davidvossel/stonith_refactoring_cleanup
Low: fencing: Expose function to add stonith device internally similar to the add function
Medium: fencing: Stonith registers devices from the cib by watching for cib updates
Medium: fencing: Introducing the concept of verified stonith devices.
High: fencing: Avoid remote operations delegating node suicide when at all possible.
Medium: fencing: Properly supplies stonith provider attribute when building devices from cib
Medium: lrmd: Now that all stonith devices are registered everywhere, lrmd keeps up with what stonith devices it 'started'.
Low: fencing: Put bounds on length of stonith device stdout that is logged
Low: fencing: Relax condition for disabling list queries
Low: fencing: All topology devices are not found unless they are present on nodes other than the target
Medium: fencing: Fixes race condition in add/remove of stonith devices from cib.
Low: fencing: LRMD 'stopped' stonith devices now lose their 'verified' status
Low: fencing: Fixes stonith_admin crash when agent's metadata cannot be retrieved
Merge pull request #162 from davidvossel/master
High: fencing: Support 'on_target' option in fencing device metadata for forcing unfence on target node
Low: fencing: Make fence_false fail during unfencing
Low: fencing: Unfencing regression tests
Low: fencing: Add 'on_target' option parsing log message to regression tests
Merge pull request #163 from davidvossel/stonith_unfence
Low: lrmd: Increase timeout when starting stonith resources in lrmd regression test.
Merge pull request #165 from davidvossel/master
Merge pull request #166 from gao-yan/pengine-mem-leak
High: fencing: Retry stonith device for duration of action's timeout period.
Low: fencing: Regression tests for action retry until timeout expires
High: fencing: On failure, only try a topology device once from the remote level.
Medium: Low: Fixes race condition caused by stonith modifying devices from cib changes that the lrmd has also registered.
High: fencing: Use async request/response when searching for capable fencing devices.
Low: fencing: Regression tests targeting async search of capable fencing devices
Low: cts: Remove FencingDup from CTS. It is no longer needed.
Low: fencing: Make sure we correctly set the number of items returned in the query
High: pengine: cl#5111 - When clone/master child rsc has on-fail=stop, insure all children stop on failure.
Low: pengine: Regression test changes for cl#5111 fix
High: mcp: Re-attach to existing pacemaker components when mcp fails
High: fencing: Terminate stonithd when cib connection fails.
Low: fencing: Fixes issue with param list in standalone config
Merge pull request #175 from davidvossel/master
Low: cts: Make sure fencing level id always starts with a letter.
Low: fencing: Downgrade cib disconnect error msg to notice message since it occurs at pacemaker shutdown.
Merge pull request #178 from davidvossel/cts-fixes
Merge pull request #179 from inouekazu/add_fencing_max_retries
Low: fencing: Fixes glib assert caused by lookup on NULL hash table
Merge pull request #183 from davidvossel/stonith_retries_fix
Medium: fencing: Accept on_target=1 or on_target=true in fence agent metadata for consistency with other options
Merge pull request #184 from davidvossel/master
Low: fencing: Make use of crm_is_true for on_target attribute
Merge pull request #185 from davidvossel/master
Low: fencing: Store agent metadata xml output on device to avoid multiple lookups
High: fencing: Automatically append 'nodeid' to fencing agent args if it is detected agent supports 'nodeid' parameter
Low: fencing: Regression tests for appending nodeid parameter
Low: cts: MaintenanceMode test
Low: fencing: Random sleep interval option for fence_true/fence_false
Low: cts: Add random sleep range to fake fencing agents
Medium: pengine: Fixes crash caused by enabling trace logging in pengine
Merge pull request #190 from davidvossel/trace_rsc_fail
Merge pull request #188 from davidvossel/cts_fence_random
Merge pull request #186 from davidvossel/stonith_scsi_agent_support
Merge pull request #189 from gao-yan/clone-from-template
Merge pull request #182 from inouekazu/bugfix
Low: doc: Add libqb-devel to documented build dependencies
Merge pull request #191 from davidvossel/master
High: tools: Fixes crm_mon crash when using snmp traps.
Medium: pengine: Fixes crash caused by enabling trace logging in pengine
Low: fencing: Random sleep interval option for fence_true/fence_false
Low: cts: Add random sleep range to fake fencing agents
Low: fencing: Store agent metadata xml output on device to avoid multiple lookups
High: fencing: Automatically append 'nodeid' to fencing agent args if it is detected agent supports 'nodeid' parameter
Low: fencing: Regression tests for appending nodeid parameter
Low: doc: Add libqb-devel to documented build dependencies
Low: fencing: Fixes regression test related to max retries of query
Merge pull request #193 from davidvossel/master
Medium: fencing: Query timeout value was not being read correctly by daemon
Merge pull request #194 from davidvossel/master
Merge pull request #187 from davidvossel/cts_maintenance_mode
Merge pull request #192 from davidvossel/crm_mon_snmp_crash
Merge pull request #197 from gao-yan/debug-format
High: cib: Remove text nodes from cib replace operations
Merge pull request #198 from davidvossel/master
High: Fencing: Only try peers for non-topology based operations once
Merge pull request #206 from davidvossel/master
Low: core: Fixes memory leak in logging discovered by valgrind.
low: lrmd: Log action at execution and completion.
Merge pull request #207 from davidvossel/lrmd_logging_improvement
Merge pull request #208 from davidvossel/master
Low: lrmd: Downgrade start/finish log tags from NOTICE to INFO
High: core: Internal tls api improvements for reuse with future LRMD tls backend.
Low: cib: Remove stderr message from being displayed on successful remote cib connection
High: core: Addition of PSK encryption helper functions for use with secure remote LRMD connections.
High: lrmd: Remote tls support for LRMD.
Low: lrmd: Regression tests for remote lrmd
Merge pull request #209 from davidvossel/lrmd_tls_init
High: crmd: Allow crmd to manage multiple lrmd connections
High: crmd: Introducing remote lrmd connection resources
Merge pull request #212 from davidvossel/connection_resources
Merge pull request #214 from davidvossel/master
Low: crmd: Fixes merging error in connection resource code
Low: tools: Allows pengine regression tests to run again after recent xml2string change.
Merge pull request #215 from davidvossel/master
Merge pull request #218 from davidvossel/master
Low: core: Fix compile errors when gnutls-devel is not installed
Merge pull request #219 from davidvossel/master
Low: lrmd: Automatically create authkey for lrmd regression tests when no authkey exists
Medium: fencing: Don't start fencing op until a promising query result has arrived
Medium: fencing: Properly calculate fencing operation timeout value in broadcast mode
Low: ipc: Fixes ipc compression assert to assert the correct value
Low: lrmd: Remote lrmd resource agent script
Merge pull request #220 from davidvossel/remote_ra_doc
Low: pengine: Refactor op digest calculation and comparison
High: pengine: cl#5025 - Automatically clear failcount for start/monitor failures after resource parameters change
Low: pengine: Regression tests to verify failcount is cleared for start failures after attr change
Low: pengine: Update pengine regression test for new failure cleanup change
Merge pull request #217 from davidvossel/failcount_cleanup
Merge pull request #223 from JamesGuthrie/master
High: pengine: Refresh after delete action is no long required.
Low: pengine: Update to pengine regression tests for refresh change
High: pengine: rhbz#902459 - Remove rsc node status for orphan resources
Low: pengine: Update regression tests for deletion of orphan resources
Merge pull request #226 from davidvossel/status_deletion
low: crmd: Report remote agent's monitor interval in callback result
Low: crmd: Cleanup remote agent callback logic
Medium: crmd: Retry lrmd remote connection for duration of timeout period
Merge pull request #227 from davidvossel/remote-lrmd-callback-issue
High: lrmd: Ensure async connect timesout correctly
Merge pull request #230 from davidvossel/master
Low: pengine: Refactor pengine node creation
High: pengine: Introducing the whitebox container resources feature
Container resource start/stop/move regression tests
Merge pull request #232 from davidvossel/whitebox-container
Low: crmd: Fixes assert failure on shutdown
Merge pull request #233 from davidvossel/master
low: tools: Fix memleak in crm_simulate
Merge pull request #234 from davidvossel/master
High: pengine: Imply completed stop actions for resources living in whitebox container after container stops.
Low: pengine: Regression tests for whitebox container failure cases
Merge pull request #235 from davidvossel/whitebox-container
High: lrmd: Systemd unit file for standalone remote lrmd
low: spec: Move lrmd and lrmd unit file to new package
Merge pull request #238 from davidvossel/lrmd_service
High: crmd: Allow remote nodes to have transient attributes
High: lrmd: Allow remote lrmd to proxy cib ipc requests to host
High: lrmd: Allow remote lrmd to proxy crmd and attrd requests to host
Low: lrmd: Cleanup of ipc passthrough after code review
Merge pull request #243 from davidvossel/lrmd_cib_tunnel
Feature: Pacemaker Remote Daemon for extending pacemaker functionality outside corosync cluster.
Low: remote: Package sysconfig file with pacemaker_remote
High: remote: Add stonith ipc passthrough to pacemaker_remote
High: remote: Pacemaker remote protocol versioning
Low: remote: Downgrade timeout error msg to debug
Low: spec: pacemaker does not require pacemaker_remote package
Low: sysconfig: Pacemaker remote environment variables
Low: remote: Code review cleanup
Merge pull request #251 from davidvossel/pacemaker_remote
Medium: pengine: Avoid attempting to nest a remote-node's container within another remote node
Low: pengine: More remote-node regression tests
Merge pull request #257 from davidvossel/pacemaker_remote
Low: pengine: Add ability to set timeout value for remote-node connections
Low: pengine: Give remote connections a reasonable start timeout
Merge pull request #260 from davidvossel/pacemaker_remote
Low: pengine: Regression tests for whitebox updates
Merge pull request #262 from davidvossel/pacemaker_remote
Merge pull request #265 from davidvossel/pacemaker_remote_docs
High: doc: Pacemaker Remote deployment and reference guide
Doc: low: Pacemaker Remote doc cleanup
Merge pull request #266 from davidvossel/pacemaker_remote_docs
Low: doc: Documention of 'require-all' option for "OR" logic in ordered sets
Low: remote: Lower async tcp connect poll interval for faster connection time
Low: remote: Add pacemaker_remote to lrmd regression tests
Low: remote: Refactor autogen authkey during regression test
Low: remote: Properly clean up pacemaker_remoted services on regression test startup
Merge pull request #269 from davidvossel/regression_test_pcmk_remote
Low: test: Add pacemaker_remote component test to buildbot tests
Low: lrmd: Regression test detects state directory during build
Low: fencing: fixes log msg
Low: fencing: Updates to stonith regression test
Merge pull request #272 from davidvossel/regression_test_updates
Merge pull request #271 from gao-yan/crm_mon_time
Merge pull request #258 from gao-yan/container-schema
Low: Doc: Update pacemaker_remote supported version
Low: doc: Add remote-node resource options to Pacemaker Explained
Merge pull request #278 from davidvossel/pacemaker_remote_docs
Low: crmd: Fixes memleak in failure cleanup
Low: remote: Fixes memleak in internal remote resource agent
Low: remote: Fixes memleak in async TCP connect
Low: remote: Fixes invalid read of freed memory
Low: crmd: Fixes hash table assert when resetting stonith failures
Merge pull request #280 from davidvossel/pacemaker_remote_memtest
Low: doc: fixes typo
Low: doc: Future plans for pacemaker-remote daemon
Low: doc: Revision update
Merge pull request #291 from davidvossel/pacemaker_remote_docs
Low: Doc: Pacemaker Remote doc fixes and updates
Merge pull request #293 from davidvossel/pacemaker_remote_docs
High: remote: Workaround for inconsistent tls handshake behavior between gnutls versions
Merge pull request #299 from davidvossel/tls_handshake_fix
Low: lrmd: Fixes service and lsb regression tests from conflicting with systemd
High: lrmd: Make lrmd and pacemaker_remoted in charge of setting up rsc state folder
Low: lrmd: Regression test to verify rsc tmp directory creation and use
Merge pull request #302 from davidvossel/lrmd_test_fixes
Low: lrmd: Fixes regression test expected return codes
Merge pull request #305 from davidvossel/master
Low: crmd: Self contain stonith_failures hash table to access functions
Low: fencing: Return -ENODEV for failed fencing operations when no devices are found
Medium: crmd: Stop trying to fence when no devices exist capable of fencing a node
Fix: attrd: Fixes deleted attributes during dc election
Merge pull request #310 from davidvossel/attrd_del_dc_election
Fix: remote: Init remote-node status when integrated
Merge pull request #312 from davidvossel/remote-node-status
High: crmd: cl5164 - Do not let remote-nodes enter crmd cluster node cache.
Merge pull request #323 from davidvossel/cl5164
Merge pull request #324 from davidvossel/master
High: crmd: cl5164 - Fixes crmd crash when using pacemaker-remote
Low: lrmd: Simplify lrmd client's usage of stonith api
Merge pull request #325 from davidvossel/master
Merge pull request #327 from davidvossel/master
High: crmd: Properly handle recurring monitor operations for remote-node agent
High: pengine: cl5164 - Fixes pengine segfault when calculating transition with remote-nodes.
Low: pengine: cl5164 - Remote connection resources always opt-in to cluster.
Low: pengine: cl5164 - Regression tests
Merge pull request #329 from davidvossel/master
High: mainloop: Force removal of ipc fd in dispatch_del callback
Low: mainloop: Hand the channel off to g_io_watch and forget it.
Merge pull request #330 from davidvossel/master
Medium: cts: Add lxc autogeneration tool for pacemaker remote testing
Medium: cts: Addition of Pacemaker Remote simple cts test
Low: cts: Keep ping resources off lxc containers
Low: cts: Make pacemaker_remote CTS tests optional, off by default
Low: cts: Update ignore logs list for RemoteSimple
Allow remote-nodes to be cached
We no longer have to care if crm_peer_uname is used with cluster or remote nodes
Low: remote: Properly handle expected late replies when TLS transport is in use.
Merge pull request #340 from davidvossel/pcmk_remote_cts
Fix: Invalid formatting of log message causes crash
Merge pull request #341 from davidvossel/master
Low: cts: Added environment verify option to lxc_autogen.sh
Low: cts: Add ability to do a clean restore to lxc_autogen.sh
Low: cts: Verify environment can run pacemaker_remote tests, skip test if it can't
Low: cts: Wait for stable cluster before exiting RemoteSimple test
Low: cts: Add multi-state resource to pacemaker_remote test
Merge pull request #343 from davidvossel/pcmk_remote_cts
Fix: pengine: Order demote/promote of resources on remote nodes to happen only once the connection is up
Low: pengine: Remote multistate ordering pengine regression test
Low: cts: Don't make noise in lxc_autogen when ping resource constraint can't be applied
Merge pull request #344 from davidvossel/pcmk_remote_cts
Low: remote: Don't remove hash table items during iteration
Low: remote: Downgrade proxy relay failure to a debug msg
Low: cts: Verify libvirtd is up before running pacemaker remote tests.
Low: cts: Don't remove lxc config files during pcmk remote tests
Low: cts: Distribute lxc config files to all nodes in cluster so monitors do not fail
Merge pull request #347 from davidvossel/pcmk_remote_cts
Fix: tools: Fixes formatting of remote-nodes in crm_mon and crm_simulate
Low: crmd: Cache baremetal remote-nodes
Low: pengine: Strip 'remote-' prefix from baremetal remote-node attributes so attributes can be consistent between baremetal and resource remote-nodes
Merge pull request #348 from davidvossel/baremetal_remote
High: pengine: Allow fencing of baremetal remote nodes
High: crmd: Remote-node fencing glue
High: pengine: Probe resources on baremetal remote-nodes
Low: pengine: Updates to regression tests to reflect probing remote-nodes
Low: crmd: Clear remote-node status AFTER start completes, not before.
Low: pengine: Refactor how we detect remote-nodes to use helper functions
Low: pengine: Initial baremetal remote-node PE regression tests
Low: cts: Baremetal pcmk remote cts test
Low: crmd: Populate remote peer cache when building new cib
Fix: pengine: Properly shutdown orphaned remote connection resources
Low: pengine: Orphaned remote resources pengine test
Low: cib: Prevent potential double free in get_remote_uuid
Merge pull request #350 from davidvossel/master
Low: pengine: Do not attempt to probe resources on remote-nodes that are shutting down
High: pengine: Map orphaned container fillers to container resource during shutdown to preserve ordering
Low: pengine: Orphaned whitebox container pengine regression tests
High: pengine: Properly handle orphaned multistate resources living on remote-nodes
Low: pengine: Regression tests for orphaned ms resources living on remote-node
Low: cts: Update expected error strings in RemoteSimple test
Merge pull request #351 from davidvossel/baremetal_remote
Low: cts: Autogenerate pcmk remote key for cts tests
Low: cts: Add RemoteBaremetal to standard test set
Merge pull request #354 from davidvossel/pcmk_remote_cts
High: attrd: Give remote-nodes ability to set attributes with attrd
High: crmd: Tell attrd when remote-node is disconnected
Merge pull request #355 from davidvossel/master
High: pengine: Probe container nodes
Low: pengine: Regression tests updated for container node probes
Low: cts: Hide stderr from probes in lxc container test
low: tools: Detect orphaned remote-nodes when setting attributes
Low: attrd: fixes log message typo
High: attrd: Avoid infinite write loop for unknown peers
Merge pull request #358 from davidvossel/container-probes
Low: cts: lxc_autogen.sh fixes
Merge pull request #361 from davidvossel/master
Merge pull request #363 from gao-yan/service-monitor-cancel
Feature: remote: Option to enable/disable probes on container resource remote-nodes
Low: pengine: Disable container probes regression test
Low: fencing: Update return codes for regression tests
Merge pull request #364 from davidvossel/container_probes_option
Low: fencing: Do not broadcast suicide if the on action is being executed
Low: fencing: Update fencing test to use fence_dummy agent
Low: fencing: Update regression test-suite to use fence_dummy
Fix: fencing: Allow fencing for node after topology entries are deleted
Low: fencing: Regression test for fence after all topology entries are removed
Merge pull request #365 from davidvossel/fencing_regression_test_update
Low: cts: Store lxc log files in /var/log/pacemaker_remote_lxc*.log
Merge pull request #366 from davidvossel/lxc_logging
Low: cts: Increase lxc ram size to 196mb and mount dedicated /dev/shm using 147mb
Merge pull request #367 from davidvossel/master
Low: pengine: Do not probe for connection resources
Low: pengine: Update regression tests to reflect not probing connection resoures
Merge pull request #368 from davidvossel/master
Low: ipc: detect new libqb api function for retrieving server client connection buffer sizes
High: ipc: Compress msgs based on client's actual max send size
Merge pull request #370 from davidvossel/ipc_buf_size_detection
Merge pull request #371 from davidvossel/remote-node-ipc-proxy-fixes
Low: crmd: Using PCMK_ipc_buffer variable for remote-node proxy
Low: crmd: Cache rsc_info retrieved from lrmd and pacemaker_remoted
Low: crmd: Pad internal lrmd rsc_info and metadata retrieval timeout
Merge pull request #373 from davidvossel/remote-node-ipc-proxy-fixes
Merge pull request #376 from gao-yan/crm_resource-reply
Fix: fencing: Deep copy current topology level list on remote op
Merge pull request #378 from davidvossel/master
Fix: ipc: Ensure buffer environment variable is read before setting max buffer len
Fix: ipc: Client buffer size must never fall below our max read size
Low: mainloop: Don't allow client buf size to be determined by hardcoded value given to mainloop
Merge pull request #377 from davidvossel/ipc_buf_size_detection
High: doc: Baremetal remote-node documentation and tutorial
Merge pull request #381 from davidvossel/pacemaker_remote_docs
High: remote: Add support for ipv6 into pacemaker_remote daemon
Merge pull request #386 from davidvossel/ipv6_pcmk_remote
Fix: Probe for baremetal remote-node connection resources.
Low: pengine: Regression test update for baremetal connection rsc probes
Merge pull request #387 from davidvossel/fix-baremetal-probes
Merge pull request #390 from davidvossel/ipc_server_buffer
High: ipc: Have the ipc server enforce a minimum buffer size all clients must use.
Low: cib: We can now guarantee all clients will have at least a minimum size event buffer
High: pengine: Disable container node probes due to constraint conflicts
High: pengine: Recover unexpectedly running container nodes.
Low: pengine: Regression tests for recovering unexpectedly running container nodes
Merge pull request #391 from davidvossel/rm-container-probes
Fix: Removes unnecessary newlines in crm_resource -O output
Merge pull request #394 from davidvossel/bz1038155_crm_resource_newline
Merge pull request #396 from inouekazu/master
Merge pull request #395 from xrobau/patch-1
Merge pull request #400 from gao-yan/install-cib_xml.py
Low: Avoid potentially misleading and inaccurate compression time log msg
Fix: pengine: Fixes memory leak in regex pattern matching code for constraints.
Merge pull request #402 from davidvossel/pcmk_cleanup
Fix: ipc: fix memory leak for failed ipc client connections.
Merge pull request #409 from gao-yan/crm_ticket-tests
High: cl#5055: Improved migration support.
Low: updated pengine regression tests for improved migration support
Low: new migration logic advanced test cases
Merge pull request #401 from davidvossel/new_migration_logic
Low: fencing: Immediately fail remote fencing operation on peer timeout
Medium: pengine: Correctly perform partial migrations when node's uname is not equal to the id.
Merge pull request #411 from davidvossel/master
Merge pull request #415 from gao-yan/remove-node-segfault
High: pengine: cl#5186 - Avoid running rsc on two nodes when node is fenced during migration
High: pengine: Force record pending for migrate_to actions
Low: pengine: Update regression tests for record pending on migrate_to
Low: pengine: Regression test to handle partial migration with src node losing membership
Merge pull request #416 from davidvossel/new_migration_logic
High: services: Do not block synced service executions
Merge pull request #421 from davidvossel/service_nonblock
Merge pull request #423 from gao-yan/remote-cib-tls
Low: pengine: Free memory allocated by find_actions()
Merge pull request #424 from davidvossel/master
Merge pull request #426 from gao-yan/xml-segfault
Low: lrmd: Send event alerting estabilished clients that a new client connection is created.
High: remote: Allow baremetal remote-node connection resources to migrate
Low: tools: Return error when attempting to delete an unspecified attribute
Merge pull request #429 from davidvossel/new_migration_logic
Low: remote: Remove redundant code
Medium: remote: Fail remote-node connection rsc monitor ops early on unexpected disconnect
Merge pull request #430 from davidvossel/new_migration_logic
Low: doc: Fix pacemaker_remote baremetal documentation index table.
Low: doc: Document use of pcs to configure corosync config
Low: doc: Update distro documented for pacemaker_remote
Low: doc: Update pacemaker_remote future features documentation
Merge pull request #431 from davidvossel/doc_updates
Low: services: Keep dbus build support optional
Low: services: Fix building dbus
Low: lrmd: Make ocf dummy agents executable with regression test in src tree
Low: lrmd: Install ping agent during lrmd regression test.
High: pengine: cl#5186 - Avoid running rsc on two nodes when node is fenced during migration
High: pengine: Force record pending for migrate_to actions
High: services: Do not block synced service executions
Low: pengine: Regression test update for record-pending=true on migrate_to
Low: services: fix building dbus support in buildbot.
Low: services: Keep dbus build support optional
Low: services: Fix building dbus
Low: services: fix building dbus support in buildbot.
Low: cts: '-A allow-anywhere' option for lxc_autogen
Low: cts: lxc_autogent distribute configs in batchmode
Merge pull request #433 from davidvossel/lxc_updates
Low: changelog: Update changelog for 1.1.11 release
Low: changelog: Update changelog for 1.1.11 release
High: controld: handling startup fencing within the controld agent, not the dlm
Fix: controld: Do not consider the dlm up until the address list is present
Merge pull request #435 from davidvossel/controld_fixes
Low: lxc_autogen: Fix ram libvirt bug workaround now that libvirt packages work
Fix: stonith_api_time_helper now returns when the most recent fencing operation completed
Low: pengine: Performance increase, stack allocated score2char
Merge pull request #447 from davidvossel/master
High: remote: Enable migration support for baremetal connection resources by default
Low: crmd: Log what node lrm processes executed on
Low: cts: Verify rsc can be placed on baremetal remote-node
Low: cts: make sure pacemaker_remote is stopped before starting cts
Low: pengine: Regression test updates
Merge pull request #452 from davidvossel/master
Merge pull request #456 from lge/for-beekhof
Fix: crmd: Ack pending operations that were cancelled due to rsc deletion
Low: crmd: Remove dead code and improve code comments
Fix: services: Do not allow duplicate recurring op entries
High: lrmd: Merge duplicate recurring monitor operations
Low: lrmd: recurring op merger regression tests
Low: lrmd: Warning msg to indicate duplicate op merge has occurred
Fix: lrmd: Cancel recurring operations before stop action is executed
Fix: mainloop: Fixes use after free in process monitor code
Merge pull request #457 from davidvossel/lrm-rsc-del
Fix: crmd: make lrm_state hash table not case sensitive
Low: membership: make caching peers by uname case insensitive
Low: fencing: make hashtable mapping hostname->alias case insensitive for hostname
Merge pull request #462 from davidvossel/lrm-segfault
Fix: pengine: fixes invalid transition caused by clones with more than 10 instances
Low: pengine: ordered 16 clone instance regression test
Merge pull request #465 from davidvossel/bz1078504-11node-clone
Low: controld: Remove '-q 0' from default dlm_controld arguments
Merge pull request #470 from davidvossel/controld_fixes
High: cib: allow setting permanent remote-node attributes
High: xml: Update xml validation to allow '<node type=remote />'
Low: pengine: properly handle remote-node without connection resource.
Low: cts: test setting/removing remote-node permanent attributes
Low: pengine: Do not fence remote-node entries left over in the node section that have no connection rsc
Low: pengine: Regression test for stale leftover remote-node <node> entries
Merge pull request #467 from davidvossel/remote-node-attr
Loq: crm_node: Set correct subsystem in node rm cache msg
Merge pull request #487 from davidvossel/free_correctly_vos
Merge pull request #481 from inouekazu/case_sensitive
Low: pengine: Only report 'migrate' in transition summary if migration is actually taking place
Low: pengine: update regression tests to reflect summary fix
Merge pull request #488 from davidvossel/container-node-migrate
Fix: xml.c: compile error
Low: xml: Remove 'bc' build dependency
Merge pull request #490 from davidvossel/master
Build: remove bc from spec and travis file
Merge pull request #492 from davidvossel/master
Low: fencing: Set 'on_target=true' for fence_dummy to allow testing unfencing
Fix: fencing: properly detect new fencing devices for v2 cib updates
Fix: pengine: Correctly search failcount (for real this time)
Merge pull request #494 from davidvossel/master
High: pengine: Allow container nodes to migrate with connection resource
Fix: pengine: handles edge case where container order constraints are not honored during migration
Low: pengine: regression tests for remote-allow-migrate
Merge pull request #489 from davidvossel/container-node-migrate
Fix: fencing: Do not unregister/register all stonith devices on every resource agent change
Low: fencing: We don't a notice msg alerting us to all the fencing devices on every cib rsc update.
Merge pull request #500 from gao-yan/update-version
Merge pull request #498 from davidvossel/stonith-dup-register
Merge pull request #501 from gao-yan/tags-memory-leak
Low: fencing: Make regression tests pass with new fence_dummy agent
Merge pull request #502 from davidvossel/unfence-tests
Merge pull request #503 from aspiers/fix-ordering-pdf-link
Merge pull request #504 from yuusuke/fix_stonithd
High: fencing: Fence using all required devices
High: fencing: Execute all required fencing devices regardless of what topology level they are at
High: fencing: default to 'off' when agent does not advertise 'reboot' in metadata
Low: fencing: move nodeid regression test requirements to fence_dummy
Low: fencing: remove duplicate test cases
Low: fencing: regression tests for automatic unfencing
Low: fencing: regression tests for required devices in topology
Low: fencing: default reboot to off regression tests
Merge pull request #505 from davidvossel/unfence-tests
Merge pull request #506 from jnpkrn/fencing-literals
Fix: membership: Avoid duplicate peer entries in the peer cache
Merge pull request #518 from jnpkrn/pengine-typo
Low: cts: more cts changes to allow pacemaker-2.0 schema
Fix: mainloop: Canceling in-flight operations should not fail if child process has already exited.
Low: cts: Use a more reliable way to set lxc network addresses for container node
Low: cts: Avoid asking for user permission during lxc_autogen script
Low: cts: properly detect when test environment is incapable of running RemoteLXC
Fix: pengine: avoid assert by searching for stop action on correct node during LogActions
Low: pengine: Don't report a resource as restarting when it is a moving multistate resource
Low: pengine: Update pengine regression test to reflect corrected LogActions output
Low: cts: fixes syntax error
Fix: pengine: properly place resource within a container when container is a remote-node.
Low: pengine: container is remote regression test
Fix: remote: Handle request/response ipc proxy correctly
Fix: pengine: properly handle fencing of container remote-nodes when the container is orphaned
Low: cts: reference ping resource by clone id in location constraints
Low: cts: fix how lxc_autogen handles placing resources on remote nodes
High: remote: clear remote-node status correctly
Low: pengine: regression test updates for recent container fixes
Fix: crmd: avoid double free caused by nested hash table removal
Fix: ignore SIGPIPE with gnutls is in use
Low: cts: Fix cts '--choose' option
Low: cts: sort partition details in a consistent way
High: lrmd: store failed operation exit reason in cib
High: crm_mon: add exit reason to resource failure output
Merge pull request #561 from davidvossel/err-reason-str
High: crmd: When container's host is fenced, cancel in-flight operations
Merge pull request #562 from davidvossel/bz1117341-cancel-incomplete
Merge pull request #564 from aspiers/graph-node-names
High: lrmd: pass exit reason prefix to ocf scripts as env variable
Low: lrmd: prevent \n from being printed in exit reason output
Low: services: restore SIGPIPE default behavior for child processes
Fix: lrmd: fix rescheduling of systemd monitor op during start
Merge pull request #567 from davidvossel/master
Fix: logging: correctly initialized default log file
Low: cts: fix building test array
Low: cts: properly output where the cts log is
Low: cts: fixes error when printing usage
Low: services: Refector services os execute function
High: cts: phd docker autogen support
High: pengine: fixes segfault caused by malformed log warning
High: crmd: Fixes crash when remote connection migration fails
High: crmd: Remote-node throttle jobs count towards cluster-node hosting conneciton rsc
Low: crmd: display lrm node when rsc registration fails
Low: crmd: throttling debug statement
High: crmd: properly update job limit for baremetal remote-nodes
Low: remote: advertise more reasonable timeout periods for remote-nodes connections
High: ipc: properly allocate server enforced buffer size on client
Low: lrmd: ensures systemd python package is available when systemd tests run
High: pengine: only fence baremetal remote when connection can fails or can not be recovered
High: crmd: do not remove connection resources during re-probe
High: pengine: allow baremetal remote-nodes to recover without requiring fencing when cluster-node fails
Low: pengine: clarify remote-node fencing error message
Low: pengine: regression tests for baremetal recovery when cluster-node fails
High: lrmd: cancel pending async connection during disconnect
Merge pull request #587 from davidvossel/master
High: pengine: resource discovery mode for location constraints
Low: pengine: resource-discovery regression test
High: pengine: per-node control over resource discovery
High: crm_resource: avoid deletion of lrm cache on node with resource discovery disabled.
Low: pengine: regression test for pengine per node rsc discovery
High: pengine: prevent disabling rsc discovery per node in certain situations
Merge pull request #589 from davidvossel/master
Fix: pengine: allow remote-nodes to be placed in maintenance mode
Low: cts: move custom RemoteBaremetal pattern into patterns.py
Merge pull request #596 from davidvossel/master
Low: cts: Make the remotebaremetal test reusable as the base for other tests
Low: cts: refactor fence watch patterns for pattern reuse
Fix: remote: report timestamps for remote connection resource operations
Low: remote: advertise default port correctly, further adjust default timeout periods
Low: cts: baremetal remote fencing test
Low: cts: baremetal remote stonithd ignore patterns
Merge pull request #597 from davidvossel/master
Low: cts: Remote connection migration test
Low: crm_resource: Allow -M location to be reinforced with -f even if rsc is already on node
Low: cts: RemoteRscFailure test verifies monitoring of rscs on remote-nodes
Low: cts: ignore patterns for RemoteRscFailure test
Merge pull request #599 from davidvossel/master
Low: build: Add ability to set custom spec version for 'make rpm'
High: pengine: add #kind=remote for baremetal remote-nodes
Low: pengine: regression test updates for #kind=remote|container|cluster
Fix: cib: tls sock cleanup for remote cib connections
Fix: dbus ref count leaks
Low: lrmd: systemd stress tests
Fix: dbus: handle dispatch queue when multiple replies need to be processed
Merge pull request #604 from davidvossel/master
High: pengine: exclusive discovery implies rsc is only allowed on exclusive subset of nodes
Low: pengine: updates regression tests
High: doc: documents resource-discovery location constraint option
Fix: pengine: make resource-discovery work with location rules
Low: pengine: update regression tests for rsc discovery with rules
Low: lrmd: fix regression test LSBdummy install
Merge pull request #609 from davidvossel/fix-travis
Low: doc: Make Clusters from Scratch compile correctly
Merge pull request #611 from davidvossel/master
Low: fencing: stonith -l option for custom log
Low: fencing: have regression test parse logfile for patterns
Merge pull request #612 from davidvossel/fence-test-updates
Merge pull request #603 from davidvossel/tls-cleanup
Fix: pengine: properly recover remote-nodes when cluster-node proxy goes offline
low: pengine: update regression tests for remote-node recovery
High: remote: Treat recv msg timeout as a disconnect
Fix: crmd: Handle remote connection failures while executing ops on remote connection
High: pengine: Fence baremetal remote when recurring monitor op fails
Fix: crmd: gracefully handle remote node disconnects during op execution
Low: pengine: regression test updates
Fix: crmd: Ensure remote connection resources timeout properly during 'migrate_from' action
Fix: pengine: properly handle ordering during remote connection partial migration
Low: pengine: new remote regression tests
High: pengine: prefer migration target for remote node connections
Low: pengine: regression tests for remote rsc partial migration
Fix: crmd: never stop recurring monitor ops for pcmk remote during incomplete migration
Low: pengine: do not recording pending migrate_to ops for pacemaker remote
Low: crmd: upgrade a log msg to info
Low: pengine: regression test updates
High: ipc: use server enforced buffer during ipc client send
Merge pull request #614 from davidvossel/master
Merge pull request #618 from kgaillot/meta
Merge pull request #620 from HideoYamauchi/patch-for-pacemakerd
Merge pull request #594 from HideoYamauchi/check1010
Low: doc: pacemaker-remote related typos
Merge pull request #621 from davidvossel/master
Merge pull request #619 from kgaillot/CFS
High: pengine: require-all feature for ordered clones
High: xml: Clone the latest constraint schema in preparation for changes"
High: xml: add 'require-all' to xml schema for constraints
Low: pengine: regression tests for require-all=false in order constraints
Merge pull request #623 from kgaillot/doc
Low: pengine: remove unnecessary nested loop
Merge pull request #625 from davidvossel/master
Fix: pengine: imply stop in container node resources when host node is fenced
Low: pengine: regression tests for imply container rsc stop on fence
Merge pull request #629 from kgaillot/CTS
Fix: pengine: fixes pengine crash with orphaned remote node connection resource
Low: pengine: orphaned remote regression tests
Merge pull request #630 from davidvossel/master
Merge pull request #636 from gao-yan/order-stateful-migratable
Merge pull request #638 from kgaillot/sig
Fix: ensure if B is colocated with A, B can never run without A
Low: pengine: enforce colocation fix regression test
Merge pull request #643 from davidvossel/colo-order-chain-fix
Fix: properly record stop args after reload
Merge pull request #653 from davidvossel/master
Fix: attrd: properly write attributes for peers once uuid is discovered
Merge pull request #657 from davidvossel/master
Fix: systemd: fix crash caused when canceling in-flight operation
Fix: pengine: remove unnecessary whitespace from notify environment variables
Low: pengine: regression test for master promotion with notify
Merge pull request #663 from davidvossel/clone-notify-env
Merge pull request #661 from kgaillot/doc
Feature: remote: pcmk remote client tool for use with container wrapper script
Merge pull request #662 from davidvossel/dbus-crash-fix
High: extra: docker container tech wrapper script for pcmk remote
High: pengine: ability to launch resources in isolated containers
Low: pengine: isolation regression tests
High: pengine: disable reloading of resources within isolated container wrappers
Low: pengine: regression tests for disable reload during isolation
High: pengine: disable migrations for resources with isolation containers
Low: spec: add docker-wrapper directory to spec file
Low: lrmd: properly handle poke requests in lrmd client when using ipc
Low: enable logging in remote_ctl.c
High: lrmd: enable ipc proxy for docker-wrapper privileged mode
Low: rename pacemaker_remote_ctl to lrmd_interal_ctl
Low: docker-wrapper: set authkey file permissions and properly set container 'node name' during start
Fix: docker-wrapper: properly separate docker and resource specific attributes
Low: lrmd: preserve exit reason string when isolation wrappers are in use
High: lrmd: isolation wrappers are moved from containers to isolation OCF namepace
High: pengine: pass node name of pcmk host into isolated resources
High: lrmd: do not let isolation wrappers inherit the wrapper attribute, or it makes a loop
Fix: attrd: fixes double free in attrd legacy
High: docker-wrapper: properly forward key value pairs during isolation
High: docker-wrapper: add demote, promote, and notify actions for multistate rsc
Merge pull request #675 from davidvossel/pcmk-docker-bff
Merge pull request #677 from davidvossel/attrd-double-free-fix
High: pengine: handle cloned isolated resources in a sane way
Low: pengine: update isolation rsc tests for isolation host feature
Low: pengine: regression test for cloned isolated primitive
High: pengine: handle isolated resource scenario, cloned group of isolated resources
Low: pengine: regression test update for new cloned isolated instance name
Low: docker-wrapper: only override exit reason if docker daemon is inactive
Merge pull request #678 from davidvossel/pcmk-docker-bff
Low: docker-wrapper: add reuse option for presistent containers
Low: docker-wrapper: refactor wrapper arg prefix
Low: docker-wrapper: for non priviledged mode, make sure to record stdout so exit reason is set
Low: docker-wrapper: update meta-data documentation
Low: docker-wrapper: per resource logging of lrmd and remote client tool
Merge pull request #682 from davidvossel/pcmk-docker-bff
Fix: crmd: include remote nodes when forcing cluster wide resource reprobe
Fix: make targetted reprobe for remote node work, crm_resource -C -N <remote node>
Merge pull request #689 from davidvossel/crm_resource_C
High: pengine: only kill process group on timeout when on-fail does not equal block.
High: common: Increment current and age for lib common as a result of APIs being added
High: services: Increment lib service's current version to indicate a new interface is being exported
Merge pull request #690 from davidvossel/pid-kill-fix
Merge pull request #691 from davidvossel/spec-fix
High: pengine: restart master instances on instance attribute changes
Low: pengine: update regression tests to account for promote restart fix
Merge pull request #692 from davidvossel/promote-restart-fix
Fix: mainloop: resolves race condition mainloop poll involving modification of ipc connections
High: lrmd: move isolation wrappers to hidden directory
Low: revert pengine test changes
Low: pengine: handle setting container attribute in the correct location
Fix: pengine: process rsc parameters during unpack
Merge pull request #695 from davidvossel/mainloop_poll_mod
Merge pull request #696 from davidvossel/pcmk-docker-bff
Dejan Muhamedagic (5):
Medium: RA: ClusterMon - Don't fail in stop if the process is missing (bnc#569957)
Low: RA: SysInfo: fix meta-data
Low: RA: SysInfo: fix conversion from/to MB
Medium: RA: SysInfo: min_disk_free parameter
Feature: lrmd: Support CIB secrets - allow storing parameters in local files (lf#2415, cl#5121, bnc#792140)
Dirk Mueller (1):
Check for compiler flags with -Werror
Ferenc Wágner (1):
Imported Upstream version 1.1.13
Florian Crouzat (1):
Low: doc: Document SNMP and email notifications
Florian Haas (2):
pkg-config: linker flag should be -lpe_status, not -lpe_state
pkg-config: correct includedir
Franky Van Liedekerke (1):
cman_pre_start fixes
Fu, Hu (1):
Fix: Remove repeated code
Gao,Yan (205):
Medium: corosync: Fix compilation when libqb is not available
High: crmd: cl#5057 - Restart sub-systems correctly (bnc#755671)
Low: RA: controld - Redirect the standard output and error on checking state (#bnc757660)
Low: RA: controld - Fix the typo which should be "globally_unique" (cl#5061, bnc#762603)
High: Core: Revert to the original process flags for the support of rolling-upgrade
Low: RA: ClusterMon - A minor tweak
High: ACL: Support ACLs with libqb
High: Tools: attrd - Fix compilation when "--with-acl" is disabled
Low: Build: Fix an implicit declaration if "--with-acl"
Low: Build: Check if gio supports GDBusProxy (glib >= 2.26)
High: ACL: Include required headers
High: Core: Include required header
High: Core: Include required header for glib compatibility code
High: fencing: Fix registration for Linux-HA fencing agents
Low: doc: Revise of "Multi-site Clusters and Tickets" chapter
Medium: crmd: cl#5057 - More actions for S_TRANSITION_ENGINE + I_FAIL (bnc#755671)
High: legacy: Correctly update node status when using the corosync plugin
High: crmd: Correctly handle scheduled node down events
High: PE: cl#5044 - migrate_to no longer requires load_stopped for avoiding possible transition loop
Low: PE: cl#5044 - migrate_to no longer requires load_stopped for avoiding possible transition loop (regression test)
Low: PE: Update regression tests for the load_stopped changes
High: Core: Bug cl#5105 - Re-expose API elements needed by thrid-parties
High: cib: Bug cl#5105 - Re-expose API elements needed by thrid-parties
Medium: Core: Bug cl#5105 - Define a common function for disconnecting from a cluster
High: legacy: Set to the minimum scheduling priority when using SCHED_RR policy (bnc#779259)
High: PE: Fix memory leak on processing message (bnc#780224)
Medium: PE: cl#5115 - Do not use obsolete lrm_resource sections to avoid fatal assert (bnc#786701)
Low: PE: cl#5115 - Do not use obsolete lrm_resource sections to avoid fatal assert (bnc#786701) (regression test)
Low: Tools: ptest - Fix for being able to be built with latest code
High: PE: Prevent double-free for cloned primitive from template
Low: PE: Prevent double-free for cloned primitive from template (regression tests)
Medium: Core: Supply the correct number of arguments for the format string
Low: PE: Reduce the error log to warning when on-fail is "stop" (#bnc790711)
High: PE: cl#5099 - Probe operation uses the timeout value from the minimum interval monitor by default (#bnc776386)
Low: PE: cl#5099 - Probe operation uses the timeout value from the minimum interval monitor by default (#bnc776386) (regression tests)
Low: PE: cl#5099 - Probe operation uses the timeout value from the minimum interval monitor by default (#bnc776386) (update regression tests)
High: PE: Support for resource container
Low: PE: Support for resource container (regression tests)
High: fencing: Correctly register devices from CIB for Linux-HA fencing agents (bnc#800306)
High: ACL: Fix compilation when "--with-acl" is enabled
High: lrmd: Support nagios remote monitoring
High: pengine: Process rsc_ticket dependencies earlier for correctly allocating resources (bnc#802307)
Low: pengine: Process rsc_ticket dependencies earlier for correctly allocating resources (bnc#802307) (update regression tests)
Fix: crmd: Correctly relay messages for remote clients (bnc#805626, bnc#804704)
Fix: ais: Prevent assertion failure on routing ais messages (bnc#805626)
Fix: pengine: cl#5142 - Do not delete orphaned children of an anonymous clone
Build: ACL: Fix compilation when "--with-acl" is enabled
Fix: ipc: Fix memory leak on sending message (bnc#807172)
Doc: Pacemaker Explained: Add nagios resource class
Medium: xml: Support resource container in the 1.2 schema
Medium: xml: Support nagios resource class in the 1.2 schema
Fix: legacy: Support "crm_node --remove" with a node name for corosync plugin (bnc#805278)
Medium: services: Improve the performance of synchronous actions
Fix: legacy: cl#5148 - Correctly remove a node that used to have a different nodeid
Fix: tools: crm_mon - Print a timing field only if its value is non-zero
Log: tools: cibadmin - Do not log for query commands by default
Fix: crmd: Do not update fail-count and last-failure for old failures
Fix: crmd: cl#5152 - Correctly clean up fenced nodes during membership changes
Fix: crmd: Do not record pending delete operations in the CIB
Bug: PE: cl#5155 - Block the stop of resources if any depending resource is unmanaged
Test: PE: cl#5155 - Block the stop of resources if any depending resource is unmanaged
Fix: PE: Mark unrunnable stop actions as "blocked"
Test: PE: Mark unrunnable stop actions as "blocked" (update regression tests)
High: PE: cl#5128 - Support maintenance mode for a single node
Low: PE: cl#5128 - Support maintenance mode for a single node (regression tests)
Fix: xml: Purge diff markers even if there's no digest
Fix: PE: Delete the old resource state on every node whenever the resource type is changed
Test: PE: Delete the old resource state on every node whenever the resource type is changed (update regression tests)
Fix: plugin: Change the function declaration in the plugin header
Fix: legacy: Handle messages in the common way in the plugin dispatch function by default
Fix: crmd: Prevent messages for remote crmd clients from being relayed to wrong daemons
Fix: crmd: Prevent PCMK_LRM_OP_NOT_INSTALLED for being mapped to PCMK_LRM_OP_ERROR
Feature: pengine: Distinguish between the agent being missing and something the agent needs being missing
Test: pengine: Distinguish between the agent being missing and something the agent needs being missing
Feature: cib: Support XML comments in diffs
Fix: crmd: Expose "expected-quorum-votes" option in the metadata only for plugin-based clusters
Fix: services: Fix the executing of synchronous actions
Fix: lrmd: Correctly cancel monitor actions for lsb/systemd/service resources on cleaning up
Fix: plugin: Reset the scheduling policy of daemons to SCHED_OTHER with nice value -10
Fix: crm_resource: Wait for a reply for each request to crmd
Fix: crm_master: Avoid sending redundant "--node" options to crm_attribute
Fix: election/fencing: Fix the casts that would result in incorrect values for big-endian systems
Fix: crmd: Send the first throttle update
Build: cts: Install cib_xml.py
Feature: pengine: cl#5174 - Allow resource sets and templates for location constraints
Test: pengine: cl#5174 - Allow resource sets and templates for location constraints
Feature: crm_ticket: Support multiple modifications for a ticket in an atomic operation
Test: tools: Update the regression tests of crm_ticket
Fix: services: Correctly reset the nice value for lrmd's children
Fix: fencing: Update stonith device list only if stonith is enabled
Fix: services: Reset the scheduling policy and priority for lrmd's children without replying on SCHED_RESET_ON_FORK
Fix: cluster: Fix segfault on removing a node
Feature: crm_mon: Display brief output if "-b/--brief" is supplied or 'b' is toggled
Fix: cib: Fix remote cib based on TLS
Fix: xml: Fix segfault in find_entity()
Fix: services: Correctly reset the nice value for lrmd's children
Fix: services: Reset the scheduling policy and priority for lrmd's children without replying on SCHED_RESET_ON_FORK
Fix: cluster: Fix segfault on removing a node
Fix: cib: Fix remote cib based on TLS
Fix: xml: Fix segfault in find_entity()
Fix: crmd: Prevent manual fencing confirmations from attempting to create node entries for unknown nodes
Fix: crmd: Prevent message relay from attempting to create node entries for unknown nodes
Feature: tools: Display pending state in crm_mon/crm_resource/crm_simulate if --pending/-j is supplied (cl#5178)
Fix: pengine: Prevent memory leak on handling pending state
Fix: pengine: cl#5187 - Prevent resources in an anti-colocation from even temporarily running on a same node
Test: pengine: cl#5187 - Prevent resources in an anti-colocation from even temporarily running on a same node (update regression tests)
Test: pengine: cl#5187 - Prevent resources in an anti-colocation from even temporarily running on a same node
Feature: pengine: Support per-resource maintenance mode
Test: pengine: Support per-resource maintenance mode
Build: Install compatibility.h
Fix: cib: Ensure 'cibadmin -R/--replace' commands get replies
Fix: xml: Fix segfault in TRACKING_CHANGES()
Fix: cib: Do not disable cib disk writes if on-disk cib is corrupt
Refactor: cib: Check if on-disk cib is corrupt in the separate on_disk_cib_corrupt() rather than in validate_on_disk_cib()
Feature: xml: Enable resource sets in location constraints for 1.2 schema
Log: attrd: Reduce log severity for normal messages
Log: corosync: Reduce log severity for a redundant message
Feature: pengine: Support id-ref in nvpair with optional "name"
Test: pengine: Support id-ref in nvpair with optional "name"
Fix: crmd: Correctly detect changes of ticket state
Build: Update versions in configure.ac and sample spec file
Feature: pengine: Support cib object tags
Test: pengine: Support cib object tags
Fix: pengine: Ignore failure-timeout only if the failed operation has on-fail="block"
Test: pengine: Ignore failure-timeout only if the failed operation has on-fail="block" (update a regression test)
Test: pengine: Ignore failure-timeout only if the failed operation has on-fail="block"
Feature: pengine: Support site-specific instance attributes based on rules
Test: pengine: Support site-specific instance attributes based on rules
Fix: pengine: Fix memory leak in expand_tags_in_sets()
Fix: xml: Correctly subtract xml with v1 patchsets
Fix: xml: Do not prune meaningful leaves from v1 patchsets
Feature: pengine: Support cluster-specific instance attributes based on rules
Test: pengine: Support cluster-specific instance attributes based on rules
Test: pengine: Support site-specific instance attributes based on rules for stretch cluster
Fix: xml: Do not prune meaningful leaves from v1 patchsets
Feature: crmd: Set "cluster-name" property to corosync's "cluster_name" by default for corosync-2
Fix: pengine: Use "#cluster-name" in rules for setting cluster-specific instance attributes
Fix: pengine: Use "#site-name" in rules for setting site-specific instance attributes
Fix: pengine: Block restart of resources if any dependent resource in a group is unmanaged
Test: pengine: Block restart of resources if any dependent resource in a group is unmanaged
Fix: pengine: Mark unrunnable stop actions as "blocked" and show the correct current locations
Test: pengine: Mark unrunnable stop actions as "blocked" and show the correct current locations
Fix: pengine: Prevent possible segfault
Fix: pengine: cl#5200 - Before migrating utilization-using resources to a node, take off the load that will no longer run there if it's not introducing transition loop
Test: pengine: cl#5200 - Before migrating utilization-using resources to a node, take off the load that will no longer run there if it's not introducing transition loop
Fix: crm_simulate: Clean up temporary shadow files on exit
Fix: cib: Prevent assert on sending cib sync requests
Fix: cluster: Prevent search of unames from attempting to create node entries for unknown nodes
Fix: xml: Prevent incorrect xml diffs from corrupting the cib
Fix: cib: cl#5222 - Prevent assert in xml_calculate_changes() on performing cib operations
Fix: lrmd: Prevent OCF agents from logging to random files due to "value" of setenv() being NULL
Fix: crmd: Remove unknown nodes with conflicting unames from CIB
Fix: crmd: Remove state of unknown nodes with conflicting unames from CIB
Feature: extra: Add logrotate configuration file for /var/log/pacemaker.log
Fix: cluster: Remove unknown offline nodes with conflicting unames from node cache
Fix: membership: Prevent use-after-free in reap_crm_member()
Fix: attrd: Remove offline nodes from node cache for "peer-remove" requests
Fix: pengine: cl#5130 - Choose nodes capable of running all the colocated utilization resources
Test: pengine: cl#5130 - Choose nodes capable of running all the colocated utilization resources
Feature: crm_diff: Support generating a difference without versions details if --no-version/-u is supplied
Log: crm_resource: Print a clear message if the attribute is not found for the resource
Fix: cluster: Invoke crm_remove_conflicting_peer() only when the new node's uname is being assigned in the node cache
Fix: xml: Prevent assert errors in crm_element_value() on applying a patch without version information
Fix: pengine: Fix the behaviors of multi-state resources with asymmetrical ordering
Test: pengine: Fix the behaviors of multi-state resources with asymmetrical ordering
Build: services: Conditionally clean up services actions
Fix: xml: cl#5231 - Unset the deleted attributes in the resulting diffs
Fix: xml: Do not prune leaves from v1 cib diffs that are being created with digests
Fix: xml: Do not change the ordering of properties when applying v1 cib diffs
Build: Fix building if ncurses doesn't provide pkg-config files
Fix: crmd: Correctly add the local option that validates against schema for pengine to calculate
Fix: crmd: Check "watchdog" daemon option before setting "have-watchdog" property
Fix: pengine: Handle ordering between stateful and migratable resources
Test: pengine: Handle ordering between stateful and migratable resources (regression tests)
Fix: crm_mon: Fill CRM_notify_node in traps with node's uname rather than node's id if possible
Fix: crmd: Add "on_node" attribute for *_last_failure_0 lrm resource operations
Feature: fencing: cl#5134 - Support random fencing delay to avoid double fencing
Build: Fix building when heartbeat support is disabled
Fix: pengine: Do not reschedule monitors that are no longer needed while resource definitions have changed
Test: pengine: Do not reschedule monitors that are no longer needed while resource definitions have changed (update regression test)
Test: pengine: Do not reschedule monitors that are no longer needed while resource definitions have changed
Fix: pengine: Resolve memory leaks
Refactor: pengine: Reschedule specific monitors when resource definitions have changed
Fix: crmd: Reset stonith failcount to recover transitioner when the node rejoins
Fix: crm_resource: Display the locations of all anonymous clone children when supplying the children's common ID
Log: crm_resource: Fix output format of the warning from --force-start
Feature: pengine: cl#5207 - Display "UNCLEAN" for resources running on unclean offline nodes
Fix: pengine: Avoid trailing whitespaces when printing resource state
Test: pengine: cl#5207 - Display "UNCLEAN" for resources running on unclean offline nodes (update regression tests)
Fix: fencing: Correctly make args for fencing agents
Build: spec: Move the service and configuration files of crm_mon to cli package
Fix: xml: Fix upgrade-1.3.xsl to correctly transform ACL rules with "attribute"
Fix: crmd: Prevent the old version of DC from being fenced when it shuts down for rolling-upgrade
Fix: pengine: Prevent use-after-free in sort_rsc_process_order()
Fix: cib: Enable cib legacy mode for plugin-based clusters
Fix: cib: Revert d153b86 to instantly get cib synchronized in legacy mode
Fix: cib: Always generate digests for cib diffs in legacy mode
Fix: xml: Do not dump deleted attributes
Fix: cib: Ability to upgrade cib validation schema in legacy mode
Fix: cib: No need to enforce cib digests for v2 diffs in legacy mode
Fix: xml: Prevent potential use-of-NULL
Fix: xml: Correctly create v1 patchsets when deleting attributes
Fix: cib: Correctly add "update-origin", "update-client" and "update-user" attributes for cib
Fix: pengine: cl#5130 - Only check the capacities of the nodes that are allowed to run the resource
Test: pengine: cl#5130 - Only check the capacities of the nodes that are allowed to run the resource (regression test)
Hideo Yamauchi (18):
Medium: crmd: Ensure the transition timer is correctly reset after a change
Correction of the registration from a difference.
Correction of the log output of the practice node of stonith.
Correction of the log output of the practice node of stonith.(2)
Support of upstart of crm_mon.
Support of upstart of crm_mon.(2)
Support of upstart of crm_mon.(3)
Support of upstart of crm_mon.(4)
Mistake of the set of the variable.
Change of the log level.
Fix: pacemakerd: resolves memory leak of xml structure in pacemakerd
Fix: crmd: resolves memory leak in crmd.
Update membership.c
Low: mcp: Add 'start on started local and runlevel' to Upstart job.
Low:Support of the start from systemd of crm_mon.(1)
Low:Support of the start from systemd of crm_mon.(2)
Low:Support of the start from systemd of crm_mon.(3)
Low : fencing : Correction of the log output of the practice node of stonith.
Hu Fu (1):
fix: spelling mistakes about "allegedly"
James Guthrie (4):
Doc: Pacemaker_Explained: Changed /etc/ais/authkey to /etc/corosync/authkey
Low: crm_report: use paths configured in autoconf.
Low: crm_report: search for corosync config if not found.
Low: crm_report: added -D and --max-depth option.
Jan Pokorný (29):
Low: doc: fix operarions/op attribute (disabled->enabled)
Low: doc: Pacemaker Explained: unify using spaces
Low: doc: Pacemaker Explained: fix (missing) tags
Low: doc: Pacemaker explained: drop stale crm-debug-origin
Low: doc: Pacemaker explained: drop crm-debug-origin in pot/ro-RO
Low: doc: Pacemaker explained: fix interlinks
Revert "Low: doc: Pacemaker Explained: fix (missing) tags"
Low: doc: Pacemaker Explained: fix tags
Low: Tools: Fix a typo in usage examples
Low: doc: Clusters from Scratch: trailing whitespaces--
Low: doc: Clusters from Scratch: trailing whitespaces-- (2)
Low: doc: Pacemaker Explained: remove "generated" attr
Low: doc: Pacemaker Explained: Ch-Basics cleanup
Low: doc: Pacemaker Explained: Ch-Basics cleanup++
Low: doc: Pacemaker Explained: Ch-Basics cleanup++
Low: xml: fix typo in dtd files
Low: extra/resources/HealthCPU: ideling->idling
Low: more consistent use of F_CRM_ macros
Low: cib: remove commented out part in createEmptyCib
Low: xml: use proper UTF-8 encoding instead of utf-8
Med: xml: fix regular expression in fencing.rng
Med: xml: fix RELAG NG spec. violation in rule.rng
Low: crm_get_msec: prevent integer overflow
Low: fencing: wider use of defined literals
Low: lrmd: wider use of defined literals
Low: using $(MAKE) variable, boost parallel build time
Low: pengine: fix typo
Low: doc: fix typos in index terms
Med: fix xml.c memleak in "missing upgrade XSLT" case
Jesse Newland (1):
fix regression in 0394da94 that made it impossible to write a new UUID
John Shi (1):
High: RA: controld - wait for dlm_controld to be ready in start phase
Kacper Kowalik (Xarthisius) (1):
Fix building with glibc-2.16
Katelyn Perry (2):
Fix: Correct spelling of superseded
Fix: spelling of 'promotion' in tools/crm_master
Kazunori INOUE (47):
Low: fencing: Add option for specifying number of fencing retries
Low: Correctly handle a return value from crm_pidfile_inuse()
Low: Correctly handle a return value from crm_pidfile_inuse()
Low: mcp: set environment variables which the cluster-glue module needs.
Log: crmd: Allow action initiations to be tracked centrally on the DC
Feature: Upstart script for Pacemaker
Low: Bump the log of unset DC to info
Low: Bump the log of an operation start of a resource to info
Low: Core: Support PCMK_logpriority to set syslog priority
Low: fencing: reset fencing fail count, when fence by stonith_admin succeeded
Fix: mcp: Update upstart job
Feature: mcp: Alternate Upstart job controlling both pacemaker and corosync
Low: Build: Support Upstart job config files
Low: crmd: Block after 10 failed fencing regardless of crmd-transition-delay
Low: Build: Fixed typo
Low: mcp: Support PCMK_fail_fast to reboot a machine at the time of process (subsystem) failure
Low: mcp: Fix so that only pacemakerd restarts, at the time of pacemakerd failure
Feature: Fencing: On timeout, clean up the agent's entire process group
Refactor: fence_legacy: stonithd now handles killing the process group for us
Make specifications of job the same as pacemaker.service (unit file)
Fix: Fencing: Apply correct score to the resource of group
Fix: crmd: Memory leak
Fix: attrd: Memory leak
Low: crmd: Change the default value of node-action-limit
Fix: services: Fixes segfault associated with cancelling in-flight recurring operations.
Low: crmd: Change the default value of node-action-limit
Fix: services: Fixes segfault associated with cancelling in-flight recurring operations.
Low: attrd: Support "--refresh" command of attrd_updater
attrd: Change the threshold of log level to LOG_INFO
Low: attrd: Increase log level of first update of attribute
Low: ipc: Decrease log level of connection and disconnection with client
Low: attrd: Decrease log level when broadcasting an attribute
Low: lrmd: clean up the agent's entire process group
Low: attrd: Make hostname hashtable case insensitive
Low: operation_finished: Reduce log level 'terminated with signal' at the time of stop
Fix: services_action_cancel: Interpret return code from mainloop_child_kill() correctly
Fix: crmd: make node_state erase correctly
Fix: pengine: Correctly search failcount
Fix: make resource ID case sensitive
Fix: crm_node: Prevent use-after-free in tools_remove_node_cache()
Fix: PE: Log when resources require fencing but fencing is disabled
Fix: crmd: Ensure crm_config options are re-read on updates
Low: fix small memory leaks
Low: mcp: Add config example to exert watchdog of corosync
Low: mcp: set environment variable which the cluster-glue module needs
Low: mcp: Avoid the error of modprobe in Upstart job
Low: mcp: Add 'stop on runlevel' to Upstart job
Keisuke MORI (3):
High: Fix memory leak in cib when writing the cib contents.
High: Fix memory leak in cib related to libqb ipc.
Low: Build: Always use a short hash for the build version.
Ken Gaillot (108):
git-ignore more autobuild/configure/build targets
document libuuid-devel as build dependency
update Clusters From Scratch for Fedora 21
minor clarifications in Clusters From Scratch
git-ignore more build targets
document more optional build dependencies
fix internal crossreference in Clusters From Scratch so it validates
CTS: make README clearer and more up-to-date
CTS: fix regression in CTSLab.py's --trunc option, and use cleaner exception handling
CTS: fix regression for cman use case in patterns.py
CTS: environment.py fixes (useless print, missing --help option, missing usage() argument)
tools: consistently use 4-space indents in crm_standby
tools: crm_standby: use numeric operators where appropriate
tools: crm_standby: update accepted options to match documentation
doc: move shared text to new subdirectory, update Clusters From Scratch to include it
doc: Pacemaker Explained: remove syntax highlighting where it's not appropriate
doc: Pacemaker Explained now includes shared intro text
doc: Pacemaker Explained: convert FAQ appendix to asciidoc
doc: Pacemaker Explained: misc. corrections
doc: Pacemaker Explained: update for current XML/command-line syntax and RFC 5737 IPs
doc: Pacemaker Explained: format option tables more consistently
Core: move I/O-related functions from lib/common/utils.c to new io.c
Refactor: move cib/io.c:sync_directory() to lib/common/io.c:crm_sync_directory()
Refactor: move file-reading code from cib/io.c:validate_cib_digest() to lib/common/io.c:crm_read_contents()
Refactor: move synchronous write code from cib daemon to libcrmcommon
doc: add doxygen comment blocks to lib/common/io.c
Core: move digest-related functions from lib/common/xml.c to new digest.c
Refactor: move digest comparison code from CIB daemon to libcrmcommon
doc: add doxygen comment blocks to lib/common/digest.c
cib: remove dead code
cib: refactor: merge on_disk_cib_corrupt() into validate_on_disk_cib() to avoid extra stat
Core: refactor: separate CIB-specific code from XML stream writer
cib: move disk validation code from CIB daemon to CIB library
cib: refactor: separate daemon-specific code from generic code
cib: move CIB writing functionality from CIB daemon to CIB library
core: new internal function crm_compat_realpath() for cross-platform realpath()
cib: file-based backends now detect whether file is live CIB
cib: cib_file.c: improve error messages and doxygen comment blocks
cib: automatically update .sig files in file backends when clients update live CIB
cib: cib_file.c: fix sense of daemon user test when writing live CIB
doc: Pacemaker Explained: update appendices on upgrading to be more generic vs. 0.6-focused
doc: Pacemaker Explained: improve stylistic consistency
doc: Pacemaker Explained: minor clarifications
doc: Pacemaker Explained: clarify section on moving resources
doc: Pacemaker Explained: clarify section on groups and clones
doc: Pacemaker Explained: clarify section on editing configuration
doc: Pacemaker Explained: clarify chapter on constraints
doc: Pacemaker Explained: clarify chapter on multi-site clusters
doc: Pacemaker Explained: clarify chapter on managing nodes
doc: Pacemaker Explained: clarify chapter on resources
doc: Pacemaker Explained: clarify chapter on STONITH
doc: Pacemaker Explained: clarify chapter on placement strategies
doc: Pacemaker Explained: minor reorganization of section headings
doc: Pacemaker Explained: clarify chapter on rules
doc: Pacemaker Explained: clarify chapter on status
doc: Pacemaker Explained: bump document version
doc: update Makefile to remove no-longer-usable crmsh option and abstract the rsync target
Feature: attrd supports private attributes (not written to CIB)
Feature: core: attrd_update_delegate() supports private attributes
Feature: tools: attrd_updater supports private attributes
Fix: update .gitignore for recent heartbeat changes
Refactor: attrd: define and use constants for attrd operation strings
tools: improve help text for attrd_updater options
Refactor: attrd: make each client operation handler its own function
Refactor: attrd: merge attrd_client_message() into attrd_ipc_dispatch()
attrd: only send ack for recognized operations
Fix: cib: assignment where comparison intended
doc: minor clarifications to Pacemaker Explained
doc: minor corrections to Pacemaker Explained
Fix: tools: typos in crm_resource help text and comments
tools: Clarify logic in crm_resource with comments and defined constant
Fix: tools: free allocated memory correctly in crm_resource
doc: minor clarifications to Pacemaker Explained's Multi-Site Clusters chapter
Refactor: tools: functionize working set logic in crm_resource
Feature: attrd supports queries
tools: attrd_updater: refactor and improve option help messages
Feature: tools: attrd_updater supports --query and --all options
Fix: attrd: b22b1fe did uuid test too early
Refactor: tools: guard crm_resource actions with booleans rather than comparing rsc_cmd
Feature: tools: add crm_resource --wait option
doc: update Doxyfile to doxygen 1.8.5
Fix: doc: correct errors in doxygen comment blocks
Fix: lrmd, services: interpret LSB status codes properly
Fix: tools: make crm_mon last updated header consistent across formats
Fix: tools: Improve crm_mon output with certain option combinations
tools: display node names more consistently in crm_mon output
tools: crm_mon prints Stopped clones only if --inactive was specified
Fix: systemd: improve unit files
Fix: update .gitignore to ignore *.tar.gz and lrmd/lrmd_internal_ctl
Fix: doc: enable doxygen comment blocks in source under lib
Refactor: lrmd: merge two overlapping functions
Fix: lrmd: guard timeb-dependent code
Fix: lrmd: report original timeout when systemd operation completes
Refactor: lrmd: let time_diff_ms() take NULL first argument as "now"
Fix: lrmd: Use original timestamps when reporting multi-step systemd operations
Build: update libtool versions for 1.1.13 release
Log: cib: restore read-and-verify logging to 1.1.12 verbosity
Build: handle RNG versioning better in xml Makefile
xml: add RelaxNG schema for crm_mon XML output
Refactor: crmd: use defined constant for magic "direct nack" RC
Fix: crmd: report operation rc as advertised instead of status
Fix: crmd: don't update fail count twice for same failure
Fix: crmd: respect start-failure-is-fatal even for artifically injected events
Refactor: crmd: remove dead code
Refactor: crmd: rename action number variable in process_graph_event()
Refactor: crmd: get full action information earlier
Fix: crmd/pengine: handle on-fail=ignore properly
Refactor: crmd/pengine: implement on-fail=ignore without allow-fail
Kristoffer Grönlund (16):
Low: report: Add support for xz compressed logs
Low: crm_report: Force grep to interpret logs as text
fix: crm_resource: Help text for -v was incorrect
Doc: Update pcs/crmsh quick reference
Update pcs-crmsh-quick-ref.md with crmsh equivalents
Update TOC
Doc: flip colocation example to match pcs version
crm_attribute: Default --lifetime to forever when using --node
crm_report: Add support for ISO8601 time format in logs
Update the readme with details on user interfaces
Fix: crm_diff: Remove mention of --filter option from usage
Fix: ping: Correctly advertise multiplier default
Low: ping: Add use_fping parameter
Low: ping: Remove incorrectly advertised migrate_to|migrate_from
Low: ping: Pass extra options to fping
Low: ping: Correct metadata for attempts parameter
Lars Ellenberg (30):
Build: fix double typedef declarations
Build: fix configure --without-corosync
Build: fix configure --with-heartbeat
Build: fix make rpm without corosync
Low: crm_mon -A: fix memleak, add g_free_list(attr_list)
Tools: teach crm_set_options and crm_help about optional command line arguments
Fix: crm_mon -A: print_attr_msg needs to lookup paramters, not meta
Feature: crm_mon --neg-location drbd-fence-by-handler
Fix: crmd: Don't add node_state to cib, if we have not seen or fenced this node yet
Fix: crmd: don't segfault on shutdown when using heartbeat
Fix: crmd: kill pengine on shutdown also when using heartbeat
Fix: iso8601: Prevent dates from jumping backwards a day in some timezones
Fix: get_cluster_type: failing concurrent tool invocations on heartbeat
mainloop: fix potential segfault for fd = 0
Low: heartbeat 3.0.6 knows to finds the daemons; drop compat symlinks
Low: use CRM_NODE_MEMBER, not CRM_NODE_ACTIVE
Medium: don't trigger a stonith_reconnect if no longer required
Medium: map heartbeat JOIN/LEAVE status to ONLINE/OFFLINE
fix crm_update_peer_proc to NOT ignore flags if partially set
Medium: fix tracking of peer client process status on heartbeat
High: fix stonith ignoring its own messages on heartbeat
Low: add debugging aid to help spot missing set_msg_callback()s on heartbeat
Low: allow heartbeat to spawn the pengine itself, and tell crmd about it
Low: native_print: report target-role as well
Low: bump 'Sign-in failed but no longer required' to info for now
Fix: provide fake meta data for ra class heartbeat
Low: amend heartbeat metadata template to use single quotes
Fix: lrmd: add back support for class heartbeat agents
Test: fix expected summary files for target-role annotation
Low: spelling fix -- heartbeart
Lars Marowsky-Bree (8):
High: ais: Make sure nodeid is set when using the plugin (bnc#803746)
Fix: plugin: remove memory leak in legacy code
plugin: fix memory leak in send_cluster_text
Low: fix a small memory leak
crm_mon: Bug cl#5167 - Only print "stopped" node list for incomplete clone sets
Low: spelling fixes in comments and log messages
Low: controld: allow the stonith-enabled check to be overridden
High: plugin: stonith-ng needs membership updates
Led (1):
Fix bashisms in LSBDummy script
Lidong Zhong (1):
Fix: controld: Return OCF_ERR_INSTALLED instead of OCF_NOT_INSTALLED
Matous Ejem (1):
Doc: Clusters from Scratch: Fix pcs commands
Michael Chapman (4):
Remove cancelled ops from history cache
tools: fix "crm_shadow --reset <name>"
mainloop: implement gio_poll_dispatch_mod
mainloop: implement gio_poll_dispatch_mod
Michal Kubecek (1):
Low: avoid list traverse restarts when stripping text nodes
Norbert Tretkowski (4):
Update security.txt
Update security.txt
Update security.txt
Update acls.txt
Oleksandr Bodnarashyk (1):
allow to build on FreeBSD
Olivier Allart (9):
added a dual layer dual PDU fencing example to ch-Stonith
updated Ch-Stonith.txt with detailed topology example
updated Ch-Stonith.txt topology example, clarified PDU/PSU and fixed apc agents IP mismatches
Reformatted pcs-crmsh-quick-ref.md with more detail
Reformatted pcs-crmsh-quick-ref.md with more detail, updated toc
Moved TOC tool link to comments
fix: pcs-crmsh-quick-ref.md: List RA info
fix: pcs-crmsh-quick-ref.md: Constraints manipulation
fix: pcs-crmsh-quick-ref.md: rsc ban,move,clear
Patkos Csaba (1):
On systems without tm_gmtoff use timezone considering that it returns the gmt offset negated and without DST adjustment.
Philipp Marek (7):
Low: Bugfix for ./configure when using Heartbeat.
Low: Build: Bugfix for ./configure when using Heartbeat
Medium: Fix uninitialized data for --with-heartbeat.
Medium: Fix a compile error for --without-corosync.
Merge branch 'master' of git://github.com/ClusterLabs/pacemaker
Fix: On cluster signoff, make sure that the g_source gets disconnected.
Low: crm_report: Add booth configuration to report.
Piotr Chmylkowski (1):
High: Fencing: Fix passing of parameters containing '='
Ralph Dehner (1):
Fix spelling error
Rob Thomas (3):
Fix overflow on SMTP subject line
Add clustermon.sh
Realised I wasn't using the 'fixed' vars.
Stanislav Bogatyrev (1):
Fix Debian init script default runlevels and stop deps
Takatoshi MATSUO (12):
Low: services: Fix compile error when not using systemd
Low: Build: Fix compile error without upstart and systemd on RHEL6
Fix: mcp: inhibit error messages without cman
Low: extra: add failure_score parameter into ping RA
Build: exclude rc version from LAST_RELEASE and NEXT_RELEASE
Build: fix "without" argument
Test: PE: add -i|--io-dir option to switch I/O directory
Test: PE: Test exits if it cannot read regression.core.sh
Low: tools: clear ncurses screen when crm_mon cannot connect
Low: tools: remove unnecessary function of crm_mon
Feature: crm_mon: add --hide-headers option to hide all headers
Low: tools: align indent of help text
Ultrabug (1):
Fix: Build: respect ncurses lib flags
Viacheslav Dubrovskyi (2):
Fix revnumber according last requirements format
Fix undefined symbols in libs
Vladislav Bogdanov (4):
Fix: Remove duplicate definition of mainloop_io_t from headers
Fix: cib: Prevent memory leak during connection setup
Feature: PE: Unlink pengine output files before writing.
Feature: Core: Allow blackbox logging to be disabled with SIGUSR2
Xia Li (5):
modified a message typo issue, changed "laoded" to "loaded"
Test : cts : automatically determine whether to run RemoteBaremetal
Test:cts: fix shutdown pattern not found issue
Test:cts: fix RemoteBaremetal step2 move resource action failed issue
Test : cts : change default cluster stack from openais to corosync
Yoshihiko SATO (1):
Low: fencing: stonith action message when the fence is successful
Yuichi SEINO (1):
Fix:Signal FD leak
Yuusuke Iida (1):
Added define of CS_USES_LIBQB to crm_config.h.in
andreas.grueninger (3):
feature test for sighandler_t added and for ncurses improved
add NULL check to avoid unnecessary error message
Merge remote-tracking branch 'beekhof/master' into upstream
davidvossel (17):
Medium: lrmd: Check rsc long name when retrieving rsc from lrmd.
High: doc: Updated "Clusters from Scratch" for dlm/gfs2 in Fedora 17.
Low: build: Ability to build separate docs based on different cli types (pcs or crm)
High: doc: Updates to Clusters from Scratch using pcs shell
Low: doc: Remove some references to missing images in cfs so it can be built with and without them.
Low: doc: Update NTP section of cfs for new fedora documentation
Low: doc: Use a consistent cluster name for CFS documentation
Low: doc: Fixes grammar error in CFS installation section
Low: doc: Fixes an incorrect reference to a pcs command to show raw cib
Low: doc: Combines documented installation of wget and httpd into one step
Low: doc: Remove reference to automated constraint as it is explicitly created at the moment with pcs
Low: doc: Move placement of a step used in stonith configuration for CFS
Low: doc: Add per rsc constraints to configuration recap
Low: doc: Update todo for pending CFS changes waiting on pcs update
Low: doc: Move pcs specific sentence into pcs if block.
Low: doc: Revise todo item so it actually makes sense
Low: doc: Demonstrate that pcs automatically discovers resource standard and provider on creation
hfu (1):
chinese document translation
ldzhong (1):
don't start dlm if stonith-enabled is false
makepkg (1):
Test: lrmd: support class heartbeat tests
renayama19661014 at ybb.ne.jp (1):
Fix: crmd: Do not start timers for already completed operations
root (1):
heartbeat: Fix a variable name
yusuke (4):
Low: Core: When a definition does not exist, it corrects so that facility of corosync may be used.
Fix: It corrected so that the reading pipe which is not closed might be closed.
Fix: fixed a memory leak in lrmd.
Fix: fixed a memory leak in crmd.
yuusuke (17):
Fix: fixed a memory leak in stonithd and crmd.
Implement the ability to change the display options during operation.
When timing is validated, it corrects so that operation may also be validated
fix reference in "data_set->ticket" of NULL.
Merge branch 'crm_mon_customize' of github.com:yuusuke/pacemaker into crm_mon_customize
Fix memory leak when used score2char()
Fix: remove the mainloop_trigger that are no longer needed.
Fixed a problem that chkconfig has not been able to get a list correctly in the environment of RHEL5.
Fixed a problem that only pacemakerd did not inherit the logfacility of corosync.conf.
Fix: memory leak found by valgrind was corrected.
Fix: crmd: Correctly detect changes to the configuration root
Fix: When data was old, it fixed so that the newest cib might not be acquired.
fixed so that change of <resources> could be handled correctly.
Fixed so that it could change while a setup of "startup-fencing" starts.
Low: fencing: Remove redundant processing
Log: Reduce the log when pacemakerd is run on command.
Fix: allows the user to specify a log level that is output to syslog
-----------------------------------------------------------------------
No new revisions were added by this update.
--
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/debian-ha/pacemaker.git
More information about the Debian-HA-Commits
mailing list