[PKG-Openstack-devel] Bug#801627: Taking over sheepdog

Hans Freitag zem at fnordpol.de
Tue Jul 17 16:18:40 BST 2018


Hi,

On 16.07.2018 22:50, Thomas Goirand wrote:
> Hans, Michael, Guido, please do take over Sheepdog. This can be done
> within or outside of the OpenStack team, as you like. Just let me know.
> It'd be nice to have version 1.x for Buster.

There is a debian directory in the 1.0 source code, and I did manage to
compile a test version, with a few issues that I submitted on github so
far.

For the moment I would like to do some more testing, because 1.0
produces a segfault when it tries to access a 0.83 repo, it seems to
work with a new sheepdog.  Somehow the cache and the webdir/swift is not
working anymore with the new Package. The systemd unit file is going to
active(exit) if the service fails, which is the wrong state.

The current upstream maintainer is only fixing bugs (which must not
neccessarily be a bad thing), but he or she has to respond to my
bugreports, too.

I still have to do a multinode setup to see if sheepdog is reasonably
stable and does not loose any of my data, I can put together a debian
package after that, if I decide that sheepdog is useable.

Oh and I did not even start with all the libvirt related issues
regarding cluster storage integration! Thats another pile of issues.

regards

    Hans



More information about the Openstack-devel mailing list