[Pkg-iscsi-maintainers] open-iscsi -13 upload?

Sun Feb 7 11:17:28 UTC 2016

Hi Ritesh,

so since we've now hopefully nailed the bug you've encountered,
assuming that it's now really fixed, I would suggest that we maybe
upload -13 to unstable. (IMPORTANT: please pull again from git before
you do so, because I pushed something else as well!)

Going through the list of open bugs in the bugtracker:

Remaining open bugs (not fixed by current git)
----------------------------------------------

#797112 [S| | ] [open-iscsi] open-iscsi: Prevent testing migration
Good thing I kept that one open, because the large changes I made that
we uploaded during DebConf did cause quite a few additional problems,
for example also your bug. But since quite a few things have now been
fixed in unstable, I'd probably close the bug after the -13 upload,
because I think it has now gotten enough testing in unstable and I'd
like to see the improvements in tessting. (Especially to check what
happens with the installer.)

#699240 [n| | ] [open-iscsi] open-iscsi: Broadcom Offload Engine is unusable because of missing iscsiuio daemon
That's something I'd like to wait for a new upstream version for.
(I'll prod Mike in the near future to ask about that.)

#732587 [n| | ] [open-iscsi] debian expert install on iscsi-device
I'll look into that before stretch, but the d-i does work for
PXE-booted iSCSI for me, so this appears to be a weird setup.

#664604 [w|+|=] [open-iscsi] "add event based mount support for open-iscsi"
This is now only relevant for sysvinit, but I do have an idea for that
and want to fix it for stretch, but it's not really urgent.

#492367 [w|M| ] [open-iscsi] open-iscsi: Enable static IP to iscsi root using initramfs and virtual machine Xen
I recently asked for more info on this one, because I'm not going to
merge a patch like this before knowing it's still relevant. (And even
then it might be better to merge that in initramfs-tools proper, but
that depends on what exactly is going on.)

#627908 [m|U+u|↝] [src:open-iscsi] open-iscsi: small typo in comments of iscsid.conf
Minor, will be fixed once new upstream version is there.

Pending bugs
------------

So there are quite a few pending bugs fixed in git that we should
really get into unstable, especially your bug (which was quite nasty
because it could lead to data loss) and also some others that pertain
to policy-like things in the archive (state files /var, merged /usr
support, etc.).

#804591 [n|P| ] [open-iscsi] open-iscsi: iscsi_auto flag should use iscsistart --fwparam_network in addition to -b
#810276 [n|+P| ] [open-iscsi] files with the same name installed in / and /usr
#810696 [n|+P| ] [open-iscsi] Please move state files from /var/ to /run
#810702 [n|+P| ] [open-iscsi] Generate initiator name on install, not first boot
#804162 [w|P| ] [open-iscsi] open-iscsi: support iscsi root by format in RFC 4173 (root=iscsi:server:prot:...)
#809320 [i|P|=] [open-iscsi] open-iscsi: no session termination on stop

Misc fixes not in the bugtracker
--------------------------------

I've also done some other work in git that isn't in the bugtracker,
because I didn't want to report every single detail. Here's an
overview over the changes + some rationale:

- d95ccde9e31f1cdad1aa3e34ca6a196bc23157e7: there was a bug when
creating the initramfs image because of the move from /usr to /

- 36be1d16f012903f25fc7d39b72afc4b5b12ccd9: some follow-up on
Martin Pitt's solution for postinst iscsi-iname

- f6246a8e47a57494aa89a560f9d32426ef42914d: use triggers to update
the initramfs and don't call update-initramfs directly, because that
will also work with dracut. Note that dracut support itself is still
something we have to look at (there is some native support in dracut,
but only via kernel parameters, I'm working on that), but this is the
first step.

- 3b3fa0026a69e16d37bea0a285b1f928d85a7947: this reecords the
network device used to configure open-iscsi in the initramfs in
/run/initramfs/open-iscsi.interface. Ubuntu already does this (they
also hook into ifupdown to not touch that interface, but I don't quite
grasp the consequences of that change yet, so I've only added the
harmless part so far to reduce the diff to Ubuntu)

- df7259e8aaa3c3bbca9734f633185dcc7f6b3bf3: fix reproducible build
failure on armhf, see [1] (well, at least I hope it fixes that, I don't
have a reproducible build setup here myself, and I don't really have
any armhf hardware)

In conclusion: I think it's time to upload -13 to unstable, unless
you want to do some additional testing or have some other issues
that I don't know about yet.

Regards,
Christian

[1] https://tests.reproducible-builds.org/rb-pkg/unstable/armhf/open-iscsi.html

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: OpenPGP digital signature
URL: <http://lists.alioth.debian.org/pipermail/pkg-iscsi-maintainers/attachments/20160207/ce8e130c/attachment.sig>