Quoting Bastian Blank (waldi@debian.org):
> Hi folks
>
> Please unblock lvm2/2.02.39-6.
>
> It introduces several bug fixes.
Let's try to make the release team easier...
That releases introduces the following changes, compared to the
release in testing:
lvm2 (2.02.39-3) unstable; urgency=low
.
* Depend against lsb-base.
* Make clvm depend against cman.
* Don't ignore locking failures in lvm2 init script.
* Only activate vgs local.
* Add clvm initscript. (closes: #336258)
* Try to activate anything in the lvm2 postinst. (closes: #506105)
lvm2 (2.02.39-4) unstable; urgency=low
.
* Revert locking change, it breaks too much. (closes: #506354)
* Disable cluster support in udeb.
* Install the binary as lvm in the initramfs. (closes: #503627)
* Build cluster locking internal.
* Adjust clvm documentation for the locking change.
lvm2 (2.02.39-5) unstable; urgency=low
.
* Install lvm.conf into initramfs. (closes: #439761)
* Also ignore locking in initramfs.
* Make piuparts-clean. (closes: 455115)
- Cleanup old files on upgrade. /lib/lvm-default, /etc/lvm/.cache.
- Cleanup dirs on remove. /etc/lvm/cache, /var/lock/lvm.
* Recommend dmsetup.
lvm2 (2.02.39-6) unstable; urgency=low
.
* Readd a vgchange binary into the initramfs. At least cryptsetup depends on
this.
So, -4 was triggering some RC bugs, such as #506536 and #506534. This
is fixed in -6
From quite far, this makes a lot of changes and, would I be in the
release team, I would feel quite uncomfortable allowing such changes
to go in testing when seeing how they can hardly break systems. The
"it breaks to much" comment is indeed quite self-explanatory here.
Notbeing a release team member, this is of course up to them to decide
but I felt some "duty" to try helping around here. Such big noodles
plate is probably not something we really want so close to a release.
Attachment:
signature.asc
Description: Digital signature