Hi,
I received bug#509184 today. It was reported against a version of
bacula in backports, but applies to the version in lenny as well (though
not the version in sid).
The bug in question is a double-free and does not appear to be a
security hole. It results in a crash of the entire backup system. It
seems to be somewhat rare, but I am not certain of that.
Upstream has issued a small patch that they have tested.
Would you approve an upload to testing-proposed-updates into lenny with
such a patch if I built it?
-- John
--- Begin Message ---
* fabien foglia <fabien.fog...@tigrou3tac.org> [2008-12-19 12:30:51 CET]:
> Version: 2.4.2-3.1~bpo40+1
As this is a backports version I would encourage you to either ask
directly on backports-users mailinglist which is the central address for
packages taken from backports, and/or at least send me a copy of the
bugreport when you're confident that it's not just a problem of the
backport. :)
> I'm using etch-backports bacula-director with postgresql support on
> Debian etch/amd64.
> Sometimes, director crashes with "ABORTING due to ERROR in
> smartall.c:194 *double* *free* from jcr.c:343"
> It's the bug #1162 in the bacula bug tracker.
>
> It's resolved with this patch :
> http://puzzle.dl.sourceforge.net/sourceforge/bacula/2.4.2-jobend-crash.patch
Ah, that's neat and convenient. The patch looks pretty tight to the
issue. I am discussing with John what's the best approach for it. For
what I understand it's a crash after the backup is finished so the
impact is pretty low?
> please merge this patch or update bacula-director to the lastest 2.4.2
> bacula revision
I'm not completely sure what you mean with "the latest 2.4.2 bacula
revision". There is only a single 2.4.2 release, not several revisions
to it ... But the patch does look like it makes sense indeed.
Thanks for the hint,
Rhonda
--- End Message ---
