-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Greetings,

I've been having important issues with my backups since the installation
of a new version of the tzdata package (Venezuelan timezone change from
UTC-0400 to UTC-0430) -- one of the backups, the largest one which has
to travel through the LAN, keeps failing at exactly 2 hours after the
start time (note: in some test setups it has failed at 1 hour, but the
main cases are repeteadly failing at 2 hours)

I've been discarding possible sources for the problem:

a) I have another Bacula setup where the timezone change was implemented
on the same way, and it is working nicely, with a bigger volume of data
being backup through the LAN.

b) I connected the Dir/SD and the FD directly via crossover cable, and
it keeps failing -- just for the record, I mounted NFS through this
cable and Bacula backups it nicely, thus discarding problems in the NIC,
~ AFAICT. (I originally thought we could be facing problems with the
network hardware linking the two hosts but this discarded it and anyway
the other backups on the same network run OK)

c) I upgraded Bacula from 1.38 to 2.2.6 (clients in 2.2.5, since I'm
using Debian Etch and the Backports for IA64 are still 2.2.5) which BTW
failed on the first try because of the update scripts for sqlite having
AUTOINCREMENT instead of SERIAL for the Media backup table -- and it was
fruitless.

d) I umounted the whole cluster and fsck'ed the filesystem (it is a GFS
filesystem, which in my other setup backups just ok) in order to prevent
any problem that might be arise by the FD stalling while reading some
specific file -- it ran OK and the problem's still there.

e) I used Heartbeat Interval = 60 in the setup, with no results.

Here's a peek of what's on my bacula/log at the Dir/SD. Take a special
look to the start time and the time of the failure, offseting by only
ONE second.

Start time: 11-ene-2008 00:39:20

11-ene 02:39 name-sd JobId 845: Error: bsock.c:444 Read error from
client:IP:36643: ERR=Connection reset by peer
11-ene 02:39 name-dir JobId 845: Fatal error: No Job status returned
from FD.
11-ene 02:39 name-dir JobId 845: Error: Bacula name-dir 2.2.6 (10Nov07):
11-ene-2008 02:39:21

I hope I'm missing the bigger picture of the problem and maybe it can
actually be solved by some simple maneuver. Currently I'm unable to
backup this service (which is MBOX-format mail, BTW) with Bacula and
it's a shame since we use it for almost everything else.

I would really appreciate some comments on this problem. Thank you very
much for your time,
Jose
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHh2nQUWAsjQBcO4IRAjetAJ9YqHy4Kii/6fILO40+kiuEFtKa5wCeLecN
i41HnECSJxh4iBjz2j8EGeQ=
=I7mK
-----END PGP SIGNATURE-----

-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users

Reply via email to