|
From: David Y. <da...@pr...> - 2010-12-14 23:11:07
|
Hi bacula-users,
We have configured disk volumes and auto-labeling in such a way that a
new file is created for every backup job, and the job ID is part of the
filename:
--
Pool {
Name = daily
Pool Type = Backup
Recycle = no
Use Volume Once = yes
AutoPrune = yes # Prune expired volumes
Volume Retention = 8 days # one week + 1 day margin
Label Format = "${JobName}-${JobId}"
Maximum Volume Jobs = 1
Volume Use Duration = 23h
ActionOnPurge = Truncate
}
--
This works great, and it's really easy to identify which file is created
by which backup, with the exception of a case where a backup fails..
because the backup volume has already been created based on the name.
I.e, say I have 2 backup jobs, "serverA" and "serverB"...
Job "serverA" runs, and auto-labels a new file as
"/store/serverA-434"... but the job fails because ServerA is
unreachable. So bacula times out, and then proceeds with backup job
"serverB", but it uses the file it's already created for serverA's backup.
Any way around this?
Thanks,
D
|