As Arno Lehmann suggested as an additional answer to my last problems
i looked a while around in the lists archive and found one thing which
alarmed me a little....
in one of the mails Kern wrote (this was a discussion about concurrent jobs):
From: Kern Sibbald <[EMAIL PROTECTED]>
Re: Fwd: Another problem with concurrent jobs 2005-04-21 23:26
Some more thoughts on this: I"m not convinced that it is necessary to use
different directories, though it might be a good idea, but most likely if you
do not use different Pools (i.e. different volumes) you will have two jobs
writing into the same disk volume, which will not work.
On Thursday 21 April 2005 18:53, Kern Sibbald wrote:
> It is perhaps not well documented, but concurrent jobs to the same Device
> resource that references a file rather than a tape device is not supported.
> To run multiple concurrent jobs, you must use different Device resources
> pointing to different directories.
>
my setup now is to have different pools for each storage and each storage has
it's own disk-device.
but what i discovered last night when some jobs run in parallel, they really
wrote into the same file! is this now (still) really a problem? did i lost
my backup-integrity?
and... if it's a problem... how can i avoid this??
additional to this here's a copy of the beginning of my bsr-files for two jobs
running in parallel and writing to the same file:
==> sonne-data.bsr <==
Volume="BackupFile-0001"
VolSessionId=2
VolSessionTime=1115045982
VolFile=0-0
VolBlock=848024679-1846476699
FileIndex=1-21818
Volume="BackupFile-0001"
VolSessionId=2
VolSessionTime=1115045982
VolFile=0-0
==> sunserv-data.bsr <==
Volume="BackupFile-0001"
VolSessionId=3
VolSessionTime=1115045982
VolFile=0-0
VolBlock=1430309864-1847895963
FileIndex=1-14567
Volume="BackupFile-0001"
VolSessionId=3
VolSessionTime=1115045982
VolFile=0-0
i'm not sure if i understand this information correctly, but when looking at the
"VolBlock"-values, i really get a bit nervous....
