server, you could watch the log files and start a script when a new lease
is handed out. Or better yet, watch for name server messages if dhcp is
set up to update the dns.
MfG, Ulrich
--
Heinz Ulrich Stille / Tel.: +49-541-9400473 / Fax: +49-541-9400450
Tammen Gmbh / Wilhelmstraße 16 / D-4907
s before that. To avoid manual intervention, run a small dummy job
to the same tape earlier.
Btw: I'm using 1.39.22. If you have < 1.38, better upgrade.
MfG, Ulrich
--
Heinz Ulrich Stille / Tel.: +49-541-9400473 / Fax: +49-541-9400450
design_d
27;ll try again with trace enabled, but that will take some days. Any
ideas what might have gone wrong? And what debug level should I use?
MfG, Ulrich
--
Heinz Ulrich Stille / Tel.: +49-541-9400473 / Fax: +49-541-9400450
design_d gmbh / Wilhelmstr. 16 /
On Friday 06 October 2006 14:11, Kern Sibbald wrote:
> On Friday 06 October 2006 12:14, Heinz Ulrich Stille wrote:
> > In some circumstances the director process eats all available memory and
> > then some. That is, 1GB ram and about 4GB swap on my machine. There is
> I'
that tape and then use it,
so I had to go and insert it again.
Would it be possible to suppress the "offline on unmount" when the
daemon is shutting down, or if the loaded tape has an error? In
that case, I'll probably want to deal with the error and then use
that tape anyway.
MfG,
ll Interval = 5 minutes
Offline on Unmount = yes
Close on Poll = yes
Spool Directory = /var/bacula/spool
Maximum Spool Size = 275 G
}
The drive is a Certance (Quantum) CL800 LTO3 streamer.
MfG, Ulrich
--
Heinz Ulrich Stille / Tel.: +49-541-9400473 / Fax: +49-541-9400450
design_d gmbh /
g something else) I'll get the success messages
of the next run interleaved with the errors of the timed out one.
For clarity I'd prefer to get the failure notices as soon as the
timeout occurs (in the logs, too). Is this possible?
MfG, Ulrich
--
Heinz Ulrich Stille / Tel.: +49-541-9400
estion: I'd prefer to use "Offline On Unmount"; but I think the tape
is unmounted after each single job, so that it would be ejected between my
"full" and "catalog" jobs? If that is the case, is there a way around it?
Jobs running each other
think the jobs "full.2006-03-27_17.30.00", "catalog.2006-03-27_17.30.01" and
"closing.2006-03-27_17.30.02" should all have been cancelled on the 28th,
15:30 (17:30 + timeout of 22 hours).
MfG, Ulrich
--
Heinz Ulrich Stille / Tel.: +49-541-9400473 / Fax: +49-5
e log says e.g.:
Mar 28 17:34:22 mail bacula-dir: mail-dir: full.2006-03-27_17.30.00 Fatal
error: Job canceled because max start delay time exceeded.
What more should there be? And that cancellation should have been at 15:30:00.
MfG, Ulrich
--
Heinz Ulrich Stille / Tel.: +49-541-9400473 / Fax: +49
2,950,158 OK 29-Mar-06 22:27 web
Running Jobs:
JobId 16 Job verify.2006-03-30_11.49.06 is running.
Verify Job started: 30-Mar-06 11:49
Files=4,882 Bytes=0 Bytes/sec=0
Files Examined=4,882
Processing
file:
/arc/taos_2002/webr/70-0001/050599/weber/deutsch/unternehmen/vertret
only ever checked when the job really is started. A watchdog ist running,
but apart from the fact that it would only check MaxWaitTime it skips the
job because JobId is 0.
No problem - functionality isn't affected. It just would be nice if the
watchdog could also handle this.
MfG, Ulrich
--
Heinz
tal pool, and the director doesn't bother ejecting
> in favor of a tape from the full pool. Jobs run as Full backups
You need to use the "Full Backup Pool" job directive. It does exactly
that: change the pool if the job is upgraded to full.
MfG, Ulrich
--
Heinz Ulrich Stille / T
ther there should be
no difference. What's up? Is somebody using bacula with version
4 or higher and the default timeout? I'd prefer not to raise the
timeout to a couple of days (long weekend, holidays...).
Or should I create a feature request for mysql_ping()?
MfG, Ulrich
--
Heinz
say - what is the definition of "using"?
My guess would be "job running and either active or waiting to become
active"? But what happens when the next job wants a tape from a different
pool? The doc refers to drive, not tape?
MfG, Ulrich
--
Heinz Ulrich Stille / Tel.: +49-541
u get
Run = Incremental Pool=Incrementals mon-thu at 1:00
The automatic upgrade only changes the level, not the pool, so that
still is determined from the override. To override that in turn you
have to use "FullPool", which is evaluated when the actual level
is "full" for whatever r
ted)
and had to wait a long time (like a weekend).
Is it possible that bacula stays connected to the DB while a job is
running?
MfG, Ulrich
--
Heinz Ulrich Stille / Tel.: +49-541-9400473 / Fax: +49-541-9400450
design_d gmbh / Wilhelmstr. 16 / 49076
eature request?
MfG, Ulrich
--
Heinz Ulrich Stille / Tel.: +49-541-9400473 / Fax: +49-541-9400450
design_d gmbh / Wilhelmstr. 16 / 49076 Osnabrück / www.design-d.de
---
Using Tomcat but need to do more? Need to support web services, security?
Ge
MediaType: DLT-8000
FirstWritten: 2006-06-02 14:51:27
LastWritten: 2006-06-06 09:39:27
LabelDate: 2006-06-02 14:51:27
VolJobs: 1
VolFiles: 13
VolBlocks: 215,309
VolMounts: 2
VolBytes: 13,890,013,961
VolErrors: 0
Vol
can see, the duplicate control only uses priorities, which
I can't set from the schedule's run directive? Can I achive this in
another way?
MfG, Ulrich
--
Heinz Ulrich Stille / Tel.: +49-541-9400473 / Fax: +49-541-9400450
design_d gmbh / Natruper Str. 15 / 49076 Osnabrück / www.design-d.de
Os
e through the inner workings of the
scheduler myself. It's especially unsatisfactory since the functionality
is /almost/ there.
Oh well, time to brush up on my Python ...
MfG, Ulrich
--
Heinz Ulrich Stille / Tel.: +49-541-9400473 / Fax: +49-541-9400450
design_d gmbh / Natruper Str. 15 / 4907
w Duplicate
Jobs = no" just doesn't work. Drat!
MfG, Ulrich
--
Heinz Ulrich Stille / Tel.: +49-541-9400473 / Fax: +49-541-9400450
design_d gmbh / Natruper Str. 15 / 49076 Osnabrück / www.design-d.de
Osnabrück HRB 19116 / Geschäftsführer: Günte
ition,
although that has been overridden by the restore command.
Is this a bug, a feature, or something I just didn't understand? The
default client e.g. is not checked.
Is there a way to force restoration to a certain client, even if
the backup is from another?
MfG, Ulrich
--
Heinz Ulrich St
as the test job contains just a
dummy directory - bacula alternates between creating a new volume and
recycling the last one. I had expected it to always create a new one; I'd even
understand always pruning empty ones, but this? Is this a bug or something
missing from the documentation?
MfG, Ulri
24 matches
Mail list logo
