Arno Lehmann schrieb:
Hello,
On 1/26/2006 10:18 AM, Stefan Schwietert wrote:
I had another bug with this situation. Bacula was doing a backup from
a workstation with spooling. After 8GB it began to despool the data
to tape and while this was done the user switched off the
workstation. So the second part could not be stored and bacula
canceled the job. Now the canceled job stuck in the queue and thus
prevented the next scheduled jobs from running.
Yes. It can take up to two hours before the SD "accepts" the timeout.
Knowing myself, I guess you weren't patient enough :-)
The job was canceled in the late afternoon and I noticed it next
morning, 12 hours later, because I was missing the usual email.
Once you got the job email that states that the job failed other jobs
should run normally. (By the way, the best way to avoid that your
backup time window is used with waiting, not with working, is to
enable multiple concurrent jobs. That's probably what you'd want in
production mode.)
That's a good idea. I didn't think of it because I only thought of
running jobs with one tape drive.
Stefan
--
Stefan Schwietert Fon +49 731 1530 44
Ulmer Volkshochschule Fax +49 731 1530 60
Systembetreuer email [EMAIL PROTECTED]
begin:vcard
fn:Stefan Schwietert
n:Schwietert;Stefan
org:Ulmer Volkshochschule
adr;dom:;;Kornhausplatz 5;Ulm;;89073
email;internet:[EMAIL PROTECTED]
tel;work:+49 731 1530 44
tel;fax:+49 731 1530 60
x-mozilla-html:FALSE
url:http://www.vh-ulm.de
version:2.1
end:vcard
