Hello,
thanx to your suggestions, I think I came up with a good configuration for my daily full backup.
But I ran into a strange situation I don't understand.
I have two same installations, with this new good config active since Monday.
On Monday night:
- 1st setup had the correct volume inserted, and backed up everything correctly.
- 2nd setup had an erroneous volume inserted (the Tuesday one). Bacula started sending email requests to insert the correct Monday volume. I had 4 jobs waiting. Each job has a MaxWaitDelay of 16h. Jobs starts at Monday 11:00pm with 5 mins offset each. One of the jobs was still requesting the correct tape at Tuesday 5pm (the others were already marked as canceled), so I had to cancel this job manually to be sure that the new nightly backup was correctly run...Why?!
I expected Bacula to cancel all jobs withing 16 hours (Tuesday 3pm).
Also, when I ran the cancel command, I was presented the choice of the job, with the list of the 4 jobs, as if they were all waiting. But when I canceled the one I presumed to be still waiting, all of them disappeared, and had no more jobs waiting.
Then comes Tuesday night.
Consider that my config has AutoMount turned on, and has the last Catalog backup job running an external script to eject the tape. Obviously on Monday the 1st setup ejected the tape, the 2nd one did not eject the tape (because jobs were waiting and I had to manually cancel).
Because I wanted to lower the MaxWaitDelay to 12 hours, I stopped and restarted the Bacula daemons on both Machine on Tuesday evening (before the new jobs times).
On Tuesday night:
- 1st setup ran everything fine.
- 2nd setup did not run the new job, waiting for human intervention! I found the status of the storage unmounted and BLOCKED. I had to manually mount the device, and after few seconds I got the Bacula message stating that the tape was being recycled and jobs were released and started. Why?!
Gabriele.
thanx to your suggestions, I think I came up with a good configuration for my daily full backup.
But I ran into a strange situation I don't understand.
I have two same installations, with this new good config active since Monday.
On Monday night:
- 1st setup had the correct volume inserted, and backed up everything correctly.
- 2nd setup had an erroneous volume inserted (the Tuesday one). Bacula started sending email requests to insert the correct Monday volume. I had 4 jobs waiting. Each job has a MaxWaitDelay of 16h. Jobs starts at Monday 11:00pm with 5 mins offset each. One of the jobs was still requesting the correct tape at Tuesday 5pm (the others were already marked as canceled), so I had to cancel this job manually to be sure that the new nightly backup was correctly run...Why?!
I expected Bacula to cancel all jobs withing 16 hours (Tuesday 3pm).
Also, when I ran the cancel command, I was presented the choice of the job, with the list of the 4 jobs, as if they were all waiting. But when I canceled the one I presumed to be still waiting, all of them disappeared, and had no more jobs waiting.
Then comes Tuesday night.
Consider that my config has AutoMount turned on, and has the last Catalog backup job running an external script to eject the tape. Obviously on Monday the 1st setup ejected the tape, the 2nd one did not eject the tape (because jobs were waiting and I had to manually cancel).
Because I wanted to lower the MaxWaitDelay to 12 hours, I stopped and restarted the Bacula daemons on both Machine on Tuesday evening (before the new jobs times).
On Tuesday night:
- 1st setup ran everything fine.
- 2nd setup did not run the new job, waiting for human intervention! I found the status of the storage unmounted and BLOCKED. I had to manually mount the device, and after few seconds I got the Bacula message stating that the tape was being recycled and jobs were released and started. Why?!
Gabriele.
|
|
|
Gabriele Bulfon - Sonicle S.r.l. Tel +39 028246016 Int. 30 - Fax +39 028243880 Via Felice Cavallotti 16 - 20089, Rozzano - Milano - ITALY http://www.sonicle.com |
