This morning, I noticed that in my backup reports from Bacula, I had the following logged from our file server:
30-Jan 01:00 FileServer1-dir JobId 240: The following 3 JobIds were chosen to be copied: 161,163,231
30-Jan 01:00 FileServer1-dir JobId 240: Job not run.
30-Jan 01:00 FileServer1-dir JobId 240: Error: Could not start migration job.
30-Jan 01:00 FileServer1-dir JobId 240: Job not run.
30-Jan 01:00 FileServer1-dir JobId 240: Error: Could not start migration job.
30-Jan 01:00 FileServer1-dir JobId 240: Copying using JobId=231 Job=FileServer1.2012-01-29_23.59.01_21
30-Jan 01:00 FileServer1-dir JobId 240: Bootstrap records written to /var/lib/bacula/FileServer1-dir.restore.18.bsr
30-Jan 01:01 FileServer1-dir JobId 240: Start Copying JobId 240, Job=FileServer1_Copy.2012-01-30_01.00.00_30
30-Jan 01:01 FileServer1-dir JobId 240: Using Device "FileServer1_Full_Copy"
30-Jan 01:01 FileServer1-sd JobId 240: Ready to read from volume
"FileServer1_Full_1262" on device "FileServer1_Full"
(/mnt/backup/Bacula/FileServer1/Full).
As you can see, Job231 is copied with no problems, but Job161 and Job163 are not copied.
When run manually this morning, I was given a little more information:
run job=FileServer1_Copy pool=FileServer1_Full storage=FileServer1_Full
30-Jan 06:55 FileServer1-dir JobId 0: Fatal error: No Next Pool specification found in Pool "Default".
30-Jan 06:55 FileServer1-dir JobId 0: Fatal error: No Next Pool specification found in Pool "Default".
30-Jan 06:55 FileServer1-dir JobId 255: The following 3 JobIds were chosen to be copied: 2,161,163
30-Jan 06:55 FileServer1-dir JobId 255: Job not run.
30-Jan 06:55 FileServer1-dir JobId 255: Error: Could not start migration job.
30-Jan 06:55 FileServer1-dir JobId 255: Job not run.
30-Jan 06:55 FileServer1-dir JobId 255: Error: Could not start migration job.
30-Jan 06:55 FileServer1-dir JobId 255: Copying using JobId=163 Job=FileServer1.2012-01-26_23.59.00_09
After investigation this morning, I seem to be being bitten by this bug:
http://bugs.bacula.org/view.php?id=1399 however, there doesn't seem to be any activity since July last year. Now, I'm not complaining that the bug is not fixed - I appreciate that this is an open source project and I am grateful for vast amounts of time people spend on the project.
What I would like to know, though, is why I am being affected by this bug. To me, it seems as if I am doing something out of the ordinary, which is highlighting the issue, however I'm not sure what I'm doing wrong.
So that you can obtain a better understanding of the way in which I have setup Bacula, I have included the necessary sections of my bacula-dir.conf:
Client {
Name = FileServer1-fd
Address = FileServer1
FDPort = 9102
Catalog = MyCatalog
Password =
File Retention = 6 months
Job Retention = 6 months
AutoPrune = yes
Maximum Concurrent Jobs = 2
}
Schedule {
Name = "FileServer1 Copy Schedule"
Run = Level=Full Pool=FileServer1_Full Storage=FileServer1_Full mon at 01:00
Run = Level=Full Pool=FileServer1_Inc Storage=FileServer1_Inc tue-sun at 01:00
}
JobDefs {
Name = DefaultCopy
Type = Copy
Level = Full
Client = FileServer1-fd
Messages = Standard
Selection Type = PoolUncopiedJobs
}
Job {
Name = "FileServer1_Copy"
JobDefs = DefaultCopy
Type = Copy
Client = FileServer1-fd
FileSet = "FileServer1 System"
Pool = Default
Messages = Standard
Schedule = "FileServer1 Copy Schedule"
Storage = Default
}
Pool {
Name = FileServer1_Full_Copy
Pool Type = *Copy
Volume Retention = 1 weeks
Recycle = yes
AutoPrune = yes
LabelFormat = FileServer1_Full_Copy_
Maximum Volume Bytes = 1G
Storage = "FileServer1_Full_Copy"
}
I am using the packages provided in the main Ubuntu repository, so I am using 5.0.1 (Dir/SD/FD). I'm aware that there is a patch available on the aforementioned bug ticket, however I'm not in a position to be patching and compiling from source on this system, as it's in production. What I'm looking for is anyone's thoughts on why this is affecting me, and ways I could possibly work around the issue with the current version of Bacula I have installed.
I hope someone can help!
Thanks,
Joe Nyland
