Sorry for the added noise on the list.
I just had a duh! Moment and decided to add the word "days" to the "Max Full
Interval = 6" entry in the job definition and voila!, next job run was an
incremental. I guess the interval defaulted to seconds and not days as I
incorrectly assumed.
Thank you for a great product and community,
Erik
-----Original Message-----
From: Erik Pinzon
Sent: Tuesday, March 25, 2014 1:29 PM
To: 'bacula-users@lists.sourceforge.net'
Subject: All incremental jobs are upgraded to Full when a full already exists
and no changes to FileSet or other job properties have been made
We are in the process of replacing a very old (1.38) Bacula installation with a
newer version of it. The new installation is version 5.2.6, but the same issue
was present while testing version 5.0.2.
Database is PostgreSQL 9.1 on Debian Squeeze 6.0.9, bacula 5.0.2 is the default
package and 5.2.6 is available via Debian backports.
The issue is that all Incremental jobs are being upgraded to Full because
Bacula claims that: "No prior or suitable Full backup found in catalog". This
happens even after running a full backup manually.
I enabled full logging on the DB and could see Bacula executing the following
statements; which should result in running an Incremental backup, but the job
was upgraded to full regardless. Any help on finding what could be wrong with
my configuration would be much appreciated.
PostgreSQL Log of the job run:
2014-03-25 12:13:42 EDT LOG: statement: SELECT ClientId,Uname FROM Client
WHERE Name='iis2-fd'
2014-03-25 12:13:42 EDT LOG: statement: INSERT INTO Job
(Job,Name,Type,Level,JobStatus,SchedTime,JobTDate,ClientId,Comment) VALUES
('iis2-job.2014-03-25_12.13.42_09','iis2-job','B','I','C','2014-03-25
12:13:40',1395764020,5,'')
Notice that the job is initially inserted in the DB as level "I" -- Incremental
2014-03-25 12:13:42 EDT LOG: statement: SELECT currval('Job_Jobid_seq')
2014-03-25 12:13:42 EDT LOG: statement: SELECT FileSetId,CreateTime FROM
FileSet WHERE FileSet='iis2-fset' AND MD5='mF/UX5/802/Vv7+2i9/joD'
This query returns 1 row:
filesetid | createtime
-----------+---------------------
1 | 2014-03-25 10:45:20
(1 row)
2014-03-25 12:13:42 EDT LOG: statement: SELECT StartTime, Job FROM Job WHERE
JobStatus IN ('T','W') AND Type='B' AND Level='F' AND Name='iis2-job' AND
ClientId=5 AND FileSetId=1 ORDER BY StartTime DESC LIMIT 1
Using the FileSet ID found in the previous query, this query returns 1 row,
which I assume it indicates that in fact there is at least one successful full
backup for the same client, job and fileset.
starttime | job
---------------------+---------------------------------
2014-03-25 12:06:22 | iis2-job.2014-03-25_12.06.19_07
(1 row)
2014-03-25 12:13:42 EDT LOG: statement: SELECT StartTime, Job FROM Job WHERE
JobStatus IN ('T','W') AND Type='B' AND Level IN ('I','D','F') AND
Name='iis2-job' AND ClientId=5 AND FileSetId=1 ORDER BY StartTime DESC LIMIT 1
Bacula seems to ignore the result from the previous query and moves on to try
to find an incremental, differential or full backup. The query returns:
starttime | job
---------------------+---------------------------------
2014-03-25 12:06:22 | iis2-job.2014-03-25_12.06.19_07
(1 row)
2014-03-25 12:13:42 EDT LOG: statement: SELECT StartTime, Job FROM Job WHERE
JobStatus IN ('T','W') AND Type='B' AND Level='F' AND Name='iis2-job' AND
ClientId=5 AND FileSetId=1 ORDER BY StartTime DESC LIMIT 1
Here Bacula re-runs the first job query. Same results as above, then it records
in the logs that no full backup was found and moves on to upgrading the job to
full.
2014-03-25 12:13:42 EDT LOG: statement: INSERT INTO Log (JobId, Time,
LogText) VALUES (4,'2014-03-25 12:13:42','mgt1-dir JobId 4: ')
2014-03-25 12:13:42 EDT LOG: statement: INSERT INTO Log (JobId, Time, LogText)
VALUES (4,'2014-03-25 12:13:42','mgt1-dir JobId 4: No prior or suitable Full
backup found in catalog. Doing FULL backup.
')
2014-03-25 12:13:42 EDT LOG: statement: SELECT
PoolId,Name,NumVols,MaxVols,UseOnce,UseCatalog,AcceptAnyVolume,AutoPrune,Recycle,VolRetention,VolUseDuration,MaxVolJobs,MaxVolFiles,MaxVolBytes,PoolType,LabelType,LabelFormat,RecyclePoolId,ScratchPoolId,ActionOnPurge
FROM Pool WHERE Pool.Name='iis2-pool'
2014-03-25 12:13:42 EDT LOG: statement: SELECT count(*) from Media WHERE
PoolId=7
2014-03-25 12:13:44 EDT LOG: statement: UPDATE Job SET
JobStatus='R',Level='F',StartTime='2014-03-25
12:13:44',ClientId=5,JobTDate=1395764024,PoolId=7,FileSetId=1 WHERE JobId=4
2014-03-25 12:13:44 EDT LOG: statement: UPDATE Job SET
JobStatus='R',Level='F',StartTime='2014-03-25
12:13:44',ClientId=5,JobTDate=1395764024,PoolId=7,FileSetId=1 WHERE JobId=4
Notice her how now the job has been upgraded to Full even though there was
clearly a recent (7 minutes old) full backup in the catalog.
Resources for this job:
Client {
Name = iis2-fd
Address = iis2.childrenstrust.pvt
FDPort = 9102
Catalog = MyCatalog
Password = "zzzzz"
File Retention = 30 days
Job Retention = 3 months
AutoPrune = yes
}
Job {
Name = iis2-job
Client = iis2-fd
Type = Backup
Schedule = Regular
FileSet = iis2-fset
Max Full Interval = 6 # this entry Should force a full backup every 7 days
Pool = iis2-pool
Messages = Standard
Write Bootstrap = "/opt/bacula/bootstrap/iis2.bsr"
Full Backup Pool = iis2-pool
Incremental Backup Pool = iis2-pool
}
Pool {
Name = iis2-pool
Storage = iis2-storage
Pool Type = Backup
Recycle = yes
Recycle Oldest Volume = yes
AutoPrune = yes
Volume Retention = 14 days # this entry should remove volumes older than 2
weeks
Maximum Volume Jobs = 1
Maximum Volumes = 16 # keep a couple of extra volumes to accommodate
2 weeks of 1 full + 6 incr per week
Label Format = "${Pool}-vol-${JobId}"
}
Storage {
Name = iis2-storage
Address = mgt2.thechildrenstrust.net
SDPort = 9103
Password = "zzzzz"
Device = iis2-device
Media Type = iis2-mtype
}
FileSet {
Name = iis2-fset
Include {
Options {
signature = MD5
Exclude = yes
IgnoreCase = yes
}
File = "C:/CrissCross"
File = "C:/internal_reports"
File = "C:/provider_reports"
}
}
Thank you,
Erik
------------------------------------------------------------------------------
Learn Graph Databases - Download FREE O'Reilly Book
"Graph Databases" is the definitive new guide to graph databases and their
applications. Written by three acclaimed leaders in the field,
this first edition is now available. Download your free book today!
http://p.sf.net/sfu/13534_NeoTech
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
