Should Consolidate consider Job or just Client and FileSet?

[Joshua Myles]

I have two AI jobs that have the same client and fileset but different pools (and storages, media types, etc.) Today I saw that when my consolidate job ran it picked up a list of job IDs to consolidate but ran them for the wrong job and thus tried to consolidate to a volume in the wrong pool. Could this be because the AccurateGetJobids function does not consider the job?

https://github.com/bareos/bareos/blob/7b4067ecb525c6fc1b26fe75794371a1ae8398a9/core/src/cats/sql_get.cc#L1355-L1368

I'd previously thought it was the combination of Job/FileSet/Client that defined uniqueness but the query for finding jobs to consolidate doesn't seem to consider the job.

Josh

[Joshua Myles]

Here aresome log snippets in case it jogs someone's memory.

During Consolidate, it finds jobs to consolidate, but these are job IDs from job fatboy-local-AI not fatboy-sshfs-AI.

25-Mar 14:59 bareos-dir JobId 8119: Looking at always incremental job fatboy-sshfs-AI
25-Mar 14:59 bareos-dir JobId 8119: fatboy-sshfs-AI: considering jobs older than 11-Mar-2026 14:59:02 for consolidation.
25-Mar 14:59 bareos-dir JobId 8119: before ConsolidateFull: jobids: 6283,7605,5519,5639
25-Mar 14:59 bareos-dir JobId 8119: check full age: full is 16-Nov-2025 01:42:54, allowed is 25-Feb-2026 13:59:02
25-Mar 14:59 bareos-dir JobId 8119: Full is older than AlwaysIncrementalMaxFullAge -> also consolidating Full jobid 6283
25-Mar 14:59 bareos-dir JobId 8119: after ConsolidateFull: jobids: 6283,7605,5519,5639
25-Mar 14:59 bareos-dir JobId 8119: fatboy-sshfs-AI: Start new consolidation
25-Mar 14:59 bareos-dir JobId 8119: Using Catalog "MyCatalog"
25-Mar 14:59 bareos-dir JobId 8119: Job queued. JobId=8120
25-Mar 14:59 bareos-dir JobId 8119: Consolidating JobId 8120 started.

In the virtual full job, it tries to consolidate to a volume in fatboy-sshfs-AI's pool. The job fails when it can't read the first volume from the fatboy-local-AI's pool (different storage, different media type).

25-Mar 14:59 bareos-dir JobId 8120: Version: 25.0.3~pre68.e4a45c2dc (23 March 2026) Debian GNU/Linux 13 (trixie)
25-Mar 14:59 bareos-dir JobId 8120: Start Virtual Backup JobId 8120, Job=fatboy-sshfs-AI.2026-03-25_14.59.03_09
25-Mar 14:59 bareos-dir JobId 8120: Bootstrap records written to /var/lib/bareos/bareos-dir.restore.26.bsr
25-Mar 14:59 bareos-dir JobId 8120: Consolidating JobIds 6283,7605,5519,5639 containing 184965 files
25-Mar 14:59 bareos-dir JobId 8120: Connected Storage daemon at bareos.onnela.feep.org:9103, encryption: TLS_CHACHA20_POLY1305_SHA256 TLSv1.3
25-Mar 14:59 bareos-dir JobId 8120:  Encryption: TLS_CHACHA20_POLY1305_SHA256 TLSv1.3
25-Mar 14:59 bareos-dir JobId 8120: Using Device "bareos-sshfs1" to read.
25-Mar 14:59 bareos-sd JobId 8120: Using just in time reservation for job 8120
25-Mar 14:59 bareos-dir JobId 8120: Using Device "JustInTime Device" to write.
25-Mar 14:59 bareos-sd JobId 8120: Moving to end of data on volume "fatboy-sshfs-AI-Consolidated-1542"
25-Mar 14:59 bareos-sd JobId 8120: Ready to append to end of Volume "fatboy-sshfs-AI-Consolidated-1542" size=262
25-Mar 14:59 bareos-sd JobId 8120: stored/acquire.cc:156 Changing read device. Want Media Type="File" have="File-sshfs"
  device="bareos-sshfs1" (/var/lib/bareos/storage-sshfs)
25-Mar 14:59 bareos-sd JobId 8120: Releasing device "bareos-sshfs1" (/var/lib/bareos/storage-sshfs).
25-Mar 14:59 bareos-sd JobId 8120: Fatal error: stored/acquire.cc:206 No suitable device found to read Volume "fatboy-local-AI-Consolidated-0052"
25-Mar 14:59 bareos-sd JobId 8120: Releasing device "bareos-sshfs-consolidated1" (/var/lib/bareos/storage-sshfs).
25-Mar 14:59 bareos-sd JobId 8120: Releasing device "bareos-sshfs1" (/var/lib/bareos/storage-sshfs).
25-Mar 14:59 bareos-dir JobId 8120: Replicating deleted files from jobids 6283,7605,5519,5639 to jobid 8120
25-Mar 14:59 bareos-dir JobId 8120: Error: Bareos bareos-dir 25.0.3~pre68.e4a45c2dc (23Mar26):

I realize this could possibly be worked around by making a copy of the file set and changing one of the job definitions. But I still wonder if this behavior is intentional, or whether the consolidate query should also consider job name.

For reference, job configurations. Pools and storages are also separate, and I can include those configs if relevant.

Job {
  Name = "fatboy-local-AI"
  Description = "fatboy shares"
  Client = "spot-enc-fd"
  JobDefs = "AlwaysIncrementalJob"

  FileSet = "fatboyShares"
  Schedule = "Daily at 22:30"

  Always Incremental Job Retention = 62 days
  Always Incremental Keep Number = 62
  Always Incremental Max Full Age = 76 days

  Pool = fatboy-local-AI-Incremental
  Incremental Backup Pool = fatboy-local-AI-Incremental
  Full Backup Pool = fatboy-local-AI-Consolidated
}

Job {
  Name = "fatboy-sshfs-AI"
  Description = "fatboy shares"
  Client = "spot-enc-fd"
  JobDefs = "AlwaysIncrementalJob"

  FileSet = "fatboyShares"
  Schedule = "Daily at 23:30"

  Always Incremental Job Retention = 14 days
  Always Incremental Keep Number = 14
  Always Incremental Max Full Age = 28 days

  Pool = fatboy-sshfs-AI-Incremental
  Incremental Backup Pool = fatboy-sshfs-AI-Incremental
  Full Backup Pool = fatboy-sshfs-AI-Consolidated
}

Josh