Bareos 23.0.1~pre7.606b211eb-33 error: Volume Header Id bad: Bacula 1.0 immortal

[Heidi van Niekerk]

HiSince the upgrade to bareos 23.0.1~pre7.606b211eb-33, we are receiving this error when attempting to do some backups (Virtual Full, Incremental) and restores:

04-Jan 08:22 bareos_server-sd JobId xxxx: Warning: stored/acquire.cc:325 Read acquire: Volume Header Id bad: Bacula 1.0 immortal04-Jan 04:14 bareos_server-sd JobId xxxx: Please mount read Volume "FullVolxxxx" for:
    Job:          servername_BackupJob.2024-01-04_03.13.00_36
    Storage:      "dev0010_volume" (/backups/dev0010)
    Pool:         servernameDefault
    Media type:   File

We have established that this has to do with the ID and version number in the header of the actual volume.  When checking the header with bls, the backups on volumes where we see the error:Volume Label:
Id                : Bacula 1.0 immortal
VerNo             : 11
VolName           : IncrVolxxxx
PrevVolName       :
VolFile           : 0
LabelType         : VOL_LABEL
LabelSize         : 214
PoolName          : servername_incremental
MediaType         : File
PoolType          : Backup
HostName          : bareos_server
Date label written: 31-Jul-2013 20:27
08-Jan 09:06 bls JobId 0: Releasing device "dev0001_volume" (/backups/dev0001).

Before 2016, we used Bacula, but migrated to bareos in that year.  The above server was one of the servers that would have had backups on bacula prior to changing to Bareos.  We can therefore only assume that that is the reason why the Id mentions Bacula.Our newer servers, ones built after the migration to Bareos, have headers similar to the one below - they don't have this issue.Volume Label:
Id                : Bareos 2.0 immortal
VerNo             : 20
VolName           : FullVolxxxx
PrevVolName       :
VolFile           : 0
LabelType         : VOL_LABEL
LabelSize         : 218
PoolName          : servername_full
MediaType         : File
PoolType          : Backup
HostName          : bareos_server
Date label written: 25-Jun-2023 18:23
08-Jan 09:04 bls JobId 0: Releasing device "dev0001_volume" (/backups/dev0001).We have looked at the changelogs for Bareos 23.0.1 pre7, but could not find relating information.  Has this version stopped support for volumes with Id Bacula 1.0 immortal?
If so, is there a safe way to change the Id and version number on the volumes so that it is compatible with versions 23.0.1~pre7.606b211eb-33 and higher?We run Virtual Full backups every 2 weeks with daily Incrementals.

Our host_template looks like this:FileSet {
  Name = SHORTHOSTNAME_FileSet
  Ignore File Set Changes = yes
  Include  {
    Options {
      compression=GZIP
      signature=MD5
      noatime=yes
    }
    INCLUDES
    Exclude Dir Containing = .backup_exclude
  }
  Exclude {
    EXCLUDES
  }
}Job {
  Name = "SHORTHOSTNAME_BackupJob"
  Type = Backup
  Accurate = yes
  Allow Duplicate Jobs = no
  Cancel Lower Level Duplicates = yes
  Cancel Running Duplicates = no
  Client = SHORTHOSTNAME
  FileSet= "SHORTHOSTNAME_FileSet"
  Full Backup Pool = "SHORTHOSTNAME_full"
  Incremental Backup Pool = "SHORTHOSTNAME_incremental"
  Messages = Standard
  Pool = SHORTHOSTNAMEDefault
  Priority = 10
  Allow Mixed Priority = yes
  Schedule = SHORTHOSTNAME_Schedule
  Storage = SHORTHOSTNAME_FStorage
  ClientRunBeforeJob = /usr/local/bin/prebackup
  ClientRunAfterJob = /usr/local/bin/postbackup
  RunAfterJob = "/usr/local/bin/poller.rb %c"
  RunScript {
    Command = "/usr/local/bin/full_backup_if_sane %c"
    FailJobOnError = yes
    RunsWhen       = After
    RunsOnClient   = no
    RunsOnFailure  = no
 }
}Job {
  Name = "SHORTHOSTNAME_RestoreJob"
  Type = Restore
  Client = SHORTHOSTNAME
  FileSet= "SHORTHOSTNAME_FileSet"
  Messages = Standard
  Pool = SHORTHOSTNAMEDefault
  Priority = 1
  Allow Mixed Priority = yes
  Storage = SHORTHOSTNAME_FStorage
  ClientRunBeforeJob = "/usr/local/bin/prerestore %n"
  RunScript {
    Command = "/usr/local/bin/postrestore %n"
    RunsWhen = After
    RunsOnFailure = yes
    RunsOnClient  = yes
    RunsOnSuccess = yes    # default, you can drop this line
  }
}Client {
  Name = SHORTHOSTNAME
  Address = FQDN
  Catalog = MyCatalog
  Password = "xxxxxxxxxxxx"
  Maximum Bandwidth Per Job = 9 Mb/s
}Pool {
  Name = "SHORTHOSTNAME_full"
  Pool Type = Backup
  Recycle = yes                       # Bareos can automatically recycle Volumes
  AutoPrune = yes                     # Prune expired volumes
  LabelFormat = "FullVol"
  Storage = SHORTHOSTNAME_VFStorage
  Maximum Volume Jobs = 1
  Maximum Volumes = 2
  Volume Retention = 3 weeks          # Keep for 3 weeks
}Pool {
  Name = "SHORTHOSTNAME_incremental"
  Pool Type = Backup
  Recycle = yes                       # Bareos can automatically recycle Volumes
  AutoPrune = yes                     # Prune expired volumes
  ActionOnPurge = Truncate
  LabelFormat = "IncrVol"
  Storage = SHORTHOSTNAME_FStorage
  Maximum Volume Jobs = 1
  Maximum Volumes = 28
  Volume Retention = 3 weeks          # Keep for 3 weeks
}Pool {
  Name = "SHORTHOSTNAMEDefault"
  Pool Type = Backup
  Recycle = yes                       # Bareos can automatically recycle Volumes
  AutoPrune = yes                     # Prune expired volumes
  LabelFormat = "DefVol"
  NextPool = "SHORTHOSTNAME_full"
  Storage = SHORTHOSTNAME_FStorage
  Maximum Volume Jobs = 1
}Schedule {
  Name = SHORTHOSTNAME_Schedule
SCHEDULE
}Storage {
  Name = SHORTHOSTNAME_FStorage
  Address = STORAGE_DAEMON
  Password = "xxxxxxxxxxxxxx"
  Device = STORAGEDEVICE_volume
  Media Type = File
  Maximum Concurrent Jobs = 20
}Storage {
  Name = SHORTHOSTNAME_VFStorage
  Address = STORAGE_DAEMON
  Password = "xxxxxxxxxxxxxx"
  Device = STORAGEDEVICE_virtualfull
  Media Type = File
  Maximum Concurrent Jobs = 20
}

We found this:
https://groups.google.com/g/bareos-users/c/_o4lSHP41oo/m/oFmTuBRuBVYJ
Suggesting setting compatibility.  If this is the way to go, where would this be set?  On bareos_fd configuration on the host, or on the bareos storage daemon configuration?

Any asisstnace will be greatly appreciated.

[Bruno Friedmann (bruno-at-bareos)]

Hi,

The compatible option was in deprecated mode since a few release, and has been removed finally in 23.

Here it look like you're recycling a very old volume since ages. The best advise would be to either try a manual relabel (when in purge status) check if the header is then correctly rewrite, otherwise retire the volume (purge, then delete, and remove for FS)

Regards.