View Issue Details

IDProjectCategoryView StatusLast Update
0001611bareos-coredirectorpublic2024-04-15 09:34
Reportertmorris Assigned Tobruno-at-bareos  
Status closedResolutionwon't fix 
PlatformLinuxOSUbuntuOS Version20.04
Product Version23.0.1 
Summary0001611: Increasing retention period for pool does not increase retention period for previously created volumes when recycled
DescriptionWhen changing the "Volume Retention" setting on a pool, this does not update volumes in the pool when they are recycled. My expectation is that after recycling a volume will have the new value for retention period.
Steps To ReproduceCreate a pool with a very short retention period. Example:
Pool {
  Name = IncrPool0
  Pool Type = Backup
  Storage = BackupStore
  Recycle = yes
  Recycle Oldest Volume = yes
  Auto Prune = yes
  Action On Purge = Truncate
  Maximum Volume Jobs = 1
  Maximum Volumes = 1000
  Volume Retention = 32
  Label Format = "Incr-"
Run a number of jobs utilizing this pool, then modify the pool such that Volume Retention = 32 days
Run more jobs utilizing the same pool and see that the earlier volumes are recycled for new jobs, but the retention period on those volumes remains 32 seconds afer recycling. I have been able to see the continuing messages in bconsole that still show a retention period of 32 seconds even after the setting was changed and configs reloaded, and the volumes were recycled multiple times afterward.
Additional InformationIf this is an expected behavior, I believe it is undocumented, and there is likewise no documentation that I have found for how to fix the situation. The docs recommend not using the "delete" command, only mentioning that it necessitates the use of bscan, which itself notes that if it is being used something has gone wrong.

Changing the retention period of volumes in a pool upon rewrite seems like necessary functionality to me as a user -- same goes for sane changing of almost any configuration value, without having to start over. This issue is also confusing to debug, as the jobs get pruned and thus disappear from the records visible in the UI. I was only able to tell what was going on by reviewing messages in bconsole.

Obviously my steps to reproduce don't include my exhaustive configuration -- I haven't set up a fully isolated installation and done a truly minimal reproduction, but let me know if I can provide any other config or system particulars, or if there are other questions I can answer. Also, I selected "Major" for severity, as this severely impedes my ability to evolve my backup system in production, and "normal" for priority as I really don't know how this relates to other issues. Thank you
TagsNo tags attached.




2024-04-15 09:33

manager   ~0005896

It is documented in this chapter in the 3rd section
.. Another important aspect of a Pool is that it contains the default attributes (Maximum Jobs, Retention Period, Recycle flag, …) that will be given to a Volume when it is created. [...]. Each of these attributes can later be changed on a Volume by Volume basis using the update command in the console program. Note that you must explicitly specify which Pool Bareos is to use with each Job. Bareos will not automatically search for the correct Pool.
The command update of bconsole is also very well documented.


2024-04-15 09:34

manager   ~0005897

No changes without any PR coming from community will be made. Feel free to help improving the documentation by changing and rewording missing part, by proposing a PR on github.
Btw Mantis will be replaced soon, by github issue.

Issue History

Date Modified Username Field Change
2024-04-11 19:56 tmorris New Issue
2024-04-15 09:33 bruno-at-bareos Note Added: 0005896
2024-04-15 09:34 bruno-at-bareos Assigned To => bruno-at-bareos
2024-04-15 09:34 bruno-at-bareos Status new => closed
2024-04-15 09:34 bruno-at-bareos Resolution open => won't fix
2024-04-15 09:34 bruno-at-bareos Note Added: 0005897