On 6/10/22 17:08, Justin Case wrote:
Hi there,I am using Bacula since some weeks now, but operationally I am a complete noob. Haven’t run through many use cases yet. Since I have included my laptop as a job the incrementals run relatively quick, but the fulls take long. Sometime I put the laptop to sleep without thinking about the backup! If that happens during a job for the laptop, then the job stops if the connect does not work for some time. This leads to an errored job that has a lot of data in volumes, but the job is incomplete. I run that job later. OK, from a safety point of vie no problem at all. But what about the space occupied by the incomplete job? What is an efficient way to reclaim this space using the Baculum UI? Deleting the incomplete job does not do the trick I suppose? What are additional steps that are required? Thanks for your experienced feedback! J/C
Hello Justin, Short answer: You can not delete data inside Bacula volumes. Long answer:To begin the process of seeing if you can reclaim these volume(s), deleting the failed laptop jobs is the correct first step. This removes any mention of those jobs and files from the catalog, which means when you look to see what jobs are stored on the volume(s) that those failed jobs used (the next step to do), you will see either no jobs associated with the volume(s), or any other jobs that may have also used them.
When you check the "Jobs on volume" in Baculum, if there are no jobs associated with a volume, Bacula will just prune/purge/recycle and re-use them as necessary once a pruning operation takes place.
If there are other jobs that used these volumes, you can just let things be and the volumes will be pruned/purged/recycled and re-used once the volume reaches its retention period - as above.
If you want to reclaim the volumes sooner, you can delete the other jobs that are on the volume(s) - of course making the right decisions about important jobs and not deleting those.
But if you keep any jobs that wrote to the volume(s) and don't delete them, they are still in the catalog and you will need to wait for the volume retention period to pass before Bacula will re-use the volume(s).
Future hints:You may want to create a special pool for the laptop and maybe even set in this pool "MaximumVolumeJObs = 1". Then, you can be guaranteed that only laptop jobs will ever be on volumes in this pool, and that only one laptop job will ever be on any given volume in this pool.
Then, when you run into this situation again and you want to clean up, it will be as easy as just deleting the failed job(s), and then the volume(s) that the deleted job(s) used will be ready for re-use as detailed in the first paragraph. :)
Additionally, if you really want/need to reclaim the actual disk space these volumes use, there are some additional steps to take:
First, I would globally disable pruning in the Director's 'Director{}' block by adding "Autoprune = no". If you are running an older version of Bacula, this option may not exist.
Then, set "ActionOnPurge = truncate" in the new pool.Next create an Admin type job that runs nightly before the backup jobs run which will run a small script which calls some bconsole commands to do some catalog maintenance, including pruning and then truncating the pruned/purged volumes.
The script that my NightlyMaint Admin job runs looks like the following:
----8<----
#!/bin/bash
bcbin="/opt/comm-bacula/sbin/bconsole"
bcconfig="/opt/comm-bacula/include/bconsole.conf"
storage="XXXX" # set your defined storage resource name here
echo
echo "========"
echo "Running update stats days=3 command..."
echo -e "update stats days=3\nquit\n" | ${bcbin} -n -c ${bcconfig}
echo
echo "========"
echo "Running .bvfs_update command..."
echo -e ".bvfs_update\nquit\n" | ${bcbin} -n -c ${bcconfig}
echo
echo "========"
echo "Running prune jobs all yes command..."
echo -e "prune jobs all yes\nquit\n" | ${bcbin} -n -c ${bcconfig}
echo
echo "========"
echo "Pruning expired volumes"
echo -e "prune expired volume yes\nquit\n" | ${bcbin} -n -c ${bcconfig}
echo
echo "========"
echo "Running truncate allpools storage=${storage} command..."
echo -e "truncate allpools storage=speedy-file\nquit\n" | ${bcbin} -n -c
${bcconfig}
----8<----
My NightlyMaint Admin job looks like:
----8<----
Job {
Name = NightlyMaint
JobDefs = AdminJobs
RunScript {
RunsWhen = before
RunsOnClient = no
FailJobOnError = yes
# This script runs the update slots command piped to bconsole
# -----------------------------------------------------------
Command = "/opt/comm-bacula/include/scripts/updateslots.sh"
# This script runs prune expired volumes, update stats, and .bvfs_update
commands
#
-------------------------------------------------------------------------------
Command = "/opt/comm-bacula/include/scripts/consolecommands.sh"
# This command performs a PostgreSQL vacuum on the bacula catalog DB
# ------------------------------------------------------------------
Command = "/usr/bin/vacuumdb -q -z -d bacula"
}
}
----8<----
And finally, if you decide to use this "ActionOnPurge = truncate" option in any of your current pools, you will need to do
the following in bconsole to make sure any existing volumes in those pools inherit this new setting:
* reload * update pool=xxxx (where xxxx is each pool you have modified) * update volume fromallpools Hope this helps! Best regards, Bill -- Bill Arlofski w...@protonmail.com
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
