On 8/31/17 1:32 AM, Julian Poß wrote:
However, do you know how bareos handles incremental backups? For
example i want to keep 14 restore points on disk, and one complete
week of all backups per tape set.

This (imo) will require two jobs per client, where one of them writes
to disk and the other to tape. But does bareos know, that the
incremental backup stored on disk should be ignored, while creating
the next incremental backup for tape? Since this will probably be
done by two seperated jobs, bareos should know the different data,
based on the job history/job state, right?

Furthermore, i guess you do not have any experiences regarding
multiple storage devices on a single director, do you? I would be
relieved, if somebody could confirm my guesses regarding the bug and
fix. Thanks again, so much!

So, I have a configuration that is fairly on point for you. I do always incremental backups to disk with 30 restore points on disk.

And each week I do off-site backups to tape using virtual full backups. This is quite neat in that the virtual full backups do not go back to the client, instead they build a synthetic full backup based upon all of the backups available on disk on the server.

The trick that I found for having Bareos then ignore the off-site backups for other operations is this tiny script that gets run after each virtual full backup via:

run after job = "/usr/meridian/share/libexec/bareos/turn_virtual-full-into-archive %i"

The script is:

#!/bin/zsh

if [ $# -eq 1 ]; then
    jobid="$1"
    shift
else
    echo Usage: $0 job-id
    exit 255
fi

/usr/local/bin/psql --dbname=bareos --echo-errors --quiet \
--command="UPDATE job SET type='A' WHERE jobid='${jobid}'"

exit 0

And I have two storage devices configured on my director, "disk" and "lto6-1".

--
You received this message because you are subscribed to the Google Groups 
"bareos-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to