----- Original Message -----
Sent: Sunday, May 04, 2008 5:11 PM
Subject: [Bacula-users] A reasonable way
to get a list of files backed up ina job?
Hi all,
We're migrating to Bacula from a fairly random
system of backups, many of which used tar jobs being run on servers and then
copying them to a "backup machine" (via NFS) where they were manually migrated
to another medium. (You don't want to hear what that other medium
is...)
To move this migration along, I'd like to have a way to
determine the files that were backed up in job, so that I can delete
them. This way, I could migrate these tar files over to our bacula
machine, and then delete them once I know they are on tape. Generally,
the names are date-tagged in some way, but this isn't something that I want to
rely on (in case the copy process gets hosed, and there are 'old' files that
actually haven't been copied by Bacula).
If there were some reasonable
way to get a list from Bacula of the files in a particular job, that would
help. I see that I can run a Verify job (say, type DiskToCatalog) and
get a list of the files that have changed or otherwise differ from the state
of those on the backup set, or I can get a list of files that were
not
saved or encountered errors using parameters in a message resource ("skipped"
or "notsaved" types) in a particular backup job. This is the converse of
what I'd like to have.
There doesn't seem to be a way to determine in
this way what files
were backed up, or haven't been changed since a
backup was made.
What I'm doing now is creating a sentinel file in the
RunBeforeJob and then deleting all files "older" than this on successful
completion (in a RunAfterJob) script. This works, but it's not something
that I'm particularly comfortable with - I can't think of how it might get
screwed up, but I'd rather again have a positive state to check rather than
the lack of a negative state. And, I presume that I could write an SQL
query of some sort and query the database - perhaps through a script talking
to bconsole - but that strikes me as again not really plain enough for
whomever takes over management of this task from me when I'm done with
it.
Is there something that I'm missing here, either in my general
approach or in how the particular code in the daemons is implemented?
I'm running 2.2.8 on OpenSUSE 10.2 with MySQL.
Thanks!
_KMP
Boston