On Thu, 13 Sep 2012 10:39:39 -0400
Win Htin <win.htin AT gmail DOT com> wrote:
> Hi folks,
>
> I need to restore files from 4+ yrs back. I know the volume(s) where
> the files are backed up but the backup records have been pruned. Since
> I do have the db catalog backed up on a daily basis following is what
> I am planning to do.
>
> 1. /etc/bacula/bin/bls -j -V <volume_name> <devicename>
> 2. /etc/bacula/bin/bextract
>
> My question is, what parameters do I give to extract the db catalog
> file? The Bacula utility document has examples for restoring files but
> as far as I am aware of, not for extracting the db catalog back to a
> directory you specify on the server.
>
> Following is an example of my catalog backup report:
>
> --------------------- start ---------------------------------------
> 12-Sep 09:48 baculaserver-dir JobId 44234: BeforeJob: run command
> "/etc/bacula/config/make_catalog_backup
> bacula bacula BACULA localhost"
[...]
You seems to be unaware of how catalog backup works in your Bacula
install. "The truth" here is that Bacula does not have any special
means to back up its catalog--the script listed in your backup job
output supposedly just calls something like `mysqldump` to dump the
specified database to a file (supposedly placed somewhere
under /var/lib/bacula and named something like "bacula.sql").
That file then gets backed in a way no different from all those other
files backed up on the same host. A matching post-job script then
deletes the file containing the database dump.
Note that this file is only backed up if the rules in the FileSet
resource used by the backup job matche that file.
So you have to go this route:
1) Read the /etc/bacula/config/make_catalog_backup script and deduce
how the file it creates is named.
2) Then use bls on the relevant volumes and save its output to a file.
3) Inspect the generated listing looking for files matching the name
you deduced on step (1).
4) If you're lucky and your search turns out to be successful,
copy and paste the name of the relevant file exactly as it is
listed to another file, say, "/tmp/include". Notice the name of the
volume which contains the relevant file.
5) Run bextract on the volume defined on step (4) and pass it the
include file using the "-i" command line argument.
After bextract finishes, you'll get your catalog dump file on disk.
6) Use a tool appropriate for your database engine to restore the
catalog (for MySQL that would supposedly be
`mysql -u bacula -p bacula </path/to/the/dump.sql`).
Be sure to first back up the existing catalog!
Having said that, I should note two things.
First, if you need to just restore several files, it mighe be more
sensible to not bother with catalog at all: just run bls on the
relevant volume(s), find the needed files, construct the include file
and run bextract to extract those files. If your files have well-known
names, then just use bextract directly.
Second, there is a pitfall related to excavation of *expired* jobs:
after certain operations, Bacula performs a run over the catalog trying
to scavenge expired job records etc, and this process will happily
delete the records you restored. To combat this, you have to
temporatily adjust your expiration policy for that client -- it must be
set to a longer period than the age of the oldest job you intend to
restore. After you have restored what's needed, revert the expiration
period back.
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users
|
