2

I am looking for a way to perform a batch (non interactive) restore for a Client that has multiple job definitions which don't share the same file set.

For a client that has a single fileset I can issue a

restore client="example.eu" jobid=99,152 where="/tmp" select current all done yes

which when piped to bconsole will schedule a restore job, for the given jobids (which belong to a single fileset).

When there are more than one filesets present for the specific client bconsole prompts:

Using Catalog "MyCatalog"
The defined FileSet resources are:
     1: example.eu fileset_1
     2: example.eu fileset_2
     3: example.eu fileset_3
Select FileSet resource (1-3):

Which prevents me from running it as a batch job.

So how can I have the same batch behavior for a client that has more than one filesets in the catalog?

I would expect something like a fileset=fileset_1 option in the restore command.

xlembouras
  • 131
  • 5

2 Answers2

1

In case somebody ends up here, after some digging I ended up splitting the process to some steps:

  • find the job ids that describe the backup I need. (one full backup and possibly a combination of a differential or/and incrementals backups)
  • find all the Files that these jobs backed up and put them to a file, on my bacula director system.
  • use restore's file=</tmp/file_list option which tells bacula to fetch the files that are listed to the file_list file.

restore client="example.eu" jobid=99,152 file=</tmp/file select current all done yes

I am not really happy with this solution, because I would like the restore to be more transparent to me.

xlembouras
  • 131
  • 5
0

Not much of pro, but could you try restoring using "bextract"? "-i" option specifies an include file with a list of foledrs and files to be extracted. A full command should be something like this: bextract -i -b /archive/device /path/to/restore