#386 Issue closed: Issue in HP Data Protector restore script

Labels: bug

revog opened issue at 2014-04-02 12:04:

Hi there

Currently I'm implementing ReaR in relation to HP Data Protector on CentOS 6 machines.
Generation of ISO worked fine. During reverse engineering of failed disaster recovery process I got stuck on the restore/../40_restore_with_dp.sh script.

I've seen, that we iterate through available sessions, backuped hosts and file systems. After they get stored under /tmp/(dp_)*.

Within the script verify/../50_select_dp_restore.sh on line 30 we query the available filesystems and store them in /tmp/dp_list_of_fs_objects:

${OMNIDB} -filesystem | grep "${HOST}" | cut -d"'" -f -2 > /tmp/dp_list_of_fs_objects

But in a later step (40_restore_with_dp.sh) we read in the file /tmp/list_of_fs_objects which contains ALL available filesystems of all backuped servers. This leads to the fact, that rear tries to restore the first in the list (which does not belong to the client beeing restored in my example) and fails.

Am I doing something wrong or shouldn't the 40_restore_with_dp.sh script take the /tmp/dp_list_of_fs_objects list instead of /tmp/list_of_fs_objects?

Thanks for helping.


gdha commented at 2014-04-02 16:09:

@revog it looks your re-engineering work was the nail on the head ;-)
looking forward to your pull request. Thanks for the debugging.

gdha commented at 2014-04-03 10:40:

@revog could you verify it works well and close the issue if possible?

revog commented at 2014-04-03 13:42:

Thx @gdha for your prompt support and merge. Just tested and it works like a charm! cheers

[Export of Github issue for rear/rear.]