[OpenAFS] Howto speedup the restore of a crashed fileserver
Mon, 01 Oct 2007 16:15:04 -0400
Jose Calhariz wrote:
> On Mon, Oct 01, 2007 at 11:25:03AM -0700, Russ Allbery wrote:
>> Jose Calhariz <email@example.com> writes:
>>> I have lost a fileserver because of a corruption on the reiserfs v3.
>>> Something like 1000 volumes and 100GB of data were lost. The backups
>>> were made using amanda with standard tar.
>>> My problem is the recovering of the data to the surviving filesystem
>>> is taking too much time. Something like 7 hours to recover 7 GB.
>>> Will I need 100hours to recover from tar 100GB?
>> What did you tar up? I can think of a lot of possibilities (the raw namei
>> filesystem, the contents of AFS archived with admin credentials, tarballs
>> of volume dumps), and the answer somewhat depends.
> I have tarballs with the contents of AFS using admin credentials.
>> If you've archived AFS via tar using admin credentials, I'm not sure
>> there's anything you can do that's faster than creating new volumes and
>> untarring the contents into those volumes. You can do that directly on
>> the relevant file server to reduce your network latency a little.
If you're not already, I recommend restoring into a local directory,
then copying into AFS. That will at least make the tape part go faster
and reduce wear and tear on the tapes and drive.