[OpenAFS] Understanding questions backup volume

Frank Burkhardt fbo2@gmx.net
Thu, 9 Feb 2006 12:29:20 +0100


On Thu, Feb 09, 2006 at 11:43:45AM +0100, Lars Schimmer wrote:
> Hash: SHA1
> Hi!
> I start using backup volumes ;-)
> It is fairly easy to create one and mount them.
> But: Where is the difference between RO copies and a backup volume?
> I know, backup volumes should be used for backup, RO for distributing
> data all over the cell.

Backup volume instances are explicitely disk space efficiently stored plus
they can be referred by a volume mount point.

Hint: additional volume instances (clones) of a RW instance can be
create using 'vos clone' but it's not easy to mount them.

> A backup should be made of the backup volumes, because this doesn't lock
> the RW volumes for a long time.

Yes. Especially if you use the backup volume for different backup strategies
(see below).

> And if I vos dump the backup volumes to a backup server (amanda-afs or
> just plain dump) I could rebuild the backup volumes. Does this help me
> in case of a lost RW volume?

If you use 'vos dump', a backup volume is not necessary. Using 'vos dump
-clone' clones the RW instance first, dumps this clone and removes it after
the dump.

> At least a RO copy could be converted to a RW volume in nearly NO time,
> but a backup volume?

A backup volume can't because it's stored as a diff against the RW which means,
it's damaged when the RW is.

> Our cell is designed to have a RO copy of every RW volume.
> And if one RO copy of a RW volume resist on a file server housed in a
> datacenter "far away" I've got a quick and easy 1-day-backup in case of
> big error here. With the ROtoRW convert the cell is back up very fast.
> So why use backup volumes?

I'm using 3 backup strategies here:
   * Backup volumes are create -> any user can restore yesterday's data
     (covering 95% of all cases of "lost" data)
   * ROs are stored on a server differing from the RW one (for disaster recovery) and are
     updated regularly
     (Actually I never needed those ROs but it's a good feeling to have them :-) )
   * the afs backup system uses the backup volumes to write full and incremental backups
     to hard disk giving a used with a deleted file ~ 3 months.

> Are backup volumes built incremental?

They are not built. Backup volumes are updated to reflect the current state of
the RW instance. They are *stored* differencially to the RW instance.