[OpenAFS] Backup-strategy

Klaas Hagemann kerberos@northsailor.de
Mon, 16 Dec 2002 10:15:13 +0100


Hi,

i tried this one out.
It works very well, but it takes too long time.
We have more than 20000 volumes and it takes round about 10 seconds for
every volume. So we cannot back up all at one day.
Even running multiple jobs at once using the "&" does not help.
Any suggestions to make this one faster or how to do multiple jobs at once
in a better way?

Thanks,

Klaas
----- Original Message -----
From: <Craig_Everhart@transarc.com>
To: <openafs-info@openafs.org>; "Klaas Hagemann" <kerberos@northsailor.de>
Sent: Friday, December 13, 2002 11:33 PM
Subject: Re: [OpenAFS] Backup-strategy


> So what you want is to make your own volumes to shadow the user's volume.
>
> Do this by creating a shadow for every user volume:
> vos dump user.foo | vos restore user.foo.shadow
> where user.foo is at the satellite site and user.foo.shadow is at the
> central one.  Do incremental dumps of user.foo data every day, restoring
> them to user.foo.shadow
> vos dump -time XXXXX user.foo | vos restore user.foo.shadow
>
> Do all your backups of user.foo.shadow data.  It doesn't even have to be
> mounted if you're using vos dump to back it up.  If you're using
> something else, you get to create a shadow mount tree.
>
> If you want to be even nicer to your users, do "vos backupsys" to make
> local (satellite) clones of all the user.foo volumes, then do your
> dump/restore from the backup volumes to the shadow ones:
> vos backup user.foo
> vos dump -time XXXXX user.foo.backup | vos restore user.foo.shadow
>
> This avoids the .readonly headaches and puts the storage where you want
it.
>
> Craig