[Gluster-devel] Gluster Recovery

[Steffen Grunewald]

On Thu, Apr 26, 2007 at 12:04:48AM -0700, Anand Avati wrote:
> > 
> > Client1 puts "hello.c" onto both Server1 and Server2 via AFR.  Client2 then changes hello.c in some way.
> > Server1 goes down; data lost, no chance of recovery and is replaced by Server3, a brand new server with fresh disks.
> 
> 'changes' are done in sync on both server1 and server2 always
> (writes()s are sent to all child nodes). when server3 comes in place
> of server1, the self-heal should detect that hello.c is missing on
> server3 and sync it from server2.

Let's continue with this scenario: what would happen if one of the clients
changed hello.c again, and the failed server came up again after some
extended repair period?

What exactly would be the criteria for file resync? (Imagine one of the
copies got corrupted during fsck, while still carrying identical metadata
information)

Cheers,
 Steffen


-- 
Steffen Grunewald * MPI Grav.Phys.(AEI) * Am Mühlenberg 1, D-14476 Potsdam
Cluster Admin * http://pandora.aei.mpg.de/merlin/ * http://www.aei.mpg.de/
* e-mail: steffen.grunewald(*)aei.mpg.de * +49-331-567-{fon:7233,fax:7298}
No Word/PPT mails - http://www.gnu.org/philosophy/no-word-attachments.html