[Gluster-devel] AFR problem with 2.0rc4

Vikas Gorur vikas at zresearch.com
Thu Mar 19 10:26:25 UTC 2009

2009/3/19 Gordan Bobic <gordan at bobich.net>:
> The subvolumes have to be listed in the same order on all nodes, because
> that is the order of precedence and failover for lock servers. Could it be
> that this was the reason why you were seing problems?

That is one reason. Another reason is that the extended attributes
'replicate' uses to keep track of the most recent copy of the file
make no sense if two different clients have differently ordered

If you did not have the order of subvolumes same on all clients,
please make sure they are try your operations on a *fresh* backend.
I've done the tests you described on our local cluster and find that
self-heal works fine.

Engineer - Z Research

More information about the Gluster-devel mailing list