[Linux-cluster] NFS HA

[Michael Schwartzkopff]

Am Freitag, 13. Februar 2015, 17:39:31 schrieb Equipe R&S Netplus:
> Hello,
> 
> I would like to setup a cluster of NFS.
> With RHCS, I use the ressource agent "nfsserver".
> 
> But I have a question :
> Is it possible manage a NFS server where NFS client "will not be aware of
> any loss of service" ? In other words, if the NFS service failover the NFS
> client don't see any change.
> 
> Actually when there is a failover, I can't access to the NFS server anymore.
> Indeed, I had the message "Stale NFS file handle".
> In client NFS log :
> <<
> NFS: server X.X.X.X error: fileid changed
> 
> 
> Is there any solution please ?
> Thank you.

Be sure that you have a virtual IP adress that migrates together with the NFS 
server in the cluster.

Timeout / NFS Server migration is a problem in NFSv3. The server should inform 
the clients that it rebooted and the clients should reclaim their locks.

This locking is much more elegant in NFSv4. So in clustered NFS Server you 
should use v4. See the resource agent description for locking timeouts.

TCP or UDP is just a transport protocol and not important for the problems you 
describe.

-- 
Dr. Michael Schwartzkopff
Guardinistr. 63
81375 München

Tel: (0162) 1650044
Fax: (089) 620 304 13
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 230 bytes
Desc: This is a digitally signed message part.
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20150214/ecff1747/attachment.sig>