Am 26.06.2008 13:20 Uhr schrieb "Geoff Dunk" unter
<[EMAIL PROTECTED]>:

> Thanks for the response, I fell my 1st explanation wasn't quit clear enough.
> 
> We have to external workstations that access our cluster system running HA.
> When one of the machines running HA failed, the external client could not see
> the mounted files system.
> 
> The workstations are using the HA IP as a mount point, and when the machine
> that failed was brought back up the NFS mount could be navigated once again.
> 
> Before failure, external workstations could access servers NFS, after failure,
> NFS was unavailable.
> 
> I expected (wanted) the client to use the 2nd HA machine, while the 1st was
> offline.
> 
> Does this help at all?
> 
> 
> Regards /  Mit freundlichen Grüßen
>  
> Geoff Dunk 
> High Performance Computing Engineer
> Mercedes-Benz HighPerformanceEngines Ltd
>  

What OS You use?
We use SLES 10 and our NFS HA Cluster works.

The important things we use are:

You need /var/lib/nfs managed by the HA cluster, the directory should not be
local on one of the NFS Servers.

Furthermore I have identically "fsid" entries in /etc/exports on every NFS
server for the corresponding NFS partitions the servers should export to
clients.

Regards,
Annette


_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to