Problems with Oracle RMAN backups via NFS hanging the Oracle Server

I guess my question would be...  what's the problem that's causing their NFS mounts to disappear?  Really, there isn't a whole ton of ways to prevent a performance issue if a resource disappears during a database activity.  This isn't really an Oracle specific thing - most apps misbehave when that sort of thing happens.

So with that being said, has there been any attempt to address their NFS issue? 

Rob

I take it that's not an option.  Let me do some digging and see what options I can come up with.

I agree this would happen with other apps than Oracle.

I agree that making their network not have problems would resolve the issue.

Does anyone else have ideas? 

Please see paper “EMC Backup and Recovery for Oracle 11g OLTP” this covers this area.

http://www.emc.com/collateral/hardware/technical-documentation/h7207-backup-recovery-oracle-clariion-dd-networker-psg.pdf

The following recommendation is made

mount -t nfs -o hard,intr,nfsvers=3,proto=tcp,bg :/backup /mountpoint

Also see Data Domain’s “Linux System Tuning Guide” for more in-depth information on tuning Linux backup servers.

https://my.datadomain.com/custom-view/integration/linux_BU_servers_775-0191-0001.pdf?query=query%3Dlinux&fsearch=1&pagenumber=1&size=10&index=0&filterids=

Finally if the issue persists the customer should open an SR with support.

is this RAC environment ?

The paper listed above is a RAC environment.

http://www.emc.com/collateral/hardware/technical-documentation/h7207-backup-recovery-oracle-clariion-dd-networker-psg.pdf

if it's for RAC, where is the recommended NFS mount parameter of "noac" or "actime=0".

The mount/proxy host in the solution is single instance. The backup is not direct against the production RAC database.