GOS rhel5u5 IO failed which is created on NFS datastore

[dandandongdong]

Hi,

Does anyone why  GOS rhel5u5 IO failed which is created on NFS datastore? Many thanks.

Testbed summary:

1   Installed VMware 4.1 U2 build 502767.

2   Added vmkernel for nfs based on Emulex 10G OCE10102 card.

3   Created vmware NFS datastore.

4   Created 4 GOS in the NFS datastore:  1   windows2003   2 windows2008    3 rhel5u5    4 sles10sp2

5   Added 2 hard disk for each GOS which is also created in the same nfs datastore.

Run IO in the GOS.

When doing storage side failover operation, the IO on rhel5u5 GOS will fail with error mis-compare.

IO on other GOS are running well.

Does anyone can give some advice? thanks.

[mcowger]

how long does your storage failover operation take?

[dandandongdong]

Hi,

the time is less than 35 secs.

[dandandongdong]

Whether there are some special settings about configuring NFS on VMware?

Thanks.

[mcowger]

35 seconds is long enough for RedHat to timeout on an IO. 

Try changing your SCSI retry time to 60s.

[dandandongdong]

Whether I can change the vmware nfs parameter like NFS.SyncRetries?

How to modify the scsi timeout? In GOS itself?

Many thanks for your help.