2012-07-06 08:31 AM
One of our partners has received the following question from a customer.
"A customer is asking me about what happens in the event of a failover in a MetroCluster, specifically for CIFS and NFS. I am pretty sure that CIFS sessions are dropped, but unsure exactly what happens with NFS. They are using NFS for VMware datastores. I read an white paper (about Cambridge University Press) that suggested that the same CIFS shares could be presented from both filer heads at the same time, then failover would be seamless. I don’t understand how that could work!
Is there a document that will help me to explain the process for each protocol (or both!)? I have searched the support site, fieldportal and Technet, but can’t find the right words."
Can someone help?
2012-07-06 09:13 AM
What kind of a fail-over are you talking about: just a storage controller, or entire site?
The former is no different from a 'normal' controller fail-over - CIFS sessions need to be re-opened & NFS should work seamlessly (providing timeouts are set correctly & fail-over doesn't take too long).
The latter requires clustering at the host level - e.g. VMware HA will do the trick for VMs.
2012-07-10 02:01 AM
I've installed a MetroCluster before and worked on testing of the solution, as Radek says the failover is no different to that of a normal controller failover.
I found this very useful.