Network and Storage Protocols

Question about takeover/giveback and cifs/nfs


I needed to reboot a controller last night to clear a false error message regarding a new shelf installation.  In the process I did a takeover, rebooted the controller, and then a giveback.  Everything appeared to move between controllers as intended and I didn't have any immediate reports of any issues.  Unfortunately, this morning I received a report of a hung Xen server that has an NFS datastore.  I was under the impression that during a takeover/giveback you would lose CIFS sessions, but not NFS.  Can someone verify please?  I'd like to clarify if I made a mistake during this controller reboot or if something is not functioning correctly.  Thanks!



I would recommend checking the disk timeout on the Xen Server - if it is too small, it would notice the delay during the failover / giveback process and cause an issue.


Also bear in mind all the networking settings must be spot on, so the partner head can be accessed by the same host(s) - sounds simple, but I've seen many situations where it was a major stumbling block




I have been on the same situation, but i had seen that in the normal situation a 1-3 second interruption is normal!

Prepare the XEN to resist that and you will be fine!