We have recently upgraded from vSphere 5.1 to .5.5 U2 at our prod site. Last week we had to perform a live DR simulation at a cold site where the following issues surfaced.
All the blades were built to 5.5 U2. We tried restoring our vCenter server first which kept failing with a login error. This was using the EMC Networker backup utility. We could log in into the target host using the vSphere client from the backup server fine. The restore job would consistently fail with the same login error even when using different host in the cluster. Seeing as the only difference from the previous DR test was the host OS being version 5.1. I therefore downgraded a blade to 5.1 and tried the restore job again. This resolved the login issue temporaraly so that we could get some restore jobs running on the 5.1 host and fire the VMs up on the 5.5 blades thereafter.
Issue number 2 popped up only after we restored vCenter and got all the blades running in a cluster. I was building a few more blades to 5.1 so that we could Q more jobs over multiple hosts. We found that one of the blades now running 5.1 could not see any of the storage volumes. The same blade previously could see the volumes and had vms running on it. When looking the the HBA the on the host via vCenter the LUNs were cleary visible when selecting the storage adapter. They were also displayed during the install process confirming the physical connection between host and LUN. The same LUNs contains volumes currently in use by multiple other hosts in vCenter however some of the hosts are missing 1 or all of the volumes. When trying to add the existing volumes to a host via vCenter nothing is available to select. When trying to add the volume while logged on directly to the host the volume is available to select but then prompts to be renamed \formatted which already has VMs on it.
Have been trolling through kb's and trying out force mount commands for similar cases with love to be had. Not finding any articles etc on the issues at hand specifically. The hardware in use at the DR site is all VMware certified. We tried a firmware upgrade on the HBAs on one of the blades but did not make a difference either.
The high level hardware config is as follows. IBM blades connected to IBM SAN storage via FC. (Will consult the vendor for specific model and firmware versions at a Later stage). In the prod environment where we do not see the same issues we are running Dell blades connected to Dell storage via FCOE.
It appears like storage volume anomaly only appears as soon as vCenter 5.5U2 is introduced to the environment. Not sure where to start digging with login issue as credentials are either right or wrong. The vSphere client logs in fine but then the version of Networker backup in use is certifiewd for 5.5U2 by the vendor. Anyone else out there come across similar issues?
Regards
Russell