I have a problem that has be stumped.
We completed a datacenter relocation. We purchased new ESXi hosts and NetApp storage for the new location and used NetApp Snapmirror to replicated the data to the new datacenter. I snapmirrored the volumes and not the entire SVM so I could have the zoning on the SAN done prior to the cutover. During the cutover from Site A to Site B, broke the snapmirror relationships, made the volumes RW and presented them to the ESXi hosts. I had to mount the LUNs with the ESXi CLI on the first host and found the LUN with Virtual Center and fired it up. I added the new hosts to VC and started trying to set up storage. When I would select a LUN it would tell me that there is an existing datastore and offered 3 paths of keep it with a New Signature, Keep Existing Signature, or format..
I chose to Keep Existing Signature. I think I chose poorly. I expected the storage to automagically show up on the other 2 hosts. It did not. I had to go into them as well and mount the LUNs with the ESXi CLI, then the storage showed up. We got all the servers up and running and everything seemed fine ... until I rebooted one of the hosts.
All of the storage on that host, except for 2 newly created LUNs evaporated.
I tried reloading that host to see if it would see all of the storage, and again, I only see the 2 new LUNs
I obviously can't go in and remount all of the storage every time I reboot a host. The only solution I have found is to storage vmotion all 58 TB on to new LUNs. I would like to find a fix that is quicker and won't bugger my naming/numbering schema.