We had an issue last week that in our vSphere cluster (6 ESXi 5.0 hosts, 3 datastores), two hosts can not see one of the datastores. The issue was caused by that a guy in my team did not follow the proper LUN operation procedure.
I tried to add the storage back in the vSphere client, but somehow it asks me to format the LUN first (I guess the improper LUN operation may damage part of the partition table). Then I SSH into the host and tried the following command which I expected the same result, but to my surprise, the command worked without any problem!!
Type ‘esxcfg-volume –l’ to get the UUID of the available LUN, then run ‘esxcfg-volume –M <UUID>’ to manually mount the LUN.