Recently, we had to deal with an issue where several esxi hosts within a vCluster appeared to have non-identical datastores presented among them.
First, we had to create an inventory of which datastores should be presented to this group of hosts and distinguish which of the hosts were unable to connect to them.
So, in a total of 19 hosts, the ones marked with red, next to the naa address, were unable to list these datastores.
Next step was to contact the storage team, to identify if these datastores are truly presented to the hosts or not.
As expected, their reply was that everything was presented, accompanied by a glorious screenshot that proved their claim.
After double-checking every WWPN and Lun_IDs for each non-showing datastore to every host, after numerous Storage and Storage Adapters scans via vCenter, we tried to simply ADD New Datastore to one of the hosts, to see if anything changes and what showed up, left us speechless.
Datastore was showing as Not Consumed.
So, all esxi hosts were able to see the missing datastores but unable to mount them.
To resolve the incident, we had to ssh into each host separately and use a couple of cli commands:
First, list the available volumes that can be mounted by using the below command:
Afterwards, we mounted the specific volumes using -M (Persistent mount) using this command:
esxcfg-volume -M|--persistent-mount VMFS UUID|label