Recently we faced the I/O choking in the SAN switch port.So planned to move the storage path to different port in HPE enclosure.
In HPE enclosure,vcem play a vital role to hold the server profile and network/SAN uplink information will be saved/configured.
Combining the two set of x ports provides the another SAN path.
Now when changing the SAN uplink port the datastore in the ESXi went to greyout and path in the host says dead.
VM was pingable and anything inside was stunned.
For testing this suitation,I created two case
- Without VM
- With VM.
For Without VM scenario,host with one datastore no VM residing.Changed the storage path, datastore went to greyout and reset the HBA to get the datastore back.
Esxcli storage San FC reset -A vmhbax
For with VM scenario,same setup but with VM provided an IP address and made it reachable via RDP.
During the failover to New path,VM become stunned and resultant error message all path down in VMkernel logs.
As per VMware,the best practice or thumb rule.During storage issue don’t do rescan or refresh the HBA.just reboot the host to recover.
Finally migrated the VM and then rebooted the host to change storage path.