Due to volume snapshots, a customer lost one of their iSCSI datastores. Because the Zombie VMs have active processes, even with that volume being back online it shows as inaccessible. A storage rescan is unable to unmount/mount that datastore and a reboot option was unavailable.
To recover the volume without a reboot:
esxcli storage filesystem list
Find the naa ID of the effected volume (you can also get this from “devices” under storage in the vClient.
vsish -e ls /storage/scsifw/devices/naa.60a98000572d5674424b43775a415565/worlds/ |sed ‘s:/::’ |while read i;do ps |grep $i;done