vCenter (5.1.0b) still triggers "Cannot connect to storage" alarms on all of the hosts in a cluster even though we unmounted the LUN device properly.
Is there a way to prevent these false alarms and only have them trigger when mounted LUN devices lose connection?
thanks
could you please check in the cluster setting _ HA , whether this data store is selected for data store Heartbeating option ? if it is selected , un-select it . more over rescan all the HBA's and data store as well
The devices are actually RDM disks, not datastores.
We delete the RDM from the VM and unmount it, but the vCenter alarm still trips when the LUN is un-presented from the hosts.
ok then run this command ( esxcli storage core device detached list) to check whether it is de-attched fully. if it is not deattched then manually de-attach ( esxcli storage core device set -d naa.xxx... --state=off).
thanks, we'll try this on the next LUN device we intend to remove.
we finally had another chance to test this and running the 'esxcli storage core device detached list' command does show the LUN as detached from the hosts.
however, when the device zoning is removed, the hosts still trigger the "Cannot connect to storage" warnings.
~ # esxcli storage core device detached list
Device UID State
------------------------------------ -----
naa.###################### off
I don't think detaching the LUN is intended to prevent this warning, so it's "normal". It unmounts the storage device but still regularly tries to contact it through SCSI sense codes, and will trigger the alarm once it disappears.
You might be able to prevent it for real with host-side path masking after it has been detached, but I personally don't think it's worth the effort:
VMware KB: Masking a LUN from ESX and ESXi using the MASK_PATH plug-in
it seems that way MKguy, LUN masking isn't convenient from powerCLI either.
thanks.