VMware Cloud Community
ivyowl
Contributor
Contributor

How to remove a DEAD LUN from ESX 3.5 instead of rebooting the host

One particular SAN LUN has been assigned to each of five ESX 3.5 U5 hosts in a HA/DRS cluster. This LUN was then mapped as a raw LUN to one of the VMs on one of the hosts. It worked fine at first, but at one point when trying to start the VM it would not power up. I then tried to unmap the LUN and remap it, at which point I got a generic error. When I asked the SAN team to unassign the LUN from all of the hosts they did so, but the hosts still see the LUN. When the SAN team remapped the lun I still experienced the same problems mapping it to the VM. I checked all of the hosts to verify that none of them are holding a SCSI reservation for that LUN. Below is the output from esxcfg-mpath -l, notice the DEAD LUN entry.

My question is how do I remove this entry for DEAD LUN without rebooting the host(s)?

Disk vmhba1:0:4 /dev/sde (1048576MB) has 4 paths and policy of Most Recently Used

FC 24:0.0 2100001b320bb718<->50050768011038e0 vmhba1:0:4 On active preferred

FC 24:0.0 2100001b320bb718<->50050768011038d9 vmhba1:1:4 On

FC 30:0.0 2100001b320b2a19<->50050768012038e0 vmhba2:0:4 On

FC 30:0.0 2100001b320b2a19<->50050768012038d9 vmhba2:1:4 On

Disk vmhba1:2:9 /dev/sdaf (500MB) has 4 paths and policy of Most Recently Used

FC 24:0.0 2100001b320bb718<->5005076801102a1c vmhba1:2:9 Dead preferred

FC 24:0.0 2100001b320bb718<->50050768011038dc vmhba1:3:9 Dead

FC 30:0.0 2100001b320b2a19<->50050768012038dc vmhba2:2:9 Dead

FC 30:0.0 2100001b320b2a19<->5005076801202a1c vmhba2:3:9 Dead

Disk vmhba1:0:18 /dev/sdf (1048576MB) has 4 paths and policy of Most Recently Used

FC 24:0.0 2100001b320bb718<->50050768011038e0 vmhba1:0:18 On active preferred

FC 24:0.0 2100001b320bb718<->50050768011038d9 vmhba1:1:18 On

FC 30:0.0 2100001b320b2a19<->50050768012038e0 vmhba2:0:18 On

FC 30:0.0 2100001b320b2a19<->50050768012038d9 vmhba2:1:18 On

Disk vmhba1:2:5 /dev/sdm (1048576MB) has 4 paths and policy of Most Recently Used

FC 24:0.0 2100001b320bb718<->5005076801102a1c vmhba1:2:5 On active preferred

FC 24:0.0 2100001b320bb718<->50050768011038dc vmhba1:3:5 On

FC 30:0.0 2100001b320b2a19<->50050768012038dc vmhba2:2:5 On

FC 30:0.0 2100001b320b2a19<->5005076801202a1c vmhba2:3:5 On

Any help would be greatly appreciated.

0 Kudos
3 Replies
weinstein5
Immortal
Immortal

Welcome to the Forums - have you rescanned the HBAs in the Sotrage Adapter tab of you ESX server configuration -

If you find this or any other answer useful please consider awarding points by marking the answer correct or helpful

If you find this or any other answer useful please consider awarding points by marking the answer correct or helpful
0 Kudos
ivyowl
Contributor
Contributor

Yes, we have rescanned the HBAs numerous times. Eventhough the LUN has been un-presented from SAN side, after re-scanning the LUN still shows up as available to the cluster.

0 Kudos
EXPRESS
Enthusiast
Enthusiast

What was the solution for this, I am now seeing Dead Luns as well... which should actually be active.






Thank you,

Express

Thank you, Express
0 Kudos