Hi,
recently I removed a LUN from my ESX hosts, and then something strange happend. The LUN, with LUN id 8, disapeard from vmhba0, but not from vmhba1. This resultet in a very high "Disk latency" and svMotion didn't work.
This message appeard in the console (and in the messages log):
"Oct 10 01:20:22 host-p-d-esx01 kernel: http://11133215.835955 sd 4:0:3:0: still retrying 799941 after 360s"
Before I discovered this problem, the LUN was unpresented from the SAN as well, so no way back there. The LUN path then appeard with status "dead". I didn't manage to remove this LUN id from vmhba1 in any other way than do a "cold boot". Normal "shutdown -r now" resultet in a hang situation when trying to umount the LUNs.
BTW: I've added and removed LUNs many times to these ESX hosts without problems.
Does anyone here experienced this, and is there anything to do other than "cold boot" ? Any recommendations ?
Enviroment/Servers:
VMware ESX 4.0.0, build 236512 (4 of them)
HITATCHI SAN
Regards,
BKiil
Hello.
You could mask the LUN using the claimrule feature discussed in the Fibre Channel SAN Configuration Guide, but a reboot might be faster and/or safer.
Good Luck!
Anytime you make changes to the SAN as far as adding or removing LUNs and things like that. Do a rescan for datastores on your hosts. If you dont it will cause issues. My VMs would blank out every view minutes when I forgot to rescan. Hope this helps.
I've run into this issue with HDS, and a cold boot was the only way to get things back in order. See this kb for a description of what may have happened.
vExpert
VMware Communities Moderator
vmwise.com / @vmwise
-KjB