ESXi 4.1 U1 server partially frozen

[SilverNZ]

Hi Everyone

I have a bit of a weird problem. I have an ESXi host which has partially fallen over. The vCenter cannot communicate with it, however I can still SSH into it.  I cannot browse to it's datastores via SSH though, the session then freezes.

SSH:

Cd /vmfs/volumes

ls -l

--freezes here--

I can still run the esxtop command, and VMs do register as consuming CPU and memory, however they are not contactable via ICMP or RDP

I attempted to restart the management agents, however it got stuck on "Running usbarbitrator restart". I also noticed the following entry:

Starting likewise site affinity servicetouch: /var/lock/subsys/netlogond: no such file or directory

(I looked in /var/lock and there is only an ISCSI folder, subsys does not exist)

Any ideas?

[Ramverma]

Please check below KB if it helps

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=101713...

[john23]

can you check memory consumption in esxtop command

-A

[wila]

When I see that my first impression is that your storage that holds the vSphere image has failed.

Try rebooting the host.

--

Wil

[SilverNZ]

A reboot did do the trick. I'm still unsure what the problem is. I could browse all files on the ESXi installation, however nothing on the datastores(connected via a raid card). I would have thought that if the raid card had caused a problem, ESXi would simply show them as disconnected?

[SilverNZ]

I checked it, and VMs were marked as consuming memory, however it was next to nothing. Not sure if that is helpful?

[wila]

Not necessarily, if you have a running ESXi host and you rip out storage where your ESXi host boots from then the host will continue to run as good as it can. However as storage has disappeared anything that involves storage or having to access local storage might die/lockup.

Without knowing all of the specifics of your specific setup, the thing I suggest is to look at the log files, check to see if anything is written in the log before and after your host started to experience problems. Are there any I/O errors?

--

Wil

[gabinun]

Can you compare that ESX host to another in the same Cluster. Are they seeing the same LUNS?

[SilverNZ]

Not quite possible as the only other ESXi 4 host in the vCenter also has local disks. (I'm not using a cluster)

[SilverNZ]

Perfect, thank you. I'll run through the syslogs and have a look. I've also installed the ESXi dump collector. Hopefully I'll be able to get something if it fully keels over next time as well