One ESX host keeps disconnecting...

[CLowe]

I have one host in a 4 host cluster (HA snd DRS on identical hardware), all current patches, that this week suddenly started showing up as "not responding" in VC. It's gotten so bad that within 30-60 seconds of doing a disconnect/connect it goes back to a "not responding" state. All the vm's on it are fine. I have tried restarting the vc agent on the server and restarting the vc service but no go. I was going to try rebooting the host but it won't stay in vc long enough to be able to even vMotion anything.

Anyone seen this behavior before?

[kjb007]

Have you changed IP addresses? When you registered the host, was it via IP or hostname? Restart the mgmt agent on the host, service mgmt-vmware restart.

-KjB

[CLowe]

IP address has not changed. Host was registered 6 months ago via hostname. Have restarted agent multiple times. Only change lately was Update 2 applied to the cluster about 3 weeks ago and the date patch applied on the 14th.

One thing I just noticed though. On 2 of the 4 hosts (including the host that is giving me fits), if you go into the configuration tab there is something called HEALTH STATUS in the Hardware section. This does not exist in the other 2 hosts. They are supposed to be identical configurations.

[SimonHuizenga]

is your DNS ok? only use lowercase names in DNS or hostfile.

[JonRoderick]

One of my 3.5 U2 hosts did a similar thing (once) this morning. What ESX build version/VC build version are you using (I'm using 110268 - the latest I think).

Jon.

[CLowe]

The hosts file is all lowercase and is identical to the other 3 hosts.

[CLowe]

According to VC, the hosts are 3.5.0 110181 and VC and client are 2.5.0 104215

[espi3030]

Sounds like a licensing issue to me. I would check the license status on this host first, then check your license server to make sure license is being properly read.

Hope this helps.

[kjb007]

This should also be reflected in the vmkernel/hostd.log

-KjB

[mike_laspina]

Hello,

Have a look at the vpxa.log or post it

cat /var/log/vmware/vpx/vpxa.log

[JonRoderick]

My vpxa log didn't seem to cover the time period of the problem - I've gone route one and raised a support call.

[CLowe]

Final update:

After disconnecting/reconnecting about 20 times I was able to finally get all the vm's off the host. I then rebooted it and have not had a problem since late yesterday. At this point I don't know what the issue could have been.

>>> espi3030 <communities-emailer@vmware.com> 8/27/2008 8:10 AM >>>

Chris Lowe,

A new message was posted in the thread "One ESX host keeps disconnecting...":

http://communities.vmware.com/message/1035376

Author : espi3030

Profile : http://communities.vmware.com/people/espi3030

Message: