Host keeps going disconnected

[mlydy]

Ok this one is starting to drive me a little crazy. I have one blade out of 14 in a resource pool that VC keeps reporting as "Disconnected". I can get the host to reconnect by doing a service mgmt-restart and service vmware-vpxa and the host will rejoin the cluster. Then a few minutes later it will be disconnected. Is there something else I need to restart or does anybody have any suggestions for what might be happening? This server has been running fine for about a year in this cluster.

[bludden]

How is the available space of the partitions of the ESX server in question? I think I saw a similar problem a long time ago that was related to an issue with the cluster agent filling a directory in /home which was not made as a separate partition at install.....

Brian

[mlydy]

Hey Brian,

Thanks for the reply. Currently there is plenty of free space on the server, that was a good thought though.

[christianZ]

Have you checked the licencing? Just a thought.

[mlydy]

Yeah I'm good on liceneses as well. 35 hosts with about 6 licenses free.

[Chamon · Accepted solution]

I've seen this as well. Our fix was to remove the host from the cluster and re add it back in. It might be an agent problem. This will uninstall the vpxuser and the agent and reinstall them when they are readded. The vms on the host will continue to funtion.

[bludden]

Yep that sounds like a good plan - it's fixed a couple of issues for me (just not this one) in the past.

Brian

[mlydy]

Well I finally managed to remove it and readd it. I'll watch it and let you know what the result is... I'll award points soon hopefully! :smileylaugh:

[admin]

It this blade on a different subnet, perhaps NAT'ed behind a firewall? I had a similiar situation before where Virtual Center would lose connectivity to 2 ESX blades that were in a DMZ after staying connected for only a minute. The other ESX blades on the same subnet had no connectivity issues.

Chris

[Chamon]

Hopefully that was all that it was If not let us know and we can see

where to go from there.

[bolsen]

I've had this problem and it was caused by resource pool configuration failing to update the VM.

Use the vic and connect to the host having the problems. Watch the activity log and look for vms not updating. (In my case many of them were 2.X VMs.)