Virtual Machine network problem

[ronmanu07]

Hi All,

I have a strange one here...

I have 8 ESX 3.0.2 hosts, they all function perfectly. The only issue I have is on 1 of the hosts if I have a vm on a certain vlan, the vm cannot be contacted yet the host can ping both the gateway and other servers on that subnet????

I have checked the configs of the 'bad' host against all the others and they are identical! My last resort will be to reboot the host, but don't really see that as a solution here. The logs look nice and clean also.

Hope this makes sense because it is a weird one...

[mcowger · Accepted solution]

On the pSwitch to which that VM is connected, is that port configured EXACTLY like the rest (same vlan access settings and stuff)?

--Matt

[ronmanu07]

When you say pSwitch (u mean Physical Switch) then yes I believe so, I have got the comms guys to check on there end also.

[Erik_Zandboer]

I am not 100% sure what you mean - if you say:

"the VM cannot be contacted" - is this by another machine, is that machine in the same subnet or not?

"yet the host can ping..." - What host? Do you mean the VM, or the SC of ESX?

As far as I can tell from your story so far, you might have a gateway incorrectly configured somewhere. This would enable you to ping the other side as long as you are in the same subnet, but not as soon as your comms try to leave the subnet.

Visit my blog at http://erikzandboer.wordpress.com

[JoeLyons]

Ronmanu07,

Is it just VMs on 1 Vlan for a patricular host?

Can VMs on that host in other VLANs connect?

If so you may just have a vlan tag misconfig.

It would be worth createing a new Vswitch and setttingup the VLAN tags again, then conenct your vm and see what happens. Sometimes you can stare at a config and it looks fine but there might be a whitespace or corruption in the settings thats throws this off.

If that fails, it might be an idea to ask your network guys(nicely) to reconfigure the vlan for that host just to be extra sure.

Joe

[ronmanu07]

Hi All,

This has been fixed, it was a strange problem because a few weeks back we had a short network outage which all the ESX hosts experienced, but only 1 of them had this issue where:

• Any VM hosted on it could not be contacted by ping. (either from the host hosting it, other hosts, other servers virtual or physical, same or different subnet)

Hence I was tearing my hair out because nothing had changed in my environment and no one had tampered with any configurations on the host as I went over everything very finely. So the last resort was a......that's right a damn 'REBOOT' and that has fixed it.

Sorry lads, this was not a straight up issue, but thank you all for your suggestions.