Rebooting ESXi build 4.0.0, 193498 Lost Network on Guests.

[doepain]

I have two ESX hosts one running esxi build 4.0.0, 193498, and the other esxi build 4.0.0, 208167. These servers are identical HP DL 385 G5p with 64GB of Physiscal RAM, and almost a terabyte of local-drive space.

esxi build 4.0.0, 208167 = Host A

esxi build 4.0.0, 193498 = Host B

Last week I rebooted both hosts, and Host A came up, and then subsequently I began to power the guest VM's one -by-one. All of the guest came up with no issues. However, the same was done with Host B, but its guest VM's seemed to have lost their place on the network. Their IP settings were fine, I could not ping ANYTHING either via IP or DNS. I restarted the Network Management Agent/Service, and to no avail.

Ultimately the fix for this was to remove the Windows Server 2003 R2 SP2 hosts from the domain, then reboot the ESXi server!, and then rejoin the domain and reboot the Guest VM host. Following this the VM Guest would be able to see the network. I had to do this 3 seperate times for each of he VM guests on the ESXi host. I do not feel what I did to fix the VM guest was legit, not sure if it was rebooting the VM, the Host, or the whole Domain remove\rejoin that fixed things.

Once I would get one Guest to work I had to do it ALL-OVER again for the other two guests, each time I reboot the ESXi server I feared I would break the other guest I just repaired. Luckily that was NOT the case and the guest once repairedremained working.

Has anyone else ever experienced this issue?

Senior Hosting Engineer

ModusLink Open Channel Solutions, Inc.

[doepain]

All of the servers are set to static, and lookin at the network configs on the Win64 guests they remained intact with the correct IP's etc.

The worse thing was I was not getting any error messages that would give me any direction on a root cause of this issue. Whether or not this was some kind of an anomoly does not set well with "managemen like" folks. They don't liketaking risks on servers etc.

THis is making them think twice on how reliable is VMware when these type s of things happen, and though we fixed it we don't know what the cause was, and its logging is ver vague.

I am just looking for answers to explain what happened to them, and what we are doing to prevent it from happening again.

Senior Hosting Engineer

[doepain]

I guess that I am alone on this...

Senior Hosting Engineer

[Stuman4u]

Your not alone - however, scenario a little different. Happened this Sunday, we lost network on Dell PowerEdge R710. Had to reboot host (over 40 VM's effected) to gain network function back. We moved VM's to another host, and they would not connect to network. Only until we moved back to original host would they get network functionality. Called VMWare (after host reboot) and all logs were gone so no troubleshooting could be done.

[doepain]

My faith in VMware has tken a serious "hit" since this occurred.

Senior Hosting Engineer

ModusLink Open Channel Solutions, Inc.