Windows Server 2003 in VMWare ESXi losing connection to AD Domain

[machwhat]

Hello,

I have five images of windows server 2003 enterprise on vmware esxi in a
MS Windows Active Directory Domain. Four of these machines keep losing their connection to the domain.
In the application event viewer log, there are a lot of 'USERENV' Event 1053 errors stating:
'Windows cannot determine the user or computer name. (Not enough storage is available to complete this operation. ).
Group Policy processing aborted.'

This happens every few days. When this occurs I cannot RDC into the servers because 'the RPC server is unavailable'.
It is my understanding that since the RPC server uses a network account, it will not start correctly if it cannot
authenticate agianst the domain controllers. Therefore this is to be expected if the client cannot connect to AD.

When this happens I can RDC into the client as the LOCAL ADMINISTRATOR. I can ping both of our domain
controllers. The firewall is not enabled on the clients nor on the domain controllers. 'gpupdate /force' results in success.

There is plenty of ram and disk space allocated to these client machines. Perfmon indicates that the machines
are not using up all of the allocated resources.

Has anyone else experienced client machines unable to find the user or computer name- effectively losing the domain?


Thanks in advance for your help,

Mike

 

[Zelandakh]

Check the resources on the host as well.

 

[machwhat]

There are three logical drives on the Host-
System = 63GB / 14.84GB free
Datastore2 = 278.75GB /80.97GB free
Datastore3 = 544.75GB /77.15GB free

There is roughly 2GB of free RAM not allocated to clients, but that shouldn't matter with how VMWare manages memory, afaik.

Could 15GB not be enough free space on the host's system drive?

Thank you,

Mike

 

[machwhat]

I checked DNS, everything looks ok. On one of the servers I even set the hostfile to include the domain controllers, to no effect.

To sum up what I have tried so far to resolve this issue:

1. Removed the vm's from the domain, deleted the computer accounts, removed vmware tools, then I rejoined the domain and reinstalled vmware tools.

2.Changed NetBios to use TCP - maybe something was blocking udp? Anyway no effect.

3. Disabled the backup client. I thought perhaps the network was working too hard and dropping the domain. No effect.

4. Altered memory usage by certain services- Event ID 1034 DOES say that it's a storage issue... Not that I believe it but..

5. Set the advanced config sched.mem.pshare.enable option to FALSE- no effect either. I thought maybe resouces were not being re-allocated to the vm's quickly enough.

I did make some changes to the registry based on this article:

http://support.microsoft.com/kb/937535

and will monitor those machines, but I don't think that's the issue here either.

 

[machwhat]

Still having issues.
VMWare has been upgraded to ESXi v4.
The swap has been placed on a dedicated partition as well.
Anyone else seen this sort of behavior from W2K3 in ESXi?

 

[ArizonaGeek]

Is your DC a virtual server? Did you P2V the DC, if it is virtual? Do you have VMTools installed on the server? Did you check the time/date and are those synced to the DC or the ESX server?

Cheers
Rob

The answer is always "PEBKAC!

 

[machwhat]

Our DC's are not virtual.
Each virtual machine get's it's time from NTP. NTP is provided by the DC's as well.
Not sure about the host though, that's worth checking out.

 

[ArizonaGeek]

Well, I am outta ideas.

The only thing I would check would be to see if your DCs are replicating to each other. The last time I had something somewhat similar to your situation, it was caused by one of our DCs freaking out and had stopped replicating to the other one.

Cheers
Rob

The answer is always "PEBKAC!