I have two ESX servers that unfortunately are frozen for the time being at ESX 3.02 U1 - they are brand new 2950's and are configured with an identical configuration straight from the CD installation. They each have three virtual servers and dedicated virtual switches and physical network ports split between the VM's and the service console. I just cutover to these servers late last week from some very ancient Dell servers running the same version of ESX.
On one of them, prd1, everything is just peachy. On the other, prd2, the VI client will randomly drop off with the message:"Connection to the server has been lost. The application will now exit." Immediate attempts to reconnect thorugh the client or putty will fail sometimes. The server remains pingable and the VM's appear to be unaffected. On the console, the mgmt-vmware service is running fine and restarting will not solve the issue. If I am patient, however, I'll be able to reconnect sometime later.
This happens every few minutes. Given that I am running Server 2000 VM's without TS enabled yet, this is a big issue and promises to be quite frustrating when attempting the in place upgrade to 2003. Further, reboots (of the ESX hosts) are extremely diffcult to schedule due to the nature of the application.
Any thoughts? I am using the version of VI client that comes with 3.02U1, but have considered using a newer version. What troubles me is that putty also appears to die...
Help would be appreciated!
update
Running a continuous ping, it appears the putty and VI client drop when the ping time jumps ever so briefly to 24MS... I am patched into the same physical switch the ESX servers are on, so this isn't a WAN latency thing, it is usually under 1 ms.
