I have 8 ESX blades servers based on IBM HS21 blades.
I have one ESX host that is causing problems. Periodically it drops off the network however the VM's on this host retain connectivity.
Eg.
ESX host IP: 192.168.3.245
ESX Vmotion IP: 192.168.3.246
VM1: 192.168.3.23
VM2: 192.168.3.29
VMx: a number of other uneffected VM's
=================
When the issue arrises I can not ping
192.168.3.245
192.168.3.246
192.168.3.29
however can still ping 192.168.3.23 and all other VM's on this host
On the host I can ping all IP'S however nothing outside this host.
The process is to Remote Dekstop to the VM's and shut them down. Clean shutdown.
It always seems to be one VM's IP and the two host IP's that are un routable but the other VM's are fine and connectable
Are there any tools I can use to diagnose the issue. I have re-built this VM from scratch. Formattted HDD. Always the same VM. Logically it seems that the issue would sit with the Blade Switch versus the ESX Virtual switch. Also previously it had the same issue running ESX 3.
From the VMs on this host that you cannot ping from outside of this host can you ping them from VMs that are on the same host? If you can then you can look at the pSwitche. the network traffic for the vms on the same host should stay in RAM
I am going to make some guesses here. You have multiple NICS but as this is a blade all those ports may well be on the same device. You could have a NIC issue (Firmware or hardware) but that does not seem likely
If this blade is the only one with the issue I would go with a reinstall before really going to the mat troubleshooting this issue. Then again I am lazy.