I have a server which is running ESXi 5.1 on a Dell R510. It has 3 Windows 2008R2 VMs on this machine:
• Domain Controller
• File/App server
ESXi runs on 3 600GB SAS disks in a RAID5 config. There is also a larger store of 3 SATA disks in a 5.5TB RAID5. This Host has 32GB of RAM and 2 Quad Xeon E5620 processors.
Every couple of weeks or so, EVERY guest on the host will max out its CPU, effectively locking the guest. Ping packets sent to any of the guests only reply intermittently. There is nothing in the windows logs that indicates any applications hanging etc. I was able to log into the guest and see the CPU maxing, but that took ages to get that far because of the performance. I couldnt click around to check much else while it was happening because the client needed to get up and running.
During the guest lockups, ESX is still responding. I can log in with the vSphere viclient.
The only way to resolve this issue so far is to reboot the host.
Do you think you can help us?