Hello, we have been troubleshooting an issue where our Dell PowerEdge servers running ESXI 5.0, 5.1, and 5.5.
Our current configuration is Dell PowerEdge R720's with 256GB of memory, internal 4 Broadcom NICS 2 dual port
Intel nics. The servers are connected to a Dell Compellent SAN using Cisco Nexus 5500 network switch. What happens is that intermittently one of the ESXI servers will hang and none of the virtual machines will be accessible. None of the virtual machines will failover to other hosts and at times not even the IDRAC is accessible. We cannot do anything with the host in Vcenter and have to power cycle the host. Once we power cycle the host all the virtual machines will come up on the other hosts just fine and after the power cycle we can move virtual machines back to the original host with no issues.
This issue occurs on our R720's here in the US and our R610's in the UK. We have worked with VMware, Copilot, and Dell to try and resolve this issue but have had no luck resolving this issue. The only thing that is visible that the VMware logs show that there was a loss of connectivity to the Compellent storage, but when we experience this issue we have seen losses of other network connectivity and it appears that the Vmotion network fails since none of the virtual machines failover to the other hosts.
Here is a list of some of the things we have done here in the US to try and resolve this issue.
Originally we were running R620's running ESXI 5.1 using Broadcom 10GB nics to connect to the Compellent Storage. We were using the Dell OEM version of ESXI. VMware had us update drivers to the Broadcom nics which did not resolve the issue. We also applied updated firmware to the R620's, this did not resolve the issue. We applied all the VMware updates which did not resolve the issue. After this we changed out the R610's with R720's switched to Intel NICS for the ISCSI and production network connection and installed ESXi 5.5 with the newest firmware and drivers available and we are still seeing the same issue.
Has someone had this same issue and were able to resolve it?
Thank you,
Richard Tracy