vCenter 4.1 unable to monitor host hardware status of ESX/ESXi hosts
We recently upgraded our vCenter environment from vCenter 4.0 to vCenter 4.1 so that we could start migrating our ESX/ESXi boxes to ESXi 4.1 since vCenter 4.1 is required to manage ESXi 4.1 hosts. We carried out an in-place upgrade and everything went on well without any issues and we started migrating ESX hosts to ESXi 4.1.
One day we noted that a disk had failed in one of our ESXi Servers but we did not get any alert from vCenter. We rely on vCenter alerts for host hardware status for our ESXi hosts since we cannot install vendor monitoring agents. Realized that the host’s hardware status tab got refreshed only when we manually refreshed or clicked on the host hardware status tab. By default , vCenter is to scan the hosts every 5 minutes for change in status of hardware components. Next we tried to connect to the ESXi Server directly to see if ESXi host was able to detect any change in status of hardware components and ESXi host was able to detect them. This meant that the issue was with Virtual Center.
We then logged a SR with VMware Technical support with whom we had to try out various combinations to rule out all possible issues .
- We tried to see if ESX host was able to detect status change in hardware components when connected via vSphere client directly. ESXi host was able to detect the status change of hardware components.
- We were asked to install vCenter 4.1 directly on a VM and see if the issue persisted ( This request was weird I thought !!) . Reason provided was that they wanted to isolate Physical Server issue and also at the same time vCenter 4.1 upgrade issue. We tried to install it freshly on a VM and the issue persisted.
- We also realized and informed them that vCenter was not able to update hardware status across all ESX/ESXi version of hosts from ESX/ESXi 4.0 to ESX/ESXi 4.1
VMware Technical Support team has finally acknowledged this as a BUG and is currently working on resolution for the same.
What this has meant is that we will have to physically monitor the hosts for any hardware component failure since vCenter hardware alerts wont work !!