A couple of days ago Runecast Analyzer has been upgraded to version 3.0.0. With that upgrade a very important beta-feature became GA: HW Compatibility and Upgrade Simulator.
I used to run the Runecast service account with readonly privileges. It has been sufficient up to version 2.7.x. Even the hardware compatibility check (beta) did work with readonly privileges. After upgrading my appliance to version 3.0.0 (GA), I found a notification. Missing privileges..
Once you open host details and click on I/O devices tab, there’s further information.
Before upgrading an ESXi host, it is best practice to to look at VMware HCL and check compatibility of host and IO devices. The combination of driver version, firmware version and ESXi release is crucial for compatibility. Even minor updates might lead to loss of HCL compatibility. A system that used to be HCL compliant at time of deployment, might no longer be compatible after e.g. the third ESXi update release. Updates can bring new driver versions which in turn might require higher firmware versions.
If you’re lucky you may have a software solution that keeps track of all your firmware and driver versions. Runecast Analyzer for example does a pretty good job and shows you current compatibility issues with a single click. Furthermore you can simulate updates/upgrades to any higher vSphere version and the resulting HCL status.
Unfortunately many customers do not have a software solution like that. In these cases you need to go back to the roots (literally) and gather all information on the ESXi shell. To do so you need to enable SSH service on all hosts you want to verify. That can be done in the vSphere-Client or more elegant and faster by a PowerCLI command.
Virtual Distributed Switches have many advantages over standard switches. Because you have a centralized configuration over all hosts they’re less error prone to configuration errors than standard switches. Call me old fashioned but I prefer to have at least the hosts management interface on a standard switch. In case something bad happens, you can still access the host and make changes on the interface.
Recently a customers host had failed. After restoring configuration, for some reason vmnics were swapped between vdSwitches and it wasn’t possible to configure that host neither with hostclient nor with vCenter. The customer was short on vmnics in the past and has configured Management Network on a distributed Portgroup on a distributed vSwitch. This is legal and usually not a problem. In that special case it was a problem. I was literally locked out of the host. Reassigning NICs in the DCUI didn’t work, because they were all claimed by Distributed-vSwitches thus not available for standard switches.
What now ?
There’s help, but you need to access the CLI of DCUI. Login to DCUI console, select “Troubleshooting Options” in the main menu.
If you’ve ever had the task to upgrade a number of ESXi servers, you’ll know that it’s a hell lot of work to be done even before you can upgrade the first host. Is the host still on the VMware Hardware Compatibility List (HCL)? Are my PCIe cards (NIC, HBA, SCSI-controller) still supported? Does the current firmware match the future driver?
Finding answers all of these questions might take hours of research. Especially if you have non-uniform clusters.
Runecast got you covered
We have good news for those of you who already use Runecast-Analyzer to scan their cluster against KB issues. The latest beta (2.7.0) of the HW compatibility tool can now simulate host upgrades. Wow cool is that!