In this post I am going to be covering the details that sometimes get missed when performing upgrades to a vCenter environment. A good example of this is making sure that the latest drivers updates are installed for a particular device, such as an NVIDIA K1 card! The VDI environment that I work on has such cards and the drivers do change from 5.5 to 6.0 so they must be installed. Now, I wasn’t sharp enough to include them in the image before building these hosts, but I am going to do it for my 32 production VSAN nodes.
Other interesting things that I’ll be looking at is the enabling of the VSAN Health check via vCenter and also upgrading the Distributed Virtual Switch in my environment!
Install latest NVIDIA GPU drivers on ESXi 6.0 U1
1. I went to the VMware KB Article that details the instructions for this process and downloaded the drivers I needed.
2. Then I extracted and uploaded the driver .VIB file to my VSAN datastore.
3. Enable SSH on the host and connect to it via an SSH Client.
4. Next I had to put the host into maintenance mode that I wanted to work on, because the old drivers need to be removed. Run the following commands to stop the graphics services:
esxcli system maintenancemode set --enabled
esxcli software vib list
5. Run the command below to remove the driver and wait for it to complete, it can take a few minutes.
6. Navigate via the CLI to the VSAN directory where the new VIB has been uploaded too. (It is important to navigate into the folder itself if the driver is stored on a VSAN datastore). Run the following commands:
esxli software vib install -v /vmfs/volumes/*YOUR_VSAN*/*FOLDERNAME*/NVIDIA-VMware_ESXi_6.0_Host_Driver
7. Once the install has complete, run the following to exit maintenance mode and start the xorg service:
esxcli system maintenancemode set --disabled
8. Repeat this for your remaining nodes.
Note: The drivers do not take effect until a reboot, but this will be done shortly in the next section.
Enable VSAN Health Check:
1. Login to the vCenter via the Web Client
2. Navigate to Cluster, Manage, Settings and to the Virtual SAN “Health” Section.
3. Select “Enable” at the top right of the screen and accept the warning:
4. The cluster will now go through a rolling maintenance mode/reboot cycle similar to that of a VUM task. VSAN warnings may appear on the dashboard whilst this happens as hosts are being configured for alerting. Wait until it has finished to investigate existing warnings.
Upgrade Distributed Switch
1. Login to the vCenter via the vSphere Web client
2. Navigate to Home, Networking, DataCenter and then find your dvSwitches in the Navigator pane. You can see in the summary it’s current version and that an Upgrade is available.
3. Right click the switch and select Upgrade and then “Upgrade Distributed Switch”.
8. On my main dvSwitch (that has a vMotion vmkernel port group) I saw a warning relating to CoS tagging. This was never setup so I assume it was a default previously set and now due to changes in CoS configuration is warning to create new traffic rules on the port groups instead. Click next and OK at the prompt.
10. Repeat these steps, as necessary, for any other dvSwitches in your environment.
That is it for the extra configuration that I had to perform for my upgrade. Nothing too spectacular or taxing but always worth considering. After this had all been done, I actually ran through the VSAN Health check on the web client to make sure that the new HCL database check was up to date and that all my components were on the latest and greatest versions. It is an excellent tool suite built-in, the final cherry on the cake would be to have the VSAN Observer built into the web client, which I heard was in the pipeline at VMworld…so I’m keeping my eyes peeled for that!
In my next post, I’ll be running through the Horizon View Upgrade. I might even sneak in an extra post about the new Access Point Appliance which I’ve also setup and tested!