Hot Shot
Hot Shot

Upgrade of VDI Environment with NVidia Grid

Hello Everyone,

I wanted to get some insight into upgrading VDI environment and how you guys go about it with the least amount of time and outage. We are planning on updating our vCenters from 6,7 U3 to 7 U2 and hosts as well. We have 2 pods one in each site. Each pod is configured on one vCenter cluster that consists of 15 hosts with Tesla NVidia cards. All hardware is Cisco UCS based. As far as a plan I have come up with the following:

1. Upgrade vCenters to 7U2 sometime after hours when there is not much activity

2. Upgrade UCSM and Fabric to supported version

3. Vacate couple of hosts and upgrade UCS firmware on them

4. Upgrade vacated hosts to 7U2

5. Install supported NVidia VIB on vacated hosts

6. Move images to vacated hosts

7. Remove AppVolumes agent and NVidia driver

8. Install newer NVidia driver and AppVolumes Agent and snap them

9. Pick a maintenance window and disable all pools on one side and remove all clones

10. Put all non-updated hosts in mainenance mode and repeat steps 3,4 and 5 on them

11. Re-enable pools and enable provisioning

Is this a route you guys take as well? With those NVidia cards everything is so much more complicated therefore wanted to see if maybe anybody came up with easier route?

Thanks in advance


Labels (1)
Tags (2)
0 Kudos
0 Replies