Patching hyper-v Hosts
Im wondering what folks are doing for patching hyper V hosts. I think like many, we dont rely on the host to take down the guests properly when a restart is necessary on the host. We are manually patching the hosts now and manually taking down all the guests. My thought is the process would be something like this.
Script forcibly log out all users from Guests (seems like a hung user can cause the guests to come down consistently.. that is just a theory to be fair)
Script taking down all guests, verify not pingable and then continue to Host patch
If one or more not pingable, script bringing up the guests that went down, abort the patch and report there was an issue
Patch host
Set all guests to restart when host comes up (not part of kace task chain)
Wait X amount of time and run a script to make sure hosts and all guests respond, possibly that a service is running.
If one or more not up, report as such with an urgent message
I have no idea exactly how Id accomplish this, but certainly would involve a task chain and multiple scripts.
Anyone doing anything like this or have any advice?
Answers (1)
how about clustering the HV-Hosts, evacuating the guests, patch the host, reboot it and allow it to "play again"?
(I don't use HV but these steps are common with the usual virtualization environments and their vendors you find out there)
Comments:
-
I've decided to use out network monitor to report status. Not sure why I didnt think of that before, it is essentially doing it now.
When you say 'clustering' the HV hosts.. what do you mean by that precisely?
In regards to evaluating the guests.. as I understand it from my team, we shouldnt rely on the host to take down the quests.. as I understand it, the host can take them down hard if they dont come down nicely. I guess that is what I was looking for, if someone has a script which takes them down, and verifies they are down before proceeding to patching the host.
Maybe Im making more out of this then I need to. I just patched a host and let it take down the guests and all was well.. but it is when that doesnt happen that concerns me. I guess we had an AD controller not come down nicely and it was rug pulled by the host winding up in a blue screen. Perhaps I should be looking at GPO settings for the hosts. IE, settings which determine when a quest is down etc.
Im just looking for what is working for folks because this is our first go around at this. The plan is to patch guests during the week and do the hosts on friday night at this point. We are also noticing depending on when things were patched last we are needing repeat deploys. I guess this shouldnt happen once we are on a monthly schedule.
Thank you for the response. So far, knock on wood, Kace has done a great job with our test servers. - barchetta 2 years ago
