I'm managing a Hyper-V cluster with three hosts and around 25 VMs, and I'm new to the patch rebooting process since we just switched from VMware, where this wasn't a big concern. I'm looking for insights on how other sysadmins handle host reboots for patching. While I'm open to migrating VMs off the host, rebooting, and bringing them back, I'd like to find a more efficient method. I've also come across Cluster-Aware Updating (CAU) but haven't used it before and am unsure of its effectiveness in our setup. Any tips or strategies would be really helpful!
3 Answers
One solid approach is to set your cluster up to run at about 2/3 of its full capacity. This way, you can migrate the VMs to the other 2/3 of the cluster, patch and restart the 1/3 that needs the update, and keep everything running smoothly. Automation can definitely help streamline this process too!
My method is to drain one of the hosts in the failover cluster. I pause that host, install the updates, reboot it, and then unpause it. After that, I move the VMs off another host and repeat the process. It keeps things organized and systematic.
If you're looking to minimize downtime, you can reboot all the VMs along with their host when you do updates. Just remember that those VM guests also need to reboot for patches, so it’s a two-for-one deal.
So when you drain a node, does that just temporarily relocate the VMs to other hosts until you bring the original host back online?