VSphere Cluster, Patching and VMs

[Elektritter]

Hi,

I need some help because it looks like I don't understand at some point the vSphere cluster environment.

Until today I thought that with DRS enabled cluster applying a patch to a cluster will end in workload:

* migrate vms from host A to host B

* apply patches to host A

* move back VMs

* do the same with vms on host B

and so on.

But now I saw that remidation of a whole cluster ends up with suspending a lot of VMs in cluster, not all. I don't understand why and why some are not suspended. Is there a rule or setting I miss?

This is a bit frustrating because this is exactly what I wanted to avoid.

E

[depping]

How did you do maintenance/applied patches?

[Elektritter]

Simply marked the baseline I wanted and then clicked "remediate" (In cluster view, NOT single host)

[depping]

Can you check your "remediation settings"? It should look like this, where the "do not change power state" of course is key.

 

[Elektritter]

Ok, you hit the point - when I click on start remidation it says "suspend" as poweroption - but I cannot change it.

Call me blind or whatever - I don't find the settings screen you posted.. I only find HOST related remediation settings in lifecycle manager but not for a cluster and I don't find something in documentation. May you point me out to the right direction?

 

[depping · Accepted solution]

Ah sorry about that:

Go to your cluster, then click on "updates", then click on "image" and then under "image compliance" you have those three dots on the right side. it says "edit remediation settings". there you can change it for the cluster.

[Elektritter]

Ok, got it.

Looks like I made a mistake in general - I did not work with images. Now I see the settings above and understand why it suspended the VMs - it used the per host settings.

But why now two hosts has newer broadcom etc. drivers than in latest image (7.0 U3c - 19193900) and so cannot remidiated is something for a new thread except you have a fast answer for it.

Many, many thanks for pointing me to the correct settings. It looks like it was a big misunderstanding on my side about how patches in a cluster gets applied.

E

[depping]

Great to hear it worked!