As I have a bunch of questions around how you can delete the vSphere Cluster Service VMs (vCLS VMs) I figured I would create a quick demo. It is pretty straight forward, and it should only be used when people are doing some kind of full cluster maintenance. This demo shows you how to get the VMs deleted by leveraging a vCenter Server Level Advanced setting (config.vcls.clusters.domain-c<identifier>.enabled). I have also written a post which has a bunch of requirements, Q&A, and considerations for the vCLS VMs, if you are interested in that read it here. Note, if you have a resource pool configuration, enabling “retreat mode” (disabling vCLS)) doesn’t impact resource pools in any shape or form, it just impact DRS load balancing. Anyway, I hope you find the demo useful.
What is a workaround when all vCLs VMs were deleted? How to recreate them?
If they have been deleted you enable vCLS again using the advanced setting shown in the demo. If for whatever reason they just disappeared, add the advanced setting with “False”, wait for 60 seconds, then change the value to “True” and they should pop up again.
Hello,
I would like to reinstall totally my vCenter. I’m going to delete my vCenter VM for that. What is the best practice concerning vCLS before that, please ?
I would recommend enabling Retreat mode so that the vCLS VMs are deleted before you add the hosts to the new vCenter Server.
Greetings Duncan! Big fan! Is there a way to programmatically grab the cluster number needed to be able to automate this with powercli … I have a 4node self managed vsan cluster, and once upgrading to 7U1+ my shutdown and startup scripts need tweaking (bc the vCLS VMs do not behave well for this use case workflow). An alternative could be to not use DRS and HA … but I would like the customer to gain the benefits of these important features … I’ve scrubbed the communities site (to no avail for my use case), and bf I call support I figured I’d ask …
William actually has examples of how to retrieve the MoRef for various objects, like the cluster: https://www.virtuallyghetto.com/2011/11/vsphere-moref-managed-object-reference.html. Hope this helps, and thanks for the compliments!
Thx Duncan. Figured it out for my use case, and posted to communities https://communities.vmware.com/t5/VMware-vCenter-Discussions/disable-vCLS-need-to-delete-vCenter-advanced-option-name/m-p/2837907/highlight/true#M42545
that is awesome, thanks for sharing!
In order to retire an old datastore that vCLS VMs were placed on I put my cluster into retreat mode. The VMs did not delete themselves. So I went ahead and manually powered off and deleted the VMs. Now, at this point, after taking the cluster out of retreat mode and turning DRS back on the vCLS VMs are not being recreated and the cluster is showing the expected vSphere DRS functionality alert. Is there any way to force the cluster to re-provision the vCLS VMs?
Thanks,
Joe
thanks for sharing. I had the problem that from the three vCLS VMs only one was available. The other two were sitting on an unavailable datastore… so I had no chance to manupualte them. So I just did your procedure: All VMs were deleted and then recreated on proper datastores.
Hello, what about disabling vCLS permently when you only have only 2 hosts in the cluster and no DRS license ? Will it impact HA ? HA relies on DRS to restart VM but since there is only 2 hosts in the cluster, if one fails it should not need much computation to know where to restart the VMs. 🙂
Sure, you can do that. But keep in mind that over time vCLS will be used by HA as well. So disabling it may cause problems in the future.
Hello Duncan, I have My vCLS on “inaccessible” status, What can I do?
Awesome. Made the jump from 6.7 to 7.0.3 and several clusters had bad vcls vm’s that were basically in emergency (im doing nothing) “Grub rescue>” . this cleaned up those clusters including the bad ones.