Removing compute nodes from cloud groups
You can remove one or more compute nodes from a cloud group when the compute node is no longer needed.
Before you begin
About this task
To remove a compute node from its cloud group, the compute node must be powered on and quiesced
before you switch it to maintenance mode. To successfully switch into maintenance mode, you must
evacuate all virtual machines running on that compute node to other compute nodes in the same cloud
group. A virtual machine will not migrate successfully if no other compute nodes in the cloud group
has sufficient resources to host it. If a virtual machine cannot be migrated, you must manually stop
and store the virtual machine before its compute node can go into maintenance mode.
Important: If you want to remove all compute nodes from a cloud group, you
must first remove all data stores associated with that cloud group.
Remember: When removing compute nodes from Virtual Manager cloud groups, you
must perform the quiesce and maintenance operations (described in Steps 1 and 2 of the following
procedure) on the compute node within each Software Workload Environment hosted by the Virtual
Manager cloud group.
Restriction: If a compute
node in a Virtual Manager cloud group is configured as a dedicated failover host in the Admission
Control section of the VMware vSphere HA settings
for the cluster, it cannot be removed from the cloud group. You must remove the host from the list
of dedicated failover hosts before you attempt to remove it from the cloud group in Cloud Pak. If the compute node is disconnected
in VMware, removing it from the dedicated
failover list might not take effect. Double check the dedicated failover hosts list before you
attempt to remove the compute node from the cloud group in Cloud Pak. If the host is still in the list of
dedicated hosts, attempting to remove it from the cloud group might result in Cloud Pak System
showing the compute node as removed, while VMware shows the compute node as still present
in the cluster.