We need a documentation item which describes how CNV behaves with oc adm drain (kubectl drain).
The problem is as follows:
Today CNV (kubevirt) can not reliably identify node drains if they are triggered through a call of "oc adm drain" (or "kubectl drain").
If these commands are called on a node of a cluster with CNV deployed, then those nodes might not go to maintenance if there are VMs running on top of that node.
The current solution is to use the Node Maintenance Operator for putting nodes into maintenance.
Fabian, could you please provide the original BZ URL for this issue? Thanks!
There was no original bug.
ROman/rmohrRoan can provide the relevant details for this item.
The content associated with this bug was published in the CNV 2.0 Release Notes: https://docs.openshift.com/container-platform/4.1/cnv/cnv_release_notes/cnv-release-notes.html
I verify with Dev, the issue should also be in 2.1 release notes.