Description of problem: The feature gate CPUNodeDiscovery needs to be enabled to allow VMs to request specific CPU model names (compatibility). This is currently disabled to not accidentally block QE testing, but we would like to release 1.4 WITH the feature flag enabled by default once we confirm it works. The feature gate can be added manually: Add CPUNodeDiscovery to the kubevirt-config: $ oc edit configmap -n kubevirt kubevirt-config
while we have seen that it is working, i think its a bit late to enable it. can we target to 2.0?
Absolutely, 2.0 only. We already agreed to not have it in 1.4 enabled by default.
last update: After installation of hco-bundle-registry:v2.0.0-10 # oc describe cm kubevirt-config -n kubevirt-hyperconverged Name: kubevirt-config Namespace: kubevirt-hyperconverged Labels: app=hyperconverged-cluster Annotations: <none> Data ==== feature-gates: ---- DataVolumes,SRIOV,LiveMigration,CPUManager Events: <none>
Posted as https://github.com/kubevirt/hyperconverged-cluster-operator/pull/89
Hi Martin, can this be moved to ON_QA? The PR was merged.
Yes, but someone from the HCO team should fill in the fixed in version field..
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHEA-2019:1850