Closed rthallisey closed 5 years ago
/assign booxter
@rthallisey I am confused now. Here in the issue you are saying that converged/operator.yaml
is downstream only (meaning, to be used when deploying on top of OCP?) while in https://github.com/kubevirt/hyperconverged-cluster-operator/pull/41#discussion_r278048417 Petr says that it's generic manifest (?) and so we shouldn't touch it.
On the other hand, evidently when I apply manifests from converged
on OCP using default manifests I still get e.g. bridge
CNI not cnv-bridge
so is the manifest really for downstream then?
Please clarify because right now I am not even sure what this issue is supposed to tackle. Thanks.
Not sure that is true. Converged is simply one big CSV for all components, should work on both U/S and D/S. Standard it the other "mode", where you deploy all components using their own CSVs. Not sure if it is used anywhere.
Converged is only going to be used d/s (released Openshift only) and checked into upstream git. For upstream deployments (k8s or master branch of Openshift), we're going to use the 'standard' manifest.
The issue has to do with SRIOV_NETWORK_TYPE
needing a different value d/s, so the downstream yaml (converged) needs the correct value. I'm assuming this is still the expected behavior @phoracek?
@tiraboschi @irosenzw this should be covered by the work you're doing.
@rthallisey with OpenShift's sriov-operator it is likely that we won't handle SR-IOV in HCO at all. It still need to be discussed though.
SR-IOV is no longer part of Network Addons Operator. We can close this issue.
deploy/converged/operator.yaml
should change to it's downstream value singeconverged
is downstream only.