Bug 1886148 - [ovs] Network operator stuck on message "the cluster operator network has not yet successfully rolled out" while downgrading from 4.6 -> 4.5
Summary: [ovs] Network operator stuck on message "the cluster operator network has not...
Alias: None
Product: OpenShift Container Platform
Classification: Red Hat
Component: Networking
Version: 4.5
Hardware: All
OS: All
Target Milestone: ---
: 4.5.z
Assignee: Casey Callendrello
QA Contact: zhaozhanqi
Depends On: 1885848
Blocks: 1886127
TreeView+ depends on / blocked
Reported: 2020-10-07 18:39 UTC by Aniket Bhat
Modified: 2021-04-20 15:53 UTC (History)
8 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of: 1885848
Last Closed: 2021-04-20 15:53:31 UTC
Target Upstream Version:

Attachments (Terms of Use)

Comment 2 Tim Rozet 2020-10-22 14:57:45 UTC
If we need to downgrade and OVS needs to be disabled/stopped we really need to do that in CNO I think. I assume downgrade has the same order of operations as upgrade, where it will run CNO upgrade/downgrade first then do MCO. In that case when we downgrade CNO and try to start the pods as container OVS, we need to stop and systemctl disable the openvswitch service on the host. Casey do you agree?

Antonio if we downgrade to 4.5 with MCO, do we boot into a previous rpm ostree? Will it not have the openvswitch RPM anyway?

Comment 3 Tim Rozet 2020-10-22 21:44:14 UTC
After speaking with Antonio 4.6->4.5 downgrade will revert to the previous rpm ostree. So openvswitch wont be present on the host. However during the downgrade we actually need to stop and disable openvswitch. This needs to be done in CNO so moving this bug to the right place.

Comment 4 zhaozhanqi 2020-10-23 12:52:42 UTC
we have one similar bug https://bugzilla.redhat.com/show_bug.cgi?id=1886127 for this.

Note You need to log in before you can comment on or make changes to this bug.