Description of problem:
The new Virtualization Overview page is visible only if the user is logged in to the cluster with kube:admin user.
If s/he is logged in with IDP and has a role binding of cluster-admin, the overview page is visible for a split second, and then the whole page turns into blank.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. log in with non kube:admin user, but with cluster-admin role assigned to the user.
2. go to Virtualization --> Overview page
the page turns blank almost immediately.
the page should be visible with contents.
OCP version: 4.10.z
CNV version: 4.10.0
Could not see the issue on OCP 4.10.4.
Are you using the dynamic plugin console?
I'm not using the dynamic plugin but the default UI. The specific OCP version is 4.10.3
I can confirm it is working only with htpasswd identity provider (non-admins can't see the virtualization overview menu entry), but it doesn't work for other identity providers, such as Google or LDAP.
FWIW, this is the UI console logs when loading the virtualization overview page when logged in with a cluster-admin user which is not the kubeadmin:
The page turns blank on line 106:
TypeError: t.labels is undefined
(In reply to Oren Cohen from comment #3)
> I'm not using the dynamic plugin but the default UI. The specific OCP
> version is 4.10.3
> I can confirm it is working only with htpasswd identity provider (non-admins
> can't see the virtualization overview menu entry), but it doesn't work for
> other identity providers, such as Google or LDAP.
thanks, I also tested it works with htpasswd identity provider.
so it's an issue for other identity providers.
It also always happens when logged in with kubeadmin, on a specific cluster (cnv.engineering.redhat.com).
Reducing the bug severity as the overview page works as expected on other clusters I've tried.
This problem also happen to kube:admin user when navigating to k8s/virtualization/~new-from-template
nvm. Was using unsupported configuration running 4.10 CNV on OCP 4.9
Tiger's use case of running OpenShift Virtualization 4.10 on OpenShift 4.9 is a valid use case, we allow people to do this to get pre-releases.
New capabilities that require upgrades in OCP 4.10 obviously will not work, but the basic functionality of creating and running a VMs should not be broken like this.
Please triage as usual.
moving to urgent because 4.10 is released, and we need to backport a fix ASP
I'm not managing the ocp-virt cluster, but the cnv.engineering.redhat.com one.
@yzamir , could you please backport the fix to 4.10.z ?
I'm upgrading this cluster only to official released versions, and I guess this is the same as in ocp-virt cluster.
the backport BZ is ON_QA:
Fix merged 4 days ago into 4.10 upstream
I was meant to comment on 4.10.z bug, but commented to this one. :)
verified on console master
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 (Important: OpenShift Container Platform 4.11.0 bug fix and security update), and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.