kubevirt deploy stuck - (install) Installing Binaries
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
KubeVirt Charm |
Triaged
|
Low
|
Unassigned |
Bug Description
I realized that using the kubevirt charm rev-12 the deployment is stuck with the status:
kubernetes-worker/1 active idle 8 10.5.3.88 80/tcp,443/tcp Kubernetes worker running.
calico/2 active idle 10.5.3.88 Calico is active
containerd/2 active idle 10.5.3.88 Container runtime available
kubevirt/1 maintenance executing 10.5.3.88 (install) Installing Binaries
The workaround is to reboot the nodes:
logs => 2023-06-28 15:07:21 INFO unit.kubevirt/
Changed in charm-kubevirt: | |
milestone: | none → 1.28 |
status: | New → Triaged |
importance: | Undecided → Low |
Changed in charm-kubevirt: | |
milestone: | 1.28 → 1.28+ck1 |
Changed in charm-kubevirt: | |
milestone: | 1.28+ck1 → 1.29 |
I spoke with Pedro about this and it seems that apt was in the process list running on the machines when this condition occurred. As stated in the juju log it was trying to install necessary packages. But I also understand from him that it wasn't finishing and restarting the machines unstuck it. It would be nice if this condition occurs again to have the logs from apt to see why it was stuck. Fundamentally this isn't something the charm can do much about unless there's some kind of preparation that apt needs from the charm.
I suppose it's POSSIBLE apt is blocked waiting on some kind of human input?
The next steps would be from charmed-kubernetes to attempt to reproduce in a different environment.
@pedrovlf was this on focal or jammy?