- Related issues: #2546 [BUG] Harvester Cloud Provider is not able to deploy upgraded container after upgrading the cluster
Category:
- Rancher integration
Verification Steps
- Prepare the previous stable Rancher rc version and Harvester
- Update
rke-metadata-config
to{"refresh-interval-minutes":"1440","url":"https://yufa-dev.s3.ap-east-1.amazonaws.com/data.json"}
in global settings - Update the ui-dashboard-index to
https://releases.rancher.com/dashboard/latest/index.html
- Set
ui-offline-preferred
toRemote
- Refresh web page (ctrl + r)
- Open Create RKE2 cluster page
- Check the
show deprecated kubernetes patched versions
- Select
v1.23.8+rke2r1
- Finish the RKE2 cluster provision
- Check the current cloud provider version in workload page
- Edit RKE2 cluster, upgrade the kubernetes version to
1.23.9-rc3+rke2r1
- Wait for update finish
- The cloud provider is upgrading
- delete the old cloud provider version pod (v0.1.3)
- Wait for newer version cloud provider have been bumped
Expected Results
After upgrading the existing RKE2 guest cluster kubernetes version from older v1.23.8+rke2r1
to 1.23.9-rc3+rke2r1
.
The Harvester cloud provider can successfully updated from v0.1.3
to v0.1.4
Cloud provider before upgrade: v0.1.3
After upgrade: v0.1.4