Cluster delete operations become stuck during TCA operations and upgrades.
2.x, 3.x
This issue is caused by the introduction of a duplicate Harbor add-on to workload clusters after using the search feature located in the VIM associations page. When using the search feature (TCA UI - > Partner Systems - > Modify Partner Systems - > VIM Associations - > Search Filter), if the user deselects a VIM and clicks on Finish then all workload clusters are disassociated from Harbor. This procedure will introduce a duplicate add-on to each workload cluster and subsequent node pool creations will be stuck in a processing state.
Delete duplicate Harbor add-ons
From the TCA UI, go to the CaaS infrastructure page.
Select a V2 cluster which shows the add-ons count as 9.
In the list of add-ons, delete any add-ons named harbor and harbor-tca-addon.
Once the delete operation finishes, click on Deploy Add-on.
Select harbor from the Select Add-on list and click on NEXT.
Click on Select registered harbor and click OK.
Select the harbor instance from the Select Harbor Repository list and click on SELECT.
Click on Next.
Click on Deploy.
Repeat these steps for each cluster that has the duplicate harbor add-on.
Prevent duplicate Harbor add-ons during workload deletion
The following procedure can be used to avoid causing the creation of a duplicate harbor add-on during a workload cluster deletion workflow.
To delete a workload cluster, go to the TCA UI, navigate to Partner Systems, and select harbor from the list of Partner Systems entries.