Cannot Remove CatalogApp

I have a catalog app (grafana) which I cannot remove. This issue is similar to Redis remove stuck and CatalogApp Stuck; however, the stuck app is preventing me from accessing the Catalog Apps page for an entire Project.

I installed grafana in the Default project and later went to remove it to change how I set it up. It did not remove and now when I go to the Catalog Apps page for the Default project I get the “loading page” icon with all apps grayed out.

The page remains that way (for weeks).

The rancher container logs show:
[tiller] 2018/08/22 06:45:28 preparing install for grafana
[storage] 2018/08/22 06:45:28 getting release history for “grafana”
[tiller] 2018/08/22 06:45:28 failed install prepare step: a release named grafana already exists.
Run: helm ls --all grafana; to check the status of the release
Or run: helm del --purge grafana; to delete it
Error: a release named grafana already exists.
Run: helm ls --all grafana; to check the status of the release
Or run: helm del --purge grafana; to delete it
E0822 06:45:28.879514 1 generic_controller.go:204] AppController project-69ggs/grafana [helm-controller] failed with : exit status 1
[main] 2018/08/22 06:45:31 Starting Tiller v2.8.0-rancher2 (tls=false)
[main] 2018/08/22 06:45:31 GRPC listening on :50984
[main] 2018/08/22 06:45:31 Probes listening on :32872
[main] 2018/08/22 06:45:31 Storage driver is ConfigMap
[main] 2018/08/22 06:45:31 Max history per release is 0
[tiller] 2018/08/22 06:45:32 preparing install for prometheus
[storage] 2018/08/22 06:45:32 getting release history for “prometheus”
[tiller] 2018/08/22 06:45:32 failed install prepare step: a release named prometheus already exists.
Run: helm ls --all prometheus; to check the status of the release
Or run: helm del --purge prometheus; to delete it
Error: a release named prometheus already exists.
Run: helm ls --all prometheus; to check the status of the release
Or run: helm del --purge prometheus; to delete it
E0822 06:45:32.055635 1 generic_controller.go:204] AppController project-ljqfw/prometheus [helm-controller] failed with : exit status 1
2018/08/22 06:49:02 [INFO] Running cluster events cleanup
2018/08/22 06:49:02 [INFO] Done running cluster events cleanup

Note that prometheus is hung as well.

If I delete the grafana and prometheus namespaces via kubectl they are removed and within 5-10 minutes recreated and grafana (app) is deployed again.

Is there a way to run helm del --purge grafana|prometheus or how can I clean the catalog apps up?

Did you try to restart Rancher ?

Also you should try to upgrade your Rancher installation I see that you are at 2.0.0 and they are now at 2.0.7. They tried to fix it in the latest version: https://github.com/rancher/rancher/issues/11991

If it still there, I think you should comment on the bug link.

Yes, I restarted rancher numerous times and also restarted the node that the services were starting on.

I will try the upgrade approach (will take time given the workload) but I’d be interested to know if there is a way to helm del --purge or other way to remove the stuck catalog apps.