-
Bug
-
Resolution: Unresolved
-
Undefined
-
None
-
4.20
-
Quality / Stability / Reliability
-
False
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
Description of problem:
Installing a 4.20 cluster on OpenStack platform via IPI method fails while bootstrapping the cluster. The cloud controller manager pods crash with following error. The CCM pods constantly restart and crash with the following error. This results into credential sync failing. ~~~ F1107 11:03:17.797998 1 openstack.go:355] Failed to create an OpenStack LoadBalancer client: failed to find load-balancer v2 public endpoint for region : No suitable endpoint could be found in the service catalog. ~~~ The master nodes do get created and join the cluster but are stuck in NotReady state. The installation works seamlessly with same clouds.yaml file and install config file for a 4.19 cluster, but fails for 4.20 installation. As a workaround the loadbalencer is manually disabled. Appending the fields below gets the cluster installed. ~~~ [LoadBalancer] enabled: false ~~~
Version-Release number of selected component (if applicable):
How reproducible:
Install an OpenStack IPI cluster using IPI method without any details of loadbalencer in cloud.yaml file
Steps to Reproduce:
1. Configure the clouds.yaml file with any loadbalencer details
2. Install a OpenStack IPI installation
3. Installation fails with CCM pods crashing
Actual results:
OpenStack IPI fails if loadbalencer details are not provided in cloud.yaml file
Expected results:
The installation should succeed with no loadbalencer details provided in clouds.yaml file.
Additional info:
Must gather of the failed attempt can be found here - https://drive.google.com/drive/folders/1kfx-ndiKHlj18tc8tD4XQ2YTGlf9CrOq?usp=sharing