Bug 1938188

Summary: LoadBalancer service for Octavia remain in Pending state until kube-controller-manager is restarted
Product: OpenShift Container Platform Reporter: Andy Bartlett <andbartl>
Component: kube-controller-managerAssignee: Maciej Szulik <maszulik>
Status: CLOSED DUPLICATE QA Contact: zhou ying <yinzhou>
Severity: medium Docs Contact:
Priority: medium    
Version: 4.6CC: anbhat, aos-bugs, cgoncalves, eduen, ihrachys, jnordell, lpeer, ltomasbo, majopela, m.andre, maszulik, mbooth, mfedosin, mfojtik, njohnston, pprinett, scohen
Target Milestone: ---Keywords: Reopened
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2022-02-15 13:12:11 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:

Description Andy Bartlett 2021-03-12 12:09:19 UTC
Description of problem:

 My customer has the following issue:

Two services of type LoadBalancer were created in our cluster that remained in Pending state and no Octavia LoadBalancers were created in Openstack. We couldn't find any relevant errors in the Openshift logging and in Openstack there weren't any issues with Octavia either. We increased the log level of the kubecontrollermanagers.operator.openshift.io, which restarted the kube-controller-manager pods in the openshift-kube-controller-manager namespace. After the restart, the LoadBalancers were created in Octavia.

Version-Release number of selected component (if applicable):

OCP 4.6.15

How reproducible:

Steps to Reproduce:

Actual results:

Expected results:

We'd expect the LoadBalancers to be created in the reconcile loop of the kube-controller-managers without needing to restart them.
We'd also expect the events in the namespace in which the LoadBalancer service is created to show more logs/error messages when creating them through the cloud provider fails.

Additional info:

Comment 21 Maciej Szulik 2021-04-09 15:13:43 UTC
The solution to this particular problem is to upgrade to OCP 4.6.19 or newer.

*** This bug has been marked as a duplicate of bug 1924785 ***

Comment 38 Martin André 2022-02-15 13:12:11 UTC
Closing as a dup of https://bugzilla.redhat.com/show_bug.cgi?id=2004542. The fix has already merged in 4.8 and above, the backport for 4.7 is in progress at https://bugzilla.redhat.com/show_bug.cgi?id=2054669 and we'll backport down to 4.6.

*** This bug has been marked as a duplicate of bug 2004542 ***