Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1620121

Summary: Service catalog 3.11.0-0.20.0 CrashLoopBackoff with unknown admission plugin
Product: OpenShift Container Platform Reporter: Mike Fiedler <mifiedle>
Component: Service CatalogAssignee: Mike Fiedler <mifiedle>
Status: CLOSED NOTABUG QA Contact: Zhang Cheng <chezhang>
Severity: high Docs Contact:
Priority: unspecified    
Version: 3.11.0   
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2018-08-22 14:40:20 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:
Embargoed:
Attachments:
Description Flags
inventory none

Description Mike Fiedler 2018-08-22 13:31:05 UTC
Created attachment 1477888 [details]
inventory

Description of problem:

Installed a cluster from puddle 3.11.0-0.20.0 and the service catalog pods would not start - stuck in CrashLoop with the error below.

Backing off to service catalog images from 3.11.0-0.16.0 worked ok.

I0822 13:21:41.633656       1 feature_gate.go:194] feature gates: map[OriginatingIdentity:true]
I0822 13:21:41.633791       1 feature_gate.go:194] feature gates: map[OriginatingIdentity:true NamespacedServiceBroker:true]
I0822 13:21:41.633818       1 hyperkube.go:192] Service Catalog version v3.11.0-0.20.0;Upstream:v0.1.29 (built 2018-08-22T01:00:39Z)
W0822 13:21:41.891525       1 util.go:112] OpenAPI spec will not be served
I0822 13:21:41.892466       1 util.go:182] Admission control plugin names: [NamespaceLifecycle MutatingAdmissionWebhook ValidatingAdmissionWebhook DefaultServicePlan ServiceBindingsLifecycle ServicePlanChangeValidator BrokerAuthSarCheck KubernetesNamespaceLifecycle]               
Error: failed to initialize admission: unknown admission plugin: KubernetesNamespaceLifecycle



Version-Release number of selected component (if applicable): Service catalog and openshift-ansible both at 3.11.0-0.20.0


How reproducible: Always


Steps to Reproduce:
1.  Install a 3.11.0-0.20.0 cluster with service catalog enabled (inventory attached)


Actual results:

Service catalog pods fail to start (see above for error messages)

Comment 1 Jay Boyd 2018-08-22 14:35:51 UTC
last week we picked up a new upstream Service Catalog binary, it required a change to the list of Admission Controllers that are run.  The name changed from KubernetesNamespaceLifecycle to NamespaceLifecycle.  The Ansible playbooks laid down at /root/openshift-ansible/roles/openshift_service_catalog/templates/api_server.j2  look right, but the daemonset.apps/apiserver was using the old value (KubernetesNamespaceLifecycle).

Comment 2 Mike Fiedler 2018-08-22 14:36:04 UTC
@jaboyd investigated and found old name for admission controller being used.   Assigning to myself to check the slave and make sure uplevel rpm in use.

Comment 3 Mike Fiedler 2018-08-22 14:40:20 UTC
Backlevel openshift-ansible in use.  Closing as NOTABUG