-
Bug
-
Resolution: Won't Do
-
Undefined
-
None
-
None
-
None
When deploying a minimal broker using the default example in the web console and adding an httpGet liveness probe, the StatefulSet is created but the Pods are not deployed. In the Namespace Events I see the following:
LAST SEEN TYPE REASON OBJECT MESSAGE
0s Warning FailedCreate statefulset/ex-aao-ss create Pod ex-aao-ss-0 in StatefulSet ex-aao-ss failed error: Pod "ex-aao-ss-0" is invalid: spec.containers[0].livenessProbe: Required value: must specify a handler type
When looking at the yaml of the created StatefulSet, the livenessProbe is missing the httpGet parameter (the readiness and startup probes are created fine):
livenessProbe: failureThreshold: 3 periodSeconds: 10 successThreshold: 1 timeoutSeconds: 5
If I manually edit to StatefulSet to add the missing httpGet.port: 8161 parameters, the Pod deployment finishes successfully.
The ActiveMQArtemis yaml used for the test was the following:
kind: ActiveMQArtemis apiVersion: broker.amq.io/v1beta1 metadata: application: ex-aao-app name: ex-aao namespace: a-test spec: deploymentPlan: image: placeholder jolokiaAgentEnabled: false journalType: nio managementRBACEnabled: true messageMigration: false persistenceEnabled: false requireLogin: false size: 2 livenessProbe: failureThreshold: 3 httpGet: port: 8161 periodSeconds: 10 successThreshold: 1 timeoutSeconds: 5 readinessProbe: failureThreshold: 3 httpGet: port: 8161 periodSeconds: 10 successThreshold: 1 timeoutSeconds: 5 startupProbe: failureThreshold: 600 httpGet: port: 8161 periodSeconds: 1 successThreshold: 1 timeoutSeconds: 1
Using tcpSocket instead for the livenessProbe works fine, only httpGet seems to be affected.
Tested on OpenShift 4.12 and 4.13 with AMQ Broker Operator 7.11.4-opr-1.
- clones
-
ENTMQBR-8685 Failed to create broker deployment when using an httpGet livenessProbe
- Closed