pull policy configuration is NOT being honored

[johnkim76]

The configurable pull policy is currently being set to 'Always', even though it's being configured to be set to "IfNotPresent".

[rthallisey]

Can you paste the logs from the broker?

[eriknelson]

I suspect this is a configuration issue and not actually a bug in the broker. Can you oc exec into a broker container that you think is not honoring the pull policy and post the /etc/ansible-service-broker/config.yaml contents?

Are you deploying with catasb? Which branch and environment?

[johnkim76]

@rthallisey @eriknelson I was trying to debug another issue, and with @jwmatthews help, we determined that the pull policy was being set to 'always'. I'm using catasb (dev branch), and I've been deleting all the images before doing a reset of the environment. Currenlty in a modified environment to complete another task, but I'll provide the requested info once I reset it back.

[eriknelson]

@johnkim76 Are you using the release-0.9 broker or the master broker?

[johnkim76]

@eriknelson I was using catasb which seemed to pull the latest image. I was not using the broker source when I hit this issue.

[jwmatthews]

I can't reproduce with latest deployment from catasb 'dev' branch.

$ oc export pod apb-4a477dee-ee0d-4524-8c6d-dbbed6ddca04 apiVersion: v1 kind: Pod metadata: annotations: openshift.io/scc: anyuid creationTimestamp: null name: apb-4a477dee-ee0d-4524-8c6d-dbbed6ddca04 spec: containers:

• args:
  • provision
  • --extra-vars
  • '{"mediawiki_admin_pass":"password1","mediawiki_admin_user":"admin","mediawiki_db_schema":"mediawiki","mediawiki_site_lang":"en","mediawiki_site_name":"MediaWiki","namespace":"demo"}' image: ansibleplaybookbundle/mediawiki123-apb imagePullPolicy: IfNotPresent name: apb resources: {} securityContext: capabilities: drop:
    • MKNOD
    • SYS_CHROOT privileged: false seLinuxOptions: level: s0:c9,c4 terminationMessagePath: /dev/termination-log terminationMessagePolicy: File volumeMounts:
  • mountPath: /var/run/secrets/kubernetes.io/serviceaccount name: apb-4a477dee-ee0d-4524-8c6d-dbbed6ddca04-token-676hm readOnly: true dnsPolicy: ClusterFirst nodeName: localhost restartPolicy: Never schedulerName: default-scheduler securityContext: seLinuxOptions: level: s0:c9,c4 serviceAccount: apb-4a477dee-ee0d-4524-8c6d-dbbed6ddca04 serviceAccountName: apb-4a477dee-ee0d-4524-8c6d-dbbed6ddca04 terminationGracePeriodSeconds: 30 volumes:
• name: apb-4a477dee-ee0d-4524-8c6d-dbbed6ddca04-token-676hm secret: defaultMode: 420 secretName: apb-4a477dee-ee0d-4524-8c6d-dbbed6ddca04-token-676hm status: phase: Pending qosClass: BestEffort

[johnkim76]

@jwmatthews @eriknelson @rthallisey I just reset everything, and I can NOT reproduce this either.

[johnkim76]

cannot reproduce.. closing.