Created attachment 1307917 [details]
heapster pod log

Description of problem:
In one new built OCP environment where metrics was not deployed, deploy metrics 3.6 via ansible, only heapster pod was generated, hawkular-metrics and hawkular-cassandra was not generated.

Note: Deploy metrics along with OCP would encounter the same issue.

# oc get po
NAME             READY     STATUS    RESTARTS   AGE
heapster-0tz1v   0/1       Running   7          1h

And only heapster image was pulled, maybe it is because hawkular-metrics and hawkular-cassandra is too large to pull within the required time caused this issue.

# docker images | grep metrics
registry.ops.openshift.com/openshift3/metrics-heapster                 v3.6                d7f375c5d7c3        6 hours ago         274.4 MB

All pods would be generated if we re-deploy metrics again.

Version-Release number of the following components:
# rpm -qa | grep openshift-ansible
openshift-ansible-lookup-plugins-3.6.173.0.2-1.git.0.f6cbecf.el7.noarch
openshift-ansible-3.6.173.0.2-1.git.0.f6cbecf.el7.noarch
openshift-ansible-filter-plugins-3.6.173.0.2-1.git.0.f6cbecf.el7.noarch
openshift-ansible-callback-plugins-3.6.173.0.2-1.git.0.f6cbecf.el7.noarch
openshift-ansible-playbooks-3.6.173.0.2-1.git.0.f6cbecf.el7.noarch
openshift-ansible-docs-3.6.173.0.2-1.git.0.f6cbecf.el7.noarch
openshift-ansible-roles-3.6.173.0.2-1.git.0.f6cbecf.el7.noarch

# openshift version
openshift v3.6.173.0.2
kubernetes v1.6.1+5115d708d7
etcd 3.2.1


How reproducible:
Always

Steps to Reproduce:
1. Deploy metrics 3.6 in one new built OCP environment
2.
3.

Actual results:
Only heapster pod was generated

Expected results:
All pods should be ready

Additional info:
Attached inventory file.