contiv-experimental/cluster

collins container stops when provisioning the node running it

Closed this issue · 0 comments

commissioning the node running clusterm itself results in collins container getting killed. This seems related to docker service getting restarted as part of commission which should ideally not happen if nothing changed.

Filling this common issue to track ansible side behavior as well any cluster-mgr side fixes.

++++ Clusterm Log snippet ++++++
Jan 05 11:44:07 cluster-node1 clusterm[14242]: level=info msg=
Jan 05 11:44:07 cluster-node1 clusterm[14242]: level=info msg="PLAY RECAP ******************************************************************** "
Jan 05 11:44:07 cluster-node1 clusterm[14242]: level=info msg="cluster-node1-0 : ok=43 changed=17 unreachable=0 failed=0 "
Jan 05 11:44:07 cluster-node1 clusterm[14242]: level=info msg=
Jan 05 11:44:07 cluster-node1 clusterm[14242]: level=error msg="error handling event "nodeConfigure: cluster-node1-0". Error: Post http://localhost:9000/api/asset/cluster-node1-0?reason=N
ode+is+alive+and+discovered+in+monitoring+subsystem&state=Discovered&status=Allocated&tag=cluster-node1-0: dial tcp [::1]:9000: getsockopt: connection refused"
[vagrant@cluster-node1 ~]$