-
Bug
-
Resolution: Done
-
Medium
-
Casablanca Release
-
OOM on Windriver lab
On OOM deployment, DMAAP failure happens from time to time. When DMAAP fails, published event can not be received by subscriber and service distribution fails.
Worse thing is DMAAP service can not be restored reliably after failure by helm deploy command like the following:
root@oom-rancher:~# helm delete dev-dmaap --purge
root@oom-rancher:~# kubectl -n onap delete pvc dev-dmaap-dbc-pg-data-dev-dmaap-dbc-pg-0
persistentvolumeclaim "dev-dmaap-dbc-pg-data-dev-dmaap-dbc-pg-0" deleted
root@oom-rancher:~# kubectl -n onap delete pvc dev-dmaap-dbc-pg-data-dev-dmaap-dbc-pg-1
persistentvolumeclaim "dev-dmaap-dbc-pg-data-dev-dmaap-dbc-pg-1" deletedroot@oom-rancher:~# helm deploy dev local/onap -f /root/integration-override.yaml --namespace onap
No known workaround to restore the service reliably at this point except reinstalling ONAP.
The message-router application.log is attached.