systemctl stop atomic-openshift-master-api does not do graceful shutdown leading to service outage
Issue
- When we stop (or restart) the master-api service, it takes up to 20 seconds until the master gets removed in the 'kubernetes' endpoint. Thus the DNS queries from all containers were going to the "stopped" master-api for 20 seconds, which resulted in a lot of errors on my check tool.
Environment
- Red Hat OpenShift Container Platform 3.3 and 3.4
- Red Hat OpenShift Container Platform 3.11
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.