DescriptionEvgheni Dereveanchin
2016-02-26 16:52:54 UTC
Description of problem:
Currently each deployment leaves a replication controller behind for rollback purposes. Eventually this leads to etcd getting filled with stale data. Also, if quotas are enforced, deployments will soon stop working due to replication controllers piling up, and the only way to clean them up is using "oadm prune"
Version-Release number of selected component (if applicable):
3.1.1
Proposed improvement:
Allow setting replication controller limit per project or per deployment.
Additional info:
upstream pull request was merged:
https://github.com/kubernetes/kubernetes/pull/19590
Confirmed with OCP , the issue has fixed:
openshift v3.3.0.27
kubernetes v1.3.0+507d3a7
etcd 2.3.0+git
Auto cleanup the first RC, only retain the latest 3 RC;
oc get rc
NAME DESIRED CURRENT AGE
history-limit-3 0 0 9m
history-limit-4 0 0 7m
history-limit-5 1 1 7m