c14d8c6514
This is to move to onDelete upgrade strategy for ceph-osd daemonsets so that osd upgrade cane be performed by failure domains as current upgrade strategy(RollingUpdate) will randomly pick the osd pods for upgrade. This will be more helpful when we have rack based failure domains on the ceph clusters. This ps will add a new job called post-apply to restart the osd pods rack by rack - post-apply job will make sure osds gets restart rack by rack which will save upgrade time. - its less/no distruptive since we are upgrading per failure domain. also this job will be enabled only when we have OnDelete upgrade strategy in values. Change-Id: I2e977e75616e08fee780f714bbd267743c42c74d |
||
---|---|---|
.. | ||
bin | ||
utils | ||
configmap-bin.yaml | ||
configmap-etc.yaml | ||
daemonset-osd.yaml | ||
job-bootstrap.yaml | ||
job-image-repo-sync.yaml | ||
job-post-apply.yaml | ||
pod-helm-tests.yaml |