Description of problem: Duplicate strategy is evicting the deploy pod in the RC which is preventing new pods from being created. Version-Release number of selected component (if applicable): 4.4.0-rc.1 How reproducible: Always Steps to Reproduce: 1. Install descheduler operator via OLM 2. Add the duplicate strategy. 3. Now create pods by running the command below "oc run hello --image=openshift/hello-openshift:latest --replicas=12" 4. While the pods are in ContainerCreating state edit the kubedescheduler object to change the interval seconds to some value. Actual results: Now user can see that all pods goes into terminating state including the deployer pod and the descheduler evicts the deployer pod due to which the pods never come up. Expected results: deployer pod should not be evicted as this would prevent more pods from creation. Additional info: This is a symptom of https://github.com/kubernetes-sigs/descheduler/issues/230
We spoke about it with Mike and I doubt we'll fix it for 4.4, moving to 4.5
Pr upstream to address this in a quick way: https://github.com/kubernetes-sigs/descheduler/pull/275
Verified with the payload below and do not see the issue happening, moving the bug to verified state. [ramakasturinarra@dhcp35-60 Downloads]$ oc get clusterversion NAME VERSION AVAILABLE PROGRESSING SINCE STATUS version 4.5.0-0.nightly-2020-05-21-072118 True False 3h9m Cluster version is 4.5.0-0.nightly-2020-05-21-072118
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2020:2409