Details
-
Feature
-
Resolution: Won't Do
-
Major
-
None
-
None
-
False
-
False
-
0
-
0%
-
0
Description
What is the problem that your customer is facing?
Use of 'requiredDuringSchedulingIgnoredDuringExecution' for elastic search pod instead of 'preferredDuringSchedulingIgnoredDuringExecution' because that is a soft policy.
Actual issue/use case where it's needed:
Normally each elasticsearch pod is placed to a different node thanks to 'preferredDuringSchedulingIgnoredDuringExecution' policy.
However, when one of node got down, the pod running on the down node is moved to another node where elasticsearch pod is already running. The pod isn't moved to the previous node even after the down node recovered. This is a problem we are facing.
What is the business impact, if any, if this request will not be made
- Medium, If the node goes down, so the pod will not moved to that node even after the down node is recovered.
What are your expectations for this feature
Use of 'requiredDuringSchedulingIgnoredDuringExecution' for elastic search pod instead of 'preferredDuringSchedulingIgnoredDuringExecution' because that is a soft policy.
