topologyspreadconstraints plugin is evicting more than expected pods

[SaiSasankKhajjayam]

What version of descheduler are you using?

descheduler version: v0.30.1

Does this issue reproduce with the latest release?
yes

Which descheduler CLI options are you using?
NA
Please provide a copy of your descheduler policy config file
Pls find the attached unit test case to reproduce the issue.

What k8s version are you using (kubectl version)?

kubectl version Output

$ kubectl version
  Client Version: v1.29.8
  Kustomize Version: v5.0.4-0.20230601165947-6ce0bf390ce3
  Server Version: v1.29.8

What did you do?
Problem: descheduler is evicting 26 pods when one rack was restarted (16nodes/4 racks/64 pods/2 topology constraints) and expected eviction should be 16.
Initial Configuration of pods distribution as below:
n1 - rack1 - 4
n2 - rack1 - 4
n3 - rack1 - 4
n4 - rack1 - 4
n5 - rack2 - 4
n6 - rack2 - 4
n7 - rack2 - 4
n8 - rack2 - 4
n9 - rack3 - 4
n10 - rack3 - 4
n11 - rack3 - 4
n12 - rack3 - 4
n13 - rack4 - 4
n14 - rack4 - 4
n15 - rack4 - 4
n16 - rack4 - 4

Entire rack3 was shut down and pod distribution looks as below:

n1 - rack1 - 6
n2 - rack1 - 6
n3 - rack1 - 5
n4 - rack1 - 4
n5 - rack2 - 6
n6 - rack2 - 5
n7 - rack2 - 5
n8 - rack2 - 5
n9 - rack3 - 0
n10 - rack3 - 0
n11 - rack3 - 0
n12 - rack3 - 0
n13 - rack4 - 6
n14 - rack4 - 6
n15 - rack4 - 5
n16 - rack4 - 5

When rack3 was back after some time the descheduler is evicting 26 pods instead of 16 pods.

simulationUnitTestCase.txt

What did you expect to see?
16 pods eviction

What did you see instead?
26 pods evicted.

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale