Redpanda Fails to Start After Reducing CPU Cores

Last updated: August 10, 2026

Overview

When running Redpanda, you may encounter a situation where Redpanda fail to start after a update that reduces the number of CPU cores. This typically results in pod status CrashLoopBackOff and errors such as "PostStartHook failed" and "PostStopHook failed".

General Issue

After reducing the number of CPU cores in your Redpanda Helm chart (for example, from 4 to 2), Redpanda pods may fail to start. Kubernetes events and logs will show errors like:

  • PostStartHook failed
  • PostStopHook failed
  • Pod status: CrashLoopBackOff

On bare metal Redpanda logs may include:

Detected decrease in number of cores dedicated to run from 4 to 2, but it is impossible because node_local_core_assignment feature flag is not yet active.

Root Cause

Redpanda does not support reducing the number of CPU cores unless on version 24.3.x or above. 

Resolution

  1. Restore the previous CPU core count: set the number of cores back to the original value.
  2. Restart the Redpanda pods after updating the core count.
  3. Upgrade to at least to at least 24.3.x, but preferably the newest Redpanda version.
  4. Reduce core count to your desire amount.

Additional Notes

  • This issue is specific to self-hosted Redpanda and does not apply to Redpanda Cloud.
  • Always test configuration changes in a non-production environment before applying them to production.

If you continue to experience issues, please contact Redpanda Support.