Issue in Migration of Pods due to PDB

Question

Hello,

I need to migrate our pods from old node groups to new ones, including a Windows unmanaged node group. However, we have a PodDisruptionBudget (PDB) applied that is restricting this process.  We want the migration to be done with zero downtime. Are there recommended steps or best practices to ensure that the migration can proceed smoothly while respecting the PDB constraints?

Any guidance or suggestions would be greatly appreciated!

Thank you.

Oleksii Bebych · Accepted Answer

Before starting the migration, scale up the new node groups to ensure they have enough capacity to accommodate all the pods from the old node groups.

If possible, temporarily relax the PDB constraints to allow more disruptions. This can help in speeding up the migration process.
```
apiVersion: policy/v1
kind: PodDisruptionBudget
metadata:
  name: my-pdb
spec:
  minAvailable: 50%
```

Use a gradual approach to cordon and drain nodes. This minimizes disruptions by respecting the PDB constraints.
```
kubectl cordon 
kubectl drain  --ignore-daemonsets --delete-emptydir-data --force
```

Continuously monitor the status of the pods and ensure they are successfully scheduled on the new nodes. Use kubectl get pods and kubectl describe pod  to check pod status and events.

Issue in Migration of Pods due to PDB

Add your answer

Relevant content