Description
Kubernetes Version | 1.28 EKS |
Couchbase Server | 7.2.5 |
Operator | 2.7.0-228 |
Issue Description
Some CB Pods are not being scheduled in spite having enough K8S nodes available.
Cluster Setup
- Each k8s node is an m5.large instance.
- 20 nodes in EKS cluster.
- All k8s nodes are labelled couchbase-node="true"
- 3 AZs: us-east-2a, us-east-2b, us-east-2c
Process followed
- Initially created a Couchbase cluster having 10 pods on EKS cluster (2 server groups: 6 pods with data service and 4 pods with index-query service)
- Next, added two more server groups: 4 pods with eventing, 4 pods with analytics via YAML.
- Now, 2 pods of analytics server groups are not being scheduled. Total 18 cb pods.
Initially the analytics pods cb-example-0012 and cb-example-0013 are not being scheduled.
Operator retries by bringing up new pods cb-example-0018 and cb-example-0019 but again it is not able to schedule.
I have attached the operator logs and the YAMLs as well.