[Bug]: Infinite topic sync in topic operator

[padilo]

Bug Description

When the specs on an unmanaged KafkaTopic resource (annotated with strimzi.io/managed="false") is different from the topic on the cluster, the topic operator enters a loop and continuously tries updates the resourceVersion.

Seems like it's trying to consolidate the state, but as is unmanaged it can't and continues in the loop forever.

Steps to reproduce

You can do the following script:

helm install strimzi-cluster-operator --set replicas=1 --version 0.42.0 oci://quay.io/strimzi-helm/strimzi-kafka-operator

# Create Kafka cluster:
kubectl apply -f - <<EOF 
apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: my-kafka
spec:
  kafka:
    replicas: 1
    listeners:
      - name: plain
        port: 9092
        type: internal
        tls: false
    storage:
      type: persistent-claim
      size: 1Gi
  zookeeper:
    replicas: 1
    storage:
      type: persistent-claim
      size: 1Gi

  entityOperator:
    topicOperator: {}
EOF
kubectl wait --for=condition=ready --timeout=10m kafka/my-kafka

# Create a topic
kubectl apply -f - <<EOF
apiVersion: kafka.strimzi.io/v1beta2
kind: KafkaTopic
metadata:
  name: sync-test
  labels:
    strimzi.io/cluster: my-kafka
spec:
  replicas: 1
  partitions: 1
  config:
    retention.ms: 432000000
EOF

sleep 5

# Annotate topic as unmanaged
kubectl annotate kafkatopic sync-test strimzi.io/managed="false"

# Path the topic with updated retention
kubectl patch kafkatopic sync-test --type='merge' -p '{"spec": {"config": {"retention.ms": "432000001"}} }'

Expected behavior

Topic operator should entirely ignore the KafkaTopic resources if it is set to unmanaged

Strimzi version

0.42.0

Kubernetes version

Kubernetes 1.30.0

Installation method

helm chart

Infrastructure

Kind, also confirmed with EKS

Configuration files and logs

all relevant stuff is already in the steps to reproduce

Additional context

I would like to work on this issue :)

[ppatierno]

@fvaleri FYI this is coming from the related Slack thread.

[fvaleri]

@padilo I can confirm the issue, nice catch and report.

I would like to work on this issue :)

Please, go on. I will be happy to review.

[scholzj]

Triaged on the Community call on 8.8.2024: This is a valid issue and should be fixed.

[scholzj]

@padilo Do you still plan to open a PR for this? It would be great to have this fixed. Thanks.

[padilo]

I will do it this weekend El jue, 8 ago 2024, 18:33, Jakub Scholz ***@***.***> escribió:

…

@padilo <https://github.com/padilo> Do you still plan to open a PR for this? It would be great to have this fixed. Thanks. — Reply to this email directly, view it on GitHub <#10380 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAF7MVZI27W4U3ZCMHKJJK3ZQOMUVAVCNFSM6AAAAABLMDGKSSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDENZWGIZDSMZRHE> . You are receiving this because you were mentioned.Message ID: ***@***.***>