Waiting for at-least one connect pod availability

Kafka Connect Self-Managed Connectors
1

Following Manage Connectors for Confluent Platform Using Confluent for Kubernetes | Confluent Documentation

" waiting for at-least one connect pod availability"

It looks like there is a connect pod that is available.

Edit: the Connect CR in the example does not have “replicas”, and the default replicas is 0.

$ kubectl  describe -f myconnector.yaml 
Name:         test-couchbase-source123123
Namespace:    confluent
Labels:       <none>
Annotations:  <none>
API Version:  platform.confluent.io/v1beta1
Kind:         Connector
Metadata:
  Creation Timestamp:  2025-03-13T22:00:58Z
  Finalizers:
    connector.finalizers.platform.confluent.io
  Generation:        1
  Resource Version:  85303
  UID:               2cb52869-e6d7-4ac0-bfa2-6fb352a8df72
Spec:
  Class:  com.couchbase.connect.kafka.CouchbaseSourceConnector
  Configs:
    couchbase.bootstrap.timeout":            10s
    couchbase.bucket:                        travel-sample
    couchbase.compression:                   ENABLED
    couchbase.event.filter:                  com.couchbase.connect.kafka.filter.AllPassFilter
    couchbase.flow.control.buffer:           16m
    couchbase.log.document.lifecycle:        false
    couchbase.metrics.interval:              10m
    couchbase.password:                      C0uchbase123!
    couchbase.persistence.polling.interval:  0
    couchbase.seed.nodes:                    couchbases://cb.grixsip3rnsvxcl0.cloud.couchbase.com
    couchbase.source.handler:                com.couchbase.connect.kafka.example.MultiRawJsonSourceHandler
    couchbase.stream.from:                   SAVED_OFFSET_OR_BEGINNING
    couchbase.topic:                         test-default
    couchbase.username:                      clientuser
  Connect Cluster Ref:
    Name:  kafka-connect-couchbase
  Name:    test-couchbase-source123123
  Restart Policy:
    Max Retry:  1
    Type:       Never
  Task Max:     2
Status:
  App State:  Unknown
  Conditions:
    Last Probe Time:       2025-03-13T22:00:58Z
    Last Transition Time:  2025-03-13T22:00:58Z
    Message:               Application state is unknown
    Reason:                AppStateUnknown
    Status:                Unknown
    Type:                  platform.confluent.io/app-ready
  State:                   ERROR
Events:
  Type     Reason   Age                   From       Message
  ----     ------   ----                  ----       -------
  Warning  Warning  97s (x26 over 7m52s)  connector  waiting for at-least one connect pod availability

admin@ip-172-31-3-82:~$ kubectl get  pod connect-0
NAME        READY   STATUS    RESTARTS      AGE
connect-0   1/1     Running   4 (20h ago)   3d3h

It’s not clear what spec.connectClusterRef.name should be. Where do I find that? Or if this is defining it, what refers to it?

apiVersion: platform.confluent.io/v1beta1
kind: Connect
metadata:
  name: kafka-connect-couchbase
  namespace: confluent
  annotations:
spec:
  image:
    application: confluentinc/cp-server-connect:7.9.0
    init: confluentinc/confluent-init-container:2.11.0
  keyConverterType: org.apache.kafka.connect.storage.StringConverter
  valueConverterType: org.apache.kafka.connect.converters.ByteArrayConverter
  replicas: 1 <------------------ need to specify at least 1
  build:
    type: onDemand                                   
    onDemand:                                        
      plugins:
        url: 
        - name: Couchbase_Kafka_Connector_4.2.7
          archivePath: https://packages.couchbase.com/clients/kafka/4.2.7/couchbase-kafka-connect-couchbase-4.2.7.zip
          checksum:   b2aff22e3020ed8115bf7e259559c90e
      storageLimit: 4G

apiVersion: platform.confluent.io/v1beta1
metadata:
  name: test-couchbase-source123123
  namespace: confluent
kind: Connector
spec:
  name: test-couchbase-source123123
  taskMax: 2
  class: com.couchbase.connect.kafka.CouchbaseSourceConnector
  configs:
    couchbase.topic: test-default
    couchbase.seed.nodes: couchbases://cb.grixsip3rnsvxcl0.cloud.couchbase.com
    couchbase.username: clientuser
    couchbase.password: C0uchbase123!
    couchbase.bucket: travel-sample
    couchbase.source.handler: com.couchbase.connect.kafka.handler.source.RawJsonSourceHandler
    couchbase.event.filter: com.couchbase.connect.kafka.filter.AllPassFilter
    couchbase.stream.from: SAVED_OFFSET_OR_BEGINNING
    couchbase.compression: ENABLED
    couchbase.flow.control.buffer: 16m
    couchbase.persistence.polling.interval: "0"
    couchbase.log.document.lifecycle: "false"
    couchbase.metrics.interval: 10m
    couchbase.bootstrap.timeout": 10s
  restartPolicy:
    type: Never
    maxRetry: 1
  connectClusterRef:
    name: kafka-connect-couchbase # from connector.yaml
2

you need to specifiy the following in your myconnector.yaml

something like
  connectClusterRef:
    name: kafka-connect-couchbase

for ref:

hth,
michael

3

I did eventually get it sorted out. But even after aligning the cluster name - there was still an issue because the Connect CR provided in the instructions does not have “replicas” - and replicas defaults to 0.

4

which instructions are you referring to?
the docs.confluent.io one?
or the one in the github repo?

5 
kind: Connect
metadata:
  annotations:
  platform.confluent.io/connector-wget-extra-args:   --- [1]
spec:
  build:
    type: onDemand                                   --- [2]
    onDemand:                                        --- [3]
      plugins:
        url:                                         --- [4]
        - name:                                      --- [5]
          archivePath:                               --- [6]
          checksum:                                  --- [7]
      storageLimit:                                  --- [8]

At least that’s what I thought the issue is. Then again the event “waiting for at-least one connect pod availability” remains as the last event even after the connector starts working. So it may have been the fix to the cluster-name that got it working. Although I though I removed replicas: 1 and it stopped working.