panic (NPE) when applying sink

[zzzz465]

Hi, I encountered panic error on deploying sink job to cluster.
from yaml, looks like selector is not configured correctly.

applied YAML

apiVersion: compute.functionmesh.io/v1alpha1
kind: Sink
metadata:
  annotations:
    meta.helm.sh/release-name: test-pulsar-resources
    meta.helm.sh/release-namespace: test-pulsar
  creationTimestamp: 2023-07-10T07:12:43Z
  generation: 3
  labels:
    app.kubernetes.io/managed-by: Helm
  name: system-container-to-es-logs
  namespace: test-pulsar
  resourceVersion: "369822483"
  uid: 681e825d-d7c2-4b4e-bab6-ffed33c8af68
spec:
  autoAck: true
  className: org.apache.pulsar.io.elasticsearch.ElasticSearchSink
  clusterName: "<cluster-name-here>"
  image: streamnative/pulsar-io-elastic-search:3.0.0.4
  input:
    topics:
      - persistent://<topic-name-here>
  java:
    jar: connectors/pulsar-io-elastic-search-3.0.0.4.nar
    javaOpts:
      - -Xms=64M
      - -Xmx=256M
  maxReplicas: 4
  minReplicas: 1
  processingGuarantee: atleast_once
  pulsar:
    authConfig:
      genericAuth:
        clientAuthenticationParameters: file:///var/run/secrets/kubernetes.io/serviceaccount/token
        clientAuthenticationPlugin: org.apache.pulsar.client.impl.auth.AuthenticationToken
    pulsarConfig: pulsar-config
    tlsConfig:
      enabled: false
  sinkConfig:
    bulkEnabled: true
    compatibilityMode: ELASTICSEARCH_7
    compressionEnabled: true
    elasticSearchUrl: <es-url-here>
    indexName: "" # use topic name as index name
  subscriptionPosition: earliest
status:
  conditions:
    HorizontalPodAutoscaler:
      action: NoAction
      condition: HPAReady
      status: "True"
    Service:
      action: NoAction
      condition: ServiceReady
      status: "True"
    StatefulSet:
      action: Create
      condition: StatefulSetReady
      status: "False"
  observedGeneration: 2
  replicas: 0
  selector: ""

panic log

panic: runtime error: invalid memory address or nil pointer dereference [recovered]
        panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x0 pc=0x12f2315]

goroutine 551 [running]:
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Reconcile.func1()
        sigs.k8s.io/controller-runtime@v0.12.3/pkg/internal/controller/controller.go:118 +0x1f4
panic({0x165ad00, 0x25c0b00})
        runtime/panic.go:884 +0x213
github.com/streamnative/function-mesh/controllers/spec.CheckIfStatefulSetSpecIsEqual(0xc000698d80?, 0xffffffffffffffff?)
        github.com/streamnative/function-mesh/controllers/spec/common.go:1796 +0x35
github.com/streamnative/function-mesh/controllers.(*SinkReconciler).checkIfStatefulSetNeedUpdate(0xc0000d75a0?, 0xc0007e8f00, 0xc000a750b0?)
        github.com/streamnative/function-mesh/controllers/sink.go:333 +0x45
github.com/streamnative/function-mesh/controllers.(*SinkReconciler).ObserveSinkStatefulSet(0xc00038cd20, {0x1a835a0, 0xc000a750b0}, 0xc000698d80)
        github.com/streamnative/function-mesh/controllers/sink.go:71 +0x41b
github.com/streamnative/function-mesh/controllers.(*SinkReconciler).Reconcile(0xc00038cd20, {0x1a835a0, 0xc000a750b0}, {{{0xc000523605, 0xb}, {0xc00004bc00, 0x1b}}})
        github.com/streamnative/function-mesh/controllers/sink_controller.go:86 +0x39f
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Reconcile(0x1a834f8?, {0x1a835a0?, 0xc000a750b0?}, {{{0xc000523605?, 0x17753a0?}, {0xc00004bc00?, 0xc0006a4cf8?}}})
        sigs.k8s.io/controller-runtime@v0.12.3/pkg/internal/controller/controller.go:121 +0xc8
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).reconcileHandler(0xc00022a280, {0x1a834f8, 0xc000401ea0}, {0x16b2ae0?, 0xc000082de0?})
        sigs.k8s.io/controller-runtime@v0.12.3/pkg/internal/controller/controller.go:320 +0x309
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).processNextWorkItem(0xc00022a280, {0x1a834f8, 0xc000401ea0})
        sigs.k8s.io/controller-runtime@v0.12.3/pkg/internal/controller/controller.go:273 +0x1d9
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Start.func2.2()
        sigs.k8s.io/controller-runtime@v0.12.3/pkg/internal/controller/controller.go:234 +0x85
created by sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Start.func2
        sigs.k8s.io/controller-runtime@v0.12.3/pkg/internal/controller/controller.go:230 +0x587

[zzzz465]

FYI: I'm using streamnative/function-mesh:v0.14.0 docker image

[jiangpengcheng]

hm, I can't reproduce it with the same yaml, I got a different error:

The SinkWebhook "system-container-to-es-logs" is invalid: spec.resources: Invalid value: v1.ResourceRequirements{Limits:v1.ResourceList(nil), Requests:v1.ResourceList(nil)}: resource requirement is invalid

after adding the required field it works, how did you deploy the function-mesh controller?

[zzzz465]

I deployed function mesh controller using the released helm chart.
https://github.com/streamnative/function-mesh/releases/tag/function-mesh-operator-0.2.17

also I disabled admissionWebhook for cluster management. I'll try installing it with webhook

[zzzz465]

I enabled admissionWebhook and the controller still panics. however, I've found that the statefulset log shows the JVM process started with -Xmx=0 and keeps failing, so I added resources.requests. the controller still raise panic but when I deleted the statefulset, the pod is now working fine and the statefulset is recreated.

I'm not sure why the controller panic on my environment unlike your comment. but I think the missing resources field is the problem

[zzzz465]

I was able to launch sink.

attempted steps to solve the issue

1. pulsar-config configmap doesn't exists so I added one
2. the configmap requires brokerURL, etc.
3. there's a few deprecated configs. eg: input