Crash during startup in reconciler #4413

aberres · 2023-03-30T12:36:45Z

Report

Let me start with a heads-up: This bug report is best-effort. Keda was in a crash loop, so getting things back up was more important than figuring out what exactly happened. I will try to add all the details I have. I it is not enough, feel free to close.

I cannot tell exactly what triggered it, but after a CI run Keda started to crash loop. Find the relevant part of the log below.
No new scaled objects were introduced. The very same config has been running for quite some time.

Looking at the scaled object, we can see that active and fallback are Unknown instead of True or False.

❯ kubectl get scaledobjects.keda.sh 
NAME                                         SCALETARGETKIND      SCALETARGETNAME                              MIN   MAX   TRIGGERS   AUTHENTICATION   READY   ACTIVE    FALLBACK   AGE
...
master-c96-planningsvc-worker-interactive    apps/v1.Deployment   master-c96-planningsvc-worker-interactive    0     10    cron                        True    Unknown   Unknown    2d2h
...

As obviously something was wrong, I kubectl deleted the scaled object and recreated it. Keda happily started right away.

Expected Behavior

Keda does not crash when "something" (whatever something is) is wrong with a metric.

Actual Behavior

Keda crashes during startup.

Steps to Reproduce the Problem

I have no idea so far. It only happened once.

Logs from KEDA operator

Observed a panic in reconciler: runtime error: invalid memory address or nil pointer dereference	{"controller": "scaledobject", "controllerGroup": "keda.sh", "controllerKind": "ScaledObject", "ScaledObject": {"name":"master-c96-planningsvc-worker-interactive","namespace":"mpp-backend-master-c96"}, "namespace": "mpp-backend-master-c96", "name": "master-c96-planningsvc-worker-interactive", "reconcileID": "23c24efb-96df-482e-acc7-ce3b2dd36986"}

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=0x2f62194]

goroutine 467 [running]:
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Reconcile.func1()
	/workspace/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:119 +0x1fa
panic({0x34bf3a0, 0x63ea940})
	/usr/local/go/src/runtime/panic.go:884 +0x212
github.com/kedacore/keda/v2/pkg/scaling/resolver.ResolveScaleTargetPodSpec({0x434be30, 0xc000fa4510}, {0x4360950, 0xc0008deea0}, {0x3a94540?, 0xc000ec0a00})
	/workspace/pkg/scaling/resolver/scale_resolvers.go:73 +0xd4
github.com/kedacore/keda/v2/pkg/scaling.(*scaleHandler).performGetScalersCache(0xc000805110, {0x434be30, 0xc000fa4510}, {0xc0000551d0, 0x4d}, {0x3a94540, 0xc000ec0a00}, 0xc000e3cf00, {0x0, 0x0}, ...)
	/workspace/pkg/scaling/scale_handler.go:347 +0x6e5
github.com/kedacore/keda/v2/pkg/scaling.(*scaleHandler).GetScalersCache(0xc000ec0600?, {0x434be30, 0xc000fa4510}, {0x3a94540, 0xc000ec0a00})
	/workspace/pkg/scaling/scale_handler.go:273 +0xf6
github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).getScaledObjectMetricSpecs(0xc000f226c0, {0x434be30, 0xc000fa4510}, {{0x4353fc0?, 0xc000fa4540?}, 0xc001260cc0?}, 0xc000ec0600)
	/workspace/controllers/keda/hpa.go:200 +0xda
github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).newHPAForScaledObject(0xc000f226c0, {0x434be30?, 0xc000fa4510?}, {{0x4353fc0?, 0xc000fa4540?}, 0x3a221a0?}, 0xc000ec0600, 0xc0012395f0)
	/workspace/controllers/keda/hpa.go:74 +0x66
github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).updateHPAIfNeeded(0xc000f226c0, {0x434be30, 0xc000fa4510}, {{0x4353fc0?, 0xc000fa4540?}, 0xc000fa4510?}, 0xc000ec0600, 0xc00087ba40, 0xc001580580?)
	/workspace/controllers/keda/hpa.go:152 +0x7b
github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).ensureHPAForScaledObjectExists(0xc000f226c0, {0x434be30, 0xc000fa4510}, {{0x4353fc0?, 0xc000fa4540?}, 0x4353fc0?}, 0xc000ec0600, 0x0?)
	/workspace/controllers/keda/scaledobject_controller.go:431 +0x238
github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).reconcileScaledObject(0xc000f226c0?, {0x434be30, 0xc000fa4510}, {{0x4353fc0?, 0xc000fa4540?}, 0xc0011bae10?}, 0xc000ec0600)
	/workspace/controllers/keda/scaledobject_controller.go:229 +0x1c9
github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).Reconcile(0xc000f226c0, {0x434be30, 0xc000fa4510}, {{{0xc000a7b740?, 0x10?}, {0xc0011bae10?, 0x40da87?}}})
	/workspace/controllers/keda/scaledobject_controller.go:175 +0x526
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Reconcile(0x434be30?, {0x434be30?, 0xc000fa4510?}, {{{0xc000a7b740?, 0x32ae080?}, {0xc0011bae10?, 0x0?}}})
	/workspace/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:122 +0xc8
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).reconcileHandler(0xc00080b720, {0x434bd88, 0xc00109e180}, {0x3619020?, 0xc000e4a4e0?})
	/workspace/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:323 +0x38f
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).processNextWorkItem(0xc00080b720, {0x434bd88, 0xc00109e180})
	/workspace/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:274 +0x1d9
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Start.func2.2()
	/workspace/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:235 +0x85
created by sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Start.func2
	/workspace/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:231 +0x333

KEDA Version

2.10.0

Kubernetes Version

1.25

Platform

None

Scaler Details

cron

Anything else?

No response

The text was updated successfully, but these errors were encountered:

JorTurFer · 2023-03-30T17:21:04Z

Hi,
Definitively KEDA shouldn't panic. This looks like something wrong with the deployment and the scaledobject github.com/kedacore/keda/v2/pkg/scaling/resolver.ResolveScaleTargetPodSpec. Could you share both please? (deployment and scaledobject).
Obviously you can anonymize it replacing all sensitive data, but don't remove any field please because it looks like something related with the spec

erez-levi · 2023-04-25T02:36:15Z

Hi, had the same issue in my environment, I also kubectl deleted the scaled object and recreated it. I'm running keda version 2.10.0, was this issue fixed on 2.10.1? Did I understood correctly? @JorTurFer

JorTurFer · 2023-04-25T06:38:55Z

Hi @erez-levi ,
No, this issue is still active, the OP didn't share the required information, so we couldn't continue with it (and that's why it's still opened)

aberres · 2023-04-25T06:59:51Z

Sorry, forgot about it. I can look into it later.

But just to make sure there are no misunderstandings: The resources have been unchanged for quite some time (I'd say a year at least) and after delete/recreate continuing to work as expected. So I am not sure if there is really an issue with the resources. I'd expect the situation to happen more often.

AndreOrlovE · 2023-04-25T11:07:41Z

Hi there! On behalf of @erez-levi , here are deployment and scaled object configs (some data were removed):

apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-deployment-name
  namespace: myns
spec:
  progressDeadlineSeconds: 600
  replicas: 34
  strategy:
    rollingUpdate:
      maxSurge: 25%
      maxUnavailable: 25%
    type: RollingUpdate
  template:
    spec:
      containers:
        image: gcr.io/image_name:1.0.0
        imagePullPolicy: IfNotPresent
        name: container-name
        ports:
        - containerPort: 8888
          name: http
          protocol: TCP
        resources:
          limits:
            cpu: "6"
            memory: 24Gi
            nvidia.com/gpu: "1"
        volumeMounts:
        - mountPath: /volume1
          name: nfs
          subPath: img
      restartPolicy: Always
      volumes:
      - name: nfs
        persistentVolumeClaim:
          claimName: nfs-server
status:
  availableReplicas: 34
  collisionCount: 1
  conditions:
  observedGeneration: 4250
  readyReplicas: 34
  replicas: 34
  updatedReplicas: 34

apiVersion: keda.sh/v1alpha1
kind: ScaledObject
metadata:
  annotations:
  finalizers:
  - finalizer.keda.sh
  generation: 3
  labels:
    scaledobject.keda.sh/name: my-deployment-name
  name: my-deployment-name
  namespace: myns
spec:
  advanced:
    horizontalPodAutoscalerConfig:
      behavior:
        scaleDown:
          policies:
          - periodSeconds: 120
            type: Pods
            value: 1
          selectPolicy: Max
          stabilizationWindowSeconds: 300
        scaleUp:
          policies:
          - periodSeconds: 120
            type: Pods
            value: 1
          - periodSeconds: 120
            type: Percent
            value: 10
          selectPolicy: Max
          stabilizationWindowSeconds: 300
      name: keda-my-deployment-name
  cooldownPeriod: 0
  maxReplicaCount: 100
  minReplicaCount: 3
  pollingInterval: 30
  scaleTargetRef:
    name: my-deployment-name
  triggers:
  - metadata:
      activationThreshold: "0"
      query: <some_metrics>
      serverAddress: http://prometheus-server:9090
      threshold: "0.01"
    type: prometheus
status:
  conditions:
  - message: ScaledObject is defined correctly and is ready for scaling
    reason: ScaledObjectReady
    status: "True"
    type: Ready
  - message: Scaling is performed because triggers are active
    reason: ScalerActive
    status: "True"
    type: Active
  - message: No fallbacks are active on this scaled object
    reason: NoFallbackFound
    status: "False"
    type: Fallback
  externalMetricNames:
  - s0-prometheus-prometheus
  health:
    s0-prometheus-prometheus:
      numberOfFailures: 0
      status: Happy
  hpaName: keda-my-deployment-name
  lastActiveTime: "2023-04-25T10:56:53Z"
  originalReplicaCount: 15
  scaleTargetGVKR:
    group: apps
    kind: Deployment
    resource: deployments
    version: v1
  scaleTargetKind: apps/v1.Deployment

the errors:

	panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x0 pc=0x2f62194]

goroutine 476 [running]:
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Reconcile.func1()
	/workspace/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:119 +0x1fa
panic({0x34bf3a0, 0x63ea940})
	/usr/local/go/src/runtime/panic.go:884 +0x212
github.com/kedacore/keda/v2/pkg/scaling/resolver.ResolveScaleTargetPodSpec({0x434be30, 0xc001725a40}, {0x4360950, 0xc0013a6d80}, {0x3a94540?, 0xc0013b8000})
	/workspace/pkg/scaling/resolver/scale_resolvers.go:73 +0xd4
github.com/kedacore/keda/v2/pkg/scaling.(*scaleHandler).performGetScalersCache(0xc000533c70, {0x434be30, 0xc001725a40}, {0xc001689ad0, 0x2c}, {0x3a94540, 0xc0013b8000}, 0xc0009baf00, {0x0, 0x0}, ...)
	/workspace/pkg/scaling/scale_handler.go:347 +0x6e5
github.com/kedacore/keda/v2/pkg/scaling.(*scaleHandler).GetScalersCache(0xc00172e200?, {0x434be30, 0xc001725a40}, {0x3a94540, 0xc0013b8000})
	/workspace/pkg/scaling/scale_handler.go:273 +0xf6
github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).getScaledObjectMetricSpecs(0xc000e2bc80, {0x434be30, 0xc001725a40}, {{0x4353fc0?, 0xc001725a70?}, 0x17?}, 0xc00172e200)
	/workspace/controllers/keda/hpa.go:200 +0xda
github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).newHPAForScaledObject(0xc000e2bc80, {0x434be30?, 0xc001725a40?}, {{0x4353fc0?, 0xc001725a70?}, 0x4374d98?}, 0xc00172e200, 0xc001e7b5f0)
	/workspace/controllers/keda/hpa.go:74 +0x66
github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).updateHPAIfNeeded(0xc000e2bc80, {0x434be30, 0xc001725a40}, {{0x4353fc0?, 0xc001725a70?}, 0xc001725a40?}, 0xc00172e200, 0xc00157fa40, 0xc000922760?)
	/workspace/controllers/keda/hpa.go:152 +0x7b
github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).ensureHPAForScaledObjectExists(0xc000e2bc80, {0x434be30, 0xc001725a40}, {{0x4353fc0?, 0xc001725a70?}, 0x4353fc0?}, 0xc00172e200, 0x0?)
	/workspace/controllers/keda/scaledobject_controller.go:431 +0x238
github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).reconcileScaledObject(0xc000e2bc80?, {0x434be30, 0xc001725a40}, {{0x4353fc0?, 0xc001725a70?}, 0xc000922740?}, 0xc00172e200)
	/workspace/controllers/keda/scaledobject_controller.go:229 +0x1c9
github.com/kedacore/keda/v2/controllers/keda.(*ScaledObjectReconciler).Reconcile(0xc000e2bc80, {0x434be30, 0xc001725a40}, {{{0xc0017a1fe0?, 0x10?}, {0xc000922740?, 0x40da87?}}})
	/workspace/controllers/keda/scaledobject_controller.go:175 +0x526
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Reconcile(0x434be30?, {0x434be30?, 0xc001725a40?}, {{{0xc0017a1fe0?, 0x32ae080?}, {0xc000922740?, 0x0?}}})
	/workspace/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:122 +0xc8
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).reconcileHandler(0xc000da6140, {0x434bd88, 0xc000b36840}, {0x3619020?, 0xc0009de360?})
	/workspace/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:323 +0x38f
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).processNextWorkItem(0xc000da6140, {0x434bd88, 0xc000b36840})
	/workspace/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:274 +0x1d9
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Start.func2.2()
	/workspace/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:235 +0x85
created by sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Start.func2
	/workspace/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:231 +0x333

--
2023-04-25T00:21:23Z	INFO	Observed a panic in reconciler: runtime error: invalid memory address or nil pointer dereference	{"controller": "scaledobject", "controllerGroup": "keda.sh", "controllerKind": "ScaledObject", "ScaledObject": {"name":"w-panoptic-segmentation-t4","namespace":"pred"}, "namespace": "pred", "name": "w-panoptic-segmentation-t4", "reconcileID": "81b99eee-c87d-44d2-9e83-5b12deba73f7"}
--
"panic: runtime error: invalid memory address or nil pointer dereference [recovered]"

zroubalik · 2023-05-02T11:48:07Z

Thanks for reporting, this seems like a duplicate of: #4389

djsly · 2023-05-15T18:17:17Z

We just got it again by this yesterday, one random ScaledObejct started crashLooping Keda, rendering ALL scaled operation un-operational

The weird part if that it's never the same ScaledObject, and the affected ScaledObject is in the cluster for a long time.

zroubalik · 2023-06-08T14:43:05Z

Closing this in favor of #4389 please continue there.

aberres added the bug Something isn't working label Mar 30, 2023

JorTurFer mentioned this issue Mar 30, 2023

Release: 2.10.1 #4374

Closed

17 tasks

tobotg mentioned this issue May 29, 2023

SIGSEGV crash in RequestScale #4557

Closed

zroubalik closed this as completed Jun 8, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Crash during startup in reconciler #4413

Crash during startup in reconciler #4413

aberres commented Mar 30, 2023

JorTurFer commented Mar 30, 2023

erez-levi commented Apr 25, 2023

JorTurFer commented Apr 25, 2023 •

edited

Loading

aberres commented Apr 25, 2023

AndreOrlovE commented Apr 25, 2023

zroubalik commented May 2, 2023

djsly commented May 15, 2023

zroubalik commented Jun 8, 2023

Crash during startup in reconciler #4413

Crash during startup in reconciler #4413

Comments

aberres commented Mar 30, 2023

Report

Expected Behavior

Actual Behavior

Steps to Reproduce the Problem

Logs from KEDA operator

KEDA Version

Kubernetes Version

Platform

Scaler Details

Anything else?

JorTurFer commented Mar 30, 2023

erez-levi commented Apr 25, 2023

JorTurFer commented Apr 25, 2023 • edited Loading

aberres commented Apr 25, 2023

AndreOrlovE commented Apr 25, 2023

zroubalik commented May 2, 2023

djsly commented May 15, 2023

zroubalik commented Jun 8, 2023

JorTurFer commented Apr 25, 2023 •

edited

Loading