[BUG]CacheRuntime controller panics with nil pointer dereference when `client` component is not defined in CacheRuntimeClass topology

[btxu-db]

What is your environment(Kubernetes version, Fluid version, etc.)
k8s version: 1.36.1-aliyun.1
fluid version: 1.1.0-alpha.11
Describe the bug
When the client component is not defined in the topology of a CacheRuntimeClass, generateRuntimeConfigData panics with a nil pointer dereference, causing the cacheruntime-controller pod to crash and restart.
The error message is as follows:

kubectl logs -n fluid-system cacheruntime-controller-868f9494b4-46xp8 --previous
2026/08/03 17:18:24 nodeExcludeSelector is empty, no node would be excluded when syncing schedule info of runtimes
2026/08/03 17:18:24 BuildDate: 2026-07-12_09:41:07
2026/08/03 17:18:24 GitCommit: 676f47a841133ba9ffcaafe9c76b44b2cb6b11a9
2026/08/03 17:18:24 GitTreeState: dirty
2026/08/03 17:18:24 GoVersion: go1.25.12
2026/08/03 17:18:24 Compiler: gc
2026/08/03 17:18:24 Platform: linux/amd64
2026-08-03T17:18:24.423+0800    INFO    setup   utils/pprof.go:33       Enabling pprof  {"pprof address": ":6060"}
2026-08-03T17:18:24.423+0800    INFO    setup   utils/pprof.go:48       Starting pprof HTTP server      {"pprof server address": ":6060"}
2026-08-03T17:18:24.438+0800    INFO    setup   app/cache.go:143        AdvancedStatefulSet controller added successfully
2026-08-03T17:18:24.438+0800    INFO    dump    dump/dump.go:61 Register goroutine dump generator
2026-08-03T17:18:24.438+0800    INFO    setup   app/cache.go:171        starting cacheruntime-controller
2026-08-03T17:18:24.439+0800    INFO    controller-runtime.metrics      server/server.go:205    Starting metrics server
2026-08-03T17:18:24.439+0800    INFO    controller-runtime.metrics      server/server.go:244    Serving metrics server  {"bindAddress": ":8080", "secure": false}
I0803 17:18:24.539967       1 leaderelection.go:250] attempting to acquire leader lease fluid-system/cache.data.fluid.io...
I0803 17:18:42.443670       1 leaderelection.go:260] successfully acquired lease fluid-system/cache.data.fluid.io
2026-08-03T17:18:42.444+0800    INFO    controller/controller.go:178    Starting EventSource    {"controller": "advancedstatefulset-controller", "source": "kind source: *v1alpha1.AdvancedStatefulSet"}
2026-08-03T17:18:42.444+0800    INFO    controller/controller.go:178    Starting EventSource    {"controller": "advancedstatefulset-controller", "source": "kind source: *v1.PersistentVolumeClaim"}
2026-08-03T17:18:42.444+0800    INFO    controller/controller.go:178    Starting EventSource    {"controller": "advancedstatefulset-controller", "source": "kind source: *v1.Pod"}
2026-08-03T17:18:42.444+0800    INFO    controller/controller.go:186    Starting Controller     {"controller": "advancedstatefulset-controller"}
2026-08-03T17:18:42.444+0800    INFO    controller/controller.go:178    Starting EventSource    {"controller": "CacheRuntimeController", "source": "kind source: *v1alpha1.CacheRuntime"}
2026-08-03T17:18:42.444+0800    INFO    controller/controller.go:178    Starting EventSource    {"controller": "CacheRuntimeController", "source": "kind source: *v1alpha1.AdvancedStatefulSet"}
2026-08-03T17:18:42.444+0800    INFO    controller/controller.go:178    Starting EventSource    {"controller": "CacheRuntimeController", "source": "kind source: *v1.DaemonSet"}
2026-08-03T17:18:42.444+0800    INFO    controller/controller.go:186    Starting Controller     {"controller": "CacheRuntimeController"}
2026-08-03T17:18:42.546+0800    INFO    controller/controller.go:220    Starting workers        {"controller": "advancedstatefulset-controller", "worker count": 3}
I0803 17:18:42.546842       1 statefulset_controller.go:243] "AdvancedStatefulSet deleted" statefulSet="default/mooncake-demo-master"
I0803 17:18:42.546862       1 statefulset_controller.go:243] "AdvancedStatefulSet deleted" statefulSet="default/mooncake-demo-worker"
I0803 17:18:42.546903       1 statefulset_controller.go:233] "Finished syncing AdvancedStatefulSet" statefulSet="default/mooncake-demo-worker" elapsedTime="85.875µs"
I0803 17:18:42.546869       1 statefulset_controller.go:233] "Finished syncing AdvancedStatefulSet" statefulSet="default/mooncake-demo-master" elapsedTime="53.418µs"
2026-08-03T17:18:42.547+0800    INFO    controller/controller.go:220    Starting workers        {"controller": "CacheRuntimeController", "worker count": 3}
2026-08-03T17:18:42.547+0800    INFO    cachectl.CacheRuntime   engine/engine.go:100    Set the syncRetryDuration       {"cacheRuntime": {"name":"mooncake-demo","namespace":"default"}, "cache engine": "cache", "id": "default-mooncake-demo", "syncRetryDuration": 15}
2026-08-03T17:18:42.749+0800    INFO    cachectl.CacheRuntime   engine/util.go:56       mountRoot       {"cacheRuntime": {"name":"mooncake-demo","namespace":"default"}, "cache engine": "cache", "id": "default-mooncake-demo", "path": "/runtime-mnt/cache"}
2026-08-03T17:18:42.749+0800    INFO    cachectl.CacheRuntime   engine/util.go:56       mountRoot       {"cacheRuntime": {"name":"mooncake-demo","namespace":"default"}, "cache engine": "cache", "id": "default-mooncake-demo", "path": "/runtime-mnt/cache"}
2026-08-03T17:18:42.850+0800    INFO    cachectl.CacheRuntime   engine/util.go:56       mountRoot       {"cacheRuntime": {"name":"mooncake-demo","namespace":"default"}, "cache engine": "cache", "id": "default-mooncake-demo", "path": "/runtime-mnt/cache"}
2026-08-03T17:18:42.850+0800    INFO    controller/controller.go:115    Observed a panic in reconciler: runtime error: invalid memory address or nil pointer dereference        {"controller": "CacheRuntimeController", "object": {"name":"mooncake-demo","namespace":"default"}, "namespace": "default", "name": "mooncake-demo", "reconcileID": "e3fc461b-2938-4b8a-ba9d-cc640297d2fd"}
panic: runtime error: invalid memory address or nil pointer dereference [recovered, repanicked]
[signal SIGSEGV: segmentation violation code=0x1 addr=0x0 pc=0x22bc421]

goroutine 211 [running]:
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Reconcile.func1()
        /go/src/github.com/fluid-cloudnative/fluid/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:116 +0x2cd
panic({0x285a600?, 0x3be7750?})
        /usr/local/go/src/runtime/panic.go:783 +0x136
github.com/fluid-cloudnative/fluid/pkg/ddc/cache/engine.(*CacheEngine).generateRuntimeConfigData(0xc0009f0900, {0x2d6a220, 0xc0009f0cc0}, 0xc000adcc08)
        /go/src/github.com/fluid-cloudnative/fluid/pkg/ddc/cache/engine/cm.go:178 +0xf81
github.com/fluid-cloudnative/fluid/pkg/ddc/cache/engine.(*CacheEngine).createRuntimeValueConfigMap(0xc0009f0900, {0x2d6a220, 0xc0009f0cc0}, 0xc000adcc08, {0xc000493ea0, 0x1, 0x1})
        /go/src/github.com/fluid-cloudnative/fluid/pkg/ddc/cache/engine/cm.go:93 +0x20b
github.com/fluid-cloudnative/fluid/pkg/ddc/cache/engine.(*CacheEngine).createRuntimeConfigMaps(0xc0009f0900, {0x2d6a220, 0xc0009f0cc0}, 0xc0004ebce0)
        /go/src/github.com/fluid-cloudnative/fluid/pkg/ddc/cache/engine/cm.go:57 +0x34f
github.com/fluid-cloudnative/fluid/pkg/ddc/cache/engine.(*CacheEngine).Setup(0xc0009f0900, {{0x2d6a5d8, 0xc0001776c0}, {{0xc0009c0226, 0x7}, {0xc0009c0230, 0xd}}, {0x2b418b2, 0xa}, 0xc000942008, ...})
        /go/src/github.com/fluid-cloudnative/fluid/pkg/ddc/cache/engine/setup.go:50 +0x405
github.com/fluid-cloudnative/fluid/pkg/controllers.(*RuntimeReconciler).ReconcileRuntime(0xc000360370, {0x2d6fe88, 0xc0009f0900}, {{0x2d6a5d8, 0xc0001776c0}, {{0xc0009c0226, 0x7}, {0xc0009c0230, 0xd}}, {0x2b418b2, ...}, ...})
        /go/src/github.com/fluid-cloudnative/fluid/pkg/controllers/runtime_controller.go:272 +0x57c
github.com/fluid-cloudnative/fluid/pkg/controllers.(*RuntimeReconciler).ReconcileInternal(0xc000360370, {{0x2d6a5d8, 0xc0001776c0}, {{0xc0009c0226, 0x7}, {0xc0009c0230, 0xd}}, {0x2b418b2, 0xa}, 0xc000942008, ...})
        /go/src/github.com/fluid-cloudnative/fluid/pkg/controllers/runtime_controller.go:181 +0x13bf
github.com/fluid-cloudnative/fluid/pkg/controllers/v1alpha1/cacheruntime.(*CacheRuntimeReconciler).Reconcile(0xc00032ede0, {0x2d6a530, 0xc000a88b40}, {{{0xc0009c0226, 0x7}, {0xc0009c0230, 0xd}}})
        /go/src/github.com/fluid-cloudnative/fluid/pkg/controllers/v1alpha1/cacheruntime/cacheruntime_controller.go:120 +0xb58
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Reconcile(0xc0002d0c80, {0x2d6a530, 0xc000a88b40}, {{{0xc0009c0226, 0x7}, {0xc0009c0230, 0xd}}})
        /go/src/github.com/fluid-cloudnative/fluid/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:119 +0x15b
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).reconcileHandler(0xc0002d0c80, {0x2d6a530, 0xc000a88b40}, {0x294e800, 0xc0004b41c0})
        /go/src/github.com/fluid-cloudnative/fluid/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:316 +0x44d
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).processNextWorkItem(0xc0002d0c80, {0x2d6a568, 0xc000286230})
        /go/src/github.com/fluid-cloudnative/fluid/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:266 +0x359
sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Start.func2.2()
        /go/src/github.com/fluid-cloudnative/fluid/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:227 +0xac
created by sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Start.func2 in goroutine 120
        /go/src/github.com/fluid-cloudnative/fluid/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:223 +0x927

What you expect to happen:
I expect that when the client component is not defined in a CacheRuntimeClass, the controller should either treat the client component as disabled and continue reconciling normally, or reject the resource with a clear validation error — instead of crashing with a panic.
**How to reproduce **

1.Create a CacheRuntimeClass that defines only master and worker components in topology — the client component is intentionally omitted. Then create a Dataset and CacheRuntime.
Note: whether the CacheRuntime spec contains a client field does not affect this bug — the panic occurs either way. The trigger is the missing client component in the CacheRuntimeClass topology:
dataset-runtime.yaml:

apiVersion: data.fluid.io/v1alpha1
kind: Dataset
metadata:
  name: mooncake-demo
  namespace: default
spec:
  placement: Shared
  accessModes:
    - ReadWriteMany
  mounts:
    - name: mc
      mountPoint: "mooncakefs:///"
---
apiVersion: data.fluid.io/v1alpha1
kind: CacheRuntime
metadata:
  name: mooncake-demo
  namespace: default
spec:
  runtimeClassName: mooncake-demo
  master:
    replicas: 1
  worker:
    replicas: 2
    tieredStore:
      levels:
        - emptyDir:
            quota: 1Gi
          high: "0.8"
          low: "0.5"

cacheruntimeclass.yaml:

apiVersion: data.fluid.io/v1alpha1
kind: CacheRuntimeClass
metadata:
  name: mooncake-demo
fileSystemType: mooncakefs
topology:
  master:
    service:
      headless: {}
    template:
      spec:
        restartPolicy: Always
        containers:
          - name: master
            image: crpi-4hkqof7tc9brc6d5.cn-hongkong.personal.cr.aliyuncs.com/mooncake1314/mooncake:v2
            command:
              - /custom-entrypoint.sh
            args:
              - master
              - start
            imagePullPolicy: IfNotPresent
            readinessProbe:
              tcpSocket:
                port: 50051
              initialDelaySeconds: 10
              periodSeconds: 5
              failureThreshold: 12
            env:
              - name: POD_NAME
                valueFrom:
                  fieldRef:
                    fieldPath: metadata.name
            ports:
              - containerPort: 50051
                name: rpc
              - containerPort: 8080
                name: metadata
              - containerPort: 9003
                name: metrics
  worker:
    service:
      headless: {}
    template:
      spec:
        restartPolicy: Always
        containers:
          - name: worker
            image: crpi-4hkqof7tc9brc6d5.cn-hongkong.personal.cr.aliyuncs.com/mooncake1314/mooncake:v2
            command:
              - /custom-entrypoint.sh
            args:
              - worker
              - start
            imagePullPolicy: IfNotPresent
            readinessProbe:
              tcpSocket:
                port: 50052
              initialDelaySeconds: 5
              periodSeconds: 5
              failureThreshold: 12
            env:
              - name: POD_NAME
                valueFrom:
                  fieldRef:
                    fieldPath: metadata.name
            ports:
              - containerPort: 50052
                name: data
              - containerPort: 9300
                name: http   

2. Apply it:

kubectl apply -f cacheruntimeclass.yaml
kubectl apply -f dataset-runtime.yaml

3. Observe the cacheruntime-controller pod:

kubectl get pods -n fluid-system -w

The controller pod crashes and restarts:

NAME                                       READY   STATUS             RESTARTS      AGE
cacheruntime-controller-868f9494b4-46xp8   1/1     Running            1 (2s ago)    6m10s
cacheruntime-controller-868f9494b4-46xp8   0/1     Error              1 (20s ago)   6m28s
cacheruntime-controller-868f9494b4-46xp8   0/1     CrashLoopBackOff   1 (12s ago)   6m39s
cacheruntime-controller-868f9494b4-46xp8   1/1     Running            2 (12s ago)   6m39s

4. Check the panic stack trace in the previous container's logs:

kubectl logs -n fluid-system <cacheruntime-controller-pod-name> --previous

The panic stack trace shown at the top of this issue can be found there.
Additional Information
Note: the container image and entrypoint script are irrelevant to this bug — the controller panics during ConfigMap generation, before any pod is scheduled. Any placeholder image works for reproduction.
Defining a client component in the CacheRuntimeClass topology avoids the panic:

  client:
    template:
      spec:
        restartPolicy: Always
        containers:
          - name: client
            image: <your-image>
            command:
              - /custom-entrypoint.sh
            args:
              - client
              - start
            imagePullPolicy: IfNotPresent
            securityContext:
              privileged: true
              runAsUser: 0

[xliuqq]

@btxu-db can you fix the issue and add an e2e test for mooncake cache system?

[btxu-db]

/assign
@xliuqq /assign Sure, I'll take both.

We discussed the direction offline. The plan is to make the controller tolerate a missing client component rather than reject the CacheRuntimeClass, since cache systems with a KV interface (Mooncake, and others in that family) have no POSIX mount semantics and therefore no client component to declare.

One thing worth noting: simply not panicking isn't sufficient. Several downstream behaviors may still assume the fuse client exists — PV/PVC creation, Dataset Bound evaluation, webhook injection into application pods, and mount-path-dependent operations such as DataLoad. I'll audit these against a build with the nil fix applied and post the findings here, so we can decide which need to be explicitly rejected and which belong in a separate design discussion.

The e2e test for the Mooncake cache system will follow once the scope is settled.

[btxu-db]

The fix in cm.go has been verified and passed on the real cluster.

NAME                                       READY   STATUS    RESTARTS      AGE
cacheruntime-controller-685c444589-twtmh   1/1     Running   0             13s
csi-nodeplugin-fluid-f2jhh                 2/2     Running   2 (72m ago)   6h46m
csi-nodeplugin-fluid-lvbhl                 2/2     Running   2 (72m ago)   6h46m
csi-nodeplugin-fluid-pvgmx                 2/2     Running   2 (72m ago)   6h46m
dataset-controller-56855b9fcd-tddv4        1/1     Running   1 (72m ago)   6h46m
fluid-webhook-5fc666f47b-hr85n             1/1     Running   2 (11m ago)   6h46m
fluidapp-controller-669d6b888d-gg8pg       1/1     Running   2 (11m ago)   6h46m
No resources found in default namespace.
No resources found in default namespace.
NAMESPACE      NAME                   DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR            AGE
fluid-system   csi-nodeplugin-fluid   3         3         3       3            3           <none>                   6h46m
kube-system    kindnet                3         3         3       3            3           kubernetes.io/os=linux   6h51m
kube-system    kube-proxy             3         3         3       3            3           kubernetes.io/os=linux   6h51m

Even without the client component, Fluid still created the PVC and PV, and their status is Bound.

NAMESPACE   NAME                                  STATUS   VOLUME                  CAPACITY   ACCESS MODES   STORAGECLASS   VOLUMEATTRIBUTESCLASS   AGE
default     persistentvolumeclaim/mooncake-demo   Bound    default-mooncake-demo   100Pi      RWX            fluid          <unset>                 69s
NAMESPACE   NAME                                     CAPACITY   ACCESS MODES   RECLAIM POLICY   STATUS   CLAIM                   STORAGECLASS   VOLUMEATTRIBUTESCLASS   REASON   AGE
            persistentvolume/default-mooncake-demo   100Pi      RWX            Retain           Bound    default/mooncake-demo   fluid          <unset>                          69s
NAMESPACE      NAME                   DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR            AGE
fluid-system   csi-nodeplugin-fluid   3         3         3       3            3           <none>                   6h56m
kube-system    kindnet                3         3         3       3            3           kubernetes.io/os=linux   7h
kube-system    kube-proxy             3         3         3       3            3           kubernetes.io/os=linux   7h
NAME                         READY   STATUS    RESTARTS   AGE   IP           NODE               NOMINATED NODE   READINESS GATES
pod/mooncake-demo-master-0   1/1     Running   0          26s   10.244.1.7   mooncake-worker2   <none>           <none>
pod/mooncake-demo-worker-0   1/1     Running   0          26s   10.244.2.7   mooncake-worker    <none>           <none>
pod/mooncake-demo-worker-1   1/1     Running   0          26s   10.244.1.8   mooncake-worker2   <none>           <none>
NAME                               TYPE        CLUSTER-IP   EXTERNAL-IP   PORT(S)   AGE     SELECTOR
service/kubernetes                 ClusterIP   10.96.0.1    <none>        443/TCP   7h      <none>
service/svc-mooncake-demo-master   ClusterIP   None         <none>        <none>    7m25s   cacheruntime.fluid.io/component-name=mooncake-demo-master,cacheruntime.fluid.io/name=mooncake-demo
service/svc-mooncake-demo-worker   ClusterIP   None         <none>        <none>    7m25s   cacheruntime.fluid.io/component-name=mooncake-demo-worker,cacheruntime.fluid.io/name=mooncake-demo
NAME                                           DATA   AGE
configmap/fluid-runtime-config-mooncake-demo   1      7m25s
configmap/kube-root-ca.crt                     1      7h
NAME                                  UFS TOTAL SIZE   CACHED   CACHE CAPACITY   CACHED PERCENTAGE   PHASE    HCFS URL   TOTAL FILES   CACHE HIT RATIO   AGE
dataset.data.fluid.io/mooncake-demo   2.00GiB          0B       2.00GiB          0.0                 Failed              0             0                 7m25s
NAME                                       AGE     STATUS
cacheruntime.data.fluid.io/mooncake-demo   7m25s   

The error message only says it timed out waiting for FUSE, without mentioning that this Dataset has no client component.

Events:  Type     Reason       Age                From               Message  ----     ------       ----               ----               -------  Normal   Scheduled    108s               default-scheduler  Successfully assigned default/consumer to mooncake-worker2  Warning  FailedMount  16s (x3 over 78s)  kubelet            MountVolume.SetUp failed for volume "default-mooncake-demo" : rpc error: code = Internal desc = timeout waiting for FUSE mount point to be ready11m         Normal    Pulled                    pod/mooncake-demo-worker-0                 Container image "mooncake-min:v3" already present on machine11m         Normal    Pulled                    pod/mooncake-demo-worker-1                 Container image "mooncake-min:v3" already present on machine11m         Normal    Created                   pod/mooncake-demo-worker-0                 Created container worker11m         Normal    Started                   pod/mooncake-demo-worker-0                 Started container worker8m30s       Normal    Killing                   pod/mooncake-demo-worker-0                 Stopping container worker8m30s       Normal    Killing                   pod/mooncake-demo-master-0                 Stopping container master8m30s       Normal    Killing                   pod/mooncake-demo-worker-1                 Stopping container worker7m59s       Normal    SuccessfulCreate          advancedstatefulset/mooncake-demo-worker   create Pod mooncake-demo-worker-0 in AdvancedStatefulSet mooncake-demo-worker successful7m59s       Normal    Pulled                    pod/mooncake-demo-worker-0                 Container image "mooncake-min:v3" already present on machine7m59s       Normal    Created                   pod/mooncake-demo-worker-0                 Created container worker7m59s       Normal    Started                   pod/mooncake-demo-worker-0                 Started container worker7m59s       Normal    SuccessfulCreate          advancedstatefulset/mooncake-demo-worker   create Pod mooncake-demo-worker-1 in AdvancedStatefulSet mooncake-demo-worker successful7m59s       Normal    SuccessfulCreate          advancedstatefulset/mooncake-demo-master   create Pod mooncake-demo-master-0 in AdvancedStatefulSet mooncake-demo-master successful7m59s       Normal    Started                   pod/mooncake-demo-master-0                 Started container master7m59s       Normal    Created                   pod/mooncake-demo-master-0                 Created container master7m59s       Normal    Pulled                    pod/mooncake-demo-master-0                 Container image "mooncake-min:v3" already present on machine7m59s       Normal    Started                   pod/mooncake-demo-worker-1                 Started container worker7m59s       Normal    Created                   pod/mooncake-demo-worker-1                 Created container worker7m59s       Normal    Pulled                    pod/mooncake-demo-worker-1                 Container image "mooncake-min:v3" already present on machine16s         Warning   FailedMount               pod/consumer                               MountVolume.SetUp failed for volume "default-mooncake-demo" : rpc error: code = Internal desc = timeout waiting for FUSE mount point to be ready