[BUG] mountpod 没有正常退出

[YunhuiChen]

What happened:
批量创建一些应用pod:

apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: dynamic-multi
  labels:
    app: dynamic-multi
spec:
  replicas: 30
  podManagementPolicy: Parallel
  serviceName: dynamic-multi
  selector:
    matchLabels:
      app: dynamic-multi
  template:
    metadata:
      labels:
        app: dynamic-multi
    spec:
      containers:
      - command:
        - sh
        - -c
          # - sleep infinity
        - "while true; do echo $(date -u) >> /data/out.txt; sleep 0.5; done"
        image: registry.cn-hangzhou.aliyuncs.com/juicedata/mount:ce-v1.2.0
        name: tools
        volumeMounts:
        - mountPath: /data
          name: dynamic-multi
          mountPropagation: HostToContainer
  volumeClaimTemplates:
  - metadata:
      name: dynamic-multi
    spec:
      accessModes: [ "ReadWriteMany" ]
      storageClassName: dynamic-ee
      resources:
        requests:
          storage: 1Gi

mountpod没有被正常删除，状态转为error：

Name:                 juicefs-stage3-pvc-48832491-6215-4091-936e-37ebeb32a083-upcxav
Namespace:            kube-system
Priority:             2000001000
Priority Class Name:  system-node-critical
Service Account:      juicefs-csi-node-sa
Node:                 stage3/172.30.117.99
Start Time:           Thu, 16 Jan 2025 11:32:17 +0800
Labels:               app.kubernetes.io/name=juicefs-mount
                      chaostest=true
                      juicefs-hash=15a68d09a1b3c9dfb67ce31e876b502d0975d6b8a2456680332b6af03696e3e
                      volume-id=pvc-48832491-6215-4091-936e-37ebeb32a083
Annotations:          cni.projectcalico.org/containerID: b5994e0628e6f331fbac981d58dac649647398eb06b1775ac13bb54ad8fd28ee
                      cni.projectcalico.org/podIP:
                      cni.projectcalico.org/podIPs:
                      juicefs-delete-at: 2025-01-22 02:54:02
                      juicefs-delete-delay: 30s
                      juicefs-uniqueid: pvc-48832491-6215-4091-936e-37ebeb32a083
                      juicefs-uuid: stage-volume1
Status:               Failed
IP:
IPs:                  <none>
Containers:
  jfs-mount:
    Container ID:  docker://81a08e3a5e05557bccbe083dd1446ffc854680afc61f6b46c6f6249bb8298491
    Image:         registry-vpc.cn-hangzhou.aliyuncs.com/juicedata/mount:ee-5.1.7-e634965
    Image ID:      docker-pullable://registry-vpc.cn-hangzhou.aliyuncs.com/juicedata/mount@sha256:264270c44064954e246733df2e231facdbd80efa2db81678f6c341c9081670e7
    Port:          <none>
    Host Port:     <none>
    Command:
      sh
      -c
      cp /etc/juicefs/stage-volume1.conf /root/.juicefs
      exec /sbin/mount.juicefs stage-volume1 /jfs/pvc-48832491-6215-4091-936e-37ebeb32a083-upcxav -o foreground,no-update,cache-size=102400,enable-xattr,enable-acl,cache-dir=/data/stage-cache
    State:          Terminated
      Reason:       Error
      Exit Code:    2
      Started:      Thu, 16 Jan 2025 11:32:18 +0800
      Finished:     Wed, 22 Jan 2025 10:56:03 +0800
    Ready:          False
    Restart Count:  0
    Limits:
      cpu:     2
      memory:  5Gi
    Requests:
      cpu:     100m
      memory:  200Mi
    Environment Variables from:
      juicefs-pvc-48832491-6215-4091-936e-37ebeb32a083-secret  Secret  Optional: false
    Environment:
      JFS_FOREGROUND:  1
      JFS_SUPER_COMM:  /tmp/fuse_fd_csi_comm.sock
    Mounts:
      /data/stage-cache from cachedir-0 (rw)
      /etc/juicefs from init-config (rw)
      /etc/updatedb.conf from updatedb (rw)
      /jfs from jfs-dir (rw)
      /tmp from jfs-fuse-fd (rw)
      /var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-qhwp6 (ro)
Conditions:
  Type                        Status
  PodReadyToStartContainers   False
  Initialized                 True
  Ready                       False
  ContainersReady             False
  PodScheduled                True
Volumes:
  init-config:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  juicefs-pvc-48832491-6215-4091-936e-37ebeb32a083-secret
    Optional:    false
  jfs-dir:
    Type:          HostPath (bare host directory volume)
    Path:          /var/lib/juicefs/volume
    HostPathType:  DirectoryOrCreate
  jfs-fuse-fd:
    Type:          HostPath (bare host directory volume)
    Path:          /var/run/juicefs-csi/15a68d09a1b3c9dfb67ce31e876b502d0975d6b8a2456680332b6af03696e3e
    HostPathType:  DirectoryOrCreate
  updatedb:
    Type:          HostPath (bare host directory volume)
    Path:          /etc/updatedb.conf
    HostPathType:  FileOrCreate
  cachedir-0:
    Type:          HostPath (bare host directory volume)
    Path:          /data/stage-cache
    HostPathType:  DirectoryOrCreate
  kube-api-access-qhwp6:
    Type:                    Projected (a volume that contains injected data from multiple sources)
    TokenExpirationSeconds:  3607
    ConfigMapName:           kube-root-ca.crt
    ConfigMapOptional:       <nil>
    DownwardAPI:             true
QoS Class:                   Burstable
Node-Selectors:              <none>
Tolerations:                 CriticalAddonsOnly op=Exists
                             node.kubernetes.io/disk-pressure:NoSchedule op=Exists
                             node.kubernetes.io/memory-pressure:NoSchedule op=Exists
                             node.kubernetes.io/not-ready:NoExecute op=Exists
                             node.kubernetes.io/pid-pressure:NoSchedule op=Exists
                             node.kubernetes.io/unreachable:NoExecute op=Exists
                             node.kubernetes.io/unschedulable:NoSchedule op=Exists

CI可复现：https://github.com/juicedata/jfs/actions/runs/12887089221/job/35970819230?pr=2322

What you expected to happen:

How to reproduce it (as minimally and precisely as possible):

Anything else we need to know?

Environment:

• JuiceFS CSI Driver version (which image tag did your CSI Driver use):
• Kubernetes version (e.g. kubectl version):
• Object storage (cloud provider and region):
• Metadata engine info (version, cloud provider managed or self maintained):
• Network connectivity (JuiceFS to metadata engine, JuiceFS to object storage):
• Others:

[zxh326]

应该和数量无关

问题发生在这个 PR 之后，应该和这个有关 cc @zwwhdls

[zwwhdls]

mount pod 的 daemon 进程收到 SIGTERM 信号后没有传递给子进程，导致 pod 没有正常退出