CODEWITHSUNDEEP
kubernetesnfs
Dolphin
Dolphin

Reputation: 38601

Error [OpenError]: IO error: lock /opt/data/yjs-storage/LOCK: Resource temporarily unavailable

when I tried to load the nfs(v4) folder in 2 kubernetes(v1.30.x) pods, one of them throw error:

│ Persisting documents to "/opt/data/yjs-storage"
│
│ /home/node/app/node_modules/levelup/lib/levelup.js:119
│       return callback(new OpenError(err))
│                       ^
│ Error [OpenError]: IO error: lock /opt/data/yjs-storage/LOCK: Resource temporarily unavailable
│     at /home/node/app/node_modules/levelup/lib/levelup.js:119:23
│     at /home/node/app/node_modules/abstract-leveldown/abstract-leveldown.js:38:14
│     at /home/node/app/node_modules/deferred-leveldown/deferred-leveldown.js:31:21
│     at /home/node/app/node_modules/abstract-leveldown/abstract-leveldown.js:38:14
│     at /home/node/app/node_modules/abstract-leveldown/abstract-leveldown.js:38:14
│ Emitted 'error' event on LevelUP instance at:
│     at /home/node/app/node_modules/levelup/lib/levelup.js:60:19
│     at /home/node/app/node_modules/levelup/lib/levelup.js:119:14
│     at /home/node/app/node_modules/abstract-leveldown/abstract-leveldown.js:38:14
│     [... lines matching original stack trace ...]
│     at /home/node/app/node_modules/abstract-leveldown/abstract-leveldown.js:38:14 {
│   [cause]: undefined
│ }
│
│ Node.js v18.20.6

it look like it did not allow the 2 pods load the same NFS folder? this is the NFS config:

apiVersion: v1
kind: PersistentVolume
metadata:
  name: nfs-texhub-server-pv-qingdao
spec:
  capacity:
    storage: 8Gi
  nfs:
    server: 60cw9b7f-osv72.cn-qingdao.nas.aliyuncs.com
    path: /k8s/reddwarf-pro/texhub-server-service
  accessModes:
    - ReadWriteMany
  claimRef:
    kind: PersistentVolumeClaim
    namespace: reddwarf-pro
    name: texhub-server-service-pv-claim-qingdao
    uid: 926e70a4-651f-467a-9a81-0b87a7b696ee
    apiVersion: v1
    resourceVersion: '1080096'
  persistentVolumeReclaimPolicy: Retain
  mountOptions:
    - vers=4.0
    - noresvport
  volumeMode: Filesystem

it it possible to let multi pod write from the same NFS directory?

Upvotes: 0

Views: 61

Answers (1)

Imran Premnawaz
Imran Premnawaz

Reputation: 267

The error you are seeing suggests that the NFS mount is being used by several processes at once and that there is a locking problem on the NFS server, as stated in the comments.

Please be noted that LevelDB is designed to be used by a single process at a time. When you are trying for multiple processes this could also be done by using kubernetes mechanisms like Pod Affinity or Pod Anti-Affinity to make sure that only one pod can access the database at a time.

Adding to that you can also use the Pod Topology Spread Constraints which might help to resolve the constraint issue.

Upvotes: 0

Related Questions