Details
-
Bug
-
Resolution: Not a Bug
-
Normal
-
None
-
4.12.z
-
Moderate
-
No
-
False
-
Description
Description of problem:
When pod is scheduled on another node, It's failing to create with error "Multi-Attach error for volume "pvc-xxx" Volume is already exclusively attached to one node and can't be attached to another"
Actual results:
Expected results:
Pod should be scheduled on another node without any error
Additional info:
-- Volume access mode is set to RWO (ReadWriteOnce : the volume can be mounted as read-write by a single node. ReadWriteOnce access mode still can allow multiple pods to access the volume when the pods are running on the same node) -- Also they are using STS (When a StatefulSet's .spec.updateStrategy.type is set to RollingUpdate, the StatefulSet controller will delete and recreate each Pod in the StatefulSet) -- application creation workflow the helm chart of this pvc is as shown below. kind: PersistentVolumeClaim name: fm-storage namespace: cran1 spec: accessModes: - ReadWriteOnce dataSource: apiGroup: snapshot.storage.k8s.io kind: VolumeSnapshot name: fm-storage-snapshot dataSourceRef: apiGroup: snapshot.storage.k8s.io kind: VolumeSnapshot name: fm-storage-snapshot resources: requests: storage: 1Gi storageClassName: ocs-storagecluster-cephfs volumeMode: Filesystem volumeName: pvc-******* the general application creation process is as below. step 1: create PVC helm release a step 2: create CNF helm release b including all PODs When deleting release a, the volume snapshots of PVCs are created accordingly. When create release a, the PVCs are restored from volumesnapshot if they exist. the issue recently happened for pod "po-cran6-rcpfm-0". It is assigned to "worker8" and the pv associated with the pod is "pvc-d5f2c46d-49b5-4ce2-af92-d183377fea50" in this latest issue only CNF was restarted, it means PODs were recreated and PVC should not be recreated. Events: Type Reason Age From Message ---- ------ ---- ---- ------- Normal Scheduled 164m default-scheduler Successfully assigned cran6/po-cran6-rcpfm-0 to worker8.hzdc-pz-10-110-10-98.ocp.hz.nsn-rdnet.net Warning FailedAttachVolume 164m attachdetach-controller Multi-Attach error for volume "pvc-d5f2c46d-49b5-4ce2-af92-d183377fea50" Volume is already exclusively attached to one node and can't be attached to another PVC :- pvc-d5f2c46d-49b5-4ce2-af92-d183377fea50 1Gi RWO Delete Bound cran6/fm-storage ocs-storagecluster-cephfs 6h17m This issue may have occurred as a result of a previous instance of a pod not having been successfully terminated prior to the new pod attempting to attach.