Minio pod failing create

[gabeduke]

Minio Pod is failing to create with the following error (pod coming up on the master node which has no NodeD):

  Type     Reason              Age                From                       Message
  ----     ------              ----               ----                       -------
  Normal   Scheduled           2m50s              default-scheduler          Successfully assigned default/minio-65c68b8ffb-wj6lb to kube-master-449c
  Warning  FailedAttachVolume  67s (x3 over 69s)  attachdetach-controller    AttachVolume.Attach failed for volume "minio-pv" : node "kube-master-449c" has no NodeID annotation
  Warning  FailedMount         47s                kubelet, kube-master-449c  Unable to mount volumes for pod "minio-65c68b8ffb-wj6lb_default(dacf858f-5418-4f8f-af81-d46a4489d78c)": timeout expired waiting for volumes to attach or mount for pod "default"/"minio-65c68b8ffb-wj6lb". list of unmounted volumes=[minio-data]. list of unattached volumes=[minio-data default-token-c729p]
  Warning  FailedAttachVolume  4s (x5 over 65s)   attachdetach-controller    AttachVolume.Attach failed for volume "minio-pv" : rpc error: code = NotFound desc = ControllerPublishVolume: the volume minio-data not exists

[andyjeffries]

Have you changed something from the default installation? Normally the master is perfectly fine to bring pods up on.

[andyjeffries]

Does this still happen after today's upgrade to Longhorn 0.7.0?

[peon-pasado-zeitnot]

I got the same issue. I managed to 'fix' it by creating pvc via longhorn UI and editing minio deployment to use it

[ruanbekker]

I had a similar issue with Jenkins.

Tested with minio and removing the PersistentVolume section and adding metadata to PersistentVolumeClaim worked for me.

With a default install:

> kubectl get pods
NAME                     READY   STATUS              RESTARTS   AGE
minio-6d8f9dd7cd-hqgtx   0/1     ContainerCreating   0          7m27s

Pod events:

Events:
  Type     Reason              Age                   From                     Message
  ----     ------              ----                  ----                     -------
  Normal   Scheduled           <unknown>             default-scheduler        Successfully assigned default/minio-6d8f9dd7cd-hqgtx to kube-node-4843
  Warning  FailedMount         2m3s (x2 over 4m17s)  kubelet, kube-node-4843  Unable to attach or mount volumes: unmounted volumes=[minio-data], unattached volumes=[minio-data default-token-wkm77]: timed out waiting for the condition
  Warning  FailedAttachVolume  8s (x11 over 6m20s)   attachdetach-controller  AttachVolume.Attach failed for volume "minio-pv" : rpc error: code = NotFound desc = ControllerPublishVolume: the volume minio-data not exists

When removing PersistentVolume:

---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: minio-pv-claim
  labels:
    type: longhorn
    app: minio
spec:
  storageClassName: longhorn
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 5Gi
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: minio
...

It goes into a running state.

After creating a bucket, I deleted the pod and waited for the pod to be assigned to another node and the data persisted.

I haven't figured out why having the PersistentVolume definition is raising the issue

[mr-smithers-excellent]

Experienced the same issue here installing Minio on a fresh cluster using:

civo applications add MinIO:5GB --cluster=cluster-name

[Akay7]

The same problem exists after installing MinIO through the web interface.