Closed ybnelson closed 6 months ago
When creating a YBCluster using the latest yugabytedb/yugabyte:2.0.7.0-b9
image with the example deployment featured in the repository, k8s 1.15 reports the same error message:
FailedScheduling 31s (x2 over 31s) default-scheduler pod has unbound immediate PersistentVolumeClaims (repeated 2 times)
But the cluster reports the PersistentVolumeClaim as not having the storageClass, accessModes, and capacity information configured with the PVC.
$ kubectl get pvc --all-namespaces
NAMESPACE NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
sitewide-dev-application datadir0-yb-master-0 Pending 19s
sitewide-dev-application datadir0-yb-master-1 Pending 19s
sitewide-dev-application datadir0-yb-master-2 Pending 19s
sitewide-dev-application datadir0-yb-tserver-0 Pending 19s
sitewide-dev-application datadir0-yb-tserver-1 Pending 19s
sitewide-dev-application datadir0-yb-tserver-2 Pending 19s
The operator does not seem to apply the configured overrides nor apply the default standard
storage class for v2 of YugabyteDB.
After poking around the yugabyte-operator repository I did notice that in the controller implementation, the getVolumeTemplates function fills out the volume mount name and storage size request but does not override the storage class or access mode properties.
cc @Arnav15 @ramkumarvs @iSignal
This is still not working in 2022 :-(
This is still not working in 2024 :-(
@meherkumar13 the note from README:
NOTE: This repository contains code and example Custom Resources for the Kubernetes Operator (legacy) for YugabyteDB v0.0.1. This operator is no longer maintained. We recommend using the actively maintained helm charts for YugabyteDB to install yugabyte on Kubernetes.
There is also a new operator being developed, you will get more details in upcoming release(s).
Given cluster deploy intended for "faster" SSD storage class: apiVersion: yugabyte.com/v1alpha1 kind: YBCluster metadata: name: example-ybcluster namespace: yb-operator spec: replicationFactor: 3 master: replicas: 3 storage: storageClass: faster size: 1Gi tserver: replicas: 3 storage: storageClass: faster count: 1 size: 1Gi
Cluster fails to bind to new storageClass for volumes and defaults to standard storageclass. kubectl describe pod <...> Events: Type Reason Age From Message
Warning FailedScheduling 2m37s (x4 over 2m40s) default-scheduler pod has unbound immediate PersistentVolumeClaims (repeated 3 times) Normal Scheduled 2m37s default-scheduler Successfully assigned yb-operator/yb-tserver-0 to gke-yb-operator-test-default-pool-46c7083c-4wgl Normal SuccessfulAttachVolume 2m27s attachdetach-controller AttachVolume.Attach succeeded for volume "pvc-dc55ad65-1c9a-11ea-a20c-42010af0005a" Normal Pulling 2m5s kubelet, gke-yb-operator-test-default-pool-46c7083c-4wgl pulling image "yugabytedb/yugabyte:1.3.2.0-b19" Normal Pulled 2m4s kubelet, gke-yb-operator-test-default-pool-46c7083c-4wgl Successfully pulled image "yugabytedb/yugabyte:1.3.2.0-b19" Normal Created 2m4s kubelet, gke-yb-operator-test-default-pool-46c7083c-4wgl Created container Normal Started 2m4s kubelet, gke-yb-operator-test-default-pool-46c7083c-4wgl Started container (base) dsqln:yugabyte-operator andrewnelson$ kubectl get pvc -n yb-operator NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE datadir0-yb-master-0 Bound pvc-dc3ce53a-1c9a-11ea-a20c-42010af0005a 1Gi RWO standard 3m16s datadir0-yb-master-1 Bound pvc-dc427473-1c9a-11ea-a20c-42010af0005a 1Gi RWO standard 3m16s datadir0-yb-master-2 Bound pvc-dc492e47-1c9a-11ea-a20c-42010af0005a 1Gi RWO standard 3m16s datadir0-yb-tserver-0 Bound pvc-dc55ad65-1c9a-11ea-a20c-42010af0005a 1Gi RWO standard 3m16s datadir0-yb-tserver-1 Bound pvc-dc5eef8b-1c9a-11ea-a20c-42010af0005a 1Gi RWO standard 3m16s datadir0-yb-tserver-2 Bound pvc-dc6644d9-1c9a-11ea-a20c-42010af0005a 1Gi RWO standard 3m16s