-
Bug
-
Resolution: Cannot Reproduce
-
Undefined
-
None
-
4.15.z
-
Important
-
No
-
False
-
This is an LVMS Bug Report:
Please make sure that you describe your storage configuration in detail. List all devices that you plan to work with for LVMS as well as any relevant machine configuration data to make it easier for an engineer to help out.
Description of problem:
Customer has a Hosted Control Plane environment, each etcd Pod (3) spawns 1 volume from the LVMS provisioner. This is set in the *HostedControlPlane* CR owned by the Cluster CR ~~~ name: hcp-automation namespace: clusters-hcp-automation spec: autoscaling: {} clusterID: a2731297-bac3-4279-90f5-f0ba9fcf891e configuration: {} controllerAvailabilityPolicy: HighlyAvailable dns: baseDomain: <DOMAIN> etcd: managed: storage: persistentVolume: size: 8Gi storageClassName: lvms-hcp-etcd type: PersistentVolume managementType: Managed ~~~ 2 etcd pods starts fine, there's one that fails to start with error: "xfs_growfs: XFS_IOC_FSGROWFSDATA xfsctl failed: No space left on device"
Version-Release number of selected component (if applicable):
- OCP 4.15.2(Baremetal) - LVM Storage 4.15.3
Steps to Reproduce:
- Not sure at the moment, but customer mentioned this happens for multiple HCP clusters: * This is reproducible for any number of clusters. I have already attached hypershift dump from 2 clusters having similar issue "hcp-cluster-small" and "hcp-automation"
Actual results:
- HCP cluster is not fully up waiting for etcd to start
Expected results:
- etcd pod to be able to generate the LogicalVolume to start and the HCP cluster to complete its provisioning phase.
Additional info:
- is triggering
-
OCPBUGS-35072 LVMS - Guideline to troubleshoot a single node in a multi-node cluster is missing
-
- MODIFIED
-