Kubernetes PVC provisioning fails via Trident after ONTAP upgrade due to transient API communication issue
Applies to
- ONTAP 9.15 and above
- Astra Trident for Kubernetes
Issue
- After performing an ONTAP upgrade and subsequent storage failover/giveback, Kubernetes fails to provision Persistent Volume Claims (PVCs) via Trident.
- The PVC may show as “Bound” in Kubernetes, but the associated Pod remains in “ContainerCreating” state and does not start.
Kubernetes Pod status:
- Pod stuck in “ContainerCreating” for extended period
- PVC shows as “Bound”, but underlying LUN does not exist on ONTAP
- Trident controller logs show errors related to API communication and missing LUNs.
time="2026-05-11T03:52:51Z" level=error msg="API invocation failed. Post \\"https://test-iscsi-prd-c1-dc2-svm-mgmt.xxxx.com/servlets/netapp.servlets.admin.XMLrequest_filer\\": read tcp xxx.xx.xx.xxx:44248->xx.xxx.xxx.xxx:443: read: connection reset by peer"time="2026-05-11T03:53:04Z" level=error msg="error publishing ontap-san driver: problem retrieving LUN info: could not get LUN by name, error: LUN not found"
- And in the volume creation workflow:
API invocation failed. Post https://<svm-mgmt-lif>/servlets/netapp.servlets.admin.XMLrequest_filer: context deadline exceeded (Client.Timeout exceeded while awaiting headers)
Pods remain stuck in “ContainerCreating” state, and manual deletion/recreation is required to restore service.
