Trident pods in CrashLoopBackOff when doing the Trident deployment
- Views:
- 694
- Visibility:
- Public
- Votes:
- 0
- Category:
- trident-kubernetes
- Specialty:
- snapx
- Last Updated:
- 4/8/2024, 4:11:18 PM
Applies to
- NetApp Astra Trident 23.10
- RedHat RHEL 9.3 Linux worker nodes
- Cilium network plugin
- Kubernetes 1.28
- Kubernetes nodes are VMs in a VMWare environment
Issue
When Trident is deployed in a kubernetes cluster, the Trident pods:
- Do not properly start
- They remain in CrashLoopBackOff status
In the events for any of the Trident pods in CrashLoopBackOff status, you can see the following traces:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning Unhealthy 29m (x29055 over 46h) kubelet Readiness probe failed: HTTP probe failed with statuscode: 503
Warning BackOff 4m35s (x11638 over 46h) kubelet Back-off restarting failed container driver-registrar in pod trident-node-linux-xxx
And in the Trident logs, for any of the hosts, the trident-registrar log will show :
IMMDD hh:mm:ss main.go:135] Version: v2.9.0
IMMDD hh:mm:ss main.go:136] Running node-driver-registrar in mode=
IMMDD hh:mm:ss main.go:157] Attempting to open a gRPC connection with: "/plugin/csi.sock"
IMMDD hh:mm:ss connection.go:213] Connecting to unix:///plugin/csi.sock
IMMDD hh:mm:ss connection.go:232] Still connecting to unix:///plugin/csi.sock
IMMDD hh:mm:ss connection.go:232] Still connecting to unix:///plugin/csi.sock
IMMDD hh:mm:ss connection.go:232] Still connecting to unix:///plugin/csi.sock
IMMDD hh:mm:ss main.go:160] error connecting to CSI driver: context deadline exceeded