CSM: CSI Driver for PowerFlex CrashLoopBackOff due to Gateway Login Timeout
Summary: This issue affects Dell Container Storage Modules (CSM) with the Dell Container Storage Interface (CSI) Driver for PowerFlex. The driver fails probes when log in to the PowerFlex Gateway exceeds a hard coded 1 second timeout. ...
Symptoms
Driver logs show a 1 second timeout:
time="2026-05-15T17:14:29Z" level=info msg="Login to PowerFlex Gateway, system=<array id>, endpoint=https://<gateway>, user=<user>\n" time="2026-05-15T17:14:30Z" level=error msg="array <array id> probe failed: rpc error: code = FailedPrecondition desc = unable to login to PowerFlex Gateway: Get \"https://<gateway>/api/login\": context deadline exceeded"
Pods can show in a CrashLoopBackOff or Error state.
oc get pods -n vxflexos ... vxflexos-node-xxxxx 1/2 Error 442 (5m22s ago) 19h vxflexos-node-xxxxx 0/2 CrashLoopBackOff 431 (103s ago) 18h
Recent changes may include upgrade, network latency, gateway performance issues, or cluster scaling.
Cause
The CSI Driver for PowerFlex driver versions 2.13.1 and 2.14.0 use a nonconfigurable probe timeout. The login call to the PowerFlex Gateway exceeds this timeout.
Resolution
Engineering has provided a change to make the probe timeout configurable in later releases of the Container Storage Interface (CSI) Driver for PowerFlex. Upgrade to a release where the probe timeout is configurable.
Per the release notes, this issue is fixed in the following versions:
https://github.com/dell/csi-powerflex/releases (External Link)
- 2.13.2
- 2.14.1
- 2.15.0 or later
- 2.16.0 or later
After upgrade, configure the probe timeout X_CSI_PROBE_TIMEOUT per the environment needs. The default is now set to 10 seconds.