Dell Container Storage Modules (CSM) Family: PowerMax node pods are failing to start

PowerMax CSI driver Node pods are failing to start as they time-out when trying to register the node on the PowerMax arrays.

CSI Driver (Node- Driver container)

In the Node driver log, multiple errors are evident when trying to connect to the iSCSI interface on the remote array. This causes a time delay when registering the node on the array.

Error discovering 10.10.10.11: signal: killed
time="2025-02-10T15:53:12Z" level=error msg="Failed to connect to the IP interface(10.10.10.10) of array(000000000001)"

CSI Driver (Node- Registrar container)

Note the two minute timeout when trying to register the node with the arrays.

I0210 15:52:41.785854       1 main.go:96] "Received GetInfo call" request="&InfoRequest{}"
I0210 15:54:41.800437       1 main.go:108] "Received NotifyRegistrationStatus call" status="&RegistrationStatus{PluginRegistered:false,Error:RegisterPlugin error -- plugin registration failed with err: rpc error: code = DeadlineExceeded desc = context deadline exceeded,}"

The Kubernetes nodes at each site do not have iSCSI connectivity to the remote PowerMax array, only the local PowerMax.

The CSI driver for PowerMax does not support a non-uniform metro cluster. (i.e. nodes on Site1 can only see volumes on Site 1, and nodes on Site2 can only see volumes on Site 2).

Resolution 

1. Ensure all hosts in the Kubernetes cluster have iSCSI connectivity to the iSCSI interfaces on both PowerMax arrays at each site.
2. CSM engineering is evaluating the requirements to address the pods failing issue in an upcoming product release.