PowerStore: Permanent Device Loss (PDL) error on VMware cluster when using LUN 0
Summary: PowerStore Logical Unit Numbers (LUNs) show as 0 bytes or unknown capacity on the ESXi operating system because VMware views it as a RAID controller. This is due to the LUNs being used as a Storage Array Controller Device (SACD) and Virtual Volume (vVol) protocol end point. ...
Symptoms
Example screenshot from Host:
PowerStore LUNs start with na.68ccf and are identified with path vmhba*:C*:T*:L0. Only the SACD with LUN ID 0 reports Permanent Device Loss (PDL).
If any other LUN ID reports PDL, this KB does not apply.
This LUN when presented by the PowerStore shows up as a size of 0 and does not have a datastore on it.
Cause
PowerStore exposes Storage Array Controller Device (SACD) by default, and it is under LUN 0.
PowerStore exposes vSphere Virtual Volumes LUN 254 and 255 for the vVol Physical Endpoint.
These LUNs are not "real" LUNs that cannot be used for any volume mapping.
Resolution
To prevent the errors logged against the Storage Array Controller Device (SACD), create and map one volume per host or host group as LUN 0.
The LUNs must be 1 TB or less each but can be small as well - for example 1GB.
Presenting a real volume/LUN as LUN 0 to the host allows the VMware host to connect to it and stop the PDL errors.
Example screenshot of volume creation:
Once the LUN is created and mapped, rescan storage on the host/cluster.
The LUNs show as added and bound to LUN 0, with the configured capacity.
Note: For a multi-appliance cluster, there are multiple SACDs. Each appliance has its own SACD, so LUN 0 needs to be created for each appliance. The LUN numbering exists on a "per-appliance" level, so a LUN 0 volume can be created for each appliance.