Error: 0x5 0x24 0x00 on ALUA RTPG (0xA3) for Huawei OceanStor followed by dead paths.

Products

• VMware vSphere ESXi

Issue/Introduction

Storage paths on a VMware ESXi host are marked as "Dead" when communicating with a Huawei OceanStor (Dorado) array.

• In /var/log/vmkernel.log, the following error appears: WARNING: VMW_SATP_ALUA: satp_alua_issueCommandOnPath:1061: Probe cmd 0xa3 failed for path "vmhba#:C#:T#:L#" (0x5/0x24/0x0).
• SCSI Sense Code: 0x5 0x24 0x0 (ILLEGAL REQUEST / INVALID FIELD IN CDB) returned for REPORT TARGET PORT GROUPS (0xA3).
• Subsequently Storage paths are marked as Dead.
• The issue may affect only one Fabric or specific HBAs (e.g., vmhba1).

 

Environment

• VMware vSphere ESXi 8.0
• Huawei OceanStor / Dorado SAN

Cause

On the host initiators on the Huawei OceanStor, the multipath_type is set to default instead of third-party. As a result, the array does not correctly support ALUA queries from the VMware ESXi VMW_SATP_ALUA plugin.

 

 

Resolution

To resolve this, update the initiator configuration on the Huawei OceanStor.

1. Reconfigure Huawei initiators and  engage Huawei support if needed.

2.Perform a storage rescan on the ESXi host to restore paths: esxcli storage core adapter rescan --all

3.Verify the path status: esxcli storage core path list

4.If paths remain Dead, reset the affected HBA to clear the error state: esxcli storage san fc reset -A vmhba#

5.If the issue persists after these changes, contact the storage vendor to verify the array-side ALUA configuration.

Additional Information

Additional Information: 

Understanding how paths to a storage/LUN device are marked as Dead