ESXI PSOD with Exception 13 in world: Fastslab with qfle3 Driver and QLogic 57810 Adapter

Products

• VMware vSphere ESXi

Issue/Introduction

• VMware ESXi host fails with a PSOD (purple screen of death)
  
  • 
• We will also observe below errors in /var/run/log/vmkernel.log
  • YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu26:20###40)ql_fcoe:vmhba67:CreateFabric:112:Info: Fabric created: 0x43#######90
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu26:20###40)ql_fcoe:vmhba67:SendFCoEDiscoverySolicitation:1208:Info: Sending FIP discovery for vlan_id = 0x3ea (0x43#######90
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu10:20###29)qfle3: qfle3_queue_alloc_with_attr:665: [vmnic0] Feature RSS requested.
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel:
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu10:20###29)qfle3: qfle3_queue_alloc_with_attr:654: [vmnic0] Feature LRO requested.
YYYY-MM-DDT10:12:12Z In(182) vmkernel:
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu10:20###29)qfle3: qfle3_rq_alloc:330: [vmnic0] Rxq 1 is leading RSS with 4 RSS queues.
YYYY-MM-DDT10:12:12Z In(182) vmkernel:
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu10:20###29)NetqueueBal: 2418: vmnic0: Requesting 3 RSS sec queues for engine type 4
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu10:20###29)NetqueueBal: 4198: vmnic0: New NetQ RSS engine created
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu26:20###40)ql_fcoe:vmhba67:SendFCoEDiscoverySolicitation:1208:Info: Sending FIP discovery for vlan_id = 0x3ea (0x43#######90
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu26:20###40)ql_fcoe:vmhba67:SendFCoEDiscoverySolicitation:1208:Info: Sending FIP discovery for vlan_id = 0x3ea (0x43#######90
YYYY-MM-DDTHH:MM:SSZ Wa(180) vmkwarning: cpu26:20###40)WARNING: ql_fcoe:vmhba67:FipDiscoveryTimeoutWork:126: Max retry exhausted for Fabric = 0x43#######90 vlan_id = 0x##a
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu26:20###40)ql_fcoe:vmhba67:StartPortLogout:1113:Info: Sess 0x43#######d0 port_id f##fe
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu26:20###40)ql_fcoe:vmhba67:CancelExchangeHandling:196:Info: Enter for Sess = 0x43#######d0
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu78:2099341)ql_fcoe:vmhba67:DeleteFabric:123:Info: Fabric 0x43#######90 000000 destroyed
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel:
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu78:2099341)ql_fcoe:vmhba67:SendFCoEVlanSolicitation:1480:Info: Sending FCoEVlanSolicitation request (0x1)
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu26:20###40)ql_fcoe:vmhba67:SendFCoEVlanSolicitation:1480:Info: Sending FCoEVlanSolicitation request (0x2)
YYYY-MM-DDTHH:MM:SSZ In(182) vmkernel: cpu26:20###40)ql_fcoe:vmhba67:SendFCoEVlanSolicitation:1480:Info: Sending FCoEVlanSolicitation request (0x3)
YYYY-MM-DDTHH:MM:SSZ Wa(180) vmkwarning: cpu26:20###40)WARNING: ql_fcoe:vmhba67:FipVlanTimeoutWork:258: FIP VLAN Max Retries reached, cur vlan and pri: 1002

 

Environment

• VMware vSphere ESXI 8.0U3 

Cause

• Memory corruption within the netPktPageCache-AnyMem slab caused by a defect or version mismatch between the qfle3 driver and the Marvell/QLogic 57810 firmware during network packet processing.

Resolution

• Update the qfle3 driver and the Marvell/QLogic 57810 firmware to the latest certified versions.
• If Fiber Channel over Ethernet (FCoE) is not utilized in the environment, disable FCoE at the BIOS/UEFI level to prevent driver initialization loops and "Max retry exhausted" errors.
• Reboot the host to apply BIOS changes and initialize the updated driver/firmware stack.