ESXi Network and Storage Path Collapse on QLogic HPE Synergy CNA

Products

• VMware vSphere ESXi
• VMware Telco Cloud Infrastructure

Issue/Introduction

VMware ESXi hosts equipped with QLogic HPE Synergy 4820c Converged Network Adapters may experience a simultaneous loss of network and storage connectivity. The adapter's Management Firmware fails to respond to ESXi driver polling, leading to an abrupt termination of FCoE storage paths and Ethernet receive queues.

The /var/run/log/vmkernel.log displays the following sequence of errors:

cpu94:2098331)[qed_mfw_fill_tlv_data:4394(sp-<REDACTED_PCI_ID>)]Can't collect TLV management info
FSS: 7441: Failed to open file '<REDACTED_FILE_HANDLE>'; Requested flags 0x5, world: 2106151 [sut], (Existing flags 0x5, world: 2106151 [sut]): Busy
WARNING: NMP: nmp_DeviceRequestFastDeviceProbe:237: NMP device "<REDACTED_NAA_ID>" state in doubt; requested fast path state update...
WARNING: FS3J: 3301: Error committing txn callerID: <REDACTED_CALLER_ID> to slot 1: Timeout
Fil6: 4094: '<REDACTED_HOSTNAME>': Fil6 file IO () : Timeout
[qedentv_multictx_remove_rx_rule:1757(vmnic6)]Removing mac:<REDACTED_MAC>, vlan_id:0x0, from fp:2, op:MAC_DEL, hw_fn:0
[qedentv_multictx_q_free:7063(vmnic6)]fp:2, is_last:0, qtype:RX, hw_fn:0

Environment

ESXi: 7.0.3

TCI: 2.7

Cause

The issue is caused by a hardware or firmware lockup on the QLogic Converged Network Adapter (CNA) ASIC or physical PCIe bus, resulting in a simultaneous crash of both the storage and network data planes.

Resolution

• Engage Hardware vendor (HPE/QLogic) to evaluate PCIe slot registers and internal adapter dump logs to check for physical faults.

• Verify that qedentv and qedf driver/firmware combinations match the VMware Compatibility Guide (vCG) (or vSAN HCL if using vSAN) for the specific ESXi version.

• Update the host system BIOS to mitigate physical PCIe bus instabilities.

Additional Information

VMware Cloud Foundation (VCF) Known Issues