ESXi 8.0 experiences I/O errors and SCSI aborts on lsi_msgpt35 HBA

Products

• VMware vSphere ESX 8.x

Issue/Introduction

VMware vSphere ESXi hosts experience severe storage performance degradation. Symptoms include virtual machine clone tasks freezing, backup tasks taking significantly longer than expected, and guest operating systems (such as <REDACTED_HOSTNAME>) reporting disk I/O errors. Additionally, vCenter Server task schedules and management paths may become unresponsive.

The vmkernel.log reports frequent lsi_msgpt35 driver errors on the affected host (e.g., <REDACTED_IP>), indicating dropped frames (H:0x2) and aborted I/O requests (H:0x5):

In(182) vmkernel: cpu15:2098055)lsi_msgpt35_0: _scsih_task_mgmt: 871: C0:T1:L2: handle(0x0017): TM(abort) request start
In(182) vmkernel: cpu15:2098055)lsi_msgpt35_0: _scsih_abort: 652: C0:T1:L2, handle(0x0017), smid(3073), io_in_time_ms(2713382506), abort_in_time_ms(2713502508), delta(120002 ms)
In(182) vmkernel: cpu96:2098418)NMP: nmp_ThrottleLogForDevice:3893: Cmd 0x28 (0x45ba082ac340, 2099608) to dev "naa.<REDACTED_SECRET>" on path "vmhba2:C0:T1:L2" Failed:
In(182) vmkernel: cpu96:2098418)NMP: nmp_ThrottleLogForDevice:3898: H:0x5 D:0x0 P:0x0 . Act:EVAL. 
...
In(182) vmkernel: cpu5:2097307)NMP: nmp_ThrottleLogForDevice:3893: Cmd 0x2a (0x45ba0bd4e2c0, 2102986) to dev "naa.<REDACTED_SECRET>" on path "vmhba2:C0:T1:L2" Failed:
In(182) vmkernel: cpu5:2097307)NMP: nmp_ThrottleLogForDevice:3898: H:0x2 D:0x0 P:0x0 . Act:EVAL.

Further analysis indicates that a secondary host (<REDACTED_IP>) connecting to the same external SAS storage array does not exhibit these errors in its logs. Additionally, storage backend status reviews show no apparent abnormalities, isolating the issue to the local hardware of the affected host.

Environment

VMware vSphere ESXi 8.0

Cause

Physical hardware degradation or failure on the local Host Bus Adapter (HBA) or associated SAS cables connecting the ESXi host to the backend storage array.

Resolution

 

• Contact the server hardware vendor to perform hardware diagnostics on the affected ESXi host.

• Inspect and replace the physical SAS cables connecting the Host Bus Adapter to the storage array.

• If cable replacement does not resolve the issue, replace the affected Host Bus Adapter (HBA).

• To isolate the fault, perform a swap test by exchanging the SAS cables between the affected host and a known-working host on the storage array side.