Error: H:0x8 SCSI_HOST_RESET and Inaccessible Local Datastore on Cisco UCS with lsi_mr3 Driver

Products

• VMware vSphere ESXi

Issue/Introduction

Symptoms:

• Virtual Machines appear as invalid/inaccessible in the vSphere Inventory.
  
  
• The storage controller registers sudden state-change or hot-plug events in /var/run/log/vmkernel.log -
  
  2026-07-05T18:26:30.100Z cpu8:2097962)lsi_mr3: megasas_hotplug_work:518: event code: 0x71.
2026-07-05T18:27:17.257Z cpu5:2097564)lsi_mr3: mfi_TaskMgmt:690: Processing taskMgmt abort for device: vmhba1:C2:T0:L0
2026-07-05T18:27:17.257Z cpu5:2097564)lsi_mr3: mfi_TaskMgmt:724: ABORT request for SN 1257101388 Wld 2097217
2026-07-05T18:27:17.257Z cpu5:2097564)lsi_mr3: mfi_TaskMgmt:739: TA : SMID 1928 IN 29814904s OUT 0s TM 29814912s Delta 8s
2026-07-05T18:27:17.257Z cpu5:2097564)WARNING: lsi_mr3: mfi_TaskMgmt:743: Abort not supported on C2:T0:L0 for SMID 1928

2026-07-05T18:27:17.645Z cpu24:2098062)ScsiDeviceIO: 4124: Cmd(0x45b8d92b6e08) 0x2a, CmdSN 0x1631dad7 from world 2097217 to dev "naa.################################" failed H:0x8 D:0x0 P:0x0
2026-07-05T18:27:19.254Z cpu7:2098096)lsi_mr3: mfi_VirtReset:633: VR : SMID 1921 IN 29814902s OUT 0s TM 29814913s Delta 11s
• Logging to CIMC (Cisco's hardware interface management web-console), shows underlying hardware degraded events under "Fault History".
  
  
  
  Events seen by hardware vendor -
  5:2026 Jul  5 19:35:33 UTC:(4.2(2a)):fault-engined:-: xxxCVxxx: %CIMC-5-EQUIPMENT_DEGRADED:[F0996][warning][equipment-degraded][sys/rack-unit-1/board] HDD12_STATUS: Storage Local disk is degraded

Environment

VMware vSphere ESXi

Cause

• The issue is caused by a physical hardware failure or degradation of the local disk(s) backing the RAID volume. When a physical disk drops offline or enters a degraded state, the lsi_mr3 controller loses communication with the logical device, triggering a continuous recovery loop of SCSI host resets (H:0x8).

Resolution

To resolve this issue, perform the following steps:

1. Hardware Validation:
   
   

   • Log in to the Cisco Integrated Management Controller (CIMC) or the server's hardware management interface.
   • Navigate to Storage > Cisco Raid Controller > Physical Drive Inventory.
   • Identify any drives marked as Degraded, Failed, or Predictive Failure.

2. Hardware Replacement:
   
   

   • Engage your hardware vendor (Cisco Support) to validate the disk health and arrange for a physical replacement of the faulty drive, if your hardware vendor concludes the disk being in unhealthy state.

3. Storage Rescan:
   
   

   • Once the physical drive has been replaced and the RAID volume has finished rebuilding/initializing, log in to the ESXi host via the Host Client or vCenter Server.
   • Navigate to Storage > Adapters and click Rescan.
   • Verify the datastore is once again visible and mounted.

4. VM Recovery:
   
   

   • If the VMs still appear as Inaccessible, right-click the VM in the inventory and select Unregister.
   • Browse the datastore, locate the VM's .vmx file, right-click it, and select Register VM.
     
     Note: Even after re-registering the VM, if the VM is not accessible then consider restoring the VM from backup.