Virtual machine with large RDMs in virtual compatibility becomes unresponsive

Products

• VMware vSphere ESXi

Issue/Introduction

Symptoms:

• A virtual machine with 2TB+ RDMs in virtual compatibility becomes hang suddenly.
• In the vmkernel.log file, you see entries similar to:

2019-09-30T06:29:05.564Z cpu3:69956)WARNING: VSCSI: 3502: handle 8194(vscsi0:2):WaitForCIF: Issuing reset;  number of CIF:2
2019-09-30T06:29:05.564Z cpu3:69956)WARNING: VSCSI: 2650: handle 8194(vscsi0:2):Ignoring double reset
2019-09-30T06:29:32.820Z cpu11:65891)VSCSI: 2896: handle 8194(vscsi0:2):Reset [Retries: 1/0] from (vmm0:TEST)
2019-09-30T06:29:32.820Z cpu11:65891)WARNING: NMP: nmpDeviceTaskMgmt:2304: Attempt to issue lun reset on device naa.11111111111111111111111111111111. This will clear any SCSI-2 reservations on the device.
2019-09-30T07:12:51.819Z cpu39:65891)VSCSI: 2896: handle 8194(vscsi0:2):Reset [Retries: 85/0] from (vmm0:TEST)
2019-09-30T07:12:51.819Z cpu39:65891)WARNING: NMP: nmpDeviceTaskMgmt:2304: Attempt to issue lun reset on device naa.11111111111111111111111111111111. This will clear any SCSI-2 reservations on the device.

Environment

VMware vSphere ESXi 6.5
VMware vSphere ESXi 6.7

Cause

This issue is caused when certain commands are sent from the VM with 2TB+ RDM in virtual compatibility.

Resolution

The issue has been addressed in the below releases of vSphere:

• VMware ESXi 6.5 P05. For more information, see VMware ESXi 6.5 P05 Release notes. VMware ESXi650-202007001 Release Notes

• VMware ESXi 6.7 P03. For more information, see VMware ESXi 6.7 P03 Release notes. VMware ESXi670-202008001 Release Notes

 

Workaround:
If this issue happens, rebooting ESXi host is needed to recover from hang state.

Workarounds are the followings:

• Use RDM disk in physical compatibility
• Use RDM disk with lower size than 2TB in using virtual compatibility