ESXi PSODs with nmp_TriggerFlakyPathEvaluation when NMP tries to claim over 32 paths.

Products

• VMware vSphere ESXi
• VMware vSphere ESXi 8.0

Issue/Introduction

ESXi host may fail with a PSOD citing nmp_TriggerFlakyPathEvaluation in the backtrace and that the NMP module was involved in a Panic.

####-##-##T##:##:##.###Z cpu3:2098063)WARNING: ScsiPath: 7685: Plugin 'NMP' had an error (Failure) while claiming path 'vmhba3:C0:T#:L##'. Skipping the path.
####-##-##T##:##:##.###Z cpu3:2098063)ScsiClaimrule: 2041: Plugin NMP specified by claimrule ##### was not able to claim path vmhba3:C0:T#:L##: Busy
####-##-##T##:##:##.###Z cpu3:2098063)ScsiClaimrule: 2545: Error claiming path vmhba3:C0:T#:L##. Failure.
####-##-##T##:##:##.###Z cpu3:2098063)WARNING: NMP: nmp_PspClaimPath:142: Claim of path 'vmhba3:C0:T#:L##' by plugin VMW_PSP_RR for device 'naa.################################' failed. Failure
####-##-##T##:##:##.###Z cpu3:2098063)WARNING: ScsiPath: 7685: Plugin 'NMP' had an error (Failure) while claiming path 'vmhba3:C0:T#:L##'. Skipping the path.
####-##-##T##:##:##.###Z cpu3:2098063)ScsiClaimrule: 2041: Plugin NMP specified by claimrule ##### was not able to claim path vmhba3:C0:T#2:L##: Busy
####-##-##T##:##:##.###Z cpu3:2098063)ScsiClaimrule: 2545: Error claiming path vmhba3:C0:T#:L##. Failure.
####-##-##T##:##:##.###Z cpu3:2098063)WARNING: NMP: nmp_PspClaimPath:142: Claim of path 'vmhba3:C0:T#:L##' by plugin VMW_PSP_RR for device 'naa.################################' failed. Failure
####-##-##T##:##:##.###Z cpu3:2098063)WARNING: ScsiPath: 7685: Plugin 'NMP' had an error (Failure) while claiming path 'vmhba3:C0:T#:L##'. Skipping the path.
####-##-##T##:##:##.###Z cpu3:2098063)ScsiClaimrule: 2041: Plugin NMP specified by claimrule ##### was not able to claim path vmhba3:C0:T#:L##: Busy
####-##-##T##:##:##.###Z cpu3:2098063)ScsiClaimrule: 2545: Error claiming path vmhba3:C0:T#:L##. Failure.
####-##-##T##:##:##.###Z cpu3:2098063)WARNING: NMP: nmp_PspClaimPath:142: Claim of path 'vmhba3:C0:T#:L##' by plugin VMW_PSP_RR for device 'naa.################################' failed. Failure
####-##-##T##:##:##.###Z cpu0:2098287)World: 3151: PRDA 0x420040000000 ss 0x0 ds 0x103 es 0x103 fs 0x103 gs 0x0
####-##-##T##:##:##.###Z cpu0:2098287)World: 3153: TR 0x758 GDT 0x45398000##88 (0xffff) IDT 0x420039169000 (0xffff)
####-##-##T##:##:##.###Z cpu0:2098287)World: 3155: CR0 0x80##0031 CR3 0x802107a000 CR4 0x142768
####-##-##T##:##:##.###Z cpu0:2098287)Backtrace for current CPU #0, worldID=#######, fp=0x43146b0dee88
####-##-##T##:##:##.###Z cpu0:2098287)0x453a2291bf50:[0x420039f03431][e-mail]#v2_12_0_0+0x9d stack: 0x0, 0x420039f03394, 0x21, 0x0, 0x0
####-##-##T##:##:##.###Z cpu0:2098287)0x453a2291bfa0:[0x42003913a234]vmkWorldFunc@vmkernel#nover+0x31 stack: 0x42003913a230, 0x0, 0x453a2291f000, 0x453a2189f100, 0x453a2291f100
####-##-##T##:##:##.###Z cpu0:2098287)0x453a2291bfe0:[0x42003942c##5]CpuSched_StartWorld@vmkernel#nover+0xe2 stack: 0x0, 0x4200390dbe00, 0x0, 0x0, 0x0
####-##-##T##:##:##.###Z cpu0:2098287)0x453a2291c000:[0x4200390dbdff]Debug_IsInitialized@vmkernel#nover+0xc stack: 0x0, 0x0, 0x0, 0x0, 0x0
####-##-##T##:##:##.###Z cpu0:2098287)VMware ESXi 8.0.2 [Releasebuild-22380479 x86_64]
#PF Exception 14 in world #######:nmp_TriggerF IP 0x420039f03431 addr 0x48
PTEs:0x0;

Module(s) involved in panic: [nmp Version 1.0.0-0, Built on: Sep  4 2023]
####-##-##T##:##:##.###Z cpu0:2098287)cr0=0x80##003d cr2=0x48 cr3=0x602000 cr4=0x14216c
####-##-##T##:##:##.###Z cpu0:2098287)FMS=06/55/7 uCode=0x5003604

Cause

When a device already has the maximum amount of supported paths claimed, new path claiming will not work, and may lead to a PSOD

Resolution

Ensure that the host does not have more than 32 paths to a device as any additional path claims may lead to a PSOD.