Target ports fail to reconnect after a link disruption at target end.

Products

• VMware vSphere ESXi

Issue/Introduction

Symptoms:

On a server running with Emulex FCoE HBAs and the brcmfcoe driver, issues related to the initiator port failing to reconnect with the target port after a target link is up might occur.

2021-11-09T18:50:44.470Z cpu21:2097809)brcmfcoe: lpfc_els_rcv_rscn:5945: 0:(0):0214 RSCN received Data: x810220 x0 x4 x1
2021-11-09T18:50:44.470Z cpu21:2097809)brcmfcoe: lpfc_els_rcv_rscn:5955: 0:(0):5973 RSCN received event x0 : Address format x00 : DID x536120
2021-11-09T18:50:51.823Z cpu21:2097809)brcmfcoe: lpfc_els_rcv_rscn:5945: 0:(0):0214 RSCN received Data: x810260 x1 x4 x1
2021-11-09T18:50:51.823Z cpu21:2097809)brcmfcoe: lpfc_els_rcv_rscn:5955: 0:(0):5973 RSCN received event x0 : Address format x00 : DID x536120
2021-11-09T18:50:54.472Z cpu21:2097809)WARNING: brcmfcoe: lpfc_dev_loss_tmo_handler:254: 0:(0):0203 Devloss timeout on WWPN 22:01:00:02:##:##:##:## NPort x536120 Data: x8 x8 x25
2021-11-09T18:50:54.472Z cpu21:2097809)WARNING: brcmfcoe: lpfc_dev_loss_tmo_handler:296: 0:(0):3298 ScsiNotifyPathStateChangeAsyncSAdapter Num x1 TID x20, DID x536120.
2021-11-09T18:50:54.472Z cpu0:2097971)NMP: nmp_ThrottleLogForDevice:3867: Cmd 0xa3 (0x45b9c158bb48, 0) to dev "naa.60002ac000000000000000be0001cdb3" on path "vmhba0:C0:T32:L190" Failed:
2021-11-09T18:51:12.962Z cpu21:2097792)brcmfcoe: lpfc_rportStats:4197: 0:(0) Compression log for fcp target 32, path is dead, IO errors: busy 0, retry 0, no_connect 2, fcperr 0, tmo 0
2021-11-09T18:51:12.962Z cpu21:2097792)brcmfcoe: lpfc_rportStats:4222: 0:(0) Compression log for fcp target 32, path is dead, Events: npr=2, devloss=1

Note: The preceding log excerpts are examples. The date, time, FC_ID, and environmental variables might vary depending on the environment.

 

Environment

VMware vSphere 7.0.x
VMware vSphere 6.x

Cause

When an RSCN is received, the brcmfcoe driver discovery state machine might fail to reconnect with the target FC_ID.

Resolution

Currently there is no resolution.

Workaround:

On the initiator port that failed to reconnect, issue the following Loop Initialization Primitive (LIP) command:

"esxcli storage san fc reset -A vmhbaXX"