iSCSI rescan takes longer time than expected when using port binding with a single VMNIC

Products

• VMware vSphere ESXi

Issue/Introduction

• You are port binding the same vmknic on two different subnets to access a iSCSI LUN, and the rescan takes longer than expected to complete.
• The /var/log/vmkiscsid.log file (ESX host) or /var/log/messages log file (ESXi) contains an error similar to:
  
  Reason: 00080000 (Initiator Connection Failure)

Environment

• VMware ESXi 4.1.x Installable
• VMware ESX 4.0.x
• VMware ESXi 4.0.x Installable
• VMware ESXi 4.1.x Embedded
• VMware ESXi 4.0.x Embedded
• VMware ESX 4.1.x

Resolution

This is expected behavior in this configuration. When using port binding, the iSCSI Initiator tries to open a full-mesh configuration and attempts to open a connection from each vmknic to each array address, which may result with a few connection failures causing rescan to take a while.

 

Notes:

• This issue affects only the rescan time and does not affect the post-rescan performance.
• As a rescan occurs during ESX/ESXi host boot, host start-up time can be affected.

Additional Information

Considerations for using software iSCSI port binding in ESX/ESXi