Only on one controller the following warning is logged every minute (and only referencing the one IP address of a vSphere 4.1 vmkernel port used for NFS)
Tue Jan 22 11:30:13 NZDT [FILER01: nfsd.tcp.close.idle.notify:warning]: Shutting down idle connection to client (xxx.xxx.xxx.xxx) where receive side flow control has been enabled. There are 0 bytes in the receive buffer.
Tue Jan 22 11:31:13 NZDT [FILER01: ems.engine.inputSuppress:warning]: Event 'nfsd.tcp.close.idle.notify' suppressed 4 times since Tue Jan 22 11:30:13 NZDT 2013.
The customer has modified the flow control settings on the vSphere vmnics.
Flow control settings are as follows for this host (note I'm aware they're not configured the same) - these flow control settings are configured the same on the other vSphere hosts - the other hosts don't have any issues (in fact I'm not aware of any issues with the host in question except the constant warnings being logged on this single controller)
Pause parameters for vmnic1:
Pause parameters for vmnic4:
Flow control is set to full on both controllers
I haven't found any info in the hosts logs in regards to these warnings unless I'm looking in the wrong logs (have checked hostd, messages etc)
Flow control on the Cisco 3750X switch ports are configured as follows for the vSphere hosts (note I'm not a network engineer - I believe these are default?)
Port Send FlowControl Receive FlowControl RxPause TxPause