This article explains the syslog message 'peer_input_pending_internal' that is observed on MX devices.

We might be seeing the below syslog messages from the backup RE, when peer-proxy connection went down.

Oct 24 09:57:57 MX104_RE1 /kernel: peer_input_pending_internal:[5251] VKS0 for peer type 7 indx 0 reported a sb_state 32 = SBS_CANTRCVMORE

Oct 24 09:57:57 MX104_RE1 /kernel: peer_input_pending_internal:[5251] VKS0 for peer type 7 indx 0 reported a sb_state 32 = SBS_CANTRCVMORE

Oct 24 09:57:57 MX104_RE1 /kernel: peer_inputs:5505 VKS0 closing connection peer type 7 indx 0 err 5

This kernel syslog message is reported by the function 'peer_input_pending_internal' which is a Junos internal function, coded to check if a remote PFE peer has pending input on the specified socket. 

It indicates that the peer-proxy connection with backup RE (type 7) went down. The primary RE kernel sends FIB state to Backup kernel over this connection and the connection tear down is mostly initiated by backup RE (per SBS_CANTRCVMORE flag). 

These logs would not affect any service.

When we see the above syslog messages, we must perform the following checks: 

• Is the Backup RE still active/healthy, or if it has any HW malfunctions? 
• Did the Backup RE experience high CPU or memory issue? 
• Was the Backup RE rebooted recently? 
• Does the Backup RE run on the same Junos Software code with Primary RE?