cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
6342
Views
0
Helpful
8
Replies

what is the description of TCP_RETRANS_RATE_XCD_THRESH on 9250i

huang chen
Level 1
Level 1
8 Replies 8

Sangeeta Patil
Level 1
Level 1

Even I am looking for an explanation of the events that we see in "show logging onboard error-stats" for FCIP connections.

There is no explanation of these alerts in any Cisco docs.

There is a bug: CSCvy14809  mentioned for alert: TCP_SRTT_XCD_CONF_RTT where configured RTT value exceeds then we get to see this alert. But what does it means and what needs to be concluded is not mentioned.

 

I was advised by Cisco TAC recently that TCP_RETRANS_RATE_XCD_THRESH is a message indicated in the last 60 seconds the packet retransmit rate was X% with Y amount of data segments retransmitted.   By default the retransmit threshold is 0.05%.

 

Many thanks for the update @dlt113.

So as per the below event in my case the default threshold of 0.05% has exceeded and it has logged an event in logging onboard errors-stats:

fcip21 |TCP_RETRANS_RATE_XCD_THRESH |0.067 |91 |08/23/21 04:15:39

What could be the cause of the issue here? Is it switch/network which needs to be checked or are they just informational?

 

 

Also need to know how to fix for these alerts "TCP_SRTT_XCD_CONF_RTT".

In my experience the most likely cause of the TCP_RETRANS_RATE_XCD_THRESH is due to the IP network having trouble delivering packets to the distant end.  If the packet's ACK does not return within a given time (or arrives too far out of order) the Transmit side will assume the packet was dropped somewhere and retransmit it.  The higher the packet retransmits rate the less overall effective circuit throughput there will be. 

 

For defect CSCvy14809  there is no listed NX-OS fixes.  I would contact Cisco TAC on the fix ETA.

I agree with what to say.
Many thanks for the update @dlt113 

Branden Varney
Cisco Employee
Cisco Employee

To have a stable FCIP connection we have indicated that we like to see the connection have less than .05% re-transmit rate.

While it is possible this could be an MDS platform issue, the majority of root cause to re-transmissions are due to the IP network transporting the FC payload.

There could be an issue with an MTU value, bad cable, etc that could be causing this to happen. We typically investigate the IP network or anything in the middle that would be leading to re-transmits and drops of network traffic.

Her is some documentation tp help with configuration and fine tuning and things to consider:

https://www.cisco.com/c/en/us/td/docs/switches/datacenter/mds9000/sw/8_x/config/ip_services/ipsvc/cfcip.html#67649

"When FCIP is sent over a WAN link, the default TCP settings may not be appropriate. In such cases, we recommend that you tune the FCIP WAN link by modifying the TCP parameters (specifically bandwidth, round-trip times, and CWM burst size)."

 

 

 

If you do not find any root cause and continue to see TCP re-transmits please open a case with the Cisco TAC to assist you further.

 

@Branden Varney 
I appreciate the explanation provided. 

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: