LFI Problems

Unanswered Question
May 9th, 2009

Hi

I tried to configure LFI but seemingly it's not working as per expectation. When I add frame-relay traffic shaping on my interface I am encountering drops on my interface and a huge number of drops. according to cisco QoS book they say i need to configure Frame-relay traffic shapping for my lfi to be effective but on this url http://blog.sazza.de/?p=433 this is what they say "Note: For interleaving to work, both fragmentation and the low-latency queuing policy must be configured with shaping disabled." This is confusing me.

My other problem is I am sending 1Mbps of traffic and I am receiving 0.16Mbps.

I have attached my configs..

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Giuseppe Larosa Sat, 05/09/2009 - 22:53

Hello Lawrence,

I see the following:

out fragmented pkts 0

in fragmented bytes 0 out fragmented bytes 0

in un-fragmented pkts 3 out un-fragmented pkts 52290

in un-fragmented bytes 252 out un-fragmented bytes 2614602

packets smaller then 160 bytes are not fragmented and these packets have an average size of 50 bytes so no packets was fragmented.

There are however different ways to do this the more modern one doesn't use FRTS at all.

in your case you have set

frame-relay cir 124872

frame-relay bc 1250

frame-relay mincir 124872

you are shaping at 124872 bps << 1000000 bps

then if you send 1Mbps over it with some traffic generator the router tries to buffer traffic until possible then when the shaping queues are full you get tail drops.

I think what you see is correct.

Hope to help

Giuseppe

mailaglady2 Sat, 05/09/2009 - 23:23

Hi Giuseppe

"packets smaller then 160 bytes are not fragmented and these packets have an average size of 50 bytes so no packets was fragmented."

How did you get the average size of 50 bytes?

another thing, I used the same configs on the sub interface for the same router, but this interface was shut.

I have attached the sh command output.

Giuseppe Larosa Sat, 05/09/2009 - 23:32

Hello Lawrence,

I made a guess making 2614602 / 52290

that is output bytes / output packets number.

The point is that FRTS config applies to all traffic for this reason you see drops when transmitting at 1 Mbps for long time.

I've found a link to the modern all MQC approach to FRTS.

see

http://www.cisco.com/en/US/docs/ios/wan/configuration/guide/wan_mqc_fr_tfshp.html

Edit:

in the new attachment you see some fragmented packets on device 1841.

C1841#sh frame-relay fragment interface s0/1/1.100 140

fragment size 160 fragment type end-to-end

in fragmented pkts 48 out fragmented pkts 113

in fragmented bytes 5808 out fragmented bytes 17039

in un-fragmented pkts 251270 out un-fragmented pkts 327

in un-fragmented bytes 12565845 out un-fragmented bytes 20058

in assembled pkts 251286 out pre-fragmented pkts 368

in assembled bytes 12571397 out pre-fragmented bytes 36501

in dropped reassembling pkts 0 out dropped fragmenting pkts

Hope to help

Giuseppe

Actions

This Discussion