Strange multilink issue

Unanswered Question
Oct 23rd, 2009

Summary of the issue is that we replaced a customer's T1 with a circuit with 2 bonded T1s and there were multiple issues encountered. Webpages would only partially load, users reported slower internet performance, and multiple POP3 download issues. Returning the circuit to the single T1 immidiately resolved the issue. The configuration is attached to the post. One strange symptom is that when we removed the "ppp multilink interleave" command users could not load webpages, this leads me to believe there is a multilink configuration issue. If anyone can provide some assistance, it would be greatly appreciated. Thank you.

Attachment: 
I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Jerry Ye Fri, 10/23/2009 - 18:31

Can you post the output of show interface s0/2/0 and show interface s0/3/0?

Regards,

jerry

johnny.schultz Mon, 10/26/2009 - 07:51

Here you go...

Voce_BSI02#show int S0/2/0

Serial0/2/0 is up, line protocol is up

Hardware is GT96K with integrated T1 CSU/DSU

Description: 360 Order# 601738

MTU 1500 bytes, BW 1544 Kbit, DLY 20000 usec,

reliability 255/255, txload 1/255, rxload 1/255

Encapsulation PPP, LCP Open, multilink Open

Link is a member of Multilink bundle Multilink1, loopback not set

Keepalive set (10 sec)

Restart-Delay is 0 secs

Last input 00:00:00, output 00:00:00, output hang never

Last clearing of "show interface" counters 3d22h

Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 20

Queueing strategy: fifo

Output queue: 0/40 (size/max)

30 second input rate 4000 bits/sec, 9 packets/sec

30 second output rate 3000 bits/sec, 8 packets/sec

4617298 packets input, 1478569214 bytes, 0 no buffer

Received 0 broadcasts, 0 runts, 0 giants, 0 throttles

2 input errors, 2 CRC, 0 frame, 0 overrun, 0 ignored, 0 abort

4577519 packets output, 1094659099 bytes, 0 underruns

0 output errors, 0 collisions, 2 interface resets

0 output buffer failures, 0 output buffers swapped out

5 carrier transitions

DCD=up DSR=up DTR=up RTS=up CTS=up

Voce_BSI02#show int S0/3/0

Serial0/3/0 is up, line protocol is up

Hardware is GT96K with integrated T1 CSU/DSU

Description: 360 Order# 601739

MTU 1500 bytes, BW 1544 Kbit, DLY 20000 usec,

reliability 255/255, txload 1/255, rxload 1/255

Encapsulation PPP, LCP Open, multilink Open

Link is a member of Multilink bundle Multilink1, loopback not set

Keepalive set (10 sec)

Restart-Delay is 0 secs

Last input 00:00:00, output 00:00:00, output hang never

Last clearing of "show interface" counters 3d22h

Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 20

Queueing strategy: fifo

Output queue: 0/40 (size/max)

30 second input rate 4000 bits/sec, 10 packets/sec

30 second output rate 5000 bits/sec, 10 packets/sec

4610939 packets input, 1471322904 bytes, 0 no buffer

Received 0 broadcasts, 0 runts, 0 giants, 0 throttles

0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored, 0 abort

4574250 packets output, 1094684298 bytes, 0 underruns

0 output errors, 0 collisions, 2 interface resets

0 output buffer failures, 0 output buffers swapped out

5 carrier transitions

DCD=up DSR=up DTR=up RTS=up CTS=up

Voce_BSI02#

Joseph W. Doherty Fri, 10/23/2009 - 19:02

You might try setting both "ip load-sharing per-destination" and "fair-queue" under each serial interface, remove "ppp multilink interleave" under the multilink interface, and change "ip route 0.0.0.0 0.0.0.0 Multilink1" to the other multilink side's IP address.

royalblues Fri, 10/23/2009 - 23:57

Joseph

Does the load-sharing command on the serial interfaces has any effect on multilink interfaces? . I had always thought multilink has its own algorithm to distribute packets across links

Johny

You might try disabling multilink fragmentation

ppp multilink fragment disable

Narayan

Joseph W. Doherty Sat, 10/24/2009 - 04:13

Narayan,

"Does the load-sharing command on the serial interfaces has any effect on multilink interfaces?"

I wouldn't think so, but it's unusual. That, and along with documentation that mentions the need for FQ for some multilink operations, is why I suggested it's removal.

"You might try disabling multilink fragmentation "

Good suggestion! Something that I had thought to suggest and then promptly forgot to add to my post. In theory (I believe), it would offer slight less throughput to any one flow, but if a flow's packets are alternated between links, should make little net difference to throughput although should reduce performance impact of the fragmentation on both ends of the multilink.

johnny.schultz Mon, 10/26/2009 - 07:51

Thanks for your suggestions, I will try those commands the next time I am on site since I locked myself out of the router once when I entered in the ppp multilink interleave command. Have you or anyone else had these strange symptoms when working with multilink interfaces? Thanks again for your help.

Actions

This Discussion