Problems with 2xE1 Multilink PPP bundle

[oj88]

I've gone as far as configuring MLP using two 2048kbps serial links. In my setup, s0/1 and s0/2 are members of Multilink 2. The problem is, s0/2 goes protocol up and down every several seconds to a minute. My understanding of MLP is that the flapping serial link will be removed from the bundle and let the traffic go through the remaining stable link (s0/1 in this case). But in my scenario, even with s0/2's status is protocol down, if I do a sh ppp multilink, I can see that Multilink 2 treats s0/2 as "active". Being so, I get all kinds of things from debug ppp multilink events.

But if I shutdown s0/2, the problem would disapear.

Please see below:

R1 and R2's MLP configs are mirrors of each other. The following pretty much sums up how both of them are configured:

interface Multilink2

ip address x.x.x.x

ppp multilink

ppp multilink group 2

!

interface Serial0/1

bandwidth 2048

no ip address

encapsulation ppp

ppp multilink

ppp multilink group 2

!

interface Serial0/2

bandwidth 2048

no ip address

encapsulation ppp

ppp multilink

ppp multilink group 2

!

==============================

sh inter s0/2

Serial0/2 is up, line protocol is down

Hardware is GT96K Serial

Description: Tiwi-Digitel

MTU 1500 bytes, BW 2048 Kbit, DLY 20000 usec,

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

Encapsulation PPP, LCP Open, multilink Open, loopback not set

.

.

==============================

sh ppp multilink

Multilink2, bundle name is BayArea01

Bundle up for 00:02:00, 2/255 load

Receive buffer limit 24384 bytes, frag timeout 1000 ms

0/0 fragments/bytes in reassembly list

395 lost fragments, 2677 reordered

38/51687 discarded fragments/bytes, 0 lost received

0x1B51 received sequence, 0x1E2A sent sequence

Member links: 2 active, 0 inactive (max not set, min not set)

Se0/1, since 00:00:23

Se0/0, since 00:01:58

==============================

debug ppp multilink events

Sep 7 16:09:22 PH: Mu2 MLP: End bit lost, discard reassembled packet

Sep 7 16:09:22 PH: Mu2 MLP: Lost fragment 19A5 (all links out of sequence)

Sep 7 16:09:22 PH: Mu2 MLP: End bit lost, discard reassembled packet

Sep 7 16:09:22 PH: Mu2 MLP: Begin bit lost, discard fragment 19A8

Sep 7 16:09:22 PH: Mu2 MLP: Lost fragment 19AB (all links out of sequence)

Sep 7 16:09:22 PH: Mu2 MLP: Begin bit lost, discard fragment 19AC

Sep 7 16:09:22 PH: Mu2 MLP: Lost fragment 19B1 (all links out of sequence)pp mul

Sep 7 16:09:22 PH: Mu2 MLP: End bit lost, discard reassembled packet

Sep 7 16:09:22 PH: Mu2 MLP: Lost fragment 19B7 (all links out of sequence)

Sep 7 16:09:22 PH: Mu2 MLP: End bit lost, discard reassembled packet

Sep 7 16:09:22 PH: Mu2 MLP: End bit lost, discard reassembled packet

Sep 7 16:09:22 PH: Mu2 MLP: Lost fragment 19BF (all links out of sequence)

Sep 7 16:09:22 PH: Mu2 MLP: Begin bit lost, discard fragment 19C0

Sep 7 16:09:22 PH: Mu2 MLP: Lost fragment 19C3 (all links out of sequence)

Sep 7 16:09:22 PH: Mu2 MLP: Begin bit lost, discard fragment 19C4

Sep 7 16:09:22 PH: Mu2 MLP: Lost fragment 19D3 (all links out of sequence)

Sep 7 16:09:22 PH: Mu2 MLP: End bit lost, discard reassembled packet

Sep 7 16:09:23 PH: Mu2 MLP: Lost fragment 19E9 (all links out of sequence)

Sep 7 16:09:23 PH: Mu2 MLP: End bit lost, discard reassembled packet

Sep 7 16:09:23 PH: Mu2 MLP: Lost fragment 19ED (all links out of sequence)

Sep 7 16:09:23 PH: Mu2 MLP: End bit lost, discard reassembled packet

.

.

(and several of the following)

Sep 7 16:07:33 PH: Se0/1 MLP: cdp packet forwarded to wrong interface

Sep 7 16:07:33 PH: Se0/1 MLP: cdp packet forwarded to wrong interface

Sep 7 16:07:33 PH: Se0/1 MLP: cdp packet forwarded to wrong interface

.

.

Also, if Multilink 2 balks (as it occassionally does), I have to do a clear interface multilink 2 to restore the link.

Any help is highly appreciated.

[vijayasankar]

Hi,

The feature that you are insisting is called Link quality monitoring in PPP.

Unfortunately this feature is not supported for multilink bundles.

You can enable LQM only on individual PPP links, to achieve automatic link monitoring.

This will not work with multilink bundles.

Here's the URL

http://www.cisco.com/en/US/products/sw/iosswrel/ps1835/products_configuration_guide_chapter09186a00804f2079.html#wp1001080

Check out this thread in the forum.

http://forums.cisco.com/eforum/servlet/NetProf?page=netprof&forum=Network%20Infrastructure&topic=WAN%2C%20Routing%20and%20Switching&CommCmd=MB%3Fcmd%3Dpass_through%26location%3Doutline%40%5E1%40%40.1ddbad08/2#selected_message

I would suggest you to perform a complete test with the flapping link with the help of the service provider and rectify the problem on this link.

Hope this helps.

-VJ

[oj88]

But the thing is, why is Multilink 2 still using s0/2 even if it's protocol down? If I shutdown s0/2, traffic becomes normal after I do a clear interface multilink 2.

Rectifyig this to the provider is SOP. But in the meantime, I want the router to use the good link with minimum or no human intervention. That, I believe is the real purpose of having two redundant links in the first place.

Speaking of LQM, can I configure ppp quality 80 on s0/1 and s0/2, knowing that they're members of Multilink 2?

[vijayasankar]

Hi,

If the line protocol goes down, the multilink shouldn't use the failed link. It should be fault tolerant in that perspective. Not sure whether you are encountering any bug in IOS.

You can check on that.

As the LQM is not supported for multilink, you cannot use it currently in the s0/1 and s0/2.

If you are removing multilink and configuring the s0/1 and s0/2 as normal links then you can use the LQM.

-VJ