Mailing List Archive

What version are you running? Can you get "debug ppp multilink frag"
for one of the problem sessions? We recently added a knob to multilink
to allow you to tune our reassembly buffers (see CSCdv80109), however
we need to know how the fragments are being received in order to
determine if tuning will help and, if so, how to tune effectively.

Dennis

Voralt [peder@voralt.net] wrote:
> We have a PRI into a 3640 and are running 128k ISDN connections. When I
> look at "show ppp multi", I see the following. xxx has lots of lost
> fragments and discards and coincidentally, says their connection seems
> "slow". yyy says it is running fine. Another bit of information is that
> user xxx is on LEC A and all of the users on LEC A have lost fragments and
> discards and appear to be slow. yyy is on LEC B and all of the users on LEC
> b have no problems and no errors. We are using LEC C for the PRI. If we
> drop all of the users and make them dial back in, it still follows all of
> the users from LEC A. To me, it seems like an issue either inside LEC A's
> network, or between LEC A and LEC C, but I'm not sure exactly where. We've
> contacted LEC C and they say "we see no errors on the PRI, have a nice day".
> Any ideas where this error might be originating? I can't imagine that it is
> on the user end.
>
>
> Virtual-Access5, bundle name is xxx
> Bundle up for 2d22h
> Dialer interface is Serial0/1:23
> 16452 lost fragments, 53922 reordered, 0 unassigned
> 7296 discarded, 0 lost received, 25/255 load
> 0x45C2F0 received sequence, 0x385538 sent sequence
> Member links: 2 (max not set, min not set)
> Se0/1:4, since 2d22h, last rcvd seq 45C2EE
> Se0/1:0, since 2d22h, last rcvd seq 45C2EF
>
> Virtual-Access6, bundle name is yyy
> Bundle up for 2w5d
> Dialer interface is Serial0/1:23
> 0 lost fragments, 23 reordered, 0 unassigned
> 0 discarded, 0 lost received, 7/255 load
> 0x19352E received sequence, 0x166A7C sent sequence
> Member links: 2 (max not set, min not set)
> Se0/1:8, since 2w5d, last rcvd seq 19352C
> Se0/1:9, since 2w5d, last rcvd seq 19352D
>
>
> _______________________________________________
> cisco-nas mailing list
> cisco-nas@puck.nether.net
> http://puck.nether.net/mailman/listinfo/cisco-nas

--
-------------------------------------------------------------------------
|| || Dennis Peng
|| || Cisco Systems, Inc. Escalation Engineer
|||| |||| 170 West Tasman Drive Phone: (408) 526-6143
..:||||||:..:||||||:.. San Jose, CA 95134 Fax: (408) 232-2343
Cisco Systems Inc. dpeng@cisco.com
-------------------------------------------------------------------------

I don't know how much you want or what you need, so here is a chunk of the
debug. We are running 12.2(12) on a 3640. 0/1:0 and 0/1:4 are in a
multilink bundle and are having problems. 6&7 are in another bundle and
they are fine.

Jun 18 18:46:01: Se0/1:0 MLP: O frag 805C0E5A size 74 direct
Jun 18 18:46:01: Se0/1:4 MLP: O frag 405C0E5B size 77 direct
Jun 18 18:46:01: Se0/1:4 MLP: I frag 806C920A size 40 direct
Jun 18 18:46:01: Se0/1:0 MLP: I frag 406C920B size 41 direct
Jun 18 18:46:01: Se0/1:7 MLP: I frag C005331B size 50 direct
Jun 18 18:46:01: Se0/1:0 MLP: O frag 805C0E5C size 70 direct
Jun 18 18:46:01: Se0/1:4 MLP: O frag 405C0E5D size 71 direct
Jun 18 18:46:01: Se0/1:4 MLP: I frag 806C920C size 40 direct
Jun 18 18:46:01: Se0/1:0 MLP: I frag 406C920D size 41 direct
Jun 18 18:46:01: Se0/1:6 MLP: O frag 800DB348 size 758 direct
Jun 18 18:46:01: Se0/1:7 MLP: O frag 400DB349 size 760 direct
Jun 18 18:46:01: Se0/1:6 MLP: O frag 800DB34A size 758 direct
Jun 18 18:46:01: Se0/1:0 MLP: O frag 805C0E5E size 74 direct
Jun 18 18:46:01: Se0/1:4 MLP: O frag 405C0E5F size 77 direct
Jun 18 18:46:01: Se0/1:4 MLP: I frag C06C920E size 70 direct
Jun 18 18:46:01: Se0/1:6 MLP: O frag 400DB34B size 760
Jun 18 18:46:01: Se0/1:7 MLP: O frag 800DB34C size 758
Jun 18 18:46:01: Se0/1:7 MLP: O frag 400DB34D size 760
Jun 18 18:46:01: Se0/1:0 MLP: O frag C05C0E60 size 70 direct
Jun 18 18:46:01: Se0/1:0 MLP: I frag C06C920F size 62 direct
Jun 18 18:46:02: Se0/1:4 MLP: O frag C05C0E61 size 70 direct
Jun 18 18:46:02: Se0/1:4 MLP: I frag C06C9210 size 50 direct
Jun 18 18:46:02: Se0/1:0 MLP: O frag 805C0E62 size 122 direct
Jun 18 18:46:02: Se0/1:4 MLP: O frag 405C0E63 size 123 direct
Jun 18 18:46:02: Se0/1:6 MLP: O frag 800DB34E size 758
Jun 18 18:46:02: Se0/1:6 MLP: O frag 400DB34F size 760
Jun 18 18:46:02: Se0/1:7 MLP: O frag 800DB350 size 758
Jun 18 18:46:02: Se0/1:7 MLP: O frag 400DB351 size 760
Jun 18 18:46:02: Se0/1:6 MLP: I frag C005331C size 50 direct
Jun 18 18:46:02: Se0/1:4 MLP: I frag 806C9212 size 48 direct
Jun 18 18:46:02: Se0/1:0 MLP: I frag C06C9211 size 62 direct
Jun 18 18:46:02: Se0/1:0 MLP: I frag 406C9213 size 49 direct
Jun 18 18:46:02: Se0/1:0 MLP: O frag C05C0E64 size 62 direct
Jun 18 18:46:02: Se0/1:4 MLP: O frag 805C0E65 size 102 direct
Jun 18 18:46:02: Se0/1:0 MLP: O frag 405C0E66 size 105 direct

----- Original Message -----
From: "Dennis Peng" <dpeng@cisco.com>
To: "Voralt" <peder@voralt.net>
Cc: <cisco-nas@puck.nether.net>
Sent: Wednesday, June 18, 2003 6:34 PM
Subject: Re: [cisco-nas] Lost Fragments


> What version are you running? Can you get "debug ppp multilink frag"
> for one of the problem sessions? We recently added a knob to multilink
> to allow you to tune our reassembly buffers (see CSCdv80109), however
> we need to know how the fragments are being received in order to
> determine if tuning will help and, if so, how to tune effectively.
>
> Dennis
>
> Voralt [peder@voralt.net] wrote:
> > We have a PRI into a 3640 and are running 128k ISDN connections. When I
> > look at "show ppp multi", I see the following. xxx has lots of lost
> > fragments and discards and coincidentally, says their connection seems
> > "slow". yyy says it is running fine. Another bit of information is
that
> > user xxx is on LEC A and all of the users on LEC A have lost fragments
and
> > discards and appear to be slow. yyy is on LEC B and all of the users on
LEC
> > b have no problems and no errors. We are using LEC C for the PRI. If
we
> > drop all of the users and make them dial back in, it still follows all
of
> > the users from LEC A. To me, it seems like an issue either inside LEC
A's
> > network, or between LEC A and LEC C, but I'm not sure exactly where.
We've
> > contacted LEC C and they say "we see no errors on the PRI, have a nice
day".
> > Any ideas where this error might be originating? I can't imagine that
it is
> > on the user end.
> >
> >
> > Virtual-Access5, bundle name is xxx
> > Bundle up for 2d22h
> > Dialer interface is Serial0/1:23
> > 16452 lost fragments, 53922 reordered, 0 unassigned
> > 7296 discarded, 0 lost received, 25/255 load
> > 0x45C2F0 received sequence, 0x385538 sent sequence
> > Member links: 2 (max not set, min not set)
> > Se0/1:4, since 2d22h, last rcvd seq 45C2EE
> > Se0/1:0, since 2d22h, last rcvd seq 45C2EF
> >
> > Virtual-Access6, bundle name is yyy
> > Bundle up for 2w5d
> > Dialer interface is Serial0/1:23
> > 0 lost fragments, 23 reordered, 0 unassigned
> > 0 discarded, 0 lost received, 7/255 load
> > 0x19352E received sequence, 0x166A7C sent sequence
> > Member links: 2 (max not set, min not set)
> > Se0/1:8, since 2w5d, last rcvd seq 19352C
> > Se0/1:9, since 2w5d, last rcvd seq 19352D
> >
> >
> > _______________________________________________
> > cisco-nas mailing list
> > cisco-nas@puck.nether.net
> > http://puck.nether.net/mailman/listinfo/cisco-nas
>
> --
> -------------------------------------------------------------------------
> || || Dennis Peng
> || || Cisco Systems, Inc. Escalation Engineer
> |||| |||| 170 West Tasman Drive Phone: (408) 526-6143
> ..:||||||:..:||||||:.. San Jose, CA 95134 Fax: (408) 232-2343
> Cisco Systems Inc. dpeng@cisco.com
> -------------------------------------------------------------------------
>