[c-nsp] L2TP DISCONNECT caused by LCP mismatch

[Oliver Boehmer (oboehmer)]

ahmad fadly <> wrote on Thursday, June 19, 2008 8:27 AM:

> Hi,
> 
> I'm currently setting up L2TP between LNS and PDSN,
> when I dial to the PDSN the dial time sometimes takes
> a long time to be connected. I debug ppp negotiation
> and debug vpdn l2x-events, they show that the LCP
> negotiation is having very long time negotiation but
> sometimes is not. The message shows to be like this,
> ppp807 LCP: Discarding: CP_CONFACK does not match
> CP_CONFREQ.

Hmm.. the LNS sends:

> 22:43:57: ppp807 LCP: O CONFREQ [Stopped] id 1 len 14
> 22:43:57: ppp807 LCP:    AuthProto PAP (0x0304C023)
> 22:43:57: ppp807 LCP:    MagicNumber
> 0x08814572(0x050608814572)

and the client replies

> 22:43:57: ppp807 LCP: I CONFACK [ACKsent] id 3 len 18
> 22:43:57: ppp807 LCP:    ACCM 0x00000000 (0x020600000000)

> 22:43:57: ppp807 LCP:    AuthProto PAP (0x0304C023)
> 22:43:57: ppp807 LCP:    PFC (0x0702)
> 22:43:57: ppp807 LCP:    ACFC (0x0802)

The client would be violating RFC1661 by ACK'ing options which haven't
been requested, but looking at the ID (id 3) it seems that this ACK is
from the CONFREQ sent by the LAC? I ignore the possibility of the client
being completely broken ;-)
It looks like some LCP negotation problem on the LAC. Is LCP completely
negotiated on the LAC?

You've enabled "lcp renegotiation on-mismatch" in the vpdn-group. Can
you try removing this or changing it to "lcp renegotiation always" (you
need to re-establish the L2TP tunnel in order for the new settings to
take effect, i.e. clear the l2tp tunnel after the change)?

	oli