The time user gives as the start of this problem is about the same
date/time I upgraded to 3.3.2b4 at Livingston's advice.
User is running Windoze 3.1, Trumpet something, and BitSurfr Pro,
dialing MLPPP to my 5-BRI card. He was unable to retrieve HTTP
documents, or send SMTP mail. Could get POP and FTP OK. Ptrace
and tcpdump output showed only a couple of TCP packets at the beginning
of the connection ever made it back to the client. I saw
that ifconfig on the PM showed an MTU of 1500, and remembered similar
weirdness that was solved by reducing the MTU.
Trouble was...we *couldn't*!
He changed the MTU in Trumpet to no effect. I configured a user
entry in RADIUS for him, and specified Framed-MTU = 1006, but it
*still* ended up at 1500. Only after we switched to V.120 were we
able to effect a change in MTU.
Question: Is there something different about negotiating/configuring
MTU with ppp async-to-sync and/or RFC1717 MLPPP that I'm missing?
Is this configured in the T/A (e.g. BitSurfr) for these Rate Adaptations?
Note that our RADIUS had been hacked before I got here to do PPP with
a prefix instead of PAP, so I don't know if that's related. But I
saw other users on other PMs running analog PPP with different MTUs so
I think RADIUS is not the problem here.
This is weird....thanks for any help....
-- David Carmean WB6YZM DC574 <dlc@silcom.com> System/Network Administration, Silicon Beach Communications Unsolicited commercial e-mail not accepted. Violators will be prosecuted.