Jonas,
Thank you, thank you, thank you! Changing it to
1500 and rebooting cured the problem.
I had it set to 296, a figure recommended by many
references [including RFC 1191] on TCP/IP as a
good tradeoff between latency and throughput
for PPP connections.
I presume this is a problem in ip_masq's IP
fragmentation handling. Do you know if 1500
is sacred to PPP, or is the requirement simply
that the MTU/MRU settings of eth0 and ppp0
must be identical?
[ note to maintainers: This is an AWFUL bug.
RFC 1191 is very clear about independence
of path MTU's ]
Mike
-----Original Message-----
From: Jonas Storholm <[EMAIL PROTECTED]>
To: Michael Ellis <[EMAIL PROTECTED]>; [EMAIL PROTECTED]
<[EMAIL PROTECTED]>
Date: Wednesday, April 14, 1999 2:45 AM
Subject: Re: [Masq] Baffled by HTTPS access problem
Just one thing, the MTU/MRU should be 1500 on the ppp link, it is right?
I had the same problem, but just not with HTTPS but everything on the
link,
changing the MTU solved it. Just a thought :)
Jonas
_______________________________________________
Masq maillist - [EMAIL PROTECTED]
http://tiffany.indyramp.com/mailman/listinfo/masq
Admin requests can be handled by web (above) or [EMAIL PROTECTED]
