Subject: ppp problems
To: None <port-i386@NetBSD.ORG>
From: Kent Polk <email@example.com>
Date: 10/11/1999 18:26:06
Last week I upgraded one of my two I386 NetBSD hosts to 1.4.1 from
1.3.3 (the other one I already upgraded).
This host acts as a gateway between a ppp connection and a local
subnet. After upgrading to 1.4.1, the ppp connection started
experiencing problems with socket connections. The link gradually
degrades from experiencing longer and longer timeouts, to DNS host
lookup failures, until it cannot perform any connections in either
The only network configuration change between 1.3.3 and 1.4.1 was
that the local ethernet card changed name from ep0 to ex0 and I
copied the ifconfig.ep0 file to ifconfig.ex0.
A few things I noticed:
1) netstat responds terribly slowly just after a reboot. As time
goes by, it simply hangs and I cannot retrieve any information to
give me a clue as to what the problem is. When it does respond,
everything appears to be configured correctly.
2) when the nfs server is started on boot, it now initiates the
ppp link. It didn't do this with 1.3.3.
3) This behavior occurs whether I have named running or not.
I thought maybe something had changed with named and it wasn't
working correctly. I used to get spurious errors with 1.3.3 that
I didn't have named set up correctly, but I couldn't see what was
wrong with it and DNS worked as expected... But even with named
turned off the link problem still occurs.
I compiled a new kernel from the 1.4.1 distribution with 'gateway'
turned on and the changes made to support a PS/2 mouse.
Any suggestions as to what to look for?
Startup - Routing tables
Destination Gateway Flags Refs Use Mtu Interface
default (ppp server) UGS 6 4823 - ppp0
localhost localhost UH 0 46 - lo0
(ppp server) (ppp client) UH 1 10 - ppp0
(subnet address) link#1 UC 0 0 - ex0