[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
DSL Internet Connection Question
- To: misc_(_at_)_openbsd_(_dot_)_org
- Subject: DSL Internet Connection Question
- From: Robert C Wittig <wittig_(_dot_)_robert_(_at_)_sbcglobal_(_dot_)_net>
- Date: Wed, 21 Dec 2005 11:31:07 -0600
- Reply-to: Robert C Wittig <wittig_(_dot_)_robert_(_at_)_sbcglobal_(_dot_)_net>
I have a fair amount of experience networking in Windows 2000, Linux
Red Hat, and FreeBSD, but I am a relative newbie to OpenBSD, just
setting up my first practice machine, which is still basically a
typical default installation.
I have an SBC DSL connection, non-static IP (Netopia Cayman
Modem/Firewall/Router), and currently am running Windows 2000, Linux
Red Hat Enterprise 3, and OpenBSD 3.7.
Win 2000 and Red Hat are up and running fine. All I had to do with RH,
when switching from my old dial-up account to DSL, was to tell it to use
I did the same with OpenBSD (added 'dhcp' to my
/etc/hostname.interface file (real name = /etc/interface.dc0 ), and
when I rebooted, I did immediately have Internet access with lynx,
whois, etc., but it was extremely s-l-o-w... taking about one minute
before returning a webpage. On a couple whois queries, I also noticed
that the data arrived in two 'chunks'...wait a minute... half the
info arrived, then a minute later, the other half arrived.
This symptom seemed to me, to be unique enough, so that someone might be
able to quickly identify what might be hanging, then timing out, and
then succeeding with whatever 'else' alternate is actually successful.
It is worth mentioning that I am running Red Hat and OpenBSD from the
same computer, using 2 different hot-swap hard drives, running from the
same primary drive bay, so both OS's are using the exact same hardware,
except, of course, for the hard drives themselves.
Traffic inside my LAN (I have an IIS server up on Windows 2000, and
the Apache server up on OpenBSD) is quick.
I have further isolated the problem...
First, I tried accessing my website (not on my LAN, on a leased box
about 1,000 miles away) by name (very s-l-o-w... about a minutes wait,
before presenting the page) and then by IP addy (fast like a rocket,
the way things should run), which clued me that it was some sort of
domain name resolution issue.
The IP addy that DHCP provides OpenBSD in DHCPOFFER during bootup, is
the IP addy that my Modem/Router describes as my LAN IP, and this IP
addy is what then gets written to /etc/resolv.conf with the label
This is exactly the same thing that happens in Red Hat, but in Red
Hat, it works fine, whereas in OpenBSD, I have the delayed response
I re-wrote /etc/resolv.conf with the nameserver values that my
Modem/Router describes as DNS-1 and DNS-2, and as soon as I saved the
file, lynx, whois, etc., worked fast like a rocket, when domain names
were used, so obviously, this 'fixed' the problem.
So I wrote a shell script that, when run, will re-write
/etc/resolv.conf with the current working DNS-1 and DNS-2 values, that
I can run after boot, to replace the faulty info that is being place
in /etc/resolv.conf, during bootup... but the 'fix' only works for 30
minutes, at which time the resolv.conf file is over-written, I would
assume, by DHCP, and everything reverts to the delay of about 1
minute, for domain name resolution.
This is a lame solution, but better than nothing, for the moment, but
I definitely have to figure out what actually needs fixing. My best
guess is that the 'one minute hang' is because something, in one of
the config files, should not be where it is, but there are a lot of
variables to play with, and I don't want to just continue blindly
messing with things, which I have been trying for the last couple
days, and having them not help, or make things worse, and then
restoring them to their original state.
Thanks in advance, for any suggestions, comments, etc.
Visit your host, monkey.org