Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 31 Aug 1997 17:33:20 +0100
From:      Brian Somers <brian@awfulhak.org>
To:        The Hermit Hacker <scrappy@hub.org>
Cc:        Brian Somers <brian@awfulhak.org>, hackers@FreeBSD.ORG
Subject:   Re: PPP 'Bursts' with newest 3.0-CURRENT... 
Message-ID:  <199708311633.RAA04047@awfulhak.demon.co.uk>
In-Reply-To: Your message of "Sun, 31 Aug 1997 04:51:08 -0300." <Pine.NEB.3.96.970831044025.12587x-100000@thelab.hub.org> 

next in thread | previous in thread | raw e-mail | index | archive | help
> On Sun, 24 Aug 1997, Brian Somers wrote:
> 
> > > 
> > > Hi...
> > > 	
> > > 	I'm really curious as to whether anyone has experience with this.  I'm 
> > > running 3.0-CURRENT, and have just upgraded to the newest source tree (as of
> > > today) in the hopes of reducing/eliminating the following 'hills and
> > > valleys':
> > > 
> > > =================
> > > # netstat -nr
> > > Routing tables
> > > 
> > > Internet:
> > > Destination        Gateway            Flags     Refs     Use     Netif Expire
> > > default            127.2.2.2          UGSc       12        0      tun0
> > > 127.0.0.1          127.0.0.1          UH          1       44       lo0
> > > 206.231.247.114    206.231.248.173    UH          0        0      tun0
> > 
> > Let me guess that your ppp.conf contains:
> > 
> >  set ifaddr x.x.x.x 127.2.2.2
> >  add 0 0 127.2.2.2
> > 
> > And your ppp.linkup *doesn't* contain
> > 
> >  delete all
> >  add 0 0 hisaddr
> 
> 	Added this, you were right, in that I was missing it for this ISP...
> had it for one of the other ones though :(
> 
> 	Still doesn't help though :(  
> 
> 	The modem is a Magitronic 33.6 (don't ask, I haven't got a clue...a
> friend gave it to me and swears it works on every other system he's used it
> in :( ).  Oh, its an internal modem, so no lights to watch flicker.  The ISP
> is Internet Canada, using its POP in Halifax, NS, which uses a Netblazer for
> its dialup pool (figured I'd throw out *as much* info as I could think of).
> 
> 	Here's what I have now:
> 
> ==========[ ppp.conf ]==========
> default:
>  set device /dev/ttyd2
>  set speed 57600 
>  set dial "ABORT BUSY ABORT NO\\sCARRIER TIMEOUT 5 \"\" AT&FM0X3&W OK-AT-OK \\dA
> TDT\\T TIMEOUT 40 CONNECT"
>  disable lqr
>  deny lqr
>  disable pred1
>  deny pred1
>  set redial 3 10
> ican-chap:
>  set authname XXXXXXXX
>  set authkey XXXXXXXX
>  set phone 4257110
>  set openmode active
>  accept chap
>  set ifaddr 127.1.1.1/0 127.2.2.2/0 255.255.255.0
>  add 0 0 127.2.2.2
>  set timeout 300
> ============
> 
> ==========[ ppp.linkup ]======
> ican-chap:
>  delete all
>  add 0 0 hisaddr
> ============================
> 
> ======[ netstat -nr ]======
> # netstat -nr
> Routing tables
> 
> Internet:
> Destination        Gateway            Flags     Refs     Use     Netif Expire
> default            206.231.247.114    UGSc        7       54      tun0
> 127.0.0.1          127.0.0.1          UH          2   154331       lo0
> 127.2.2.2          127.1.1.1          UH          0        0      tun0
  ^^^^^^^^^
The "delete all" in ppp.linkup should have zapped this route :-(

> 206.231.247.114    206.231.248.147    UH          8        0      tun0
> =====================
> 
> > What sort of sustained throughput do you get with ftp ?  That's 
> > usually a good indication of how the line's really behaving.  I'm no 
> > expert on this though.  My ISP is drifting into another phase of 
> > lousy service (doing big commits scares me 'cos I don't know if I'm 
> > going to lose the line due to a 300 second timeout).
> 
> 	I just used Netscape to grab the jdk 1.0.2 binaries, with no other
> traffic on the line, and I was lucky to get 1k/sec :(  I also tried using
> ncftp to send up a file to my server, and it got one file sent and then closed
> the connection to to "time out"...

I'm based in the UK and get lousy throughput most of the time to 
anywhere in the US.  1k/sec is about expected, and interactive 
traffic is a joke.  But to my ISPs ftp machine, I get about 4k/sec. 
This indicates to me that my ISPs routes to the US aren't fat enough, 
but I haven't looked into it yet (it's been like this for a number 
of months now, but was fine before).

I'm surprised with the upload timeout.  Is this the norm ?  Are 
things better with pppd or with other OSs ?  You could try removing 
the Predictor1 deny/disable lines.  There should be no problems there 
these days.

> Marc G. Fournier                                
> Systems Administrator @ hub.org 
> primary: scrappy@hub.org           secondary: scrappy@{freebsd|postgresql}.org 
> 

-- 
Brian <brian@awfulhak.org>, <brian@freebsd.org>
      <http://www.awfulhak.org>;
Don't _EVER_ lose your sense of humour....





Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199708311633.RAA04047>