home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
kermit.columbia.edu
/
kermit.columbia.edu.tar
/
kermit.columbia.edu
/
newsgroups
/
misc.20021006-20030409
/
000020_hvanclee@nyx.net_Thu Oct 17 10:04:59 EDT 2002.msg
< prev
next >
Wrap
Text File
|
2003-04-08
|
4KB
|
66 lines
Article: 13783 of comp.protocols.kermit.misc
Newsgroups: comp.protocols.kermit.misc
Subject: Re: Problems transferring over ppp connection
References: <1034801647.495773@irys.nyx.net> <aokm12$7lv$1@watsol.cc.columbia.edu> <1034815854.830840@irys.nyx.net> <aol3ol$sch$1@newsmaster.cc.columbia.edu>
Organization: Nyx, Spirit of the Night
X-Newsreader: trn 4.0-test76 (Apr 2, 2001)
From: hvanclee@nyx.net (Henry van Cleef)
Originator: hvanclee@nyx.net (Henry van Cleef)
Message-ID: <1034844239.214811@irys.nyx.net>
Cache-Post-Path: irys.nyx.net!unknown@nyx3.nyx.net
X-Cache: nntpcache 2.3.3 (see http://www.nntpcache.org/)
NNTP-Posting-Host: 206.124.29.6
Date: 17 Oct 2002 02:43:06 -0600
X-Trace: omega.dimensional.com 1034844186 206.124.29.6 (17 Oct 2002 02:43:06 -0600)
Lines: 47
Path: newsmaster.cc.columbia.edu!phl-feed.news.verio.net!iad-feed.news.verio.net!ord-feed.news.verio.net!stl-feed.news.verio.net!news.cc.ukans.edu!logbridge.uoregon.edu!HSNX.atgi.net!cyclone-sf.pbi.net!216.218.192.242!news.he.net!dimensional.com!pulsar.dimensional.com!omega.dimensional.com!not-for-mail
Xref: newsmaster.cc.columbia.edu comp.protocols.kermit.misc:13783
In article <aol3ol$sch$1@newsmaster.cc.columbia.edu>,
Jeffrey Altman <jaltman@watsun.cc.columbia.edu> wrote:
>In article <1034815854.830840@irys.nyx.net>,
>Henry van Cleef <hvanclee@nyx.net> wrote:
>: In article <aokm12$7lv$1@watsol.cc.columbia.edu>,
>: Frank da Cruz <fdc@columbia.edu> wrote:
>: >In article <1034801647.495773@irys.nyx.net>,
>: >Henry van Cleef <hvanclee@nyx.net> wrote:
>: >:
>: >: The TCP/IP connection uses STREAMING, and I suspect that my problem
>: >: lies in the fact that it's much faster than my modem setup can handle.
>: >: What's the proper way to deal with this.
>: >:
>: >The culprit is most likely a lack of effective flow control.
>: >
>: Thanks for the quick response, giving me an idea where to look.
>
>I should point out that if the connection was truly using STREAMING
>transfers then if even a single error was to occur, the transfer
>would fail. Therefore, I must assume tht STREAMING transfers are
>not being used.
>
Jeffrey, you are absolutely correct on this. I was thinking of what
happens when I transfer between local machines on a network.
I verified that RTS and CTS are being explicitly set in the modem init
string. Also played around with some of the pppd options. What
seems to have had an effect is disabling software compression, and
it's obvious that wasn't buying me anything. Netscape 4.76 has no
trouble downloading on the same connection. Locking everything to
18.2 KB had no effect with compression enabled, and trying to run with
xon/xoff only made matters worse. I'm still not completely satisfied
that I've pinned things down completely, or that there aren't problems
with the remote end. Also, the Sun version of ppp that is in the
10/01 distribution wasn't really documented, and has a distinctly beta
flavor to it. I might try the iPlanet version, which is on the
installation disks. I also think that the public domain version on
samba (2.4.1) will compile for Solaris and might be worth
investigating. However, I think that the ppp newsgroup is the place
to pursue such questions. My real question here was whether the flow
control was supposed to be in the kermit running over ppp or in the
pppd's, and I think you're saying that the pppd's are the place to
look. The same kermit objects run fine over a telephone link and over
a 100baseT link through a hub between machines, using ssh.
Hank