home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Columbia Kermit
/
kermit.zip
/
newsgroups
/
misc.20030409-20031118
/
000365_fdc@columbia.edu_Sat Nov 1 11:38:56 2003.msg
< prev
next >
Wrap
Internet Message Format
|
2020-01-01
|
4KB
Path: newsmaster.cc.columbia.edu!not-for-mail
From: Frank da Cruz <fdc@columbia.edu>
Newsgroups: comp.protocols.kermit.misc
Subject: Re: Protocol Error: NAK out of window
Date: 1 Nov 2003 16:15:20 GMT
Organization: Columbia University
Lines: 54
Message-ID: <slrnbq7n0o.7lp.fdc@sesame.cc.columbia.edu>
References: <Dg_nb.118720$h61.101506@news01.bloor.is.net.cable.rogers.com> <Pine.HPX.4.44.0310311650430.13666-100000@fog.ccsf.cc.ca.us>
Reply-To: fdc@columbia.edu
NNTP-Posting-Host: sesame.cc.columbia.edu
X-Trace: newsmaster.cc.columbia.edu 1067703320 24380 128.59.59.56 (1 Nov 2003 16:15:20 GMT)
X-Complaints-To: postmaster@columbia.edu
NNTP-Posting-Date: 1 Nov 2003 16:15:20 GMT
User-Agent: slrn/0.9.7.4 (SunOS)
Xref: newsmaster.cc.columbia.edu comp.protocols.kermit.misc:14620
In article <Pine.HPX.4.44.0310311650430.13666-100000@fog.ccsf.cc.ca.us>,
Mark Sapiro wrote:
: On Thu, 30 Oct 2003, computer person wrote:
:> I am pretty new to kermit but pretty computer savey. I built a script to
:> connect to ministry of health to send health #'s for validity check. The
:> file transfer fails around 25-30% of the time on error "Protocol Error:
:> NAK out of window". Any ideas would be greatful. Do you think adding
:> "robust" would do anything positive? I only get a few chances to submit a
:> transfer per day so I try to keep the guesses to a minimum.
:
: I was hoping Frank would respond to this one as his reply would be much
: better than mine, I'm sure...
:
Not necessarily, but something must be wrong with our news service because
the original post never came in here. Or at least I missed it somehow --
hard to say, too many things changing at once at our place.
: Yes, I think "robust" would help. So might "set window-size 1". The
: error message seems clear. I think the sender is receiving a NAK to
: a packet that has already been released from the window, presumably because
: it was already ACKed. This would indicate at least one side of the
: transfer does not implement the protocol correctly. Your logs indicate
: the sender (your side) is C-Kermit 8.0.209. I would guess the problem
: is with the receiver. What Kermit is running there?
:
That's the key question of course. We have not had very good experience
with third-party Kermit implementations, as you can see from:
http://www.columbia.edu/kermit/ckermit70.html#x4.22
http://www.columbia.edu/kermit/ckermit80.html#x15
http://www.columbia.edu/kermit/kermit.html#notslow
: This seems to indicate the error doesn't occur until after the send
: completed. I don't know why this would be, but it seems clear that
: at least one side is doing something wrong. Is the file received
: completely in this case?
:
If you can collect a packet log (tell C-Kermit to "log packets") I can
take a look and see how the error comes about.
In any case, read at least the first two links above and try some of the
suggestions. You can also consult Chapter 10, "Solving File Transfer
Problems" of "Using C-Kermit".
As Mark suggests, the most likely culprit is a faulty sliding-windows
implementation in the other Kermit, in which case "set window 1" should
eliminate the error. Of course it might also slow down the transfers, but
a slower transfer is still faster than one that failed.
"Robust" is the last resort, because this results in the slowest throughput.
If you can get by with long packets but one window slot, the performance
shouldn't be too bad, unless you're going through a satellite or something.
- Frank