dale@lamont.LDGO.Columbia.edu (dale chayes) (11/18/87)
We are using Telebit TrailBlazers with Berkley 4.3 uucp to link our lab network with our geophysical research vessel (Robert D. Conrad.) The TrailBlazers have version 3.0 EPROMS. We set up a reasonable test period in the lab, and then over a local (very noisey) phone line and got good results (which I reported a couple of weeks ago.) We are now running between the lab and the ship (at sea) through an INMARSAT satelite link and having intermittent problems. In general, we see "pkcget: alarm n" and eventually the link fails. We are sucessfully transfering files, but the data rate is highly variable, and it is not reliable enough to let it run automagicly. My present hypothesis is that the extra delay induced by the sat. link is causing time outs....but I haven't been able to sleuth out what to change. I have spent hours looking at the code, (and seconds reading the profuse comments (:-)) and would really like to talk to someone who is familiar with uucp. Dale Chayes usmail: Lamont-Doherty Geological Observatory of Columbia University Route 9W Palisades, N.Y., 19064 voice: (914) 359-2900 extension 434 fax: (914) 359-6817 (all kinds) usnet: ...philabs!lamont!dale ------ -- are we having fun yet?....
pete@octopus.UUCP (Pete Holzmann) (11/18/87)
In article <222@lamont.LDGO.Columbia.edu> dale@lamont.LDGO.Columbia.edu (dale chayes) writes: >We are now running between the lab and the ship (at sea) through >an INMARSAT satelite link and having intermittent problems. In general, >we see "pkcget: alarm n" and eventually the link fails. We are sucessfully >transfering files, but the data rate is highly variable, and it is >not reliable enough to let it run automagicly. We just received our Trailblazer, running firmware level 3.01. I can also get the link into a 'pcget: alarm 1' infinite loop. It almost always happens if I force the serial line between my cpu and the modem to lose a few characters (either by overrunning the FIFO in the cpu, or by glitching the serial line with a breakout box. Since this doesn't happen if I'm not running the Trailblazer's 'g' protocol spoof, I'm ready to guess that there's a bug in the modem's error recovery support. I'm working hard on this right now! Has anybody else seen this problem? Pete -- OOO __| ___ Peter Holzmann, Octopus Enterprises OOOOOOO___/ _______ USPS: 19611 La Mar Court, Cupertino, CA 95014 OOOOO \___/ UUCP: {hplabs!hpdsd,pyramid}!octopus!pete ___| \_____ Phone: 408/996-7746