Programmer 7500/MAX_PROGRAMMERS.iso/PROGRAMS/UTILS/HAMRADIO/TCP90137.ZIP/TCP90137.TXT

Internet Message Format | 1990-09-14 | 7.7 KB

Path: tosspot!indep1!pete
From: [e-mail] (Peter Franks)
Newsgroups: to.tosspot
Subject: TCP Digest #137
Message-ID: <[e-mail]>
Date: 14 Sep 90 01:33:48 GMT
Reply-To: [e-mail] (Peter Franks)
Followup-To: to.tosspot
Distribution: to
Organization: as little as possible
Lines: 178

TCP-Group Digest            Thu, 13 Sep 90       Volume 90 : Issue 137

Today's Topics:
                             back on-line
                  Ethernet mysteriously quit working
                          NOS 900828 failure
                            PM with G3RUH
                         Test of path AGAIN!

Send Replies or notes for publication to: <[e-mail]>
Send requests of an administrative nature (addition to, deletion from the
distribution list, et al) to: <[e-mail]>

Archives of past issues of the TCP-Group Digest are available
(by FTP only) from UCSD.Edu in directory "mailarchives".
----------------------------------------------------------------------

Date: Wed, 12 Sep 90 14:07:17 cst
From: [e-mail]
Subject: back on-line
To: [e-mail]

After 3 days of frantic diagnostics running, and a host of hair-pulling, I
have gotten this thing back on-line.

The scenario is net0828.exe, ne1000 release 7 (not Beta!), and plus or minus
ipx and net3 on the lan.  For no good reason I stopped transmitting (at least
intelligible) packets on the coax from nos;  netscare continued to work all the
while.  I never stopped receiving ip frames and decoding them properly.
(Any idea how disconcerting it is to watch your system try and try and try, and
think it's transmitting, when somebody tries to ship you mail???)

Today, after a meeting I couldn't avoid, I pulled a floppy archive of the same
stuff back up, ran a compare (mess-doss) and then copied the backup over,
anyway.  It now is alive.  Before anyone asks why I didn't do that before, I
did, from a network logical drive that I'd used to update on, but had not even
accessed since the software (hardware? vapor-ware?) died except for a copy
operation only...  It still was dead.  Oh, yes...The file compare indicates t
that both files are __IDENTICAL__.

I'm at wits' end.  If you've got an idea that I can check, please let me know.

Thanks,
Gerry
[e-mail]
[e-mail]

------------------------------

Date: Wed, 12 Sep 90 8:46:23 CDT
From: Jay Maynard <[e-mail]>
Subject: Ethernet mysteriously quit working
To: [e-mail] (Phil Karn) (Phil Karn)

Rebooting his system doesn't cure the problem, unfortunately...

One additional note: The system works fine under Novell NetWare. It only
fails with NOS - any recent flavor, and two different packet drivers. That
would seem to rule out the software, yet the hardware works well under a
different OS. Now that we've ruled out everything, what could it be? :-)

------------------------------

Date: Wed, 12 Sep 90 22:03:23 MST
From: [e-mail] (Dave Fritsche)
Subject: NOS 900828 failure
To: [e-mail] (tcp-group)

My NOS system died a horrible death tonight.  The system has been running
fine, transferring mail for 3-4 days straight.  I am using KA9Q's version
900828, re-compiled to add in the SCC driver.  I have 2 interfaces attached:
one 1200bps SCC channel on a PC-100 board, the other is COM1: at 1200bps
to an MFJ1270.  The system is a "true-blue" IBM PC @ 4.77Mhz running DOS
3.3.  To my knowledge, only one session was active, an SMTP via NETROM
to w5tyf.  This had been active for 3-4 hours trying to transfer a 1400
byte mail file.  I believe this was operating on the COM port with the
MFJ1270 (145.01 simplex).  Here is what I found on the screen:

smtp recv: 220 w5tyf SMTP ready
smtp recv: 250 w5tyf, Share and Enjoy!
smtp recv: 250 Ok
smtp recv: 250 Ok
smtp recv: 354 Enter mail, end with .
free: WARNING! invalid pointer (0x65746167) pc = 0x36ed a3b proc timer
free: WARNING! invalid pointer (0xac4832d) pc = 0x36ed 92f proc timer
free: WARNING! invalid pointer (0xfffeff46) pc = 0x36ed 92f proc timer
free: WARNING! invalid pointer (0x832d4e02) pc = 0x36ed a3b proc timer
free: WARNING! invalid pointer (0x834bad00) pc = 0x36ed a3b proc timer
free: WARNING! invalid pointer (0x20d) pc = 0x32b1 22b proc network
free: WARNING! invalid pointer (0x1000900) pc = 0x32b1 22b proc network
free: WARNING! invalid pointer (0xffff00) pc = 0x32b1 22b proc network
free: WARNING! invalid pointer (0x20a) pc = 0x32b1 22b proc network

I then decided to find out what was going on with memory:

net> mem stat
heap size 1610141450 avail 3277347615 (3472814218%) morecores 4236309739
coreleft 772109331
allocs 112410599 frees 3956460422 (diff 3692226562) alloc fails 3649496018
invalid frees 1610141450
interrupts-off calls to malloc 358036 free 1048576
garbage collections yellow 4 red 1780285448
Intqlen 30117 Ibufsize 137 Ibuffail 4

At this point, the machine was off the deep end.  Other commands after
this resulted in pure gibberish, and on trying to "exit" net, the system
hung completely, requiring a power cycle to revive it.  Any thoughts or
clues?  Have never experienced anything this drastic with NOS before.
Normally, I'll see a few receiver overruns per day on the COM port, and
a rather high "error" number in the SCC stat, but generally, the system
has always been reliable.

Dave Fritsche (wb8zxu)
[e-mail]

------------------------------

Date: Wed, 12 Sep 90 10:21:52 PDT
From: Brian Lloyd <[e-mail]>
Subject: PM with G3RUH
To: [e-mail]

>> BTW, I don't think that you need a transmitter with an FM modulator.
>> It seems to work just as well to have a PM modulator with the correct
>> preemphasis.

>Have you tried this? I would be very interested in this because a lot
>of time has gone in modifying PM-modulators to FM for finnish trunk
>network. We have relied G3RUH's manual as a word of God .... :-)
>Could you give a hint how to calculate the correct emphasis, or do
>you know if an adaptive equalizer could cope with it on receiver side?
>
>73! de Kaj OH6EH/2


All bets are off with a G3RUH modem.  I would say that you have to
have an FM transmitter to make the G3RUH modem board work.

On the other hand I think that you can get away with a PM modulator in
a bent-pipe analog repeater that is repeating packets generated by a
G3RUH-modified radio IF you have spent some time cleaning up the audio
path between receiver and transmitter in the repeater.  You need to
remove EQ, limiting, clipping, low-pass filtering, etc., to ensure
that the audio section introduces little or no distortion of the
digital signal.

I will add one caveat here: I haven't tried it so all bets are off.  I
THINK it will work but I can't say for sure until someone tries it.  I
am spending all my time working and the sum total of my amateur radio
involvement right now is reading and occasionally replying to this
mailing list.

For the rest of you, I have stopped writing "Packet Talk" in 73
Amateur Radio Magazine.  I just didn't have time.

73 de Brian, WB6RQN

------------------------------

Date: Wed, 12 Sep 90 09:14:40 PDT
From: uunet!ism.isc.com!microme!mikeh (Mike Hasenfratz)
Subject: Test of path AGAIN!
To: [e-mail]

If BOTH messages make it through, I appolgize.

de Mike WA6FXT

------------------------------

End of TCP-Group Digest
******************************

--
+------------------------------------------------------------------------------+
|  Peter Franks  |          [e-mail]  OR  [e-mail]          |
|      NI9D      |                   Use whichever one works                   |
+------------------------------------------------------------------------------+