QRZ! Ham Radio 3/QRZ Ham Radio Callsign Database - Volume 3.iso/digests/tcp/930248.txt

Internet Message Format | 1994-06-04 | 16KB

Date: Sat, 25 Sep 93 04:30:01 PDT
From: Advanced Amateur Radio Networking Group <[e-mail]>
Errors-To: [e-mail]
Reply-To: [e-mail]
Precedence: Bulk
Subject: TCP-Group Digest V93 #248
To: tcp-group-digest


TCP-Group Digest            Sat, 25 Sep 93       Volume 93 : Issue  248

Today's Topics:
                        backoff bugs (3 msgs)
                      pmnos new version where ?
                      PMNOS Version 1.1 (5 msgs)
                                 RIP

Send Replies or notes for publication to: <[e-mail]>.
Subscription requests to <[e-mail]>.
Problems you can't solve otherwise to [e-mail].

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

We trust that readers are intelligent enough to realize that all text
herein consists of personal comments and does not represent the official
policies or positions of any party.  Your mileage may vary.  So there.
----------------------------------------------------------------------

Date: Fri, 24 Sep 1993 10:59:22 -0400 (EDT)
From: [e-mail] (Mike Bilow)
Subject: backoff bugs
To: [e-mail], [e-mail]

Yes, attaching time outs to the server sockets for non-activity was the
original idea under discussion.  There would have to be some minor
changes made to the TCP control block structure to hold the time of
last activity, but that is comparatively minor.

As I said, I had some reservations about having NOS close down a TCP
circuit to a server socket when a new TCP circuit is opened to the same
socket from the same host.  However -- and I am not enough of a Unix
sendmail expert to form an opinion on this -- why would a host under
normal circumstances open multiple TCP circuits to the same socket on
the same host automatically?  I could see this as normal is a real
human opens multiple FTP or telnet sessions, but there would seem to
be no merit to this kind of behavior for SMTP.  All pieces of mail for
mailboxes at any particular host ought to be sent sequentially over
one TCP circuit to the SMTP server socket, shouldn't they?

-- Mike Bilow, [e-mail]
               N1BEE @ WA1PHY.#EMA.MA.USA.NA

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

Date: Fri, 24 Sep 93 11:33:04 CDT
From: [e-mail] (Andy Warner)
Subject: backoff bugs
To: [e-mail] (Mike Bilow)

> Yes, attaching time outs to the server sockets for non-activity was the
> original idea under discussion.  There would have to be some minor
> changes made to the TCP control block structure to hold the time of
> last activity, but that is comparatively minor.

Why does this need a change to the TCP control block ? The server
should set an alarm call (or is this unavailable under NOS ?) that
interrupts the read().

> As I said, I had some reservations about having NOS close down a TCP
> [...]
> be no merit to this kind of behavior for SMTP.  All pieces of mail for
> mailboxes at any particular host ought to be sent sequentially over
> one TCP circuit to the SMTP server socket, shouldn't they?

Due to the way sendmail is written, mail is spooled on a "per
item basis", not a "per host basis", it is not uncommon for me to
see 3 or 4 smtp connections to the same machine from my unix box.
While that is really beside the point, it does serve as a good
example of why you shouldn't start making assumptions about usage
patterns.

I repeat - just fix the NOS smtp server (adding the ability to
interrupt a read() by a timer,if necessary.) The FTP server could
probably do with the same kind of treatment. Make the timeouts
configurable, and everyone will be happy (sendmail included.)
--
andyw. N0REN/G1XRL

[e-mail] Andy Warner, Cray Research, Inc. [phone]

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

Date: Fri, 24 Sep 93 18:35:59 EDT
From: "Ross Patterson" <[e-mail]>
Subject: backoff bugs
To: [e-mail]

On Fri, 24 Sep 1993 10:59:22 -0400 (ED, Mike Bilow wrote:

>                            However -- and I am not enough of a Unix
>sendmail expert to form an opinion on this -- why would a host under
>normal circumstances open multiple TCP circuits to the same socket on
>the same host automatically?  I could see this as normal is a real
>human opens multiple FTP or telnet sessions, but there would seem to
>be no merit to this kind of behavior for SMTP.  All pieces of mail for
>mailboxes at any particular host ought to be sent sequentially over
>one TCP circuit to the SMTP server socket, shouldn't they?

That's an implementation decision, and both the TCP and SMTP protocol
specifications are silent on it. If you choose to implement SMTP in
a fashion that requires a separate connection for each piece of mail,
that's your choice.  It's stupid, but legal.

"Be liberal in what you accept, conservative in what you send."
-- Jon Postel et al.

73,
Ross N4YYH

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

Date: Fri, 24 Sep 93 15:49:25 CET
From: BARRY TITMARSH <[e-mail]>
Subject: pmnos new version where ?
To: TCP-GROUP <[e-mail]>

Simple question
what ftp site and what dir ??
is the new pmnos 1.1 to be found.
thanks Barry..

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

Date: 24 Sep 1993 08:48:39 -0500 (EST)
From: Mike Murphree <[e-mail]>
Subject: PMNOS Version 1.1
To: [e-mail]

Sounds like the rest of you are getting further than I am...

One question though, are you using it over wire links or radio?

Anytime I try a telnet/tty to someone else the TNC sends out frames
but no response...

Thanks Lyman, I received your autoexec.nos file. Looks like you have
some of the same problems that I do:

Will not except "ax25 digi on", in fact crashes during load if in autoexec.

Not confirmed, but no forward server.

You are using netrom exclusively, we've purged our local netrom thanks to X1J,
maybe this is why you can make connects and I can't.

Its funny that the "isat" command is still supported, OS/2 doesn't run on XTs.

I downloaded the source and have the IBM C/Set++ compiler, maybe I can rebuild
it and get better results.

  ******************************************************
  * 73 de Mike, N4CNW@W4DPH.#TPA.FL.USA.NA             *
  *    Amprnet: [e-mail] [44.98.0.151]     *
  *   Internet: [e-mail]     *
  ******************************************************

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

Date: Fri, 24 Sep 1993 10:51:38 -0400 (EDT)
From: [e-mail] (Mike Bilow)
Subject: PMNOS Version 1.1
To: [e-mail], [e-mail]

All of the newer NOS variants are starting a session when they send a ping,
and do not trace inside a session.  So, when the ping command is given,
the console is switched to a session virtual console where the ping lives,
the ping frame is sent, and then the session virtual console is closed.
The effect is that you never trace the first ping.  This is the behavior
in JNOS, so I assume it simply carried over into PMNOS.

The "stat" command develops its list of open files throught the use of an
undocumented facility in MS-DOS.  As a matter of fact, running the DOS
version of N1BEE NOS in an OS/2 DOS window will generate a message that
says that the files list is unavailable under OS/2 DOS emulation.  I
don't know if there is any way for an OS/2 program to find out the names
of open files on a global basis, but I doubt that there is.

-- Mike Bilow, [e-mail]
               N1BEE @ WA1PHY.#EMA.MA.USA.NA

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

Date: Fri, 24 Sep 93 12:24:04 EST
From: [e-mail]
Subject: PMNOS Version 1.1
To: [e-mail], [e-mail]

> The mailbox 'operator' command causes a disconnect.  Also, when going into
> the conference mode from the mailbox, sometimes the connection is immediately
> terminated, sometimes it'll work.  However, an explicit
> 'telnet <local hostname> 3600' from the net prompt will work all the time.

Tony, I pretty much documented the operator page problem, and I think the
mailbox conference is the same thing.
Not only will telnet <host> 3600 work consistently but so will ttylink <host>
3600.

As I have mentioned earlier, I spent months (not full time) trying to get the
mailbox functions to work. I have no idea why they dont, but I decided not to
hold the release any longer than I had.
-Walt



*********************************
*     [e-mail] or     *
*    [e-mail]            *
* The home of OS/2 NOS and PMail*
*********************************

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

Date: 24 Sep 1993 15:00:08 -0500 (EST)
From: Mike Murphree <[e-mail]>
Subject: PMNOS Version 1.1
To: [e-mail]



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

Date: Fri, 24 Sep 93 22:51:48 EST
From: [e-mail]
Subject: PMNOS Version 1.1
To: [e-mail]

Mike (n4cnw) claims to be having problems as some others of you are.
He claims for instance ax25 digi on crashes the system.  I use it here with
no problem.
He requested my autoexec. Since I can not reach him directly and this may be of
value to others I am including my autoexec.nos.

If this is of no interest to you, sorry for the bandwidth.
-Walt

#----------------------------------------------------------------------------
#  - A U T O E X E C . N O S  ***
#----------------------------------------------------------------------------
#
isat on
#
hostname kz1f.ampr.org
#
ax25 mycall kz1f
#
ip address [44.104.0.23]
#
# COM1 KISS
#attach asy 0x3f8 4 ax25 ax0 2048 256 9600
attach asy 0x2f8 4 ax25 ax0 4096 256 9600
#att asy 03f8 1 slip sl0 1024 1024 19200
#
# COM2 SLIP LINK
att asy 03f8 1 slip sl0 3000 1534 19200
#ifconfig sl0 ipaddress xxx.xx.xx.xx netmask 0x0000ffff mtu 1534
#route  add xxx.xx.xx.xx sl0
domain addserver [132.239.1.1]
comm sl0 "ats0=1"
#attach asy 0x3f8 3 slip s10 2048 512 19200
#
# COM2 KISS
# attach asy 0x2f8 3 ax25 ax1 1024 256 4800
#
remote -s cg
start remote
#
#attach netrom
#netrom interface ax0 NSRIIP 192
#netrom nodetimer 1800
#netrom obsotimer 3600
#netrom bcnodes ax0 s10
#netrom alias IPSCIT
#netrom verbose on
#netrom promiscuous on
#
# Timing Parameters for ax0
#
param ax0 1 33 # was 65 / 40 at w1cg
param ax0 2 70 # was 70 / 192 at w1cg
param ax0 3 1 # was 20 / 12 at w1cg
#
domain suffix ampr.org.
#
# Routing Statements
route addprivate default ax0 [44.104.0.2]
route addprivate w1cg-5 ax0
#route addprivate [44.56]/16 ax0 wb6nil
#route add [44.104.0.15] ax0
#route add [44.104.0.19] ax0
#route add [44.104.0.20] ax0
#
#arp publish wa1oaj-2 ax25 wa1oaj
#arp add 44.104.0.5 ax25 w1cg
#arp add 44.255.255.255 ax25 qst-0
#arp add 44.104.0.2 ax25 w1cg-9
#arp add 44.104.0.20 ax25 n1bee
#
tcp mss 216 #was 216
tcp window 8192  #was 432
tcp irtt 10000
ip ttl 32
ip rtimer 60
icmp echo on
icmp trace off
attend off
tcp timertype linear
log \nos\nos.log
start ax25
start smtp
start convers
#popmail addserver giskard.uthscsa.edu 1800 POP3 kz1f kz1f walt
#start pop
#pop mailhost w1cg
#pop user kz1f kz1f
#pop mailbox kz1f
#pop ti 3600
#smtp usemx off
smtp timer 1800
#smtp gateway bbs.wa1phy
smtp trace 0
#start netrom
start ttylink
start telnet
start ftp
start echo
start discard
start finger
smtp maxclients 40
#
# ping 44.104.0.30 60
# ping 44.56.0.64 3600
# ping 44.56.8.113 3600
# ping 44.56.8.115 3600
# ping 44.56.8.116 3600
#
ax25 timertype linear
ax25 blimit 10
ax25 irtt 5000
ax25 digipeat on
ax25 bcinterval 0
ax25 bctext "Converse server running, use telnet function C"
ax25 bc ax0
ax25 maxframe 1
ax25 paclen 256
ax25 retry 10
ax25 window 1024
ax25 t3 0
#
#hopcheck maxttl 16
#hopcheck maxwait 60
#hopcheck queries 4
#hop trace on
#
domain translate no
#
# rip merge on
# rip trace 2
# rip add n2cey 1800 1
# rip add n1btq 1800 1
# rip add wa1oaj 1800 1
# rip add w1cg-3 1800 1
# rip add switch.w1cg-9 1800 1
# rip accept n2cey
# rip accept n1btq
# rip accept wa1oaj
# rip accept w1cg-3
# rip accept switch.w1cg-9
ifconfig ax0 broadcast 44.255.255.255
#ifconfig s10 broadcast 44.255.255.255
rspf mode d
#rspf interface ax0 8 1
#rspf interface s10 8 32
rspf rrhtimer 900
rspf suspecttimer 2000
rspf timer 900
rspf message "OS/2 2.0 and PM NOS for OS/2 running here"
#
motd "OS/2 2.0 and PM NOS running here"
attend on
mbox attend on
# END


*********************************
*     [e-mail] or     *
*    [e-mail]            *
* The home of OS/2 NOS and PMail*
*********************************

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

Date: Fri, 24 Sep 93 16:53:14 -0700
From: [e-mail] (Chuck Bland)
Subject: RIP
To: [e-mail]

At one time I remember seeing some config notes on RIP and RSPF in NOS.

Anyone have pointers where they might be ?

CHuck

[e-mail]

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

Date: (null)
From: (null)
>   On JNOS, when I do a connect and then switch to the trace
>session, I never see the initial connect request packet get
>sent. It's a timing problem. I do see any retry, connect
>request. Can you see and subsequent connect request get
>sent out?

 No, this is after I have started the trace and it has its
 window up. I assumed it was a separate process and would
 update even in the background unlike JNOS.

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

Date: (null)
From: (null)
>What command differences? Its identical to the earlier versions in that
>respect, save bootp and bootpd.

 Different from JNOS and some of the other NOS variants, I don't remember
 all the specifics, but some of them were timer and callsign setups for
 the conference server.

>> got it running, but it is still not working properly...

>excuse me? Works fine here. trace shows all the packets flowing into or out
>of the port. No problem with connects or icmp.

 Like I mentioned before it's spitting packets out the TNC, but no response
 from neighboring stations. This is on a 386DX/33 system with OS/2 2.1 &
 jnos1.10x1-9 that has been running successfully for at least 2 months in
 this configuration. The system has 8 meg of RAM and 210 meg of SCSI drive
 space. The TNC is connected to a STB-4 board on COM3 with the standard VCOM
 and COM.SYS drivers.

 If I issue the "ax25 digipeat on" command, it gives me all the choices
 for "on", i.e. off 0 1 yes no on n y , etc. If I put the same command in
 the autoexec.nos, it dies when I bring it up. I'm not casting stones here
 Walt, I'm just geniunely having a hard time getting the program to come
 up. I've gotten locked out from the command processor, i.e. PMNOS windows
 will move from front to back as normal, but no response to keyboard input
 in the command window or the 'Exit' command in the 'File' menu. Other
 programs and desktop items continue to operate normally. Maybe a copy of
 your autoexec file would help...


  ******************************************************
  * 73 de Mike, N4CNW@W4DPH.#TPA.FL.USA.NA             *
  *    Amprnet: [e-mail] [44.98.0.151]     *
  *   Internet: [e-mail]     *
  ******************************************************

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

End of TCP-Group Digest V93 #248
******************************
******************************