x86: restart !@$?^* ttymon without rebooting?

x86: restart !@$?^* ttymon without rebooting?

Post by Mark McWiggi » Wed, 13 Apr 1994 16:29:37



As any modem-using x86 user must know by now, ttymon periodically
dies. The live periods can be days, or, like today, hours, which can
be very frustrating just when I have OpenWin set up the way I want it.

The only way I've found to get the modem back to answering the phone
is by rebooting the system. If there's another way, I'd certainly love
to hear it!

Thanks in advance.

--
Mark McWiggins      Hermes & Associates               +1 206 632 1905 (24 hrs.)

 
 
 

x86: restart !@$?^* ttymon without rebooting?

Post by Roger Fuj » Fri, 15 Apr 1994 06:30:02



>As any modem-using x86 user must know by now, ttymon periodically
>dies. The live periods can be days, or, like today, hours, which can
>be very frustrating just when I have OpenWin set up the way I want it.
>The only way I've found to get the modem back to answering the phone
>is by rebooting the system. If there's another way, I'd certainly love
>to hear it!

I've got mine working quite reliably for a while.  It seems that
you have 2 choices:

1) Use it for uni-directional traffic
2) If you DO use it bidirectionally, you must make sure that the only
thing the dialOUT program does is open/set speed etc.../ close.  It
*MUST NOT* try to hang up the line by lowering DTR (stty 0).  

Failing these two, you can bump the restart count to ttymon to something
abnormally large (10K) and kill the ttymon - that will restart it again.

-r
--
Roger Fujii - Media *netics         Phone: (301)495-3305 x259


 
 
 

x86: restart !@$?^* ttymon without rebooting?

Post by Mark McWiggi » Sat, 16 Apr 1994 17:05:11


Cindy Swearingen from Sun sent a suggestion that mostly works:

      pmadm -d tag -s port
      pmadm -e tag -s port

Eventually the port monitor got "worn out" or something, but this
worked several times while I was fooling with a recalcitrant UUCP feed
and the port monitor quit answering the phone.

I think this is supposed to be fixed in 2.4 ...

Cindy, I tried replying to your email to thank you, but the mail bounced.
Probably cosmic ray noise or interference from Microsoft. ;-)
--
Mark McWiggins      Hermes & Associates               +1 206 632 1905 (24 hrs.)

 
 
 

x86: restart !@$?^* ttymon without rebooting?

Post by Willard Daws » Sun, 17 Apr 1994 23:17:53



>As any modem-using x86 user must know by now, ttymon periodically
>dies. The live periods can be days, or, like today, hours, which can
>be very frustrating just when I have OpenWin set up the way I want it.
>The only way I've found to get the modem back to answering the phone
>is by rebooting the system. If there's another way, I'd certainly love
>to hear it!
>Thanks in advance.

On my Dell SVR4 2.2 box, the same problem exists with ttymon; I suspect
the problem lies in the original AT&T ttymon/sacadm/portmon code, as
I've seen complaints about just about every known port of SVR4 with
these utilities.

You may have two different problems:

  1)  The ttymon process actually dies.
  2)  The status of your service is "disabled."

In case 2, the ttymon process is still running, but the modem is offline,
DTR is low.  For me, the pmadm utility is useful in re-enabling the
modem, e.g.:

  pmadm -e -p ttymon3 -s 02h

You will obviously have to change the arguments to -p and -s to fit your
own configuration.

Case 1 is a bit more problematic, but just a bit.  You can tweak sacadm
to watch the ttymon process using the -n option.  From the Dell sacadm
man page:

  Set the restart count to count.  If a restart count is not
  specified, count is set to 0.  A count of 0 indicates that
  the port monitor is not to be restarted if it fails.

So, set the restart count to any positive restart count.  I currently
use 3, but 5 or 10 or larger might be a good idea.

On the other hand, you might just disable ttymon for the serial port in
question, and use uugetty spawned from /etc/inittab, and totally avoid
the buggy SVR4 sacadm/portmon/ttymon code.