Re: [hylafax-users] Modem hung indefinately... saga continues

["Brendin Emslie" <brendin@dreamland.ab.ca>]

Interesting that you are having the same problem with hylafax using a
different version than I.  I don't think that upgrading your HylaFAX would
help.  I am using 4.1beta2.  I experienced this problem with 4.1beta1 as
well.  I am running on Redhat Linux 6.1 with 3COM 56K external modems and a
Comtrol 8 Port Rocket port isa serial card.

My hung tty's occur daily, sometimes twice daily, never on the same tty.
This may have something to do with the volume of faxes that I am pumping out
(1300 in a 10 hour period on 4 lines).  How many are you sending out?  If
there was only some way to get rid of these jobs without rebooting.  This
definately should not be required especially under a UNIX OS.

I am running my modems with a CLASS1 config file.  How about you?


-----Original Message-----
From: Alastair Broom [mailto:abroom@valleyt.co.uk]
Sent: Thursday, July 20, 2000 1:51 AM
To: Brendin Emslie
Cc: hylafax-users@hylafax.org
Subject: Re: [hylafax-users] Modem hung indefinately... saga continues


On Wed, 19 Jul 2000, Brendin Emslie wrote:

> Can anyone sugest other things to try that may help wake up this brain
dead
> tty?
>
Well, we have the same problem...

We look after three sites running hylafax-v4.0pl2 (patched for 56k modems,
y2k and gcc-2.8) running on SCO Unix 5.0.4
One of the sites (only one...) has the same problem as you - it runs for
days fine, then a fax job gets stuck on a tty and cannot be killed.
This is outgoing only - faxgetty is not running, and the modem receives no
incoming calls.

In trying to track down the problem we've:
a) replaced the modem (Multitech MT2834ZDXK)
b) moved the modem onto it's own "normal" phone line (as opposed to
   running off a POTS port on the back of one of our ISDN routers)
c) upgraded the drivers for the Stallion multiport card that the modem
   ran off.
d) moved the modem off the Stallion multiport card and back onto the tty
   on the back of the system box.

We end up with a faxsend job that cannot be killed (suggesting that the
problem is somewhere in the unix kernel space waiting for something to
happen on the tty - if it was in user space at the time of the kill, the
program would die).
We've found that *sometimes* if you leave it for a couple of days (say,
over a weekend), the job sometimes exits on it's own, and hylafax
recovers, meanwhile all the other jobs in the queue have timed out :-(

We have downloaded the latest cvs sources and compiled them on SCO, but,
to be honest, I'm a bit afraid of installing them as:
a) our client faxes out all the Purchase Orders for their warehouses using
   this system, so having to reboot their server once a month is actually
   less hassle than seriously breaking the faxing by installing a new
   version...
b) their site is 300miles from where I'm sat - sometimes you just *need*
   to be able to see what the modem lights are doing :-)

Since you seem to be having the same problem and you're running on a
different OS, then it seems the problem must be in hylafax...
Hmmm, if the sources reached an "official" beta or release I'd be happier
about installing them, then trying to track any outstanding problem...
--
Alastair Broom               ValleyT Ltd, Edinburgh, Bonnie Scotland
abroom@valleyt.co.uk




____________________ HylaFAX(tm) Users Mailing List _______________________
 To unsub: mail -s unsubscribe hylafax-users-request@hylafax.org < /dev/null