Re: flexfax: Big problem with sendfax

[Nico Kadel-Garcia <raoul@cirl.meei.harvard.edu>]

-----BEGIN PGP SIGNED MESSAGE-----

On Wed, 8 Dec 1999, Willis Gregory wrote:

> I downloaded hylafax-4.1beta2.tar.gz and loaded it on a Dell running Red 
> Hat 6.0.  I'm using a MultiTech MT2834ZDXb modem.  I'm only using it to fax 
> outgoing  reports.

You probably want to go to RedHat 6.1, but that's probably not your problem.

> After trying it successfully with a few outgoing faxes, I turned on the 
> flood gate and sent all of my outgoing faxes to the Hylafax server.  A 
> shell file runs sendfax on each file ftp'd to the received directory of the 
> Linux server.

??? Why does this sound a bit odd? What is this "shell script running
sendfax", and why do I suspect it of screwing up your system?

> Now hylafax is totally backed up.  If I try to run sendfax manually, it 
> totally blocks and won't queue the fax.

Gack. Not good. 

> When I do "faxstat -s", I get a large list.   Many of the JID's appear to 
> duplicates to the same phone number.   The status line at the top says 
> "Modem ttyS0:  Waiting for modem to come ready.  Of the approx. 240 JID's 
> in the -s output, all but about 10 have a status of "Blocked by concurrent 
> job".  Ten have a blank status.

Definitely sounds like your script has bombed badly. Start nuking
*ALL* of them, making local duplicates of the contents of the docq directory
if necessary.

> How do I tell what the "current job" is so I can kill it?  Is there a 
> maximum number of outgoing faxes in the queue which prevents sendfax from 
> taking any new requests?

There shouldn't be: but you could easily overload the memory and/or
the disk use if your system has the HylaFAX software on too small a
partition: this often happens when people try to partition their
Linux system the way they are used to with Sun or BSD systems.

> The modem itself is not doing anything at the present time.  I rebooted the 
> system and turned the modem off, but the modem is still doing nothing.
> 
> When I did "faxrm" on the first JID listed in "sendfax -s", it 
> blocked.  Finally broke out after nearly 30 minutes.  Is there a  way to 
> clean up the entire send queue?

Go in by hand to the $SPOOLDIR/doneq or $SPOOLDIR/docq directories
and start ripping out the files left there.

> If I do "faxstat -d", it also shows a large list, with most of the JID's 
> "Blocked by concurrent job".  I thought -d showed completed jobs.  Why 
> would they be shown as completed if they were blocked?

Because they are in the "doneq" directory. They've completed, but
unsuccessfully. Take a look at the "faxqclean" for details on cleaning
up this debris.

			Nico Kadel-Garcia
			Senior Engineer, CIRL 
			Mass. Eye and Ear Infirmary
			raoul@cirl.meei.harvard.edu

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQCVAwUBOE59jT/+ItycgIJRAQF5dgP9H/PDlrU5pAUwhggmwbNkzQTwsyTt6fq3
bQyNm8sXIiXXcknV436zrFzcVbj8KkAkNWy2SVUqpTI64LnI5EpLK0sa2okdaSuD
F3o+4jTock1Ur0xDiJEIY49ni0LejeEfDhWcZLHWZ7p4c1Ntc2h6li6Ai1IFET4A
zRtTHw9hmwU=
=2Yos
-----END PGP SIGNATURE-----