HylaFAX The world's most advanced open source fax server |
We've been having a problem with Hylafax for quite some time. I'd say 1/3rd of our attempts to send faxes result in either NO CARRIER or NO ANSWER errors. I've tried all the testing and tweaking I could think of. Hardware/software flow control. Changing baud rate, etc., etc. I've done plenty of searching of the web and mailing lists, and haven't found anything remotely similar... The few who report NO CARRIER errors seem the have a simple solution. We're using on-board and add-in serial port cards, with 5 different modems, mostly various USR, but an old generic 14.4 as well, with no difference in behavior between them. All configured exactly the same. This problem is prevalent on all modems tried. I've fully tested the modems and serial ports, and they appear to behave perfectly 100% of the time no matter how much I try to stress them (at least with normal PC data connections). The relevant phone lines also haven't demonstrated any problems at all in my testing. At this point, my suspicion is either a bad Hylafax binary, or bad modem initialization/reset strings, but lacking a serial port sniffer, I'm not even sure how to debug this further. Changing either would be just fine, if there's some reason for it more concrete than my guessing at the causes. I'm hoping this is a known problem that someone else has seen. If not, some direction on debugging further would be most appreciated. Software is RPM of hylafax-4.3.3-1rhel4 on CentOS 5.0 Log is attached for one example of such problems. I've got hundreds, but they all look exactly the same, just a different number, job ID, and alternating between NO CARRIER and NO ANSWER.
Attachment:
c000012386
Description: Binary data
Attachment:
c000012389
Description: Binary data