Re: [hylafax-users] Unknown problem (check modem power) again

[Dmitry Bely <dbely@mail.ru>]

obladen@komed.de (Jan Obladen) writes:

> We have a problem with Hylafax on our Mandrake 7.1-box. It is running 
> kernel 2.2.16 and hylafax-server-4.0pl2.rjc11-3mdk.rpm. We are using a 
> Sedlbauer Speedfax+ (ISA) - a semi-active ISDN-adapter that is able to 
> run as class1-modem after uploading a firmware from isdn4k-utils.
> 
> The problem occurs in some cases when we send a fax. Sometimes 
> everything runs ok but sometimes there is a problem that appears as 
> "Unknown problem (check modem power)" or in other cases as "No answer 
> (T.30 T1 timeout)" in the comm-log. The error can be reproduced by 
> faxing to the same fax machine.
> 
> Workarounds and solutions from previous postings concerning "check 
> modem power" did not do the job. The phone number 8099 we call is a 
> internal fax machine (I think it is a Bosch 373) and acts ok to normal 
> fax machines. 
> 
> Maybe it is important to mention that we get the kernel message "ISAR: 
> HDLC frame too short" on console tty0 occasionally. But the 
> kernel-message occurs even in cases where hylafax sends faxes without 
> problems.
> 
> A sample session-trace is attached. Additionally, we straced the 
> process and attached the output. 
> 
> Thanks
> Marc Hildebrand  /  Jan Obladen.
> 
> 
> 
> 
> 
> Begin session trace 
> ----------------------------------------------------------
> 
> Aug 09 11:32:12.92: [19334]: SESSION BEGIN 00000247 492218099
> Aug 09 11:32:12.92: [19334]: SEND FAX: JOB 63 DEST 8099 COMMID 00000247
> Aug 09 11:32:12.92: [19334]: MODEM set DTR OFF
> Aug 09 11:32:12.92: [19334]: DELAY 2600 ms
> Aug 09 11:32:15.52: [19334]: MODEM set DTR ON
> Aug 09 11:32:15.52: [19334]: MODEM set baud rate: 19200 baud, input flow NONE, output flow NONE
                                                                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Too bad; the flow control is a must for reliable fax operation.

> Aug 09 11:32:15.52: [19334]: MODEM flush i/o
> Aug 09 11:32:15.52: [19334]: <-- [34:ATZ&E180S0=2+FCLASS=1E0V1Q0S0=0H0\r]

Useless init string. No necessary flow control setup (modem-dependent).

> Aug 09 11:32:15.53: [19334]: --> [2:OK]
> Aug 09 11:32:15.53: [19334]: <-- [14:ATS8=2S7=60M0\r]
> Aug 09 11:32:15.54: [19334]: --> [2:OK]
> Aug 09 11:32:15.54: [19334]: <-- [12:AT+FCLASS=1\r]
> Aug 09 11:32:15.55: [19334]: --> [2:OK]
> Aug 09 11:32:15.55: [19334]: <-- [5:ATM0\r]
> Aug 09 11:32:15.56: [19334]: --> [2:OK]
> Aug 09 11:32:15.56: [19334]: MODEM input buffering enabled
> Aug 09 11:32:15.57: [19334]: <-- [12:AT+FCLASS=1\r]
> Aug 09 11:32:15.67: [19334]: --> [2:OK]
> Aug 09 11:32:15.68: [19334]: DIAL 8099
> Aug 09 11:32:15.68: [19334]: <-- [10:ATDT8099@\r]
> Aug 09 11:32:28.08: [19334]: --> [7:CONNECT]
> Aug 09 11:32:28.08: [19334]: MODEM input buffering disabled
> Aug 09 11:32:30.01: [19334]: --> HDLC<40:FF C0 04 00 00 88 01 00 51 92 
> 26 96 B6 A6 76 CE 96 F6 76 86 36 04 E2 B6 46 12 00 F2 00 01 01 1B 71 80 
> 80 80 80 20 69 14>

got NSF frame, it's non-final, so tried to get next HDLC frame

> Aug 09 11:32:30.01: [19334]: --> [2:OK]
> Aug 09 11:32:30.01: [19334]: <-- [9:AT+FRH=3\r]
> Aug 09 11:32:30.02: [19334]: --> [7:CONNECT]
> Aug 09 11:32:33.81: [19334]: MODEM TIMEOUT: receiving HDLC frame data

Something went wrong here. We expected HDLC frame, but had not received it
before 3.1 sec timeout was expired; try again

> Aug 09 11:32:33.81: [19334]: <-- data [1]
> Aug 09 11:32:34.01: [19334]: MODEM TIMEOUT: reading line from modem
> Aug 09 11:32:34.01: [19334]: MODEM <Timeout>
> Aug 09 11:32:34.01: [19334]: DELAY 200 ms
> Aug 09 11:32:34.21: [19334]: <-- [9:AT+FRH=3\r]
> Aug 09 11:32:34.22: [19334]: --> [7:CONNECT]
> Aug 09 11:32:41.21: [19334]: MODEM TIMEOUT: waiting for HDLC flags

failed again

> Aug 09 11:32:41.21: [19334]: <-- data [1]
> Aug 09 11:32:41.41: [19334]: MODEM TIMEOUT: reading line from modem
> Aug 09 11:32:41.41: [19334]: MODEM <Timeout>
> Aug 09 11:32:41.41: [19334]: DELAY 200 ms
> Aug 09 11:32:41.61: [19334]: <-- [9:AT+FRH=3\r]
> Aug 09 11:32:48.61: [19334]: --> [0:]
> Aug 09 11:32:48.61: [19334]: MODEM <Empty line>
> Aug 09 11:32:48.61: [19334]: MODEM TIMEOUT: waiting for v.21 carrier
> Aug 09 11:32:48.61: [19334]: <-- data [1]
> Aug 09 11:32:48.81: [19334]: MODEM TIMEOUT: reading line from modem
> Aug 09 11:32:48.81: [19334]: MODEM <Timeout>
> Aug 09 11:32:48.81: [19334]: DELAY 200 ms
> Aug 09 11:32:49.01: [19334]: <-- [9:AT+FRH=3\r]
> Aug 09 11:32:56.01: [19334]: --> [0:]
> Aug 09 11:32:56.01: [19334]: MODEM <Empty line>
> Aug 09 11:32:56.01: [19334]: MODEM TIMEOUT: waiting for v.21 carrier
> Aug 09 11:32:56.01: [19334]: <-- data [1]
> Aug 09 11:32:56.21: [19334]: MODEM TIMEOUT: reading line from modem
> Aug 09 11:32:56.21: [19334]: MODEM <Timeout>
> Aug 09 11:32:56.21: [19334]: DELAY 200 ms
> Aug 09 11:32:56.41: [19334]: <-- [9:AT+FRH=3\r]
> Aug 09 11:33:00.52: [19334]: --> [10:NO CARRIER]

[...]

Now remote seems to hang up; all other AT+FRH gave ERROR until T1
timeout was expired.

I don't see here any fax protocol violation from our side, although we
could give up earlier. Yes, you did not setup flow control properly, but
this should not affect HDLC handshake... Maybe the modem is
"overconfigured" by wrong init string? I don't know. Does Hylafax *always*
behaves so with this specific fax machine? Have you tried other Class1 modem?

Hope to hear from you soon,
Dmitry




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