Summary:ASTERISK-20242: Error sending fax: FAX_FAILURE_PROTOCOL_ERROR
Reporter:Steven Wheeler (swheeler)Labels:
Date Opened:2012-08-16 10:34:53Date Closed:2014-01-03 18:25:36.000-0600
Status:Closed/CompleteComponents:Applications/app_fax Resources/res_fax
Versions: Frequency of
Environment:CentOS 5.6 FinalAttachments:( 0) 1344975108.133524.pcap
( 1) 1344975108.133524.wav
( 2) console_output.txt
Description:We are often seeing this error 'FAX_FAILURE_PROTOCOL_ERROR' message when sending and receiving faxes.  There are not many details, but I will attach the fax debug and t.38 capture output for one example.  Please let me know if there is any more information you need.
Comments:By: Steven Wheeler (swheeler) 2012-08-16 10:43:24.559-0500

The asterisk console output.

By: Matt Jordan (mjordan) 2012-08-21 15:14:54.072-0500

Can you provide the full T.38 RTP stream in the context of the SIP messages as well?

And why is all of the T.38 traffic occurring on localhost?

By: Steven Wheeler (swheeler) 2012-08-21 15:53:48.322-0500

All I have is the output generated by Asterisk's feature: fax set t38cap on
Each time I have used this I have noticed that the src and dst are always localhost:1 and localhost:2.

By: Rusty Newton (rnewton) 2012-08-24 11:16:43.217-0500

Can you use tcpdump to grab a full capture covering all SIP and T.38 traffic?

By: Steven Wheeler (swheeler) 2012-08-24 11:24:21.097-0500

I can try, this is a production environment so there is a great deal of traffic.  It may take a couple days to catch it.

By: Rusty Newton (rnewton) 2012-09-19 20:45:35.758-0500

Steven, is this problem still occurring? Were you able to get the full packet capture?

By: Steven Wheeler (swheeler) 2012-09-20 09:18:32.136-0500

Yes I am still trying to get a capture of this.  It is proving more difficult than I originally thought.

By: Steven Wheeler (swheeler) 2012-11-08 16:44:11.599-0600

I am not sure I will be able to get a capture with the SIP as well.  The problem is I have to leave the packet captures running for days on a production system while waiting for the problem to occur.  But the capture files grow to be several gigabytes and Wireshark is unable to handle files that large.  Breaking the captures into chunks small enough for Wireshark to process causes the call to be distributed across several files.  Is there nothing we can tell from the T.38 alone?

By: Walter Doekes (wdoekes) 2012-11-09 03:34:29.290-0600

tcpdump -C 20 -W 100 will break up pcaps into 100 rotating 20MB pcaps.

mergecap will join pcaps as necessary.


By: Rusty Newton (rnewton) 2012-12-13 17:51:14.848-0600

Matt Jordan and I looked at this. We couldn't find anything obvious. You'll need a full SIP,T.38 and RTP capture for a *working* fax and a *failing* fax from the same endpoint. With that we should be able to pinpoint what is going on. I'll suspend the issue until you have got the debug, then we can re-open. Feel free to shoot me an E-mail, or ping a bug marshal on #asterisk-bugs.

By: Rusty Newton (rnewton) 2013-11-12 16:38:08.894-0600

Steven do you happen to have an Asterisk log showing verbose and debug messages output by Asterisk during the call shown in the pcap?

Also, to clarify:

bq. The fax servers trained down to 9600 before we were able to successfully receive the pages.

Did you successfully receive the pages despite Asterisk reporting it as a failure? Or did the fax not actually complete?


By: Steven Wheeler (swheeler) 2013-11-13 13:04:25.978-0600

According to the packet capture the pages were received but Asterisk did not generate a tiff file and the status was set to FAILED.

By: Matt Jordan (mjordan) 2013-12-07 20:53:16.871-0600

First, you'll need to hit "Send Back" when entering input. Otherwise, bug marshals often won't see the input and the issue may not get addressed.

Second, please attach a full DEBUG log showing the partially received fax. The VERBOSE and FAX messages show some of what occurred, but unfortunately, they only show that the T.38 stack finished the fax but not why it thought it was a partial receive.

By: Rusty Newton (rnewton) 2014-01-03 18:25:22.548-0600

Suspended due to lack of activity. Please request a bug marshal in #asterisk-bugs on the IRC network irc.freenode.net to reopen the issue should you have the additional information requested.  Further information can be found at http://www.asterisk.org/developers/bug-guidelines