[Freeswitch-users] inbound T.38 re-invite transaction stalls

[Jeff Pyle]

Hello,

This is on 1.10.6-release-18-1ff9d0a60e~64bit, the latest release version
from the repo.

I run the following API command to send a TIFF file with txfax:

originate
{sip_h_P-CustomHeader=value,ignore_early_media=true,fax_ident='Fax
Test',origination_caller_id_name='Fax
Test',origination_caller_id_number=2165551111,absolute_codec_string=PCMU at 20i
,fax_verbose=true,fax_enable_t38=yes,fax_enable_t38_request=yes,fax_use_ecm=no,fax_end_page=1,jitterbuffer_msec=200:200:200}sofia/external/
4045551111 at 1.2.3.4:5060 &txfax(/usr/local/etc/faxtiffs/smiley-fine.tif)

The INVITE goes out as expected, and the call connects normally.  One
second later, the far side sends a re-invite for T.38.  Local FreeSWITCH
shows the remote SDP on the console, sends a 100 Trying, and this appears
on the console:

[DEBUG] switch_core_media.c:5192 sofia/external/4045551111 at 1.2.3.4:5060 T38
ACCEPT on request
[DEBUG] switch_core_media.c:5296 sofia/external/4045551111 at 1.2.3.4:5060 T38
IS POSSIBLE on request

And then nothing.  No 200 OK, 488 Not Acceptable Here, nothing past the 100
Trying.  I can't tell why.

spandsp.conf.xml is vanilla with only ident, header and spool-dir updated.

The environment is relatively simple with all static public IPs, no NAT,
etc.  This particular FS instance serves only as an endpoint for testing
like this fax example.

This configuration has been working in general for years and specifically
on this box for almost a year.  Apparently something has changed somewhere
but I can't see what.  I'm hoping someone might be able to point me in a
direction to diagnose and solve this.


Regards,
Jeff
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freeswitch.org/pipermail/freeswitch-users/attachments/20210806/c3c72793/attachment.html>