[OpenSIPS-Users] B2B is causing an ERROR:tm:partial_dlg_matching: INV/From not parsed

[Bela H]

Hello,

I am using the OpenSIPS 3.2.11 on Debian 11 and using the B2BUA for hair pining to the own OpenSIPS.  With sngrep both the original and the B2B calls have clean SIP signalling.
However, after hang up the call there is at least (it depends if A or B subscriber ends the call) the original call dialog hangs up with call state: "state": 3 , from GUI: “Confirmed Not Acknowledged Call” and the syslog shows many different errors, e.g. “ERROR:tm:partial_dlg_matching: INV/From not parsed”

If I delete the hanging dialog either from GUI or with cli OpenSIPS crashes:
(cli): mi dlg_end_dlg dialog_id=f0c.57deb687
ERROR: communication exception for 'dlg_end_dlg' returned: could not decode json: ''
ERROR: Is OpenSIPS running?
(cli):

Mar  2 17:34:19 sbc1t kernel: [97206.071119] traps: opensips[8143] general protection fault ip:7f1671388250 sp:7ffcc4013100 error:0 in tracer.so[7f167137c000+13000]
Mar  2 17:34:19 sbc1t /usr/sbin/opensips[8143]: CRITICAL:db_mysql:wrapper_single_mysql_stmt_execute: driver error (1048): Column 'from_tag' cannot be null
Mar  2 17:34:19 sbc1t /usr/sbin/opensips[8143]: ERROR:acc:acc_db_request: failed to insert into acc table
Mar  2 17:34:19 sbc1t /usr/sbin/opensips[8143]: ERROR:dialog:build_dlg_t: invalid cseq
Mar  2 17:34:19 sbc1t /usr/sbin/opensips[8143]: ERROR:dialog:send_leg_bye: failed to create dlg_t
Mar  2 17:34:19 sbc1t /usr/sbin/opensips[8143]: CRITICAL:tracer:proto2str: unsupported proto 0
Mar  2 17:34:19 sbc1t /usr/sbin/opensips[8143]: CRITICAL:tracer:proto2str: unsupported proto 0
Mar  2 17:34:19 sbc1t /usr/sbin/opensips[8143]: CRITICAL:core:sig_usr: segfault in process pid: 8143, id: 2
Mar  2 17:34:19 sbc1t /usr/sbin/opensips[8141]: INFO:core:handle_sigs: child process 8143 exited by a signal 11
Mar  2 17:34:19 sbc1t /usr/sbin/opensips[8141]: INFO:core:handle_sigs: core was not generated
Mar  2 17:34:19 sbc1t /usr/sbin/opensips[8141]: INFO:core:handle_sigs: terminating due to SIGCHLD


In the script the B2B part is very simple:

loadmodule "b2b_entities.so"
modparam("b2b_entities", "script_req_route", "b2b_request")
modparam("b2b_entities", "script_reply_route", "b2b_reply")
modparam("b2b_entities", "b2b_key_prefix", "B2B1")
modparam("b2b_entities", "db_mode", 1)
#modparam("b2b_entities", "replication_mode", 0)

#### B2B_Logic ####
loadmodule "b2b_logic.so"
modparam("b2b_logic", "init_callid_hdr", "Init-CallID")
modparam("b2b_logic", "server_address", "sip:own_address")
modparam("b2b_logic", "db_url", "mysql://xxxxxxx@yyyyyyyy")

A certain call scenario triggers the B2B call:
b2b_init_request("top hiding");

request and reply routes are just xlogs.

The issue can be simulated any time with the same call scenario.
Is anyone able to help or just report it as a bug?

Cheers,
Bela
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensips.org/pipermail/users/attachments/20230302/b47775bb/attachment.html>