April 2018 - FreeSWITCH-users - FreeSWITCH Mailing Lists

Freeswitch crash after reloadxml
by Nikolay Zaytsev 05 Apr '18

05 Apr '18

Hi everybody, My freeswitch machine sometimes crashes when I do 'reloadxml' and there is nothing neither in the freeswitch logs nor in system logs. I am using the following configuration: - Centos 7 on Amazon EC2 m4.large instance - Freeswitch 1.6.2 with a maximum load of 250 concurrent calls - Dialplan is written in python and working via ESL Does anybody have ideas? -Best regards, Nikolay Zaytsev

2 1

Extract variables from 300 redirect
by Joseph Waite 04 Apr '18

04 Apr '18

Hi Guys I am wondering if its possible to extract a custom variable from a 300 Redirect response from a billing system? Effectively I am trying to add a custom variable to the 300 message to give a maximum call length to then use in the dial plan for a Scheduled Hangup. Any pointers of how to do this or the magic words to search confluence/google with? Regards

3 2

Using lcr in lua session and fail_on_single_reject problem
by Mimiko 04 Apr '18

04 Apr '18

Hello. I want to originate a call using lcr from lua script: freeswitch.Session("{ignore_early_media=true,originate_timeout=15,fail_on_single_reject=^^:NO_ANSWER:CALL_REJECTED:NORMAL_CLEARING:USER_BUSY}lcr/default/"..number) Using this command, originate_timeout is not taken into account and all routes are tried regardless of response reason, for example NO_ANSWER. If I remove fail_on_single_reject, then originate_timeout works but same problem: all routes are tried, except when the call was answered. In diaplan <action application="lcr" data="${destination_number} default"/> <action application="bridge" data="{fail_on_single_reject=^^:NO_ANSWER:CALL_REJECTED:NORMAL_CLEARING:USER_BUSY}${lcr_auto_route}"/> this works fine. Maybe its due to using bridge, not lcr directly. But in lua ${lcr_auto_route} variable is not available. Any suggestion? FreeSWITCH Version 1.5.6b+git~20130928T022323Z~6b9382290d (git 6b93822 2013-09-28 02:23:23Z) Yes, it is old, but can't update at this time.

1 1

mod_verto error when using Postgres DSN - no such table: json_store
by Michael Avers 04 Apr '18

04 Apr '18

Hello, I'm using Postgres in the core and everywhere else. Relevant modules such as callcenter, fifo, etc. all have the ODBC defined and working with the tables just fine. The only issue is Verto - I have json_store table in Postgres but Verto doesn't appear to be accessing it at all whenever mod_verto is loaded or reloaded. Getting this error: 2018-04-04 07:05:31.695917 [CONSOLE] switch_loadable_module.c:2034 mod_verto unloaded.*2018-04-04 07:05:31.695917 [ERR] switch_core_sqldb.c:587 NATIVE SQL ERR [no such table: json_store]** **select name from json_store where name=''* 2018-04-04 07:05:31.695917 [INFO] mod_verto.c:4749 Secure key and cert specified It says no such table because I did not have the sqlite json.db file, but after I put it back in it's able to read from it. So clearly it looks like it's not even trying to access json_store in the Postgres DB. I even added a <param name="odbc-dsn".. to verto.conf.xml even though it's not documented anywhere, still no luck. Any ideas? Is mod_verto not able to access the table when using Postgres? FYI this is on latest master and also 1.6.20 production release Thanks Mike

1 0

Freeswitch Crashed- Segmentation Fault
by Achintha 03 Apr '18

03 Apr '18

Hi all, Our freeswitch server handles 3000 – 4000 calls per day. We have custom module to rout calls. Anyway freeswitch was crashed two times I have attached core dump on “ https://pastebin.freeswitch.org/view/4463025f ” Our freeswitch version is “1.6.20-37-987c9b9~64bit (-37-987c9b9 64bit)” Please advise to solve this matter asap. -- Best Regards.. Achintha Udukumbura

4 4

Bridge and+or carrier failover
by Shane Mitchell 03 Apr '18

03 Apr '18

Hi, I'm trying to implement carrier failover with the bridge application, specifically when working with multiple simultaneous destinations. For a single destination, I can use OR (|) and fail_on_single_reject to appropriately failover. For example (simplified): {fail_on_single_reject=user_busy}sofia/gateway/GW1/ENDPOINT1|sofia/gateway/GW2/ENDPOINT1 However with multiple simultaneous destinations, I'm having issues with approach. I've tried the following (simplified): {fail_on_single_reject=user_busy}sofia/gateway/GW1/ENDPOINT1|sofia/gateway/GW2/ENDPOINT1,sofia/gateway/GW3/ENDPOINT2|sofia/gateway/GW4/ENDPOINT2 This terminates all new channels when any one of them fails. What I'm trying to do is to stop bridge attempting a specific OR chain when one fails, but to continue executing the rest of the ANDs. Is there a way to achieve bridging to multiple destinations simultaneously, with each destination having its own list of failover gateways, and having failover for a destination only occur on certain cause codes (i.e. we don't want to fail over on user_busy/etc, but we do on other types of failures)? Thanks & regards, Shane Mitchell.

3 2

Conference auto-outcall to a callcenter number with verto agents
by Adridan Tudor 03 Apr '18

03 Apr '18

Hello, The outcall works and agents join the conference but the verto communicator agents don't see the other participants. mod_verto is not sending the conference-liveArray-join verto.event to the agents. If I auto-outcall directly the agents with (user/xxx@${domain}) everything works. My setup is like this: <action application="conference_set_auto_outcall" data="{conference_member_flags=moderator}sofia/internal/799@${domain}"/> <action application="conference" data="conf-${caller_id_number}@video-mcu-stereo"/> video-mcu-stereo profile has: <param name="conference-flags" value="video-floor-only|rfc-4579|minimize-video-encoding|livearray-json-status|json-events|livearray-sync"/> <extension name="my_callcenter"> <condition field="destination_number" expression="799"> <action application="set" data="hangup_after_bridge=true"/> <action application="set" data="continue_on_fail=true"/> <action application="set" data="originate_continue_on_timeout=true"/> <action application="set" data="call_timeout=15"/> <action application="callcenter" data="my_callcenter(a)x.x.x.x"/> </condition> </extension> in directory/default.xml I have: <param name="jsonrpc-allowed-methods" value="verto"/> <param name="jsonrpc-allowed-event-channels" value="demo,conference,presence,conference-liveArray,conference-mod"/> Any help is highly appreciated! Thanks, Tudor

1 1

postgres db contains non existent channels
by Roman Dissauer 01 Apr '18

01 Apr '18

Hi, we have about 800 connected users, 150 simultaneous calls and some user agents send sip messages with non UTF-8 characters. When FreeSWITCH tries to alter the channel information in the postgres db it throughs an error: 2018-03-29 15:47:46.794346 [ERR] switch_pgsql.c:656 Error executing query: ERROR: invalid byte sequence for encoding "UTF8": 0xf6 0x63 0x68 0x6c The Problem now is that we see some calls in the database which do not exist anymore (uuid_exists returns false). This are other calls, not the one which cause the error above! It seems that everytime the error happens, Freeswitch cannot delete other channels from db when the bye is arriving +/- 100ms near the error. We tried to set client_encoding in the postgres connection string to some options e.g. LATIN1 but that didn’t solve our problem. Whichever client_encoding we used, we always had some user agents with some unknown characters. How should we deal with that? Thanks, Roman RDI SOLUTIONS e.U. Hollenthon 105 2812 Hollenthon, Österreich T: +43 1 3530349 - 10 F: +43 1 3530349 - 99 roman.dissauer(a)rdi.at www.rdi.at

2 1