sr-users

sr-users@lists.kamailio.org

39 participants
37513 discussions

High performance
by Muborakshohi Masoud 27 Apr '22

27 Apr '22

Hello! I have kamailio 5.5 configured with digest authentication. DB Mysql, 2 rtpengine one on the same server with kamailio, the other is separate. Rtpengins work in bridge mode, internal/external/external-ipV6. Kamailio listen 3 address: internal/external/external-ipV6 work in mixed mode: ipV4-ipV4, ipV4-ipV6, ipV6-ipV6. OS Debian 10, CPU 1 core, Ram 4gb, Disk not ssd. internal address is used for the gateway model of the connection to the PSTN. Before sending a call to PSTN, a request is sent to Billing by Diameter (request over the network, ims_charging ), Billing responds in the region of 10-17 milliseconds. usrloc db_mode = 0; UAS contacts are received according to the PUSH scheme (we send PUSH via http and get a REGISTER). children = 8 for each; When running synthetic tests with sipp, I get 36 cps as a maximum, but there are a lot of retrans. This is a test environment. Product: I plan to get 100 cps; 20k concurrent calls, hardware will be 128GB RAM ddr4-2400; Intel Xeon Silver 4216, Optane SSD. 2 such servers, kamailio/rtpengine are configured exactly according to the scheme of the test environment. Question: Can you please tell me how many children, shm/pv do I need? Is it worth going in the direction of asynchrony, or is the standard one enough for my purpose? I will be very glad for any help. Thank you! -- Отправлено из Mail.ru для Android

2 1

Intermittent error seen when using a python script in kamailio
by Wojciech Kopciewicz 27 Apr '22

27 Apr '22

Hi, Intermittently we are observing the following error when checking if given header is present in SIP INVITE: SystemError: returned a result with an error set The complete stack trace is (some names redacted): 11(28) ERROR: {1 195 INVITE nMHZDN.10nOIDt4ldOwz5cqz0XnMepA9} app_python3 [python_support.c:156]: python_handle_exception(): apy_exec: modify_headers((null)): Unhandled exception in the Python code: TypeError: an integer is required (got type NoneType) The above exception was the direct cause of the following exception: Traceback (most recent call last): File "/etc/kamailio/kamailio.py", line 25, in modify_headers self.modify_header("SomeHeader") File "/etc/kamailio/kamailio.py", line 29, in modify_header present = KSR.hdr.is_present(header_name) SystemError: returned a result with an error set Kamailio version (kamailio -v output): version: kamailio 5.4.6 (x86_64/linux) flags: USE_TCP, USE_TLS, USE_SCTP, TLS_HOOKS, USE_RAW_SOCKS, DISABLE_NAGLE, USE_MCAST, DNS_IP_HACK, SHM_MMAP, PKG_MALLOC, Q_MALLOC, F_MALLOC, TLSF_MALLOC, DBG_SR_MEMORY, USE_FUTEX, FAST_LOCK-ADAPTIVE_WAIT, USE_DNS_CACHE, USE_DNS_FAILOVER, USE_NAPTR, USE_DST_BLACKLIST, HAVE_RESOLV_RES, TLS_PTHREAD_MUTEX_SHARED ADAPTIVE_WAIT_LOOPS 1024, MAX_RECV_BUFFER_SIZE 262144, MAX_URI_SIZE 1024, BUF_SIZE 65535, DEFAULT PKG_SIZE 8MB poll method support: poll, epoll_lt, epoll_et, sigio_rt, select. id: unknown compiled with gcc 8.3.0 Operation system: Linux kamailio-5c7484c74d-vhhr8 5.4.0-1056-azure #58~18.04.1-Ubuntu SMP Wed Jul 28 23:14:18 UTC 2021 x86_64 GNU/Linux Are you able to help me find the root cause of this issue? Pozdrawiam / Best regards, Wojtek -- *For more information on how and why we collect your personal information, please visit our Privacy Policy <https://www.motorolasolutions.com/en_us/about/privacy-policy.html?elqTrackI…>.*

3 9

LCR issues (request_uri)
by Antoine DEBARRE 26 Apr '22

26 Apr '22

Hi, I'm trying to use LCR with kamailio v5.5.4 gw, rule and rule_target are stored on mysql. for lcr rules, do you know if prefix is mandatory ? I first wanted to start using request_uri instead and leave the prefix column null/empty because prefix seems not to support regexp, I don't want to add the same rule with different prefixes. but I'm facing issues my database is fulfilled with : lcr_gw +----+--------+--------------+---------------+----------+------+--------+------------+-----------+-------+--------+-------------+-------+---------+ | id | lcr_id | gw_name | ip_addr | hostname | port | params | uri_scheme | transport | strip | prefix | tag | flags | defunct | +----+--------+--------------+---------------+----------+------+--------+------------+-----------+-------+--------+-------------+-------+---------+ | 1 | 1 | carrier1 | 10.78.51.3 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | fd=carrier1 | 0 | NULL | | 2 | 1 | carrier2 | 10.78.51.3 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | fd=carrier2 | 0 | NULL | | 3 | 1 | carrier3 | 10.192.50.212 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | NULL | 0 | NULL | | 6 | 1 | carrier4 | 10.192.50.3 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | fd=carrier4 | 0 | NULL | | 10 | 1 | carrier5 | 10.192.50.3 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | fd=carrier5 | 0 | NULL | | 18 | 1 | carrier6 | 10.192.50.214 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | NULL | 0 | NULL | | 20 | 1 | carrier7 | 10.192.50.210 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | NULL | 0 | NULL | | 30 | 1 | carrier8 | 10.192.50.211 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | NULL | 0 | NULL | | 31 | 1 | carrier9 | 10.192.50.216 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | NULL | 0 | NULL | | 32 | 2 | media1 | 10.192.50.232 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | NULL | 0 | NULL | | 33 | 2 | media2 | 10.192.50.233 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | NULL | 0 | NULL | | 34 | 2 | media3 | 10.192.50.234 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | NULL | 0 | NULL | | 35 | 2 | media4 | 10.192.50.235 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | NULL | 0 | NULL | | 36 | 2 | media5 | 10.192.50.240 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | NULL | 0 | NULL | | 37 | 2 | media6 | 10.192.50.241 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | NULL | 0 | NULL | | 38 | 2 | media7 | 10.192.50.242 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | NULL | 0 | NULL | | 39 | 3 | xivo-231 | 10.192.50.231 | NULL | 5060 | NULL | NULL | 1 | NULL | NULL | cps=5;cc=50 | 0 | NULL | +----+--------+--------------+---------------+----------+------+--------+------------+-----------+-------+--------+-------------+-------+---------+ lcr_rule +----+------------------------+-----------------------------------+------------------+--------+----------+-----------+---------+---------+ | id | desc | request_uri | lcr_id | prefix | from_uri | mt_tvalue | stopper | enabled | +----+------------------------+-----------------------------------+------------------+--------+----------+-----------+---------+---------+ | 1 | Default | .+ | 1 | NULL | NULL | NULL | 0 | 1 | | 2 | France GSM | (0|(|00|\+)33)[67]\d{8} | 1 | NULL | NULL | NULL | 1 | 1 | | 3 | France | (0|(|00|\+)33)[1-59]\d{8} | 1 | NULL | NULL | NULL | 1 | 1 | | 4 | International default | (00|\+)[1-9].+ | 1 | NULL | NULL | NULL | 1 | 1 | | 5 | Italy | (00|\+)39.+ | 1 | NULL | NULL | NULL | 1 | 1 | | 14 | to_oldmedias | .+ | 2 | NULL | NULL | NULL | 0 | 1 | +----+------------------------+-----------------------------------+-------------------+--------+----------+-----------+---------+---------+ lcr_rule_target +----+--------+---------+-------+----------+--------+ | id | lcr_id | rule_id | gw_id | priority | weight | +----+--------+---------+-------+----------+--------+ | 25 | 1 | 3 | 3 | 1 | 1 | | 26 | 1 | 3 | 18 | 2 | 1 | | 28 | 1 | 3 | 30 | 3 | 1 | | 29 | 1 | 2 | 30 | 4 | 1 | | 30 | 1 | 2 | 3 | 2 | 1 | | 31 | 1 | 2 | 18 | 3 | 1 | | 32 | 1 | 2 | 20 | 1 | 1 | | 33 | 1 | 1 | 30 | 1 | 1 | | 34 | 1 | 12 | 1 | 1 | 1 | | 35 | 1 | 12 | 10 | 1 | 1 | | 37 | 2 | 14 | 32 | 1 | 1 | | 38 | 2 | 14 | 33 | 1 | 1 | | 39 | 2 | 14 | 34 | 1 | 1 | | 40 | 2 | 14 | 35 | 1 | 1 | | 41 | 2 | 14 | 36 | 1 | 1 | | 42 | 2 | 14 | 37 | 1 | 1 | | 43 | 2 | 14 | 38 | 1 | 1 | | 45 | 1 | 2 | 10 | 3 | 1 | +----+--------+---------+-------+----------+--------+ to test gws, I use kamcmd for example : kamcmd lcr.load_gws 1 s:0600000001 s:02111111111 s:0600000001 arg 1 : prefix or $rU arg 2 : from_uri arg 3 : request_uri result is OK with 5 gateways selected but in real test situation with the following config file : kamailio.cfg : request_route { ... if(!load_gws(1,$rU,$fU)) { sl_send_reply("500", "Cannot load gateways"); exit; } ## - dumping the list of lcr gateways $var(i) = 0; while( $(avp(lcr_gw_uri)[$var(i)]) != $null) { xlog("L_ERR","loaded gw_uri_avp[$var(i)]=’$(avp(lcr_gw_uri)[$var(i)])’\n"); $var(i) = $var(i) + 1; } if(!next_gw()) { sl_send_reply("503", "No gateways found"); exit; } } ..... => we receive a 503 "No gateways found", which indicate that load_gws() is OK, but wasn't able to parse avp lcr_gw_uri (we have no xlog output) Then I noticed that some calls are OK and others are KO (with same pattern for $rU) => lcr rule id 2 So I with debug=3 *call to 0611111111 => load_gws OK and next_gw OK* lcr [lcr_mod.c:3018]: do_from_gw(): request did not come from gw lcr [lcr_mod.c:3018]: do_from_gw(): request did not come from gw lcr [lcr_mod.c:3004]: do_from_gw(): request game from gw lcr [lcr_mod.c:3009]: do_from_gw(): added tag_avp <cps=5;cc=50> lcr [lcr_mod.c:2264]: ki_load_gws_furi(): load_gws(1, 0611111111, 0211111111) lcr [lcr_mod.c:2313]: ki_load_gws_furi(): from uri <0211111111> did not match to from regex <AMD> lcr [lcr_mod.c:2349]: ki_load_gws_furi(): request uri < sip:0611111111@10.192.50.227:5060> did not match to request regex <(00|\+)39.+> lcr [lcr_mod.c:2349]: ki_load_gws_furi(): request uri < sip:0611111111@10.192.50.227:5060> did not match to request regex <(00|\+)[1-9].+> lcr [lcr_mod.c:2349]: ki_load_gws_furi(): request uri < sip:0611111111@10.192.50.227:5060> did not match to request regex <(0|(|00|\+)33)[1-59]\d{8}> lcr [lcr_mod.c:2369]: ki_load_gws_furi(): added matched_gws[0]=[4, 0, 3, 2329862] lcr [lcr_mod.c:2369]: ki_load_gws_furi(): added matched_gws[1]=[5, 0, 1, 5553909] lcr [lcr_mod.c:2369]: ki_load_gws_furi(): added matched_gws[2]=[8, 0, 3, 7256660] lcr [lcr_mod.c:2369]: ki_load_gws_furi(): added matched_gws[3]=[7, 0, 2, 3129551] lcr [lcr_mod.c:2369]: ki_load_gws_furi(): added matched_gws[4]=[6, 0, 4, 2234577] lcr [lcr_mod.c:2085]: add_gws_into_avps(): added gw_uri_avp <6|sip:|0|||3543318538||5060||;transport=udp|0|2> with weight <2234577> lcr [lcr_mod.c:2085]: add_gws_into_avps(): added gw_uri_avp <4|sip:|0||fd=carrier5|53657610||5060||;transport=udp|0|2> with weight <2329862> lcr [lcr_mod.c:2085]: add_gws_into_avps(): added gw_uri_avp <8|sip:|0|||3593650186||5060||;transport=udp|0|2> with weight <7256660> lcr [lcr_mod.c:2085]: add_gws_into_avps(): added gw_uri_avp <7|sip:|0|||3560095754||5060||;transport=udp|0|2> with weight <3129551> lcr [lcr_mod.c:2085]: add_gws_into_avps(): added gw_uri_avp <5|sip:|0|||3526541322||5060||;transport=udp|0|2> with weight <5553909> lcr [lcr_mod.c:2610]: generate_uris(): r_uri <sip:0611111111@10.192.50.210:5060;transport=udp>, dst_uri <> lcr [lcr_mod.c:2904]: ki_next_gw(): added ruri_user_avp <0611111111> lcr [lcr_mod.c:2950]: ki_next_gw(): added tag_avp <> lcr [lcr_mod.c:2965]: ki_next_gw(): added defunct_gw_avp <3526541322> lcr [lcr_mod.c:2610]: generate_uris(): r_uri <sip:0611111111@10.192.50.212:5060;transport=udp>, dst_uri <> lcr [lcr_mod.c:2950]: ki_next_gw(): added tag_avp <> lcr [lcr_mod.c:2965]: ki_next_gw(): added defunct_gw_avp <3560095754> lcr [lcr_mod.c:2610]: generate_uris(): r_uri <sip:0611111111@10.192.50.214:5060;transport=udp>, dst_uri <> lcr [lcr_mod.c:2950]: ki_next_gw(): added tag_avp <> lcr [lcr_mod.c:2965]: ki_next_gw(): added defunct_gw_avp <3593650186> lcr [lcr_mod.c:2610]: generate_uris(): r_uri <sip:0611111111@10.192.50.3:5060;transport=udp>, dst_uri <> lcr [lcr_mod.c:2950]: ki_next_gw(): added tag_avp <fd=carrier5> lcr [lcr_mod.c:2965]: ki_next_gw(): added defunct_gw_avp <53657610> lcr [lcr_mod.c:2610]: generate_uris(): r_uri <sip:0611111111@10.192.50.211:5060;transport=udp>, dst_uri <> lcr [lcr_mod.c:2950]: ki_next_gw(): added tag_avp <> lcr [lcr_mod.c:2965]: ki_next_gw(): added defunct_gw_avp <3543318538> the right gws are selected and call are routed to them *call to 0600000001 => load_gws OK and next_gw KO* lcr [lcr_mod.c:3018]: do_from_gw(): request did not come from gw lcr [lcr_mod.c:3018]: do_from_gw(): request did not come from gw lcr [lcr_mod.c:3004]: do_from_gw(): request game from gw lcr [lcr_mod.c:3009]: do_from_gw(): added tag_avp <cps=5;cc=50> lcr [lcr_mod.c:2264]: ki_load_gws_furi(): load_gws(1, 0600000001, 0211111111) lcr [lcr_mod.c:2313]: ki_load_gws_furi(): from uri <0170619309> did not match to from regex <AMD> lcr [lcr_mod.c:2349]: ki_load_gws_furi(): request uri < sip:0600000001@10.192.50.227:5060> did not match to request regex <(00|\+)39.+> the process stops after trying to match rule with regexp <(00|\+)39.+> Don't know if it's a bug or if I misunderstood something ! Thanks in advance ! Best regards, -- <https://app.livestorm.co/comdata-group/metaverse-what-will-cx-look-like-in-…>

3 8

Local Registration Cache
by Linux Vince 26 Apr '22

26 Apr '22

Hello, I am trying to implement scenario described in this thread by Daniel-Constantin Mierla. I am able to cache registration using save but i am not able to handle unregistration. Unregister function of registrar module is not available in onreply routes and I cannot unregister until upstream registrar returns 200 OK on Expire 0 message. Is there any workaround to fix this? Thanks,

3 5

SSL errors with websocket clients
by Vinicius Kwiecien Ruoso 22 Apr '22

22 Apr '22

Hi all! How are you? I have a kamailio instance and some users are having issues connecting via websocket using TLS. The logs show SSLv3 errors. Cannot find why that error would show up if SSLv2/3 is not enabled. Double checked it via SSLLabs that only TLSv1.2 is allowed in the service. Error sample: ``` 15(36) ERROR: tls [tls_util.h:42]: tls_err_ret(): TLS accept:error:14094416:SSL routines:ssl3_read_bytes:sslv3 alert certificate unknown 15(36) ERROR: <core> [core/tcp_read.c:1512]: tcp_read_req(): ERROR: tcp_read_req: error reading - c: 0x7fafc8768190 r: 0x7fafc8768278 (-1) ``` This instance has close to 400 websocket connections opened during the day. I have a high log level enabled. I was not able to identify any reason why this happens. The certificate is a Letsencrypt certificate. The issue is that sometimes clients have a hard time connecting to the websocket and that maybe related. Below are a few more details about the config. Please let me know if you have any pointers on how to debug those connection issues. 8<----- TLS config: modparam("tls", "tls_method", "TLSv1.2+") modparam("tls", "verify_certificate", 0) modparam("tls", "require_certificate", 0) modparam("tls", "low_mem_threshold1", 0) modparam("tls", "low_mem_threshold2", 0) modparam("tls", "private_key", "/etc/certs/tls.key") modparam("tls", "certificate", "/etc/certs/tls.crt") version: kamailio 5.3.9 (x86_64/linux) 8<----- This is related to the issue https://github.com/kamailio/kamailio/issues/3085 (as pointed out it is not a problem in Kamailio code). Thanks, Vinicius

2 2

How to disable video in invite with both audio and video
by Simone de Blasiis 22 Apr '22

22 Apr '22

Hello to everyone, we have a Kamailio 5.5 + RTPEngine deployment. Some devices will send an invite with both audio and video in the offer and we are trying to disable the video portion of the invite by either setting the video port to 0 or by setting a=inactive in the video portion of the SDP packet. We know that sdpops has a function that would be able to remove the video portion but it violates standards so we can't use it. We need to disable the video in the SDP offers because one of the pstn providers we are using doesn't accept invites that contain a video offer. I've spent days searching but I can't find an obvious way of achieving this. We were originally doing it with textops but we've been having some issues where the sdp wouldn't be modified correctly in reinvites after the original call is in place so we're looking for alternatives. Can someone please help us? Thanks in advance Simone de Blasiis

3 4

DMQ sync issue between 5.4 and 5.5
by David Villasmil 20 Apr '22

20 Apr '22

Hello guys, We have a cluster of about 10 kamilios 5.4 with several million htable records. Restarting any server is not a problem, CPU goes ip for a second but everything comes back down to normal. We added a 5.5 by mistake and it won't start, it'll just eat up memory until it bails out with qm_malloc(): qm_malloc(0x149dd5ebb000, 16016) called from tm: h_table.c: build_cell(334), module: tm; Free fragment not found! Is there any known issue between DMQ 5.4 and 5.5? Regards, David Villasmil email: david.villasmil.work(a)gmail.com phone: +34669448337

1 1

"Re: TOPOS (5.1.[23]) breaks PRACK"
by Simone de Blasiis 20 Apr '22

20 Apr '22

Hello, we noticed the exact behaviour on Kamailio 5.5.4 and topos. If we disable topos the UPDATE messages are being recognized as loose in the WITHINDLG route and everything works. When topos is enabled the UPDATE messages fail the loose_route() check and even if we force the record_route() the message is still not being forwarded. I had just posted my own thread here <https://lists.kamailio.org/pipermail/sr-users/2022-April/114710.html> because I didn't see this one. I'll proceed opening an issue on GitHub so we can track this. Best regards Simone

1 1

Re: [SR-Users] TOPOS (5.1.[23]) breaks PRACK
by Simone de Blasiis 20 Apr '22

20 Apr '22

1 0

Problems with topos and UPDATE messages
by Simone de Blasiis 20 Apr '22

20 Apr '22

Hello to everyone, we are experiencing some issues with UPDATE messages being sent by the PSTN after an early media announcement. To reproduce the issue I have to call a specific number that has an early media announcement, since it's a paid phone number it has to announce how much it would cost before picking up the call. In this initial part the call is being negotiated in g711a. As soon as the announcement ends, the PSTN sends an UPDATE message containing the new desidered codec from the callee which is g729. At this moment Kamailio doesn't route the UPDATE message to the destination and it actually replies with a 404 message. I've attached a pcap with anonymized IPs to make it harder for bots. Here is a legend for the IPs: - - 12.34.123.123 is Kamailio - 62.23.19.74 is the PSTN - 1.2.130.40 is the client making the call I found a similar report <https://github.com/kamailio/kamailio/issues/2659> in the GitHub issues of the repository and we tried forcing the record_route() as suggested but, as the OP reported in his last message, it's not working. As soon as we disable topos, the UPDATE message is correctly relayed to the caller. In the documentation we can't find anything that could give us a hint on what's happening. Does anyone have any idea on what could be happening? Thanks in advance Simone

1 0

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

sr-users