sr-users

sr-users@lists.kamailio.org

46 participants
37553 discussions

Hyperthreading and high request volumes in modern times
by Brooks Bridges 30 May '25

30 May '25

So, I have a platform that handles in the order of around 2500 or more call attempts per second. There are 2 servers that are built that handle this traffic, "old" and "new". The "old" design is a bit heavy, has multiple redis calls to external IO, does a lot of "work" on each invite before passing it along, etc. The "new" design is extremely lightweight, only a couple of htable comparisons and a dispatcher check on each invite before passing it along. The 2 servers are running different versions of Kamailio, however they're both in the 5.x train and I've found nothing in the changelogs that I believe would explain a significant difference in how child process CPU scheduling is performed, especially a detrimental change. I'm more than happy to be proven wrong about that though! The "old" system is running 5.4.8: This system is running 4 x 8 core CPUs (32 cores) with no hyperthreading (4 x E5-4627 v2 to be exact) and has multiple external IO interactions with things like redis, lots of "work" being done on each invite checking various things before allowing it to proceed. This is currently running 32 child processes and again, has no issues handling the requests whatsoever and has been running at this level and higher for literal years. The "new" system is running 5.8.3: This system is running 2 x 8 core CPUs (16 cores) *with* hyperthreading (2 x E5-2667 v4 to be exact) and has no external IO interactions with anything, and only does a couple of checks against some hash table values and dispatcher before allowing it to proceed. This is also currently running 32 child processes, however I am experiencing spikes in context switching and odd memory access issues (e.g. backtraces show things like "<error: Cannot access memory at address 0x14944ec60>" in udp_rvc_loop) with similar loads. What I'm looking for here is empirical knowledge about things specifically like context switching and if hyperthreading is detrimental in cases where I'm doing very high volume, low latency (e.g. no waiting on external replies, etc) work. Is it possible there are additional issues with the way hyperthreading works and the concepts behind it if the individual calls for CPU resources are so "quick" as to overwhelm the scheduler and cause congestion in the CPU's thread scheduler itself? Have you run into anything like this? Have you discovered that very low latency call processing can incur more hyperthreading overhead than the benefit you get from it? Do you have additional data I may be missing to help me justify to the C suite to change out the system architecture to provide a higher physical core count? Can you share it? Thanks!

6 14

Sca in cluster/load balancer
by mauro.celli76＠gmail.com 29 May '25

29 May '25

We have two kamailio in load balancing, mysql is in cluster mode with galera. We need to use sca, but kamailio db_cluster is not supported (i need to specify connection string to sql not cluster name) and if phone1 is registered on kamailio1 and phone2 on kamailio2, sca not work. It is possibile to implement cluster?

3 5

How to count the first re-register request?
by James Morrison 26 May '25

26 May '25

I want to count the first REGISTER request during an initial register or re-register from UE. In my scscf instance, I have defined att_init_reg and att_rereg counters to count the number of initial registers and re-registers respectively. In my early tests, the following piece of logic worked as expected to filter the very first register request (one that is sent from UE before getting challenged). However, more tests showed that during re-registers, UE sends a REGISTER request, containing an authorization header and non-empty response, in that case the code is unable to distinguish a re-register attempt and the related counter fails to increment. Is there a way to fix this issue? #!ifdef WITH_XHTTP_PROM /* Only count the first register request */ if (!is_present_hf("Authorization") || search_hf("Authorization","response=\"\"","f")) { if ($var(impu_registered) != 1) { prom_counter_inc("att_init_reg", "1"); } else { prom_counter_inc("att_rereg", "1"); } } #!endif

3 5

5.7.7 release, tag and docker build
by Henning Westerholt 23 May '25

23 May '25

Hello, I have noticed that the git tag for release 5.7.7 was missing, I've just set it. The 5.7.7-bullseye docker image is also missing from https://github.com/kamailio/kamailio-docker/pkgs/container/kamailio Not sure if this is done automatically, if not - please can somebody please trigger the workflow to create it? Thank you, Henning

1 0

Help Adding WebRTC to Existing SIP Setup
by Fernando Lopes 22 May '25

22 May '25

Hi everyone, I currently have a working SIP setup, but I'm trying to get WebRTC working alongside it and haven’t had much success so far. Does anyone have tips or best practices for enabling both SIP and WebRTC in the same configuration? Is it possible to handle both within a single Kamailio config? Or would I need some kind of conditional logic (like if/else) to manage how the config handles different types of clients? My setup looks like this: Router → Kamailio → Kubernetes (Asterisk inside the cluster) I was having some issues with the Contact header, but I managed to resolve that. Now I'm stuck figuring out how to make WebRTC work cleanly alongside SIP. Any advice would be appreciated! Thanks!

2 1

Kamailio v5.7.7 Released
by Daniel-Constantin Mierla 21 May '25

21 May '25

Hello, Kamailio SIP Server v5.7.7 stable release is out. This is a maintenance release of the latest stable branch, 5.7, that includes fixes since the release of v5.7.6. There is no change to database schema or configuration language structure that you have to do on previous installations of v5.7.x. Deployments running previous v5.7.x versions are strongly recommended to be upgraded to v5.7.7. Note that 5.7 is now the third last stable branch, v5.7.7 being the last planned released in 5.7.x series. The latest two stable branch are 5.8 and 6.0, with v6.0.1 being released a while ago. For more details about version 5.7.7 (including links and guidelines to download the tarball or from GIT repository), visit: * https://www.kamailio.org/w/2025/05/kamailio-v5-7-7-released/ RPM, Debian/Ubuntu packages will be available soon as well. Many thanks to all contributing and using Kamailio! Cheers, Daniel -- Daniel-Constantin Mierla (@ asipto.com) twitter.com/miconda -- linkedin.com/in/miconda Kamailio Consultancy, Training and Development Services -- asipto.com

1 0

Releasing v5.7.7
by Daniel-Constantin Mierla 21 May '25

21 May '25

Hello, I am considering to release Kamailio v5.7.7 (out of branch 5.7) on Wednesday, May 21, 2025, as the last one from the 5.7.x series, to mark the end of support for it. Currently 5.8.x and 6.0.x are the latest two maintained branches. Cheers, Daniel -- Daniel-Constantin Mierla (@ asipto.com) twitter.com/miconda -- linkedin.com/in/miconda Kamailio Consultancy, Training and Development Services -- asipto.com

1 1

Re: stir/shaken integration with clearip redirect method
by JR Richardson 20 May '25

20 May '25

I appreciate all the feedback Fellas, checking the docs, there is no need to change the RURI, they do what some E164 fixup presented without the +. I'll follow up after testing, got a bit to chew one here. Thanks! JR Richardson

1 0

stir/shaken integration with clearip redirect method
by JR Richardson 19 May '25

19 May '25

Hey Folks, I'm integrating STIR/SHAKEN using kamailio, but not using self hosted certs with the secsipid module, just using a redirect from clearip on the SIP Calls. I've got the basic connection to clearip using t_relay_to_tcp for testing, but I'm struggling with the checks and timeouts. I have call progress past the initial request where I need to wait for a response. Should I put the clearip invite in its own route block use return? Has anyone successfully integrated the redirect method with kamailio to clearip and can give me some example route checks in the kamailio configs? Thanks. JR

4 6

Server closing TLS connection
by Ihor Olkhovskyi 16 May '25

16 May '25

Hello, I'm having a strange issue, when Kamailio(?) is closing TLS connection just in the middle of session. Usually I see a sort of timeout ~ 600 secs. But it does not ring a bell for me. With strace I've found, that it's actually not Kamailio, but underlying SSL library is doing this, according to line I've got from strace [2025-05-16 10:00:30] PID=1337952 FD=0 SYSCALL=close -> [No socket info] Raw: sendto(3, "<183>May 16 10:00:30 /usr/local/sbin/kamailio[1337952]: DEBUG: tls [tls_server.c:712]: tls_h_tcpconn_close_f(): Closing SSL connection 0x7f01a1ab1870\n", 150, MSG_NOSIGNAL, NULL, 0) = 150 As I don't see any explicit calls of tls_h_tcpconn_close_f in the source code, I assume it's a callback from the external OpenSSL(?) module. What would be the best way to debug why OpenSSL(?) is terminating connection, cause on Wireshark I just do see usual TCP connection with keepalives [image: image.png] kamailio -v version: kamailio 5.8.3 (x86_64/linux) 6f8a04 flags: USE_TCP, USE_TLS, USE_SCTP, TLS_HOOKS, USE_RAW_SOCKS, DISABLE_NAGLE, USE_MCAST, DNS_IP_HACK, SHM_MMAP, PKG_MALLOC, MEM_JOIN_FREE, 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_BLOCKLIST, HAVE_RESOLV_RES, TLS_PTHREAD_MUTEX_SHARED ADAPTIVE_WAIT_LOOPS 1024, MAX_RECV_BUFFER_SIZE 262144, MAX_SEND_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. openssl --version OpenSSL 3.2.2 4 Jun 2024 (Library: OpenSSL 3.2.2 4 Jun 2024) enable_tls=yes enable_sctp=no tls_threads_mode=2 tcp_accept_no_cl=yes tcp_max_connections=63536 tls_max_connections=63536 tcp_accept_aliases=no tcp_async=yes tcp_connect_timeout=30 tcp_conn_wq_max=63536 tcp_crlf_ping=yes tcp_delayed_ack=yes tcp_fd_cache=yes tcp_keepalive=yes tcp_keepcnt=5 tcp_keepidle=30 tcp_keepintvl=10 tcp_linger2=30 tcp_script_mode=1 tcp_rd_buf_size=80000 tcp_send_timeout=10 tcp_wq_blk_size=2100 tcp_wq_max=10485760 open_files_limit=63536 Many thanks in advance! -- Best regards, Ihor (Igor)

1 1

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