[SR-Users] Memory configuration

Joe Baran jbaran at whitelabelcomm.com
Sat Sep 23 13:38:00 CEST 2017


I was editing /etc/default/kamailio and changing the SHM_MEMORY parameter,
however it does not appear to be doing anything.  When I modified the
PKG_MEMORY setting in that file, it did seem to take effect.  I’m not clear
on where I change the SHM_MEMORY setting to have it actually take effect.



*From:* Joe Baran [mailto:jbaran at whitelabelcomm.com]
*Sent:* Monday, September 18, 2017 8:27 AM
*To:* 'miconda at gmail.com' <miconda at gmail.com>; 'Kamailio (SER) - Users
Mailing List' <sr-users at lists.kamailio.org>
*Subject:* RE: [SR-Users] Memory configuration



This particular system is a test system, so there’s not a TON of activity.
Though the system itself has maybe 8 GB of RAM.  So we’re talking a handful
of calls throughout a given day.  The service appears to crash maybe once
every couple of days.  I bumped the SHM_MEMORY up to 512 over the weekend
and restarted to see what impact that has.



Which specific command are you looking for the output of, I’m blanking on
which to run.



*From:* Daniel-Constantin Mierla [mailto:miconda at gmail.com
<miconda at gmail.com>]
*Sent:* Monday, September 18, 2017 3:17 AM
*To:* Joe Baran <jbaran at whitelabelcomm.com>; Kamailio (SER) - Users Mailing
List <sr-users at lists.kamailio.org>
*Subject:* Re: [SR-Users] Memory configuration



If you have a lot of traffic and registered devices, you can run out of it
quickly. What is the number of registered devices and call setups per
second/active calls at peak?

When you do a restart, can you grab the stats and see what are the values
for shared memory?

Cheers,
Daniel



On 17.09.17 21:12, Joe Baran wrote:

# Amount of shared memory to allocate for the running Kamailio server (in
Mb)

SHM_MEMORY=64



# Amount of private memory to allocate for the running Kamailio server (in
Mb)

PKG_MEMORY=8







version: kamailio 5.0.2 (x86_64/linux)

flags: STATS: Off, USE_TCP, USE_TLS, USE_SCTP, TLS_HOOKS, DISABLE_NAGLE,
USE_MCAST, DNS_IP_HACK, SHM_MEM, 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

ADAPTIVE_WAIT_LOOPS=1024, MAX_RECV_BUFFER_SIZE 262144, MAX_LISTEN 16,
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 on 11:34:41 Jun 28 2017 with gcc 4.8.5



*From:* Daniel-Constantin Mierla [mailto:miconda at gmail.com]
*Sent:* Friday, September 15, 2017 10:53 AM
*To:* Kamailio (SER) - Users Mailing List <sr-users at lists.kamailio.org>;
Joe Baran <jbaran at whitelabelcomm.com>
*Subject:* Re: [SR-Users] Memory configuration



What is the value of shared memory you set for kamailio (-m parameter)?
Maybe it is not sufficient for what you need to do.

Then what version of kamailio do you run (kamailio -v)?

The process that crashed (29438) has a different pid than the one (29468)
printing the memory errors, so may or may not be related.



On 14.09.17 20:36, Joe Baran wrote:

I feel fairly confident I’m running into a memory tuning issue, but I’m
struggling to determine the next step.  Can someone give me some guidance
on how to adjust settings to allow the system to stop crashing kamailio?



Sep 14 18:25:15 server1 kamailio[29468]: ERROR: <core>
[core/mem/q_malloc.c:292]: qm_find_free(): qm_find_free(0x7fa9eeeea000,
144); Free fragment not found!

Sep 14 18:25:15 server1 kamailio[29468]: ERROR: <core>
[core/mem/q_malloc.c:425]: qm_malloc(): qm_malloc(0x7fa9eeeea000, 144)
called from kazoo: kz_amqp.c: kz_amqp_async_query(1357), module: kazoo;
Free fragment not found!

Sep 14 18:25:15 server1 kamailio[29468]: ERROR: kazoo [kz_amqp.c:1359]:
kz_amqp_async_query(): failed to allocate kz_amqp_cmd in process 29468

Sep 14 18:25:15 server1 kamailio[29468]: INFO: <script>:
kGddp3V0bA6nitnk25dB4w..|log|failed to send Kazoo query for authentication
xxxxxxxxxxxxxxxxxx

Sep 14 18:25:15 server1 kamailio[29468]: ERROR: tm [t_reply.c:494]:
_reply_light(): cannot allocate shmem buffer

Sep 14 18:25:15 server1 kamailio[29468]: ERROR: sl [sl.c:271]:
send_reply(): failed to reply stateful (tm)

Sep 14 18:25:16 server1 kamailio[29483]: CRITICAL: <core>
[core/pass_fd.c:277]: receive_fd(): EOF on 108

Sep 14 18:25:16 server1 kamailio[29360]: ALERT: <core> [main.c:743]:
handle_sigs(): child process 29438 exited by a signal 11

Sep 14 18:25:16 server1 kamailio[29360]: ALERT: <core> [main.c:746]:
handle_sigs(): core was not generated



Thanks,



Joe Baran



_______________________________________________

Kamailio (SER) - Users Mailing List

sr-users at lists.kamailio.org

https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users



-- 

Daniel-Constantin Mierla

www.twitter.com/miconda -- www.linkedin.com/in/miconda

Kamailio Advanced Training - www.asipto.com

Kamailio World Conference - www.kamailioworld.com



-- 

Daniel-Constantin Mierla

www.twitter.com/miconda -- www.linkedin.com/in/miconda

Kamailio Advanced Training - www.asipto.com

Kamailio World Conference - www.kamailioworld.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.kamailio.org/pipermail/sr-users/attachments/20170923/238e6704/attachment.html>


More information about the sr-users mailing list