[Freeswitch-users] Enterprise/Production Quality?
Miguel Jesús López Valverde
mjlopez at smartic.es
Wed May 30 16:46:57 UTC 2018
Hello Geoff and Freeswitch-users:
Currently I have a Freeswitch server in development mode on Amazon EC2 instance and another, working in production, as a physical machine. Both are mounted on Linux Debian 8 - 64 bits and the Freeswitch version is 1.6.18 64bit in development and 1.6.19 64bit in production.
Also I’ve another Freeswitch servers in development and production, but this servers only work with TCP/UDP, no WSS and they haven’t this trouble.
These systems keep Sip and WSS-SipJs clients connected through their internal context. In both cases, they are connected to Proxies where they register and take calls with other external carriers.
All user accounts, connection information with proxies, etc. they are maintained through externals Mysql databases, (different for each system).
Everything works correctly but, on occasion, any of the Freeswitch systems working with WSS stops working without, apparently, offering information about the reason. This occurs every 2-3 weeks and it can happen to one of the two servers, the state that keep in this case the following:
The Freeswitch instance remains up.
The console maintains the connection and accepts the input of commands.
The internal and external interface appear to be up when the sofia status statement is executed.
There is connectivity with the MySQL databases.
No alerts or error messages are displayed in the console.
THERE ARE NO REGISTERED USERS, when executing the sofia status profile internal reg command, 0 registered users appear.
When executing the sofia status, I can’t see the internal profile but I can see the others profiles.
When executing the sofia status profile internal command, usually, the CLI is bloked and I need to exit with Ctrl+C.
Through netstat, it can be seen that the ports and protocols are kept correctly raised. I have compared netstat results with the affected and functioning instances and there are no differences.
Obtaining traces, for example, with ngrep, it is observed that Freeswitch does not respond to requests for registration or invites made by SIP clients on the intenal interface. It receives the requests correctly but does not return a response. SIP and WSS clients can not register either.
Obtaining traces on the port of the external interface, it is observed that Freeswitch sends registration requests to the proxy, which are answered by this with the typical unauthorized message and information to do the digest proces, which is not managed by Freeswitch and this returns to send new registers requests without negotiating digest of the previous answers.
If one of the interfaces is restarted, for example, with the sofia profile internal restart command, on some occasions the console remains dead without activity and, in others, the sentence is executed and ended correctly, but when consulting the state of the interfaces with, for example, sofia status, the internal context has disappeared. It also happens with the external context.
The only way to get the Freeswitch instance working correctly is to perform a reset of this, using the systemctl stop freeswitch / systemctl star freeswitch linux command, for example. After raising the bank again, everything goes back to work properly, the registration is made against the proxies and the clients Sip and Wss re-register correctly and the calls are processed again.
Checking the log of logs, (/usr/local/freeswitch/logs/freeswitch.log), there are no warnings, errors or anything that shows clues about the reason for the fall. Simply repeat attempts to register clients and against the proxy.
After restarting the instance and running, the time until the next incident is repeated is random. It can be a few hours or weeks. I have no way to reproduce it or induce the fault.
If you think of any reason that might lead Freeswitch to have this behavior or if you have suffered similar incidents, please, I would be very grateful if you can throw some light on me.
Miguel J. Lopez.
De: FreeSWITCH-users [mailto:freeswitch-users-bounces at lists.freeswitch.org] En nombre de Abaci B
Enviado el: miércoles, 30 de mayo de 2018 17:30
Para: FreeSWITCH Users Help <freeswitch-users at lists.freeswitch.org>
Asunto: Re: [Freeswitch-users] Enterprise/Production Quality?
While your issues definitely should be troubleshooted to find the cause of your problem, I will just answer your question if anyone has different experience, I have been using freeswitch in production for over 10 years with no major issues even with much higher load, and so have many others.
On Wed, May 30, 2018 at 10:45 AM, Geoff Mina <gmina at connectfirst.com <mailto:gmina at connectfirst.com> > wrote:
Is anyone out there actually using FS successfully in an enterprise environment?
We have deployed a handful of servers in an extremely simple configuration to allow standard SIP infrastructure to communicate with WebRTC clients.
We run ~150 concurrent users per host and we can’t go a week without something in the core of Sofia failing. We have seen hung profiles that simply don’t respond to REGISTER requests (tried every suggested tweak to no avail) as well as seeing FS hang every call in a RINGING state without actually ringing the end client.
These seem like pretty fundamental components we are struggling with. Our use case seems quite simple - yet the software (1.6.19) seems like it has never even seen a production deployment.
Anyone out there have a drastically different experience?
Professional FreeSWITCH Consulting Services:
consulting at freeswitch.org <mailto:consulting at freeswitch.org>
Official FreeSWITCH Sites
FreeSWITCH-users mailing list
FreeSWITCH-users at lists.freeswitch.org <mailto:FreeSWITCH-users at lists.freeswitch.org>
El software de antivirus Avast ha analizado este correo electrónico en busca de virus.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the FreeSWITCH-users