Re: [OpenSBC] OpenSBC scaling in B2BUA mode

[Joegen E. B. <joe...@gm...>]

I would check the max file handles per process setting.   However, it 
would be good if you could also test gradual rise to see if it is the 
CPS that OSBC can't handle or is it the concurrent calls.    From 
experience  the optimum CPS for opensbc is between 15-20.   Beyond that, 
the queue can grow beyond recovery.  Increasing CPS is one of the major 
deliverables in OpenSBC version 2.  A lot of redesign is happening at 
this very moment to make this happen.

Joegen

craig guy wrote:
> Hi,  I am running an instance of OpenSBC 1.1.5 and opensipstack pulled from CVS on 26 January.  I'm running it in B2BUA mode with 'always proxy media' set.  The instance of OpenSBC has two interfaces, one external on the internet and one internal.  There are 2 x Asterisk servers sitting on the internal interface with incoming calls being routed to them in round robin fashion.  I sent it 26 concurrent calls (setup in about 1 second) in its default configuration successfully.  I then upped the ante by sending it 120 concurrent calls (all hitting OpenSBC within a second) and it just stopped working (process died).  What settings should I configure in order to support 120 concurrent calls with media proxying?  Each call will have 4 rtp sessions and hence 480 total rtp sessions.  The default logs don't really show anything useful, they are fine until they stop at the point OpenSBC dies.  OS is Centos 5.
>
>
>  Craig
>
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Opensipstack-osbcdevel mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/opensipstack-osbcdevel
>
>
>