[Icecast] How analyze performance (buffering) problems ?

EISELE Pascal pe at letsgozic.com
Fri Nov 18 09:34:05 UTC 2005


EISELE Pascal a écrit :

> Hello,
>
> I'm looking for a method to track stream performance problems, I mean :
> why does listeners got buffering ?
> We are using two "front relay" that are hosted by differents providers.
> The front relays are connected to our main server (which encode) that is
> using two ADSL connections from two differents providers.
> How can I do to know if there is a problem between "Front relay 1" and
> "Main server using provider DSL connection 1" or between "Front relay 1"
> and listener ?
> How avoid buffering problems ? Have you got somes tips and tricks ?
>
> Bests regards,
> Pascal
> _______________________________________________
> Icecast mailing list
> Icecast at xiph.org
> http://lists.xiph.org/mailman/listinfo/icecast

I've tryed to load one of our relay with the shell scripts used on 
http://icecast.org/loadtest1.php.
Settings was :
# max concurrent curls to kick off
max=100
# how long to stay connected (in seconds)
duration=99999999
# how long to sleep between each curl, can be decimal  0.5
delay=100
Here is our mountpoint definition on the tested relay :
  <limits>
        <clients>100</clients>
        <sources>4</sources>
        <threadpool>20</threadpool>
        <queue-size>512000</queue-size>
        <client-timeout>10</client-timeout>
        <header-timeout>10</header-timeout>
        <source-timeout>60</source-timeout>
    </limits>
 <relay>
        <server>xxxxxxxx</server>
        <port>yyyyyyyyy</port>
        <mount>/letsgozik</mount>
        <local-mount>/letsgozik-HiQ</local-mount>
    </relay>
We are using icecast 2.3.0 under Debian/GNU Linux (kernel 2.4.24-grsec) 
on a AMD Athlon(tm) XP 2000+ with 512 Mo of memory.
Any idea ?

Bests regards,
Pascal



More information about the Icecast mailing list