Have you tried running ser in full debug mode, no-fork to see if you get a
crash? Maybe you have some inconsistency somewhere causing processes to go
down where a certain function is called.
Maybe you as an experiment should download the Getting Started SER
source package and compile using the make_and_install script? It could be
a mysql library problem. make_and_install will create a file with the
output. grep for 'error' and warning to catch any problems (caps too).
g-)
Matt Schulte wrote:
Ok, here's a little more info. It seems that the
procs are getting
knocked off one by one at different places in the script. Sometimes
it's a register, sometimes it's a reply (ack), random stuff. I have
break; at each ending point, any suggestions?
-----Original Message-----
From: Greger V. Teigre [mailto:greger@teigre.com]
Sent: Thursday, April 07, 2005 1:30 PM
To: Matt Schulte; serusers(a)lists.iptel.org
Subject: Re: [Serusers] Child procs
Matt,
Why do you believe that you have been running out of processes? At
any given time you may have processes without CPU load, but over time
the processes should have some CPU time.
Have you run serctl ps? AFAIK, the standard number of processes for
each
protocol (udp/tcp) is 8 unless you specify children=x.
If a ser process blocks due to problems with mysql or a radius
failover situation, you may get a problem where all processes blocks.
g-)
Matt Schulte wrote:
bueller? someone must have some clue
-----Original Message-----
From: Matt Schulte
Sent: Tuesday, April 05, 2005 7:26 AM
To: Matt Schulte; serusers(a)lists.iptel.org
Subject: RE: [Serusers] Child procs
Does anyone have any opinions on this? What I'm worried about is of
course running out of procs and thus hanging SER, I believed this to
have happened several times..
Thanks..
-----Original Message-----
From: Matt Schulte
Sent: Monday, April 04, 2005 8:45 AM
To: serusers(a)lists.iptel.org
Subject: [Serusers] Child procs
Do increasing child procs actually do anything? When I do a 'ps axu'
I
> only show 4 of the procs having CPU time, the rest having no cpu time
> at all. Any thoughts on this? Thanks, Matt
>
> _______________________________________________
> Serusers mailing list
> serusers(a)lists.iptel.org
http://lists.iptel.org/mailman/listinfo/serusers
>
> _______________________________________________
> Serusers mailing list
> serusers(a)lists.iptel.org
http://lists.iptel.org/mailman/listinfo/serusers