On Tue, Apr 14, 2020 at 11:49 AM Thomas Plant
<thomas@plant.systems.invalid> wrote:
>
> Hi,
>
> do not know if this applies to your problem, but I had the same error a
> few months ago, in a server acting as 'normal' http server, not a proxy
> setup. I'm using MPM worker and an httpd, 2.4.6 from CentOS, and had to set
>
> MaxRequestsPerChild 0
>
> in the worker configuration. On CentOS 7 it is under
> /etc/httpd/conf.modules.d/00-mpm.conf.
> Maybe this helps in your case too.
It could very well be related. If processes are trying to exit, but
can't fully exit due to some kind of hard loop/hang, they will count
against the scoreboard but not actually be usable workers.
Later versions of event are slightly better behaved.
Ultimately I suggest finding what exiting httpd processes are busy
doing and seeing if timeouts can be added. Or if some buggy code is
looping.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx
For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx
|