On Tue, 3 Jul 2001, Cliff Woolley wrote:
> On 3 Jul 2001 [EMAIL PROTECTED] wrote:
>
> >
> > - if (idle_thread_count > max_spare_threads * ap_max_daemons_limit) {
> > + if (idle_thread_count > max_spare_threads * total_non_dead) {
> > /* Kill off one child */
> > char char_of_death = '!';
> > if ((rv = apr_file_write(pipe_of_death_out, &char_of_death, &one)) !=
>APR_SUCCESS) {
> > @@ -1011,7 +1011,7 @@
> > }
> > idle_spawn_rate = 1;
> > }
> > - else if (idle_thread_count < min_spare_threads * ap_max_daemons_limit) {
> > + else if (idle_thread_count < min_spare_threads * total_non_dead) {
> > /* terminate the free list */
> > if (free_length == 0) {
>
> Phew, okay, that looks a lot better to me. I knew there was something
> fishy about multiplying by ap_max_daemons_limit. =-)
Yeah, it fixed the symptom, but not the cause. This has it I think,
because now when I start the server, it starts 3 child processes, and
kills 2. Which is correct if you look at my minsparethreads directive.
:-) Plus, it grows to handle the 10000 requests I am throwing at it 500
at a time :-)
Ryan
_____________________________________________________________________________
Ryan Bloom [EMAIL PROTECTED]
Covalent Technologies [EMAIL PROTECTED]
-----------------------------------------------------------------------------
