On Wed, 25 Apr 2001 [EMAIL PROTECTED] wrote: > I recently began to experience many lingering > mod_perl processes which slow down my tcp connection. > > I use Apache::Registry for all my mod_perl need. > Not the "real" handler. > > The practical solution I found was to cut down the > number MaxRequestPerChild and kill off the processes > after a few requests. > > I understand that Apache::PerlRun has the overhead > of compiling each script with a new request. I would > like to know whether there is a compilation overhead > if I set say MaxRequestsPerChild 1 ?? Not if you preload your scripts at the server startup. But then you pay the price of the time to spawn a new process, which has a very little overhead under low load since Apache preforks processes (assuming that you have MaxSpareServers set to a reasonable number). Under very high load your machine will spend a lot of CPU, spawning processes, which is not good. > I was also thinking that because the script is loaded > from the disk with each request there will be performance > degradation because of disk data ransfer rate will be > order of magnitude less than if the script were > to remain in the memory all the time. true, see above > That leads me to think about using very cool RAMFS that > comes linux-2.4. What does everyone think about using > RAMFS to reduce the performance degradation due to > loading script from the disk if there is one. > I have over 100 scripts altogether but the total size of them > is less than 1 MB and I have plenty of memory. don't. preload the code at the server startup. _____________________________________________________________________ Stas Bekman JAm_pH -- Just Another mod_perl Hacker http://stason.org/ mod_perl Guide http://perl.apache.org/guide mailto:[EMAIL PROTECTED] http://apachetoday.com http://eXtropia.com/ http://singlesheaven.com http://perl.apache.org http://perlmonth.com/