Dear Members, Mongrel Handlers seem to take up about 110 MB for each instance and sometimes upto 20% cpu usage on an 8 core system.. Though this has a lot to do with the application, any suggestions for improvement at the software stack level? [especially for the RAM?] regards Asif Ali -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://rubyforge.org/pipermail/mongrel-users/attachments/20090402/694af824/attachment.html>
Asif Ali <azifali at gmail.com> wrote:> Dear Members, > Mongrel Handlers seem to take up about 110 MB for each instance and > sometimes upto 20% cpu usage on an 8 core system..> Though this has a lot to do with the application, any suggestions for > improvement at the software stack level? [especially for the RAM?]Avoid slurping. Don''t read large files into memory all at once, LIMIT all your SELECT statements. That''s the biggest problem I see in all sorts of apps. Other than that there could be a lot of app specific things... Go by process of elimination and figure out where/when/if memory jumps are taking place and even rip out chunks of code that are suspect :) -- Eric Wong
Eric, Thank you for those suggestions. We are reviewing those select statements...again. But do some patches [like the forking patch] help? Asif Ali On Thu, Apr 2, 2009 at 7:06 AM, Eric Wong <normalperson at yhbt.net> wrote:> Asif Ali <azifali at gmail.com> wrote: > > Dear Members, > > Mongrel Handlers seem to take up about 110 MB for each instance and > > sometimes upto 20% cpu usage on an 8 core system.. > > > Though this has a lot to do with the application, any suggestions for > > improvement at the software stack level? [especially for the RAM?] > > Avoid slurping. Don''t read large files into memory all at once, > LIMIT all your SELECT statements. That''s the biggest problem > I see in all sorts of apps. > > Other than that there could be a lot of app specific things... > Go by process of elimination and figure out where/when/if memory > jumps are taking place and even rip out chunks of code > that are suspect :) > > -- > Eric Wong > _______________________________________________ > Mongrel-users mailing list > Mongrel-users at rubyforge.org > http://rubyforge.org/mailman/listinfo/mongrel-users >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://rubyforge.org/pipermail/mongrel-users/attachments/20090402/99ff1271/attachment.html>
Nope, that looks about normal, based on my experience. Our app would consume 80Mb once mongrel had started, and would jump to well over 120Mb after serving a request. On 02/04/2009, at 5:39 AM, Asif Ali wrote:> Dear Members, > Mongrel Handlers seem to take up about 110 MB for each instance and > sometimes upto 20% cpu usage on an 8 core system.. > > Though this has a lot to do with the application, any suggestions > for improvement at the software stack level? [especially for the RAM?] > > regards > > Asif Ali > > _______________________________________________ > Mongrel-users mailing list > Mongrel-users at rubyforge.org > http://rubyforge.org/mailman/listinfo/mongrel-users
Asif Ali wrote:> Dear Members, > Mongrel Handlers seem to take up about 110 MB for each instance and > sometimes upto 20% cpu usage on an 8 core system.. > > Though this has a lot to do with the application, any suggestions for > improvement at the software stack level? [especially for the RAM?]try the MBARI patches, if you''re using 1.8.x Cheers! -=r -- Posted via http://www.ruby-forum.com/.