2009/1/9 julio33 <[email protected]>: > > II moved from prefork to MPM, and it´s incredible, everything is fast, > when in the past i made some test to the wsgi app with ab(from apache) > the four processor start working at 100%(on top) as long the test > works, > by now when i take the same test, the max load on any of the > processors > is 9%, also everything worked fine, the only thing i noted it´s that > now i > have like 25 process of apache, much more than before, however apache > only takes 40MB and after making some process only takes 55MB, before > after working was about 100MB and stays there until i restart it
Huh, if you have moved from prefork to worker MPM, something isn't right if you have 25 processes running. That sounds like you are still running prefork. > I´m using the default mpm worker configuration: > <IfModule mpm_worker_module> > StartServers 2 > MaxClients 150 > MinSpareThreads 25 > MaxSpareThreads 75 > ThreadsPerChild 25 > MaxRequestsPerChild 0 > </IfModule> Which says, to start only 2 child worker processes initially, but if load picks up, allow up to a total of 6 child worker processes. When load drops off, Apache will start reaping process and should come back down to 2 or 3 processes. Thus, for default worker configuration, the most Apache processes you will see is 7. That is, the Apache parent process, and 6 child worker processes. What operating system (type/version) are you using? I could be going senile, but I recollect that monitoring tools in old Linux systems would show each thread within a process separately. So, you sure you aren't counting threads and not processes. > And i decided to not go to daemon mode, because i already have what > i want, however i will try that mode if you recommend me that. Whether you use daemon depends on how much memory your WSGI application uses. If running a fat application, you may be better of using daemon mode as then can restrict it to 1 fat process, rather than every Apache child worker process being fat. If the application is lightweight, then probably no harm in using embedded mode, but would still use more memory than needs to. Graham > Really thanks, you made a wonderful work with the mod_wsgi. > > Best Regards, > Julio > > On 7 ene, 23:16, Graham Dumpleton <[email protected]> wrote: >> 2009/1/8 julio33 <[email protected]>: >> >> >> >> >> >> > Hello, i will deactivate the PHP module, i was about to do that >> > yesterday, but >> > i forget that, i will disable today. >> >> > This is the apache2 -V command, it seems i use prefork: >> >> > r...@julio33:~# apache2 -V >> > Server version: Apache/2.2.8 (Ubuntu) >> > Server built: Jun 25 2008 13:54:43 >> > Server's Module Magic Number: 20051115:11 >> > Server loaded: APR 1.2.11, APR-Util 1.2.12 >> > Compiled using: APR 1.2.11, APR-Util 1.2.12 >> > Architecture: 64-bit >> > Server MPM: Prefork >> > threaded: no >> > forked: yes (variable process count) >> > Server compiled with.... >> > -D APACHE_MPM_DIR="server/mpm/prefork" >> > -D APR_HAS_SENDFILE >> > -D APR_HAS_MMAP >> > -D APR_HAVE_IPV6 (IPv4-mapped addresses enabled) >> > -D APR_USE_SYSVSEM_SERIALIZE >> > -D APR_USE_PTHREAD_SERIALIZE >> > -D SINGLE_LISTEN_UNSERIALIZED_ACCEPT >> > -D APR_HAS_OTHER_CHILD >> > -D AP_HAVE_RELIABLE_PIPED_LOGS >> > -D DYNAMIC_MODULE_LIMIT=128 >> > -D HTTPD_ROOT="" >> > -D SUEXEC_BIN="/usr/lib/apache2/suexec" >> > -D DEFAULT_PIDLOG="/var/run/apache2.pid" >> > -D DEFAULT_SCOREBOARD="logs/apache_runtime_status" >> > -D DEFAULT_LOCKFILE="/var/run/apache2/accept.lock" >> > -D DEFAULT_ERRORLOG="logs/error_log" >> > -D AP_TYPES_CONFIG_FILE="/etc/apache2/mime.types" >> > -D SERVER_CONFIG_FILE="/etc/apache2/apache2.conf" >> >> If you do not need PHP, then move to worker MPM. Doing that will also >> cut down memory usage as Apache will not create as many worker >> processes to handle requests. >> >> Then ensure you use mod_wsgi daemon mode. Use the defaults for >> WSGIDaemonProcess, which is 1 process with 15 threads. This is likely >> going to be more than enough for what you need. >> >> If you really want to try and cut down even further on memory, run a >> separate nginx web server for serving static files. Have that same >> nginx proxy the dynamic requests through to Apache/mod_wsgi and turn >> off keep alive in the Apache server. This configuration seems to be a >> favourite of people using VPS systems as it gives quite a low memory >> profile as well as other benefits. >> >> > If the python raise an error, the wsgi turns slower? or restart or >> > something?, by other side it´s bad using raise inside modwsgi? >> >> No. If the exception is not caught by the WSGI application, then all >> that will happen is that mod_wsgi will return a HTTP 500 error >> response indicating an internal server error and log to Apache error >> log the details of the Python exception that occurred. The process is >> kept running as there is no need to shut it down and restart when this >> occurs. >> >> BTW, you might find other discussion happening of interest: >> >> http://groups.google.com/group/modwsgi/browse_frm/thread/3a9c53dffd6b... >> >> Graham > > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "modwsgi" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/modwsgi?hl=en -~----------~----~----~----~------~----~------~--~---
