Thanks so much for the help. There were 2 sites with debug set to true. I will also check out #django for the sitemap things. I will let you know if it balloons up again.
On Feb 26, 6:34 pm, Graham Dumpleton <[email protected]> wrote: > 2009/2/27 Jason Broyles <[email protected]>: > > > > > Sorry about that, I am using Django. I do not know of any URL requests > > that may have caused it to jump. > > Do you have the obvious of having 'DEBUG' in Django settings file disabled. > > > The only thing I can think of is > > maybe a search engine bot, unsure though. I do have a sitemap that has > > about 100,000 pages in it. It is for a housing search application. Is > > there a way I can tell this? > > If you think it may be to do with sitemap, suggest you try and get > onto David Cramer on #django IRC. Can supply email address if need be. > I remember him complaining about overhead in large sitemaps. Something > about it all being generated in memory before being sent back. You may > even find reference to it in #django logs. > > http://www.google.com/search?hl=en&client=safari&rls=en&q=site%3Abotl... > > In particular, one grumble at end of: > > http://botland.oebfare.com/logger/django/2008/09/30/ > > > The maximum requests was set low to reset the Apache process so the > > memory stays down. A just in case thing I did. What should it be? I > > will go ahead and take out the process option and add the group in. > > > On a completely unrelated note. I just saw this in the Apache > > access.log.1 file. I have locked the server down via ssh and use an > > RSA key to access it. It would be impossible for anyone else to be on > > it. But it looks like someone used lynx in the log and tried to access > > localhost admin. Strange.... > > > "GET /admin HTTP/1.0" 404 303 "-" "Lynx/2.8.6rel.4 libwww-FM/2.14 SSL- > > MM/1.4.1 GNUTLS/2.0.4" > > If you run: > > /usr/sbin/apachectl status > > Then it tries to use lynx to connect to the server. I don't know what > URL it tries to use though. > > Graham > > > On Feb 26, 6:07 pm, Graham Dumpleton <[email protected]> > > wrote: > >> 2009/2/27 Jason Broyles <[email protected]>: > > >> > I have my server setup like the below. For some reason one of the > >> > apache processes keeps jumping up to 263mb of ram. Does anyone know > >> > why this might be? Also, once I restart Apache, the total memory usage > >> > is around 40mb. Then when I access sites each site uses about 24mb of > >> > ram. This memory is never released though. Usually if everything is > >> > running normal it can run for weeks like that. This is besides the > >> > problem of the 263mb process. > > >> Can you provide more information about what the application is > >> implemented in, eg. Django, and what the application is doing. Have > >> you been able to identify specific URL requests as causing the jump in > >> memory usage? > > >> See also comments blow. > > >> > Apache info > > >> > Architecture: 64-bit > >> > Server MPM: Worker > >> > threaded: yes (fixed thread count) > >> > forked: yes (variable process count) > > >> > Example of one site from httpd.conf > > >> > KeepAlive Off > >> > ServerLimit 2 > > >> > <VirtualHost *> > > >> > ServerNamewww.somesite.com > >> > ServerAlias *somesite.com > > >> > WSGIDaemonProcess somesite user=webuser group=webuser processes=1 > > >> Don't provide option 'processes=1'. It will default to one process > >> anyway and setting this flag has side affect of marking > >> wsgi.multiprocess as True rather than False. Yeah, sounds back to > >> front, but done for a reason. > > >> > threads=10 maximum-requests=100 > > >> Why is maximum requests so low. > > >> > inactivity-timeout=172800 > > >> Add the option: > > >> display-name=%{GROUP} > > >> This will allow you to use 'ps' command to properly identify the > >> process as being the WSGI daemon process. > > >> Otherwise ensure you have LogLevel directive in Apache set to 'info'. > >> That will allow you to get pid for WSGI daemon process from Apache > >> error logs and identify it as the fat process. > > >> > WSGIProcessGroup somesite > > >> > WSGIScriptAlias / /home/webuser/somesite.com/somesite.wsgi > > >> > </VirtualHost> > > >> 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 -~----------~----~----~----~------~----~------~--~---
