2010/1/7 qrilka <[email protected]>: > I have tried to set logging to info bu could not find anything > meaningful. > I will try to get error log later to post it here (maybe I missed > something).
I am interested in the proximity of any messages indicating that daemon mode process are being restarted. If running VirtualHost specific error logs, you may also have to look in main Apache error log, specially if using mod_wsgi 2.8 and not mod_wsgi 3.1. Graham > Best regards, > Kirill Zaborsky > > On Jan 7, 1:20 pm, Graham Dumpleton <[email protected]> > wrote: >> 2010/1/7 qrilka <[email protected]>: >> >> >> >> > Hello, >> > I have read quite a lot of discussions about problems leading to >> > 'PReamature end of script headers' line in apache log but I still can >> > not find a way to resolve it completely. >> > On avastep.ru we have a Django app (main functionality done with PIL). >> > Site works OK but from time to time I get a line in error log like: >> > --- >> > [Wed Jan 06 21:15:16 2010] [error] [client 85.175.67.142] Premature >> > end of script headers: factory.wsgi, referer:http://www.avastep.ru/ >> > --- >> > and it appears that some of site users have problems. >> > I could not find any way to reproduce this problem otherwise than on >> > that working server. >> >> > httpd.conf for the site -http://dpaste.com/141757/ >> > Software installed: CentOS release 5.2 (Final), Python 2.4.3, mod_wsgi >> > 2.8 (fromhttp://codepoint.net/attachments/mod_wsgi/) >> >> > Is there any way to diagnose what is wrong with server configuration >> > and how could it be fixed? >> >> Set: >> >> LogLevel info >> >> in Apache configuration. >> >> If the messages coincides with mod_wsgi daemon mode process restarts, >> it may be because there are active long running requests at time when >> 'maximum-requests=50000' option of WSGIDaemonProcess is triggered. >> These may also be stuck requests. >> >> This could occur because when that process recycling is triggered, it >> will only wait so long, default 5 seconds, for any other active >> requests handled by the process to complete. If they don't complete >> the process is forcibly killed off. This killing off of the process >> would result in Apache server child process proxying the request >> seeing connection close and thus a premature end of the headers being >> returned by mod_wsgi daemon process. >> >> I realise the message is a bit confusing, but so happen to be using a >> bit of existing Apache code normally used when dealing with CGI >> scripts. Thus the message is actually same message as can get when CGI >> scripts don't return headers even though is a WSGI script handled by >> mod_wsgi. >> >> So, do you have long running requests which would take a long time >> before they generate the headers for the response? >> >> 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. > > > >
-- 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.
