Hello, Running Zope 2.2.2 on Dual PIII 933 with 1GB of RAM. This machine is not fast enough to handle my requests (about 800 Users using this system). It's a very complex document managing system with about 5000 documents actually and growing every day. Search requests are with Python backend. The backend sends lists of objects with methods and data. I print the data in DTML sequences calling the objects methods. System is a 2 server system, which do the data syncing over xmlrpc. One system is a Windows NT and the other is Linux. The high load results on Windows NT in Dr. Watson system crashes and on Linux only zope goes down. Windows NT crashes very often (about every hour), while Linux is crashing every few month. 1.) I tried to optimize it with simple objects, that only have the data in it and then print it with DTML (<dtml-var "variablename">). This is faster, because of the simpler objects. 2.) I tried running zope with apache's cgi, fastcgi over sockets, fastcgi over TCP/IP. Result: Slower than only Zope. 3.) Is there a way of avoiding error 'competing conflicts' in the internal Zope Classes ? This crashes the NT box very often. So the admin have to restart the service all the time :-( And it is not sure that it comes up again. Sometimes a reboot is the only way. Any Idea ? It doesn't matter how stupid it sounds. ciao. Stefan
On Tue, 3 Apr 2001, Stefan Bambach wrote:
2.) I tried running zope with apache's cgi, fastcgi over sockets, fastcgi over TCP/IP. Result: Slower than only Zope.
Have you tried with a proxy in front? Perhaps try with Squid or Apache as an accelerating proxy in front, this may help with seving images which might free the rest of the server up to do other things. -Matt -- Matt Hamilton matth@netsight.co.uk Netsight Internet Solutions, Ltd. Business Vision on the Internet http://www.netsight.co.uk +44 (0)117 9090901 Web Hosting | Web Design | Domain Names | Co-location | DB Integration
participants (2)
-
Matt Hamilton -
Stefan Bambach