>> > - the first one is a project for a cocoon-based web site >> with 3 dozens >> >tables and a couple SQL Statements using ESQL for each page >> request. I >> >served 50 simulatnous user without acceptable performance degradation >> >and _without_ caching (only XSL Stylesheets were cached). >> CPU time and >> >memory consumption remaind constant. >> > >> > - the second project is a web application with combination >> of Cococon2 >> >and JSP-Pages, very complex SQL Statements using ESQL and complex >> >transformations, in combination with static XML files (XML >> files and XSL >> >stylesheet cached). In this configuration i have very poor >> performance, >> >a 1GHz Server with 512MB RAM can seems to be full loaded at 10 users! >> >> But this seems to be a problem of the pipeline processing. Maybe the >> different transformers or the serializers have a problem. > >Any idea how to narrow down the problem? There are only XSP generators >using ESQL (usually 2-4 of them, aggregated together) and XML File >generators (1-2 per match), one XSLT transformer and one serializer >(HTML or FOP). Hmmm, that's like searching a niddle in a haystack. I think white box testing can help us. Maybe we can seperate the different modules generating->transforming->serializing->caching->storing and stress it seperatly. My idea is to break open the processing chain and start testing with the generator, then I will grab sequential the other modules to the chain and stress it again. When you now hit a leak, then you now which component causes it. But this is just an idea. Cheers Gerhard -- Gerhard Fröhlich [EMAIL PROTECTED] "black holes are, when GOD is dividing by zero" GMX - Die Kommunikationsplattform im Internet. http://www.gmx.net --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, email: [EMAIL PROTECTED]
