Or maybe it was faster and processed more messages, so that the memory leak came faster ?
On 9/20/07, netflexity <[EMAIL PROTECTED]> wrote: > > The default java 1.5 implementation. > We haven't measured exactly, but memory wise it helped the server live for > at least double the time. > Let me collect more data. > > Thank you. > > > gnodet wrote: > > > > How much faster was the JAXP one ? > > And using which implementation ? > > > > On 9/20/07, netflexity <[EMAIL PROTECTED]> wrote: > >> > >> Even though JAXP-based XSLT component was faster and more efficient than > >> SAXON for XSLT 1.0, SAXON was not the cause of the memory leak. It was a > >> custom ContentEnricher (that supports InOut exchanges) that was not > >> issueing done(targetME). So, please be aware of this when creating custom > >> service engines or binding components. > >> > >> > >> > >> netflexity wrote: > >> > > >> > My guess would be the use of Templates. > >> > > >> > > >> > netflexity wrote: > >> >> > >> >> Have anybody experienced memory leaks with XSLT service units? We are > >> >> running into out of memory errors during load testing. If comment out > >> >> just XSLT conversion, everything performs excellent. > >> >> > >> > > >> > > >> > >> -- > >> View this message in context: > >> http://www.nabble.com/SAXON%28XSLT%29-SU-memory-leak-tf4429185s12049.html#a12799678 > >> Sent from the ServiceMix - User mailing list archive at Nabble.com. > >> > >> > > > > > > -- > > Cheers, > > Guillaume Nodet > > ------------------------ > > Blog: http://gnodet.blogspot.com/ > > > > > > -- > View this message in context: > http://www.nabble.com/SAXON%28XSLT%29-SU-memory-leak-tf4429185s12049.html#a12801919 > Sent from the ServiceMix - User mailing list archive at Nabble.com. > > -- Cheers, Guillaume Nodet ------------------------ Blog: http://gnodet.blogspot.com/
