Re: [rules-users] Drools Classloader Leak?

Interesting. Is there a reason for that? That doesn't sound too efficient. How can I get around it so that everything in the drools API uses the same default application classloader? Is there a setting or configuration somewhere that would tell drools to just use the default classloader? That would speed up the application, and require less perm gen memory. Thanks. Sent from my iPhone On May 7, 2010, at 4:30 PM, Mark Proctor<mproctor(a)codehaus.org> wrote: > Each kbase has it's own root classloader, and then an additional child > classloader per package in the kbase. > > Mark > On 07/05/2010 22:08, malkhafaji wrote: > >> Hello, >> >> I noticed (using JConsole) that my application keeps loading more >> and more >> classes. I have over 105K classes loaded! I did a lot of debugging >> and >> isolation until I found the line that brings the number of classes >> loaded >> from 20K to 105K. It is this line: >> >> this.knowledgeBase.addKnowledgePackages(packages); >> >> packages is an array list of 8 DRLs that are compiled successfully. >> I am >> calling the line above exactly 99 times. However, I do have 99 >> instances of >> this.knowledgeBase. So, I have a total of 99 instances of >> "knowledgeBase", >> and each instance is called once (line above) with 8 successfully >> compiled >> DRLs. >> >> I am starting to think that each instance of a KnowledgeBase >> somehow uses >> its own Classloader? Do I need to manually get the current >> application >> Classloader and pass it on to each instance of KnowledgeBase to >> ensure that >> no classes are loaded more than once? Is this a leak in the >> Classloader that >> comes inside the KnowledgeBase.addKnowledgePackage method? I did post >> somewhere else about delays on the line above, and no wonder it >> sometimes >> takes a couple of seconds to execute that line (the largest rules >> file I >> have is 150 KB with no more than 100 rules) if it creates a new >> Classloader >> and loads the same loaded classes AGAIN. Any insights on this? >> >> P.S. I did a search on the forum and couldn't find anything related >> to this. >> If you have better searching skills and found a post in that >> regards, please >> kindly point me to it. This is a very important issue as we are >> going to a >> live production environment in 3 weeks and I am not too happy with >> the >> performance, specifically around the line above. >> >> Thank you, and have a great weekend. >> >> > > _______________________________________________ > rules-users mailing list > rules-users(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/rules-users > _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users