To answer your last question, yes as far as I have been able to test. There are an interesting blog discussion taking up a similar problem/situation http://blog.exis.com/colin/archives/2005/08/23/i-put-a-spell-on-you-because-youre-mine-aka-why-is-tomcat-holding-onto-jars/ although the solutions seems to point in the direction of tomcat 5.5. Still I have the feeling maybe something could be done with drools here.
Kristofer On 3/14/06, Michael Neale <[EMAIL PROTECTED]> wrote: > > yeah I don't think that bug was related (there was no file locking > involved). > So its happening with groovy as well as jython? > > On 3/14/06, Kristofer Eriksson <[EMAIL PROTECTED]> wrote: > > > > yes, it is wierd, but is definitively happening, and as I've tried to > > describe before, only when any of the "scripting semantics" are used. > > > > My only guess is that it somehow has to do with classloaders and "where" > > classes are being generated/compiled. > > I am not sure but there were some indentified problems with classloading > > and > > hot deploy for JBoss and Weblogic: > > http://jira.codehaus.org/browse/DROOLS-374. > > > > Anyhow, it is a serious and real problem for me. > > > > Kind Regards > > > > Kristofer Eriksson > > > > > > On 3/12/06, Michael Neale <[EMAIL PROTECTED]> wrote: > > > > > > I have no clues about that unfortunately, it does sound wierd. What > > would > > > cause a jar to be locked? Its almost as if tomcat hasn't finished > > > classloading or something. > > > > > > On 3/10/06, Kristofer Eriksson <[EMAIL PROTECTED]> wrote: > > > > > > > > yes, the problem is exactly that. The core drools jar files (and > only > > > > those) > > > > are beeing held by tomcat (locked) when the application restarts > which > > > > prevents them from being deleted. > > > > > > > > Thx > > > > > > > > Kristofer > > > > > > > > > > > > On 3/10/06, Michael Neale < [EMAIL PROTECTED]> wrote: > > > > > > > > > > so the problem is with the jar files? (locked?) > > > > > > > > > > On 3/10/06, Kristofer Eriksson <[EMAIL PROTECTED]> > wrote: > > > > > > > > > > > > Hi all, > > > > > > > > > > > > i am having problems using drools together with tomcat ( 5.0). > > This > > > is > > > > > what > > > > > > happens: When I deploy an application with drools using the > java, > > > > groovy > > > > > > or > > > > > > python semantics, the tomcat server and the application starts > up > > > > > without > > > > > > problems. But when I make a hot deploy on the application, the > > > tomcat > > > > > > server > > > > > > seems to keep a lock on the core drools jar files preventing > them > > > from > > > > > > being > > > > > > deleted, and therefore making it impossible to redeploy. > > > > > > > > > > > > When using only the native (java) way of defining rules and > > > rulebases, > > > > > > everything works fine. I did read in jira and this mailing list > > > > archive > > > > > > about some classloader issues on appservers that was solved, but > > no > > > > > > mentioning of tomcat. > > > > > > > > > > > > This problem occurs with both drools 2.1 and 2.5, just tried the > > > > latter > > > > > > one > > > > > > today. > > > > > > > > > > > > We are using tomcat 5.0.28 (upgrade not possble at the moment). > I > > am > > > > > > personally no big fan of hot deploy but it is heavily used in > the > > > > > > development phase in a quite large project so it sort of has to > > > work. > > > > > > > > > > > > The questions if of course if anyone experienced the same, am I > > > doing > > > > > > something stupid, does drools have to be loaded outside the > > > > applications > > > > > > > > > > > classloader, or is it actually something of a problem with > Drools? > > > Is > > > > > > there > > > > > > maybe a workaround? Many questions that is! > > > > > > > > > > > > Any information is appreciated. > > > > > > > > > > > > Thanks > > > > > > > > > > > > Kristofer Eriksson > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >
