Yah, classloaders and, more specifically singletons that keep references to classloaders (like the infamous JDBC driver manager) tend to make "complete" flushes of a given memory context very problematic inside of a JVM. So a tomcat redeploy of most any "real" as opposed to sandbox app is going to leak memory. Increasing your memory size just delays the reckoning.
--- Pat > -----Original Message----- > From: Henri Dupre [mailto:[EMAIL PROTECTED] > Sent: Monday, November 28, 2005 10:43 AM > To: Tapestry users > Subject: Re: java.lang.OutOfMemoryError: PermGen space > > I thought the underlying issue was that some classloaders cannot be GCed. > So everytime you redeploy, it will increase the memory necessary for the > app. > > > On 11/28/05, Bryan Lewis <[EMAIL PROTECTED]> wrote: > > There was an earlier tip on this list about increasing the size of the > > PermGen space. I use this in my JAVA_OPTS on my development machine > (where > > I do lots of redeploying) and don't get OOME's any more. > > > > -Xms256m -Xmx256m -XX:MaxNewSize=256m -XX:MaxPermSize=256m > > > > > > > > ----- Original Message ----- > > From: "Martin Strand" <[EMAIL PROTECTED]> > > To: <tapestry-user@jakarta.apache.org> > > Sent: Monday, November 28, 2005 10:24 AM > > Subject: java.lang.OutOfMemoryError: PermGen space > > > > > > > I keep getting this error from Tomcat: > > > > > > java.lang.OutOfMemoryError: PermGen space > > > > > > It seems to happen randomly. I read this article on TSS: > > > http://theserverside.com/news/thread.tss?thread_id=36743 > > > "The permanent generation is special because it holds meta-data > describing > > > user classes (classes that are not part of the Java language)." > > > As I understand it, the article implies that this error can occur when > you > > > use a large number of classes. I've got less than 50 classes right now > for > > > all components, pages and business logic but I'm still getting that > error. > > > > > > Perhaps this PermGen space error is because of a large number of > enhanced > > > classes? Or perhaps "hot code replace" leaves a number of old replaced > > > classes hogging the PermGen space? > > > Anybody else experiencing this? > > > > > > Martin > > > > > > --------------------------------------------------------------------- > > > To unsubscribe, e-mail: [EMAIL PROTECTED] > > > For additional commands, e-mail: [EMAIL PROTECTED] > > > > > > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: [EMAIL PROTECTED] > > For additional commands, e-mail: [EMAIL PROTECTED] > > > > > > > -- > Thanks, > > Henri. > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]