tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "John McClain" <jmccl...@tcshealthcare.com>
Subject Why are there 200000 Finalizer objects at OOME???
Date Thu, 06 Jul 2006 21:44:22 GMT
I have was profiling my application using yourkit profiler. The statisxtics
are stange and I need help to understand them.

Here is the scenario:
1) started Tomcat @ 1:00 profiling using yourkit profiler - only 1 context,
our application. I set the profiler to take a heap snapshot on 90% of heap
usage. I set the heap to 64Megs because I knew there would be only 2 users
on it, and I wanted to get to a point of OOME faster
2) took a snapshot before anyone logged in, and after Tomcat booted up
3) took a few other snapshots in the first hour of usage while 2 people
using system
4) everyone left ~ 5:00pm and logged out of our system. The only thing
running was Tomcat
5) @ ~ 2:00am the next morning, a snapshot was taken because 90% of the heap
was in use (???)
6) the next morning, all the heap was used and Tomcat would only respond
with OOME - java.lang.heap out of memory error


I compared the last heap snapshot I took during step 3 above and the low
memory snapshot and found that 73% of memory was being held by
java.lang.ref.Finalizer. WHAT THE HECK IS THAT? I have looked all over the
web and have found no information on what it is or what it does?
The closest I have come to any information is articles about java finalizers
vs. destructors, and what java.lang.Object.Finalize method. I also saw a
reference to the java.lang.ref.Finalizer api, but there are no javadocs
describing its functinoliaty. My questions are:
1) How are Finalizer and java.lang.Object.Finalize related - or are they?
what is java.lang.ref.Finalizer and what do they do?
2) WHY would I have over 200000 Finalizer objects occupying 73% of the heap
at step 5 (see above), when in my final snapshot of step 3 I had 0???
3) we are also using org.apache.commons.dbcp and the sql server JTDS driver
for our db connection pool in the following manner. Does anyone know if
there is a problem with this?

DriverAdapterCPDS cpds = new DriverAdapterCPDS();

try

{

cpds.setUrl( databaseURL );

cpds.setUser( user );

cpds.setPassword( password );

cpds.setDriver( databaseDriver );


SharedPoolDataSource tds = new SharedPoolDataSource();

tds.setConnectionPoolDataSource( cpds );

tds.setMaxActive( iMaxActive ); // Connection pool Size

tds.setMaxWait( iMaxWait ); // Max connection wait time (in miliseconds)


ds = tds;

}

catch( Exception e )

{

e.printStackTrace();

}


I will be repeating this test today, but will take heap snapshots on an
hourly interval - HOPEFULLY, I can find what is going on

Is there any one out there who can help us find this memory leak? Fee for
Service could be acceptable?

John McClain
Senior Software Engineer
TCS Healthcare
jmcclain@tcshealthcare.com
(530)886-1700x235
"Skepticism is the first step toward truth"

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message