Bug 153485 - VM hangs when out of perm gen space
Summary: VM hangs when out of perm gen space
Status: RESOLVED DUPLICATE of bug 92250
Alias: None
Product: Platform
Classification: Eclipse Project
Component: Runtime (show other bugs)
Version: 3.2   Edit
Hardware: PC Windows XP
: P3 major with 1 vote (vote)
Target Milestone: ---   Edit
Assignee: platform-runtime-inbox CLA
QA Contact:
URL:
Whiteboard:
Keywords: vm
: 144823 146800 149948 150744 152072 153010 153505 157997 158916 160310 161659 161701 161821 168978 171830 195675 196162 205701 (view as bug list)
Depends on: 92250
Blocks:
  Show dependency tree
 
Reported: 2006-08-10 15:24 EDT by John Arthorne CLA
Modified: 2008-02-27 16:28 EST (History)
28 users (show)

See Also:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description John Arthorne CLA 2006-08-10 15:24:25 EDT 
I am opening this bug as a fork of bug 92250.  That bug refers to OutOfMemoryErrors that occur when a Sun VM runs out of permanent generation (permgen) memory.  Another apparent side-effect when running out of permgen memory is that the VM sometimes hangs indefinitely during class loading.  The symptoms are 100% CPU usage forever (or 50% on a dual processor machine).  The stack trace generally looks like this:

 at java.lang.ClassLoader.defineClass1(Native Method)
 at java.lang.ClassLoader.defineClass(Unknown Source)
 at org.eclipse.osgi.internal.baseadaptor.DefaultClassLoader.defineClass(DefaultClassLoader.java:160)

This is believed to be a VM bug, but keeping this open here to keep track of its occurence in Eclipse, and to record workarounds, etc.  The current suggested solution is to upgrade to the latest VM (1.5.0_07), and increase the size of the permanent generation using the command line argument:

-XX:MaxPermSize=128M (or more, for a very large application)
Comment 1 John Arthorne CLA 2006-08-10 15:25:33 EDT 
*** Bug 146800 has been marked as a duplicate of this bug. ***
Comment 2 John Arthorne CLA 2006-08-10 15:26:33 EDT 
*** Bug 153010 has been marked as a duplicate of this bug. ***
Comment 3 John Arthorne CLA 2006-08-10 15:27:48 EDT 
*** Bug 152072 has been marked as a duplicate of this bug. ***
Comment 4 John Arthorne CLA 2006-08-10 15:29:27 EDT 
*** Bug 149948 has been marked as a duplicate of this bug. ***
Comment 5 John Arthorne CLA 2006-08-10 15:41:32 EDT 
*** Bug 144823 has been marked as a duplicate of this bug. ***
Comment 6 Paul Webster CLA 2006-08-14 08:03:53 EDT 
*** Bug 153505 has been marked as a duplicate of this bug. ***
Comment 7 Thomas Watson CLA 2006-08-25 17:40:33 EDT 
*** Bug 150744 has been marked as a duplicate of this bug. ***
Comment 8 Boris Bokowski CLA 2006-09-07 17:13:15 EDT 
*** Bug 156007 has been marked as a duplicate of this bug. ***
Comment 9 Boris Bokowski CLA 2006-09-07 17:13:57 EDT 
Copied from bug 156007 comment #13:

One suggestion, though: update the readme section that talks about this
setting, to actually describe the problem it solves. Currently, it only
mentions OOM errors - but that's what developers see; regular users (like me)
only see the hangup, and therefore fail to make a connection between
MaxPermSize setting and their problem. This leads to this bug being reported
over and over again.
Comment 10 Thomas Watson CLA 2006-09-21 13:13:49 EDT 
*** Bug 157997 has been marked as a duplicate of this bug. ***
Comment 11 Thomas Watson CLA 2006-09-29 08:05:20 EDT 
*** Bug 158916 has been marked as a duplicate of this bug. ***
Comment 12 Thomas Watson CLA 2006-10-10 10:09:20 EDT 
*** Bug 160310 has been marked as a duplicate of this bug. ***
Comment 13 Barthel Steckemetz CLA 2006-10-13 05:01:56 EDT 
The first and easiest thing to do would be to deliver a eclipse.ini file
containing the additional lines:

-XX:MaxPermSize=128M

This should help most of the users. Users who overwrite these setting
by using own vmargs at the command line should get some sort of warning.
Comment 14 John Arthorne CLA 2006-10-13 10:14:40 EDT 
Barthel, we cannot do that because this is not a standard VM argument. See bug 92250 comment 28.
Comment 15 John Arthorne CLA 2006-10-18 17:02:17 EDT 
*** Bug 160310 has been marked as a duplicate of this bug. ***
Comment 16 Dani Megert CLA 2006-10-20 02:41:02 EDT 
*** Bug 161659 has been marked as a duplicate of this bug. ***
Comment 17 Kim Horne CLA 2006-10-26 11:33:50 EDT 
*** Bug 161821 has been marked as a duplicate of this bug. ***
Comment 18 John Arthorne CLA 2006-10-27 14:59:15 EDT 
*** Bug 155829 has been marked as a duplicate of this bug. ***
Comment 19 Thomas Watson CLA 2007-01-02 22:57:15 EST 
*** Bug 168978 has been marked as a duplicate of this bug. ***
Comment 20 Cameron Bateman CLA 2007-01-26 13:26:33 EST 
*** Bug 171830 has been marked as a duplicate of this bug. ***
Comment 21 John Arthorne CLA 2007-02-27 13:57:28 EST 
*** Bug 161701 has been marked as a duplicate of this bug. ***
Comment 22 Johan Walles CLA 2007-04-05 07:29:35 EDT 
(In reply to comment #0)
> This is believed to be a VM bug, but keeping this open here to keep track of
> its occurence in Eclipse, and to record workarounds, etc.

Do you have a link to the bug report in SUN's bug reporting system?
Comment 23 Thomas Watson CLA 2007-07-11 11:57:45 EDT 
*** Bug 196162 has been marked as a duplicate of this bug. ***
Comment 24 Thomas Watson CLA 2007-07-24 14:22:48 EDT 
*** Bug 195675 has been marked as a duplicate of this bug. ***
Comment 25 Thomas Watson CLA 2007-10-08 08:53:20 EDT 
*** Bug 205701 has been marked as a duplicate of this bug. ***
Comment 26 Oleg Besedin CLA 2008-02-27 16:28:35 EST 
All of the bugs mentioned here and in duplicates were resolved after addition of -XX:MaxPermSize=256m VM argument.


*** This bug has been marked as a duplicate of bug 92250 ***