Community
Participate
Working Groups
At different points in time, Eclipse Platform hangs (not a simple slowdown - I waited for 1 hour and it hasn't recovered). I was not able to identify the particular component that causes it. The hangups are quite frequent, about once in an hour. Version: 3.2.0 Build id: M20060629-1905
Could you please run with the -consoleLog command line argument (you can add "-consoleLog" to your eclipse.ini file in a separate line)? This will open a console window along with Eclipse. Then, next time you experience a hang, press Ctrl-Break in the console window and attach the resulting thread dump to this bug. You might have to increase the line buffer (console window menu->properties->layout) to be able to copy the whole thread dump. Thanks!
Aye. Will post the results as soon as I have any.
For some reason, I'm unable to start Eclipse this way. When I append -consoleLog to the eclipse.ini (it currently contains "-vmargs -Xms256m -Xmx512m"), Eclipse gives a message "Could not create the Java virtual machine" and refuses to start. I tried to put eclipse.consoleLog=true into config.ini instead - this way, Eclipse starts but there's no console you mentioned (other than the regular Console View). Am I doing something wrong? I'll attach the config files I use, just in case there's anything wrong there (should be stock installation though).
Created attachment 49397 [details] RAR archive containing eclipse.ini and config.ini
Put -consoleLog in the first line of eclipse.ini (and use Wordpad, not Notepad to edit the file).
Created attachment 49426 [details] Thread Dump Complete thread dump for the hangup bug
Created attachment 49434 [details] Yet another dump from one more hangup
In both cases, the main thread is doing work, and other threads are active as well. Could it be a problem with your heap size that causes excessive garbage collection? Have a look at the section called "Allocating enough memory and solving OutOfMemoryErrors" in readme/readme_eclipse.html, located in your Eclipse install directory. If this does not help, have you tried running with a different VM to see if the problem is related to the VM?
The VM is "Sun Java 2 Standard Edition 5.0 Update 6 for Microsoft Windows" - one of the Eclipse reference platforms, so I'd not count on any problems with this VM (not that I have any control over running a different one, though). The heap size is already set to 512M (see the config I attached previously) - however, I'll try the -XX:MaxPermSize=128m setting to see if it helps in any way. I'll let you know the progress.
John, Michael, any ideas as to what might be going on here?
Given that threads in both cases appear to be loading classes, I would suggest you ask someone on the Runtime team (like Tom) if they can provide any insight.
The thread dumps look like bug 153485. If using the -XX:MaxPermSize=128m VM option helps then this bug is a duplicate of bug 153485.
Two days without a single hangup... looks like the MaxPermSize is the solution. 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.
*** This bug has been marked as a duplicate of 153485 ***
I'm going to reopen this to implement the suggested improvement to the documentation.
Created attachment 50168 [details] Added more details in Eclipse readme
Sonia, can you release this patch to readme_eclipse.html in HEAD?
Patch released and feature.map updated.
Thanks Sonia. Marking fixed.