Community
Participate
Working Groups
I started using eclipse 3.2.1 a few days ago and I alway have the same problem. I'm writing code and suddenly it exits with the following message: JVM terminated. Exit code=6 /usr/bin/java -Xms40m -Xmx256m -jar /usr/share/eclipse/startup.jar -os linux -ws gtk -arch x86 -launcher /usr/share/eclipse/eclipse -name Eclipse -showsplash 600 -exitdata 8190001 -vm /usr/bin/java -vmargs -Xms40m -Xmx256m -jar /usr/share/eclipse/startup.jar I tried using JSDK1.5.0_09-b03, JRE1.5.0_09-b03 on a variety of machines (Windows XP, Windows Vista and RedHat EL 3). All of them with the same problem. What is the problem? My files are on AFS. Version 3.1 of Eclipse and earlier versions that I used worked fine for 2 years now. Am I missing something? Please help me because I need to use Eclipse for a lot of my projects. Thanks, Panickos
Move to Platform/Runtime
When the Java VM crashes like this, there is usually a VM crash log that contains stack trace information. It's usually found in the current working directory of the eclipse process (usually the eclipse install directory). Look for a file there called something like core... javacore... or hs_err..., and attach it here.
Created attachment 53150 [details] Error log from crashing Eclipse run
Created attachment 53509 [details] VM crash log
Comment on attachment 53509 [details] VM crash log I am seeing the same issue. Running Eclipse 3.2.1 on RedHat ES 3
I can't make much sense of these crashes... there seems to be nothing particularly interesting going on. Andrew: any reports on your end for problems like this in REL 3 with Eclipse 3.2.1?
This looks like it might be the same problem as bug #159616.
Try adding -XX:MaxPermSize=128m to your eclipse.ini
This line in the error stacks looks the same as the one I see on crashes on my WinXP computer with Sun VM: compacting perm gen total 35840K, used 35651K [0x659b0000, 0x67cb0000, 0x859b0000) the space 35840K, 99% used [0x659b0000, 0x67c80f58, 0x67c81000, 0x67cb0000) In about 80% of VM crashes I've encountered recently logs say that one of the heap areas or the PermGen area is at 99% and being compacted. For me the number of crashes went down drastically after I added VM arguments to specify the high maximum and *iniital* size of the perm gen memory: -Xmx768m -XX:MaxPermSize=256m -XX:PermSize=256m (Of course, for your needs numbers might be different.) Please let us know if using setting the initial perm gen size and/or maximum perm gen size helps.
I agree with Ben that this is a dup of bug 159616. Marking as dup of the bug that bug 159616 is marked a dup of. (if that makes any sense :) *** This bug has been marked as a duplicate of bug 161310 ***
According to the comment 9 it seems to be a VM memory management problem, according to comment 10 it is a mizilla/swt Linux libraries problem... The bug was reported against different platforms (not only Linux), so how it is linked together and why this but is a duplicate of 161310?