JDK1.4.1 and runaway processes.

[Jonathan Chen]

Hi,

I've been using the native JDK1.4.1 for a while now, and am happy to
report that it performs very well - most of the time. However, there
have been cases where the JVM appears to spawn off a new process that
runs amok.

The cases where this happens appears most commonly with Tomcat 4.1+,
when Tomcat has to compile a jsp page. The JVM spawns off a separate
process to compile a page, which then hangs and chews up processor time.
On killing everything (only a -9 works), and then restarting Tomcat, a
2nd attempt to compile the page will probably succeed.

Another case where this runaway process could happen is with code
dealing with Printers. When a new java.awt.PrintJob is requested using
java.awt.Tookit.getPrintJob, there have been times when the JVM hangs.
A quick look at top(1) reveals a runaway process.

What can I do to help track down this problem? It'd be nice to get a
core-dump for analysis, but the processes don't seem to die unless a
kill -9 is applied to it. Am I the only person to see this?

System info:
    FreeBSD grimoire.chen.org.nz 4.8-STABLE FreeBSD 4.8-STABLE #0: Sun Apr 20 11:43:08 NZST 2003     root at grimoire.chen.org.nz:/usr/obj/usr/src/sys/GRIMOIRE  i386

Running jdk-1.4.1p3_3.

Cheers.
-- 
Jonathan Chen <jonc at chen.org.nz>
----------------------------------------------------------------------
             "A person should be able to do a small bit of everything,
                                        specialisation is for insects"