Strange deadlock: JVM 1.4.2-p6

[freebsd at spatula.net]

This might be part of the problem:

"tcpConnection-6802-247" daemon prio=5 tid=0x0887f600 nid=0x8870800
waiting for monitor entry [8909000..8909cdc]
        at java.util.Date.getTimezoneOffset(Date.java:1095)
        - waiting to lock <0x2cacb228> (a java.util.GregorianCalendar)
        at
org.postgresql.jdbc1.AbstractJdbc1Statement.setTimestamp(AbstractJdbc1St
atement.java:1311)
        - locked <0x2df16fe0> (a java.lang.StringBuffer)
        at nl.profdata.syne.Log.log(Log.java:92)
        - locked <0x2cabd420> (a nl.profdata.syne.Log)

There's a lock around your log, and the log ends up getting a lock around
the GregorianCalendar.  But it looks like the other thing waiting for the
lock around the log doesn't already have a lock on the GregorianCalendar.

This probably doesn't answer your question entirely, but maybe it's
something to start with...

On Fri, 12 Nov 2004, Sebastiaan van Erk wrote:

> Hi,
>
> I have the a very strange deadlock in the JVM on a server running a web
> application on resin.
>
> System:
> FreeBSD xxx 4.10-PRERELEASE FreeBSD 4.10-PRERELEASE #1: Wed May  5
> 15:56:39 CEST 2004   root at xxx:/usr/src/sys/compile/PROFDATA  i386
>
> Java:
>
> java version "1.4.2-p6"
> Java(TM) 2 Runtime Environment, Standard Edition (build
> 1.4.2-p6-root_21_feb_2004_22_04)
> Java HotSpot(TM) Client VM (build 1.4.2-p6-root_21_feb_2004_22_04, mixed
> mode)
>
> In the attached deadlock, a bunch of threads are waiting in the private
> method getField(int) of java.util.Date, locked on the staticCal field.
> However, there is not 1 thread which actually owns the lock. How can
> this possibly happen?
>
> Thanks in advance,
> Sebastiaan van Erk
>
>
>

-- 
"The aptly-named morons.org is an obscenity-laced screed..."
	-- Robert P. Lockwood, Catholic League director of research
Nick Johnson, version 2.1                             http://web.morons.org/