On 03/10/2011 7:16 PM, Joern Huxhorn wrote:
On 30.09.2011, at 18:02, Ceki Gülcü wrote:
On 29/09/2011 5:44 PM, Joern Huxhorn wrote:
This issue is still creeping me out.
First of all, I'm perfectly fine with your workaround.
OK, good to hear. Thanks.
Nevertheless, we should probably create an app that produces the IllegalMonitorStateException and deliver it to the Java developers. If this is a Java bug - and I have no other explanation - then it's a really critical one...
Although I do not have the time to work on this myself, I support your initiative.
Hint taken. ;)
My only problem is that I don't have any system that's able to reproduce this bug (only Mac or Linux) and I'd rather not want to submit this already very obscure issue without being sure that the attached example/application is really reproducing the problem.
On the lbcore224 branch before commit eef137b0c6ca92cd4 [1], running on Windows 7 64 bits and AMD Phenom II X4 965 processor, I am able to reproduce the bug 6 or 7 times out of every 10 runs. So it's pretty reproducible, at least on my machine. The fact that the IllegalMonitorStateException is thrown within the ZooKeeperServer.shutdown method could be relavant. I was also wondering whether the IllegalMonitorStateException could be caused by Thread.interrupt calls. [1] https://github.com/ceki/logback/commit/eef137b0c6ca92cd4
Joern.