Marshall.
Thank you, as I am using it I will pursue the Java 6 link.
I fear I may not be seeing a deadlock but rather the absence (because it
has already happened) of a notification to proceed - and not because a
process has died, but rather it has completed and sent the notifyAll()
message but the intended receivers were not at the wait() stage. In
short, the receivers will now be waiting for a message that will not arrive.
I admit to not knowing enough about Java concurrency to say this with
any conviction that it is even plausible let alone likely. But I am
reading several books and the source code and trying to arrive at a
hypothesis.
Bill
Marshall Schor wrote:
Hi William -
This behavior could be due to a deadlock, or it could be due to
something as simple as sending a request to a process and waiting for
the response, but the process died for some reason.
For deadlocks, many Java implementations have tools that can help
diagnose this by printing out what locks are being held. For
instance, the IBM Java 1.4.x has info about creating a "Javadump"
which includes information about locks; see
http://publib.boulder.ibm.com/infocenter/javasdk/v1r4m2/topic/com.ibm.java.doc.diagnostics.142/html/interpretjavadump.html#interpretjavadump
and in general, look here:
http://www-1.ibm.com/support/docview.wss?rs=3182&uid=swg27007948
If you're using Java 6 from Sun, this page may be helpful:
http://java.sun.com/developer/technicalArticles/J2SE/monitoring/
Look at the tools for diagnosis of common problems - Deadlocks
-Marshall Schor
William Surowiec wrote:
Adam,
I've spent some time with the code in the eclipse debugger and can
consistently cause a condition where all the threads are running but
no work happens. When you then suspend one of the application threads
it consistently stops at line 166 in
org.apache.uima.internal.util.ResourcePool (the statement
"wait(aTimeout);" I have extracted the method and list it below.)
This is the call stack when I suspend one of the running, but not
advancing the execution, threads:
Thread [Thread-1] (Suspended) Object.wait(long) line: not
available [native method] ResourcePool.getResource(long) line:
166
AnalysisEnginePool.setResultSpecification(ResultSpecification) line:
155
MultiprocessingAnalysisEngine_impl.setResultSpecification(ResultSpecification)
line: 122
MultiprocessingAnalysisEngine_impl(AnalysisEngineImplBase).process(CAS,
ResultSpecification) line: 200
MultiprocessingAnalysisEngine_implTest$ProcessThread.run() line: 363
The extracted method in ResourcePool:
public synchronized Resource getResource(long aTimeout) {
long startTime = new Date().getTime();
Resource resource;
while ((resource = getResource()) == null) {
try {
wait(aTimeout);
} catch (InterruptedException e) {
}
if (aTimeout > 0 && (new Date().getTime() - startTime) >=
aTimeout) {
// Timeout has expired
return null;
}
}
return resource;
}
I will continue puzzling over this problem but admit to re-opening
several of Doug Lea's books to renew my (never really put to use)
acquaintance with the concepts,
Bill
<snip/>
