Operating System: All
Not sure if it is relevant, but I was runing
tail -f server.log
at the same time while running the CTS smoke tests.
Created an attachment (id=5099)
The thread dump.
Also, I have set the felix logging configuration to DEBUG (4)
Do you see the deadlock when Felix framework logging is set to ERROR (1) ?
...only with DEBUG felix logging
I think this is related to
Yes, seems like it is the same issue. However, I feel quite strongly against
classifying a deadlock issue as P3.
seems to be a duplicate of 13232, reassigning to naman for further investigation.
Sorry. Revert back to the previous issue owner.
Reducing the priority. As discussing with Jerome.
Making priority to P2 and will plan to fix the same. The issue is coming only
with felix log level set to 4. So it might be the issue the way logging is done
The Felix framework uses stdout for logging. It doesn't seem like that should
cause an issue. Last I looked the framework hung trying to print to stdout
because it required a lock that someone else held, but the framework code itself
isn't grabbing the lock, PrintStream or something else was.
Made changes in the code to avoid direct usage of PrintStream.
Can somebody confirm the following deadlock is the same please. I wasn't restarting server, I can't really say when it happened. I saw the problem when I restarted the client-side swing application.
Thread dump with jConsole :
Edit: the deadlock seems to have appeared when calling multiple time the same method in a stateless EJB (from multiple instances)
Reopening as discussed in http://forums.java.net/node/797659
Assigning to Sandeep.
This should be fixed with revision 59642.