https://issues.apache.org/bugzilla/show_bug.cgi?id=51888
--- Comment #10 from Sebb <[email protected]> 2011-09-26 16:03:55 UTC --- What appears to be the issue is that the interrupt may(In reply to comment #5) > Hello, > For me the issue is an HTTPClient one. > It is not a problem that appears to be running threadFinished when interrupt > is > still active. Actually, I think this is a general problem; however so far only HC4 seems to be affected by it. > The explanation for this would be this kind of sequence: > 1) Stopper thread (Thread-77 in my Thread Dump) first calls > StandardJMeterEngine.tellThreadsToStop which calls stop() => Thread is > suspended here > 2) "Thread Group 1-32" gets stop call , run thread is stopped and we enter > threadFinished => Thread is suspended > 3) Stopper thread enters interrupt call and calls abort which is suspended > just > before AbstractClientConnAdapter.abortConnection > 4) "Thread Group 1-32" locks > org.apache.http.impl.conn.SingleClientConnManager > through shutdown call The explanation I have is slightly different; threads are S (engine) and T (JMeterThread); numbers are time sequence: T1) active sample in progress S2) interrupt active sample T3) sample completes S4) [interrupt cleanup in progress] T5) thread completes S6) [interrupt cleanup still in progress] T7) shutdown invoked I tried adding a short pause between S6 and T7 and that fixed the problem for me, as it allowed the interrupt to complete. The fix in comment 6 works because it causes the JMeterThread code to wait until the engine interrupt has finished. I think the fix actually needs to be applied to JMeterThread instead so it applies to all samplers, not just HC4. -- Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
