Hi Kevan,

yes, we have only seen this on IBM J9. It does not occur on hudson and I have 
also failed to reproduce this locally (MacOS) in many attempts.

The reason I picked on the transaction manager thread as the likeliest 
candidate is that version 2.2 of the transaction manager by the looks of it 
introduces the problem and the shown thread I could see with any relation to 
the transaction manager ... 

Valentin 


On 8 Nov 2010, at 01:02, Kevan Miller wrote:

> 
> On Nov 5, 2010, at 1:02 PM, Valentin Mahrwald wrote:
> 
>> Investigating this a bit further there seems to be a Geronimo transaction 
>> manager thread hanging around in the itest at question, which could be the 
>> cause that the application.
>> 
>>     "Thread-18" J9VMThread:0x0B0CE700, j9thread_t:0x0A9B0E98, 
>> java/lang/Thread:0x97F159C8, state:CW, prio=5
>>           (native thread ID:0x6931, native priority:0x5, native 
>> policy:UNKNOWN)
>>           (native stack address range from:0x03BDA000, to:0x03B99000, 
>> size:0x41000)
>>          Java callstack:
>>               at java/lang/Thread.sleep(Native Method)
>>               at java/lang/Thread.sleep(Thread.java:851)
>>               at 
>> org/apache/geronimo/transaction/manager/TransactionTimer$CurrentTime.run(TransactionTimer.java:41)
>> 
>> Anyone seen this before?
> 
> That's a Daemon thread. So, I don't see any reason why the thread would be 
> responsible for a hang.
> 
> Is the hang only occurring on an IBM JSE? I don't see a hang, but do see some 
> test failures on Mac OS with both geronimo-transaction 2.1.3 and 2.2.
> 
> --kevan

Reply via email to