[ https://issues.jboss.org/browse/JBTM-1211?page=com.atlassian.jira.plugin.... ] Michael Musgrove commented on JBTM-1211: ---------------------------------------- As alluded to in an earlier comment the test is as follows: - client creates references to two remote objects (HammerObject) - a HammerObject performs arithmetic operations on a numeric value - the client hands off the test to two threads (DHThreadObject2) each of which does a 100 random ops. A worker (DistributedHammerWorker2) starts a transaction and invokes the op on both remote objects - each remote object does interposition and starts a subtransaction and performs the op - at the end of the test the client makes sure that the resulting value maintained by both remote objects are consistent The issue in the latest batch of failures is that the remote object does not return until the transaction times out (after about 4 mins). My guess is that setlock (in HammerObject) is hanging and the transaction timing out unwedges the thread. To test this theory I have turned on trace logging in the particular QA test. I'll also investigate how to trigger stack traces before the timeout (using byteman or otherwise). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira