The only thing I can do is http. I don't see the use of that. We had the deadlock because the busypool can change between the current time stamp and the clean process of "unclosed" connections in the checkTime(). The issue is not that the statement takes to long to close, but that the connection is closed when it is active and not ready to be closed.
-----Original Message----- From: Kees Jongenburger [mailto:kees.jongenburger@;omroep.nl] Sent: Thursday, November 14, 2002 9:57 PM To: [EMAIL PROTECTED] Subject: Re: deadlock on JDBCProbe and MMobjectBuilder On Thursday 14 November 2002 09:52 am, [EMAIL PROTECTED] wrote: > There is only 1 JDBCProbe thread in the thread dump and no other threads > are executing checkTime > We didn't change to DijkstraSemaphore. > > Actually this code is not blocking by itself, but the oracle driver does. > When the Statement object tries to retrieve data from oracle through the > connection object the connection object is locked by the close() call. And > the connection close call is waiting for the Statement object to stop > retrieving and close. Shall we add stats about the time taken to close the statement? can you talk on irc?
