Here it is. On Fri, Jun 21, 2013 at 6:21 PM, Nandika Jayawardana <jayaw...@gmail.com>wrote:
> Hi Sathwik, > > Can you please resend the workaround bpel. > > Regards > Nandika > > > On Fri, Jun 21, 2013 at 1:45 PM, Sathwik B P <sathwik...@gmail.com> wrote: > > > Hi Guys, > > > > PFA the changed bpel which works as expected. > > > > regards, > > sathwik > > > > > > On Fri, Jun 21, 2013 at 1:06 PM, Sathwik B P <sathwik...@gmail.com> > wrote: > > > >> Hi guys, > >> > >> The event handler is initialized even before the first receive is > >> initialized as it is defined within the process scope. Since the first > >> receive which initializes the correlated instance hasn't been processed > >> there is no way for the event handler to know the correlation value. So > the > >> onEvent message will never get routed to the right instance. > >> > >> I doubt we can fix anything here. if other ODE experts think otherwise > >> let me know. > >> > >> Now as a workaround for this problem I would suggest a change in process > >> design. Put the first receive within a scope and the remaining > activities > >> within another scope and define the event handler on the second scope. > This > >> way the process instance would have been correlated and when the > execution > >> of the second scope starts the even handler would have the correlation > >> details. > >> > >> what do you think? > >> > >> regards, > >> sathwik > >> > >> > >> On Wed, Jun 19, 2013 at 1:25 PM, Nandika Jayawardana < > jayaw...@gmail.com>wrote: > >> > >>> Hi sathwik, > >>> > >>> This is exactly the behavior I observed.Finding the correlation keys > for > >>> the OnEvent in EH_EVENT SELECT run() method fails for some reason and > >>> hence ode assumes the session keys used for implicit correlation as the > >>> correlation key. > >>> > >>> Regards > >>> Nandika > >>> > >>> > >>> > >>> On Wed, Jun 19, 2013 at 11:55 AM, Sathwik B P <sathwik...@gmail.com> > >>> wrote: > >>> > >>> > Hi Keheliya, > >>> > > >>> > I see there is some race condition. I have been testing on the trunk > >>> and > >>> > able to see this behaviour. I haven't been able to identify the root > >>> cause > >>> > yet. I can point out that there is some problem in the message routes > >>> > getting registered after the arrival of the first message followed > >>> > immediately with the second onEvent message. > >>> > > >>> > Here is the data from ODE_MESSAGE_ROUTE, 8 is the value of the > >>> correlation. > >>> > Record bearing ID 8354 has a bizzare correlation_key value which > >>> > corresponds to the onEvent message. The record bearing ID 8355 is > that > >>> of > >>> > your second receive which is perfect. > >>> > > >>> > 8354 @2[-1~hqejbhcnphr8dggwax9v7q] 15 0 7952 one > 6656 > >>> > 8355 @2[CorrelationSet~8] 31 0 7952 > >>> one > >>> > 6655 > >>> > > >>> > Due to this inappropriate correlation routing data the onEvent > message > >>> > doesn;t get routed. Will keep posted if I find anything. > >>> > > >>> > regards, > >>> > sathwik > >>> > > >>> > On Tue, Jun 18, 2013 at 9:49 PM, Keheliya Gallaba < > >>> > keheliya.gall...@gmail.com> wrote: > >>> > > >>> > > Hi Tammo, > >>> > > > >>> > > I'm using OpenJPA. Because I'm manually invoking the operations via > >>> > soapUI, > >>> > > there's always a delay between the calls. > >>> > > > >>> > > Thanks, > >>> > > Keheliya > >>> > > > >>> > > On 18 June 2013 21:42, Tammo van Lessen <tvanles...@gmail.com> > >>> wrote: > >>> > > > >>> > > > Hi Keheliya, > >>> > > > > >>> > > > do you use OpenJPA or Hibernate? Does this also happen when you > >>> add a > >>> > > small > >>> > > > delay between the first and the second call? > >>> > > > > >>> > > > Thanks, > >>> > > > Tammo > >>> > > > > >>> > > > > >>> > > > On Tue, Jun 18, 2013 at 8:28 AM, Keheliya Gallaba < > >>> > > > keheliya.gall...@gmail.com> wrote: > >>> > > > > >>> > > > > Hi devs, > >>> > > > > > >>> > > > > I have deployed a process with simple correlation and an event > >>> > handler > >>> > > in > >>> > > > > ODE 1.3.5 (the process archive attached). Sometimes when I > >>> invoke the > >>> > > > > onEvent operation as the second call to the process instance, > it > >>> does > >>> > > not > >>> > > > > get properly correlated. Process instance remain in the Active > >>> state > >>> > > > > forever. Note this occurs only for some instances. Please help > me > >>> > with > >>> > > > some > >>> > > > > insight on this problem. > >>> > > > > > >>> > > > > Thanks & Best Regards, > >>> > > > > Keheliya > >>> > > > > > >>> > > > > -- > >>> > > > > Keheliya Gallaba > >>> > > > > More about me: http://about.me/keheliya > >>> > > > > > >>> > > > > >>> > > > > >>> > > > > >>> > > > -- > >>> > > > Tammo van Lessen - http://www.taval.de > >>> > > > > >>> > > > >>> > > > >>> > > > >>> > > -- > >>> > > Keheliya Gallaba > >>> > > More about me: http://about.me/keheliya > >>> > > > >>> > > >>> > >> > >> > > >