My question is: is it true that CachingConnectionFactory causes problem in Camel OSGi environment? On Jan 3, 2014 7:21 PM, "Claus Ibsen" <claus.ib...@gmail.com> wrote:
> On Fri, Jan 3, 2014 at 1:18 PM, Muhammad Ichsan <ich...@gmail.com> wrote: > > I'm using org.springframework.jms.connection.CachingConnectionFactory > > instead of org.apache.activemq.pool.PooledConnectionFactory. Is that > > problem? > > > > If you are using blueprint, then yeah you need to call its start|stop > method also - the method names may be destroy / init or whatever, so > you need to check that out. > > If you use spring xml file, then spring automatic should call its > start|stop methods. > But in blueprint this does not happen. > > > > While using org.apache.activemq.pool.PooledConnectionFactory with > > proper start and stop (init-method="start" destroy-method="stop"), I > > have no problem. > > > > On Fri, Jan 3, 2014 at 6:32 PM, Claus Ibsen <claus.ib...@gmail.com> > wrote: > >> If you use a connection pool then make sure that this pool is also > >> shutdown when you stop the bundle. > >> > >> eg usually that is to ensure the pool start|stop methods is invoked. > >> > >> See details at: > >> http://camel.apache.org/activemq > >> > >> On Fri, Jan 3, 2014 at 12:07 PM, Muhammad Ichsan <ich...@gmail.com> > wrote: > >>> I have routes defined in an OSGi bundle. The routes use activemq > >>> component to process data. To be exact, I have camel listen into 3 > >>> different queue with concurrency = 10. > >>> > >>> As I started it in servicemix (apache-servicemix-4.4.0-fuse-00-27), I > >>> see in the Active MQ Web console that the queues are consumed as > >>> expected: > >>> - queue.a = 10 concurrent consumers > >>> - queue.b = 10 concurrent consumers > >>> - queue.c = 10 concurrent consumers > >>> > >>> Without doing any transaction, I stop the bundle. But strangely I have > >>> the following fact: > >>> - queue.a = 10 concurrent consumers > >>> - queue.b = 9 concurrent consumers > >>> - queue.c = 1 concurrent consumers > >>> > >>> where I expect all of them to be 0 concurrent consumers. > >>> > >>> This is a big problem, since when I start the bundle back, it creates > >>> more than 10 concurrent consumers for each queue. And my transaction > >>> can not run properly, because the old consumers (which remains), > >>> interfere the transaction messages. > >>> > >>> My question is, what's wrong here? Is a bug? Or I should do something > >>> to make it as expected? > >>> > >>> FYI, this is my ActiveMQ log level = INFO on Servicemix: > >>> http://pastebin.com/AcgQApDK > >>> > >>> Thanks > >>> > >>> -- > >>> ~The best men are men who benefit to others > >>> http://www.michsan.web.id 一緒に勉強しましょう！ > >>> > >>> Yang berkualitas memang beda rasanya! > >>> http://rizqi-cookies.com > >> > >> > >> > >> -- > >> Claus Ibsen > >> ----------------- > >> Red Hat, Inc. > >> Email: cib...@redhat.com > >> Twitter: davsclaus > >> Blog: http://davsclaus.com > >> Author of Camel in Action: http://www.manning.com/ibsen > >> Make your Camel applications look hawt, try: http://hawt.io > > > > > > > > -- > > ~The best men are men who benefit to others > > http://www.michsan.web.id 一緒に勉強しましょう！ > > > > Yang berkualitas memang beda rasanya! > > http://rizqi-cookies.com > > > > -- > Claus Ibsen > ----------------- > Red Hat, Inc. > Email: cib...@redhat.com > Twitter: davsclaus > Blog: http://davsclaus.com > Author of Camel in Action: http://www.manning.com/ibsen > Make your Camel applications look hawt, try: http://hawt.io >
