users@glassfish.java.net

Re: Glassfish V2: ResourceAdapter JavaBean of JCA-Connector instantiated twice

From: Sivakumar Thyagarajan <Sivakumar.Thyagarajan_at_Sun.COM>
Date: Tue, 26 Jun 2007 16:21:17 +0530

Hi

> Question: Why is the first instance stopped and a second instance
> created?

This ideally should not happen and I tried simulating this on a sample
resource adapter in my GlassFish setup [beta-3 branch b50-f] and
couldn't reproduce this.

Could you please provide more details on this issue? Could you for
example, add a "new Throwable().printStackTrace()" in your
ResourceAdapter implementation class's constructor and/or start() method
and send the 2 stack traces you find in the server.log?

Thanks
--Siva.


glassfish_at_javadesktop.org wrote:
> Using Sun Java System Application Server 9.1 (build b41d-beta2) and our JCA CICS connector I observed that during startup the application server shows the following strange behaviour:
> - the AS instantiates the ResourceAdapter JavaBean of the JCA-connector, sets all properties, calls the start() method of the ResourceAdapter JavaBean, and initializes the managed connection factories defined in the ra.xml,
> - then the AS stops the JCA-connector by calling the stop() method of the ResourceAdapter JavaBean, and
> - immediately creates a new instance of the ResourceAdapter JavaBean
>
> Question: Why is the first instance stopped and a second instance created?
>
> The JCA 1.5 spec says in chapter 5.3.1: "The application server must instantiate exactly one ResourceAdapter JavaBean per functional resource adapter instance."
>
> Considering this statement in the spec, isn't the observed behaviour of Glassfish V2 erroneous?
>
> CTee
> [Message sent by forum member 'ctee' (ctee)]
>
> http://forums.java.net/jive/thread.jspa?messageID=223928
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>