[Installation, Configuration & Deployment] - Re: Bug?: 4.2GA & MySQL => SpyJMSException
by dietice
Bump!
Am I really the only one who is using MySQL? Or, asked in an other way: Is anyone using MySQL 5.0.x together with JBoss 4.2.0 GA without getting this warning? If so, please post your datasource configuration.
@JBoss Devs: *PLEASE* have a look at this! I spend the last days looking for a solution but found nothing - except the JIRA issue mentioned in my first post.
If I don't get a response I'll open a new issue but I really don't think that I'm the only one who's using MySQL?!
Again: To repoduce this just replace /server/default/deploy/hsqldb-ds.xml with /docs/examples/jca/mysql-ds.xml, adapt it to your configuration, startup the server and you will see this exception (even with a clean installation and NO applications installed).
Last but not least it's probably worth mentioning that the connection is working because the necessary tables get created but I'm still worried why this is happening!
View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4049661#4049661
Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4049661
18 years, 10 months
[JBoss jBPM] - please help: concurrent paths execution
by ax666
Hi all, I'm really stuck with this problem now, and after reading for hours it didn't get any better. I'm using jbpm 3.2, here's a simple example of what I want do do:
| <process-definition>
| <start-state name="start1">
| <transition name="to-fork" to="fork1"/>
| </start-state>
|
| <fork name="fork1" async="true">
| <transition name="to-proc1" to="proc1"/>
| <transition name="to-proc2" to="proc2"/>
| </fork>
|
| <process-state name="proc1" async="true">
| <sub-process name="doThatThing"/>
| <variable name="p1.doWhat" access="read" mapped-name="doWhat"/>
| <transition name="1to-join" to="join1"/>
| </process-state>
|
| <process-state name="proc2" async="true">
| <sub-process name="doThatThing"/>
| <variable name="p2.doWhat" access="read" mapped-name="doWhat"/>
| <transition name="2to-join" to="join1"/>
| </process-state>
|
| <join name="join1" async="true">
| <transition name="to-end" to="end"/>
| </join>
|
| <end-state name="end">
| </end-state>
| </process-definition>
|
all of this is running in a tomcat with a JobExecutor having more than 1 thread. I'm having several issues with this scenario which due to lack of deep experience with jbpm/hibernate I can't solve.
First, JobExecutor tries to work on the same job within different threads.
Second, the join node runs into a hibernate sync problem because both tokens work with the parent token.
Note: the join node problem occurs only if the two sub-processes finish at the same time.
I think I'm missing something here in general. Any help is appreciated to point me in the right direction. All I want is to run the two (complex) sub-processes simultaneously. Again, reading all the old threads did not really help.
Thanks, Alex
View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4049659#4049659
Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4049659
18 years, 10 months
[Remoting] - Re: Problem calling EJB3 method via HTTP only
by MarcReis
Hello again,
I just wanted to post some more detail to this. Hope this is ok, since we both have the same exception...
I have debugged the client part and had a look at the server source.
The Server throws an exception when it casts in the following lines
?(MarshalledInvocation) ois.readObject()?:
| // Get the invocation from the post
| ServletInputStream sis = request.getInputStream();
| ObjectInputStream ois = new ObjectInputStream(sis);
| mi = (MarshalledInvocation) ois.readObject();
| ois.close();
|
This happens in the org.jboss.invocation.http.servlet.InvokerServlet as the exception states.
On my Client side I stepped into it at the org.jboss.aspects.remoting.InvokerRemoteInterceptor around line 62
try
| {
| client.connect();
| org.jboss.aop.joinpoint.InvocationResponse response = (org.jboss.aop.joinpoint.InvocationResponse)client.invoke(invocation, null);
| invocation.setResponseContextInfo(response.getContextInfo());
| return response.getResponse();
| }
|
As far as I can tell with my humble knowledge, going through the interceptor chain and peeking at the values, it all seemed ok.
The Object to pass (invocation) is of org.jboss.aop.joinpoint.Invocation (actually a MethodInvocation). In the MicroRemoteClientInvoker the Marshaller?s are set to HTTPMarshaller and HTTPUnMarshaller and it sets the payload to the invocationReq. After the calls of the HTTPClientInvoker it returns the response, which holds nonsense (the server throws the error when processing the request at the above mentioned line when casting).
The Result is made up of 27 chars: "’##sr#$org.jboss.invocatio". The #'s here replaces the control characters. Casting this results obviously leads to the mentioned client exception (java.lang.ClassCastException: java.lang.String).
So it seems that for some reason reading from the input stream on the server side results in an object which cannot be interpreted into the expected MarshalledInvocation. Unfortunately, I have not had time to debug the server?
Sincerely
Marc
View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4049654#4049654
Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4049654
18 years, 10 months