what is the maximum " Beans in Use count" wil be allowed for an ejb in weblogic server

[Kumar]

Hi
The following are the points regarding the problem about "EJB Client invokes
the Bean only 12 times, 13th time client not connects to server, in such a
way that for each remote method call is blocked at server side". What is the
maximum " Beans in Use count" will be allowed for an ejb in weblogic server?

-- Client side error is below, it shows clearly that problem is with
weblogic server side. Server is unable to delegate the request.
==============================================
Starting the test client program.
Lookup failed: null
javax.naming.CommunicationException. Root exception is
java.net.ConnectException: t3://localhost:7001: Destination unreachable;
nested exception is:
java.io.InterruptedIOException: Read timed out; No available router
to destination
at weblogic.rjvm.RJVMFinder.findOrCreate(RJVMFinder.java:155)
at weblogic.rjvm.ServerURL.findOrCreateRJVM(ServerURL.java:207)
at
weblogic.jndi.WLInitialContextFactoryDelegate.getInitialContext(WLInitialCon
textFactoryDelegate.java:311)
at
weblogic.jndi.WLInitialContextFactoryDelegate.getInitialContext(WLInitialCon
textFactoryDelegate.java:211)
at
weblogic.jndi.WLInitialContextFactory.getInitialContext(WLInitialContextFact
ory.java:149)
at
javax.naming.spi.NamingManager.getInitialContext(NamingManager.java:665)
at
javax.naming.InitialContext.getDefaultInitCtx(InitialContext.java:246)
at javax.naming.InitialContext.init(InitialContext.java:222)
at javax.naming.InitialContext.<init>(InitialContext.java:178)
===============================================

-- Sometimes the number 12 may vary, When I am testing on NT, the error
always occurred for 14th request instead of
13th request. Probably this is the number for NT. This has been tested on
NT.

-- This happens because, If weblogic server reaches a stage where if the
parameter for ANY BEAN "Beans in Use count"
is 13 in NT. That's it, it wont be able to process any request after that.
It is not only with EJB client, even weblogic console also fails to connect
to server. You can also see the above error. "Beans in Use count" will be
increased when server is executing the method still on one EJB. That means
if a method is blocked, then it increases the "Beans in use count". if it
reaches the "13" then server wont process any request. This has been tested.

-- In general, if any ejb client connects and makes a call on ejb, then if
server finds "idle ejb" instances for that ejb. Then it takes the one "idle
ejb" instance and executes the remote method. During the execution of a
remote method "idle ejb count" will be reduced by 1 and "Beans in Use count"
will be increased by 1 for that particular EJB. once if the method execution
finishes then "Beans in use count" will be reduced by 1 and "idle ejb count"
will be increased by 1.

I wonder weblogic server supports more "Bean in use instance" rather than
poor number 12. Please correct me if I am wrong and let me know if this
tests are controversy with any of you have more relevant information on
this...

Thanks
Kumar

[Dimitri I. Rakitine]

Did you make sure that this bean actually works and there are no locking
issues? It sounds like all WebLogic execute threads are blocked - default
number is 15 - 2 (I think) socket readers on Win32.

Try to make thread dump (Ctrl-break on Windows) when this happens.

> Thanks
> Kumar

--
Dimitri

[Kumar]

Hi Dimitri,

Thanks for response. But is this solved by using Deployment descriptor,
something like below. I will try that

<pool>
<max-beans-in-free-pool>500</max-beans-in-free-pool>
<initial-beans-in-free-pool>250</initial-beans-in-free-pool>
</pool>

I am just eager to know, I suppose this solves only the pre configuration of
how many bean instances need to create at server side. It wont solve the
Threads increasing at server side?.

Thanks
Kumar

"Dimitri I. Rakitine" <dim...@dima.dhs.org> wrote in message
news:3dc8...@newsgroups.bea.com...

[Kumar]

Thanks,
I found "ThreadCount" attribute in config.xml. I will try that option to
increase the threads.

Kumar

"Dimitri I. Rakitine" <dim...@dima.dhs.org> wrote in message
news:3dc8...@newsgroups.bea.com...

> It's the numbers you mentioned in your previous message sounded
suspicious -
> if server stops responding after 13th request, it usually means that all
> execute threads are busy/blocked, assuming default thread pool size.
>
> If this happens again, you really should make a thread dump and post it
here.

> --
> Dimitri

[Dimitri I. Rakitine]

It's the numbers you mentioned in your previous message sounded suspicious -
if server stops responding after 13th request, it usually means that all
execute threads are busy/blocked, assuming default thread pool size.

If this happens again, you really should make a thread dump and post it here.

> Thanks
> Kumar

--
Dimitri