RMI Timeout

For a while I've tried to find solution to this. So, I have a web application that uses RMI to access business module, that si, session ejb. It's using typical
Context ctx;
ctx=getInitialContext();
Facade ejb=(Facade)ctx.lookup("Facade");
etc.
where getInitialContext looks up a remote server.
private static Context getInitialContextStandalone() throws NamingException {
Hashtable env = new Hashtable();
// Standalone OC4J connection details
env.put( Context.INITIAL_CONTEXT_FACTORY, "oracle.j2ee.rmi.RMIInitialContextFactory" );
env.put( Context.SECURITY_PRINCIPAL, "oc4jadmin" );
env.put( Context.SECURITY_CREDENTIALS, "adminadmin" );
env.put(Context.PROVIDER_URL, "ormi://localhost:23791/ejb1");
return new InitialContext( env );
And of course, I have a situation: what happends if connection breaks so I cant connect to a server? Of course, I handle naming exceptions, but I didn't find any way to limit timeout. Timeout occurs after about 40 seconds in here.
So, is there any way to set value for this timeout? Can it be done from the program or is there some server or java parameter that can be used?

Hi,
Did you ever manage to get a resolution to this?
I'm hitting a similar issue when I'm invoking a BPEL using the IDevlieryService interface, and I could really do with increasing the timeout value.
Cheers,
Martin

Similar Messages

How to increase RMI-timeout

Hi there,
I have an session ejb method, running quite a while. Calling this method, results in
0rg.omg.CORBA.COMM_FAILURE: vmcid: SUN minor code: 209 completed: No
Can this timeout be increased? How to?
regards
Heiner

Hey Andy,
Thanks, I have the same problem with our "reborn" client app that now uses RMI/IIOP (fully refactored for JAAS by the way, thanks for the help previously).
I can't find this setting on the IIOP tab in the console, so I tried to set it through <-Dweblogic.iiop.IdleConnectionTimeout=6000>, but the dreaded CORBA.COMM_FAILURE still occurs on long running operations, about 4 minutes in. The server continues to completion, but the client has already given up.
Any ideas?
Best regards,
Peter
PS - server is WLS 8.1sp3.
Message was edited by premster at Sep 1, 2004 1:31 AM

Configure RMI Timeout?

Is there any way to control the timeout for a RMI call in 10.1.2.2 (heck, in any version of the Oracle app server)?
We have several business services exposed using SessionEJBs on a backend business server. However, if the call to these methods take more than about 30 seconds to complete we receive the following exeption (even if it the operation completed successfully):
com.evermind.server.rmi.OrionRemoteException: Transaction was rolled back: timed out;
How can we set the length of time before this timeout error will be generated? I'd like to ensure that in the event of long-running operations the request will not timeout and the function can complete normally. It makes sense to set some kind of a timeout but seems there should be some way to control this value based on the specific operation being performed.

Michael,
We use the [System] property "rmi.client.connection.timeout".
You can set it when you launch OC4J stand alone using the "-D" option of the java command, for example:
java -Drmi.client.connection.timeout=10 -jar oc4j.jarShould set the timeout to ten (10) seconds.
Good Luck,
Avi.

EJB client timeout

WL6.1 SP2.
When a client(either in process or standalone) makes a call to a
stateless session bean, the client is blocked until the ejb method
returns. Is there any way for the synchronus call to timeout and throw
either a RemoteException or
TimeoutException after a specified period?
Thanks in advance!
Jimmy

Jimmy <[email protected]> wrote:
Thanks for your replies.
I've tried to set the timeout attibute in the RMI node of the
file s2t2uhEOImplRTD.xml in the temp ejb dir of weblogic
and it does not work. Where do you specify the timeout attribute?
What does RMI DD mean?I don't know what's the official name of xxx_EOImplRTD.xml is - that's what I meant by
RMI DD. You can edit it and put it back into ejb-jar.
Thanks!
Jimmy
"Dimitri I. Rakitine" <[email protected]> wrote in message news:<[email protected]>...
I didn't see any way to specify it using ejbc, but adding it to the
RMI DD works (for RMI timeout) :
<rmi ...
<method name="..." timeout="..."/>
</rmi>
and the remote invocation fails when timeout is exceeded:
weblogic.rmi.extensions.RequestTimeoutException: RJVM response timed out after: '1000' milliseconds.
at weblogic.rmi.internal.BasicOutboundRequest.sendReceive(BasicOutboundRequest.java:85)
at weblogic.rmi.cluster.ReplicaAwareRemoteRef.invoke(ReplicaAwareRemoteRef.java:262)
at weblogic.rmi.cluster.ReplicaAwareRemoteRef.invoke(ReplicaAwareRemoteRef.java:229)
at weblogic.rmi.internal.ProxyStub.invoke(ProxyStub.java:35)
at $Proxy2.testTimeout(Unknown Source)
Rob Woollen <[email protected]> wrote:
There is a RMI timeout parameter, but I'm not sure if it's exposed in
6.1 or even 7.0. You might be better off asking support. That being
said, I believe it only times out the calling thread. We can't stop the
server thread doing the timed-out work.
-- Rob
Jimmy wrote:
WL6.1 SP2.
When a client(either in process or standalone) makes a call to a
stateless session bean, the client is blocked until the ejb method
returns. Is there any way for the synchronus call to timeout and throw
either a RemoteException or
TimeoutException after a specified period?
Thanks in advance!
Jimmy
Dimitri

JMX Timeouts

Hi,
In weblogic you invoke JMX via t3/rmi, so any timeout that is possible should
be at the rmi/t3 level......
so my question is: is there any flag available to accomplish the timeout ???
Or how can we have a timeout of any JMX query that takes a lot of time sometimes
-Mike

Rob:
Are you looking for the client to time out, the work on the server to
time out, or both?I want BOTH.
hmmm, JMS !!!!
Rob Woollen <[email protected]> wrote:
Are you looking for the client to time out, the work on the server to
time out, or both?
I can't recall if RMI timeouts are public/supported or not. You might
want to search the docs or post in a RMI newsgroup.
To my knowledge, JMX doesn't have timeouts on its queries. You might
ask in the management newsgroup.
You might consider architecting this with JMS instead. Post your query
on a JMS queue. A consumer can dequeue, issue the query, and post a
response to a response destination. Your client could use standard JMS
receive timeouts.
-- Rob
mike wrote:
Hi,
In weblogic you invoke JMX via t3/rmi, so any timeout that is possibleshould
be at the rmi/t3 level......
so my question is: is there any flag available to accomplish the timeout???
Or how can we have a timeout of any JMX query that takes a lot of timesometimes
-Mike

Getting an EJB to Timeout

What is the best way to get a Stateless EJB to timeout?
If my application makes a call to an EJB and I want the
maximum length of the call to take X seconds, how do I
accomplish this. If X seconds has passed, I want the EJB
to timeout and return to the client.
Is there some RMI timeout feature? Do I have to create threads
on my own to perform the timeout? What is the best method?
Thanks.

Are you using Container Managed Transactions? If your business method
is declared to execute within a TX (by setting it's TX attribute to
"Required", for example), you can then specify the timeout period for
the transaction in your weblogic-ejb-jar.xml file:
<transaction-descriptor>
<trans-timeout-seconds>60</trans-timeout-seconds>
</transaction-descriptor>
Hope this helps,
Tim

RMI Timout for EJB Stubs

I can get RMI timeouts working by manually adding timeout="N" to the
XXX_EOImplRTD.xml file in the compiled EJB jar. For example:
<rmi Name="ejb.XXX_EOImpl">
<cluster clusterable="true" />
<method name="*" timeout="500" />
</rmi>
Is there a way to set the timeout by passing a parameter to ejbc or
adding a tag to the weblogic-ejb-jar.xml?
Thanks,
Bob

I didn't find any
Bob Lee <[email protected]> wrote:
I can get RMI timeouts working by manually adding timeout="N" to the
XXX_EOImplRTD.xml file in the compiled EJB jar. For example:
<rmi Name="ejb.XXX_EOImpl">
<cluster clusterable="true" />
<method name="*" timeout="500" />
</rmi>
Is there a way to set the timeout by passing a parameter to ejbc or
adding a tag to the weblogic-ejb-jar.xml?
Thanks,
Bob--
Dimitri

URGENT: pls advise. Timeout in RMI?

I cant believe it ... if my client hangs ... the server waits forever for the client to respond and practically dies? Is this typical behaviour. How can I tell the server to timeout EVEN if there is no RemoteException within a specified period in time? Source code appreciated.
details
The question says it all - I'm looking for the simplest way to implement a timeout in RMI. So if a call is made to a remote object and a response is not received in a certain amount of time, it times out.
FYI: The problem I am trying to deal with is the software (where the remote object lies) may hang and WILL therefore NOT throw a ConnectException or RemoteException. I would like to have an automatic timeout, to prevent blocking IO stalling my server.
Thanks!

Currently RMI client can not specify timeout period for a remote method invocation by setting any sun.rmi.xxx property, so RMI client is blocked until a response is received from the remote server object. However, RMI client can specify a particular client socket factory for creating a socket to communicate with remote server object. Client socket factory can set a timeout period on the socket created to the appropriate value before returning it to the RMI transport layer.
a) Create a customized client socket factory by implementing RMIClientSocketFactory.
class MyClientSocketFactory implements RMIClientSocketFactory
Socket createSocket(String host, int port)
Socket socket = new Socket(host, port);
socket.setSoTimeout(60000L) // 1 Min
return socket;
b) Export the remote object by specifying this particular client socket factory.
In this example, if the client does not receive a response from the remote server object with in 1 min, remote method invocation on the stub will result in interrupted exception. Setting the following properties may not help in detecting timeout from the client side after initiating a remote method on the remote server object.
sun.rmi.transport.tcp.readTimeout represents the time (in milliseconds) used as an idle timeout for incoming RMI-TCP connections. This value is passed to setSoTimeout on the socket created by the RMI server transport layer after accepting a connection request from the remote client. This timeout helps the server tranport layer in detecting the unclosed connections from the client. Normally RMI client tranport layer closes the established connections if they are in idle state more than the time period specified by the property sun.rmi.tranport.connectionTimeout.
RMI client tranport layer try to reuse the established RMI socket connections in future remote method invocations.
sun.rmi.transport.connectionTimeout represents the period (in milliseconds) for which RMI socket connections may reside in an "unused" state, before the RMI runtime will allow those connections to be freed.
-- Srinath Mandalapu

Can I specify a timeout value for an RMI call?

Is it possible to specify a timeout value when a client makes a remote
RMI call to a WL
EJB? If the server method hangs does the client call hang forever with it?
- Rick Holland OO/Distributed Systems Consultant
- ObjectAge Inc.
- [email protected]
"Logic is a wreath of pretty flowers, that smell bad"

That's what I did to try this. weblogic.ejbc knows how to pass -dispatchPolicy to
the weblogic.rmic, so in the RMI descriptor it creates:
<method
name="*"
dispatch-policy="queueName"
>
</method>
so, maybe there is a command-line option for the timeout as well. Anyway, extracting
xxxRTD.xml and adding
<method
name="*"
timeout="xxx"
>
</method>
works just fine.
Rick Holland <[email protected]> wrote:
So all I have to do is build my EJB jar, extract the *RTD.xml file, hack
it, stuff it back into the jar and deploy (all within my Ant script).
Sounds fairly straight forward.
Thanks again
[email protected] wrote:
It gets there from the server. If you look inside ejbc'ed ejb-jar you'll find
RMI descriptors there (they are named like xxx.RTD.xml)
Rick Holland <[email protected]> wrote:
Thank you Dimitri, but I have one more dumb question, since I need these
timeouts
on the client side where does an RMI descriptor go? I have never used
one of these.
Thanks
[email protected] wrote:
AFAIK, you can specify it in the RMI descriptor :
<rmi Name=".....">
<cluster
     clusterable="true"
</cluster>
<method
     name="methodname"
     timeout="timeoutinmilliseconds"
</method>
</rmi>
And, if there is timeout, method invocation will fail with exception like this:
weblogic.rmi.extensions.RequestTimeoutException: RJVM response timed out after: '1000' milliseconds.
at weblogic.rmi.internal.BasicOutboundRequest.sendReceive(BasicOutboundRequest.java:85)
at weblogic.rmi.cluster.ReplicaAwareRemoteRef.invoke(ReplicaAwareRemoteRef.java:262)
at weblogic.rmi.cluster.ReplicaAwareRemoteRef.invoke(ReplicaAwareRemoteRef.java:229)
at weblogic.rmi.internal.ProxyStub.invoke(ProxyStub.java:35)
at $Proxy2.testTimeout(Unknown Source)
Rick Holland <[email protected]> wrote:
Is it possible to specify a timeout value when a client makes a remote
RMI call to a WL
EJB? If the server method hangs does the client call hang forever with it?
- Rick Holland OO/Distributed Systems Consultant
- ObjectAge Inc.
- [email protected]
"Logic is a wreath of pretty flowers, that smell bad"
- Rick Holland OO/Distributed Systems Consultant
- ObjectAge Inc.
- [email protected]
"Logic is a wreath of pretty flowers, that smell bad"
- Rick Holland OO/Distributed Systems Consultant
- ObjectAge Inc.
- [email protected]
"Logic is a wreath of pretty flowers, that smell bad"
Dimitri

RMI Server Instance Timeout

Hi
I have a Server instance which is used by clients once in 2 days.
what i want is ,if the instance is not used by any client for 1 day,it should be cleaned up.And if the clients give a request on second day,a new instance should be made and serve the clinets.
Is there any timeout feature for RMI Server side object.

I would use an Activatable server so you can reactivate automatically, and implement an Unreferenced.unreferenced() method in which you call Activatable.inactive() if the object has been idle for too long.

RMI - How to set a client timeout?

Hey folks,
please don't crucify me, but i googled for "rmi client timeout" and i only found confusing and / or really dated information.
Could somebody enlighten me on how to set a timeout for an RMI-client?
Even an RTFM would be fine, but which?

Hi,
thanks for your insistence.....:-)
ejp wrote:
1. Are you using rmic? UnicastRemoteObject? or some JBoss API?Ehm, for the client?
The client's sourcecode for connecting to the Jboss-Server / my EJB-application is:
Hashtable<String, String> props = new Hashtable<String, String>();
props.put(Context.INITIAL_CONTEXT_FACTORY,"org.jnp.interfaces.NamingContextFactory");
props.put(Context.PROVIDER_URL,"jnp://" + "myhost");
props.put(Context.SECURITY_PRINCIPAL, "myuser");
props.put(Context.SECURITY_CREDENTIALS, "mypassword");
Context ctx = new InitialContext(props);
// get RMI stub
ISendCommand sendCommand = (ISendCommand) ctx.lookup(myjndiname);
String result = sendCommand.sendCommand(myArgs);So i think this corrensponds to "UnicastRemoteObject" from your enumeration?
ejp wrote:
2. How do you know JBoss terminates the connection?Because i have a timeout configured in my EJB-application, but in the example i try to get working, the client shall use its own timeout which is before the EJB-timeout.
The EJB-timeout works, but not the client timeout.....(as described above)
ejp wrote:
3. What happens at the client? An exception? Nothing?"Nothing", i.e. the client terminates when the connection times out (from the JBoss / EJB-side!)
ejp wrote:
4. Are you ignoring any exceptions at the client?No.
Shouldn't this client timeout be totally independent from any server / Jboss / whatever stuff?
Any more ideas?

RMI Client timeout set

I want to set a client timeout for RMI withing WLS. Is there a way to do
this? Any examples?
thanks,
Mike

Sorry.. I thought you were talking about rmi/iiop.
There is some time-out mechanism in Denali (5.1.x), but nothing in 4.5.x, as the neccessary infrastructure was not in place unitl Denali.
Let me know if you want me to look into how this is done in 5.1.x.
Michael Lee wrote:
I have 4.5.1 at the customer site now. Can this work with 4.5.1.
BTW, thanks.
----- Original Message -----
From: "Eduardo Ceballos" <[email protected]>
Newsgroups: weblogic.developer.interest.rmi-iiop
To: "Michael Lee" <[email protected]>
Sent: Wednesday, June 07, 2000 6:13 AM
Subject: Re: RMI Client timeout set
The idle disconnect timeout property will be aviable in service pack 4 and5.1.1.
Michael Lee wrote:
I want to set a client timeout for RMI withing WLS. Is there a way to do
this? Any examples?
thanks,
Mike

Rmi method call timeout

i was wondering if there was any way to timeout a rmi method call.
the situation i had was that a client would envoke an rmi method, and
the method would deadlock in weblogic. the client making the call will
hang forever.
thanks

Thank you for this enlightening addition. I regret I failed to perceive
earlier that your comments were to be interpreted as evidence of sincere
commitment to solving this problem.
Regarding 7.0, you'll note that I described the timeout attribute solution
as "improving stability" rather than as a complete cure. It should be borne
bear in mind that this is an alternative to a completely locked system. When
stress tested, this change did allow the application to recover and run for
the rest of the day.
I'm obliged to you for your concern regarding any liability I might be
incurring in relation to the BEA licensing agreement. Fortunately I think I
may be able to put your mind at rest on this matter relatively easily. If I
may direct your attention to the BEA support site, my experience is that on
entering the search terms "rmi descriptor timeout", the relevant information
will be revealed. Also, I suspect that, as a fellow engineer, it has not
escaped your attention that the RMI descriptor is a valid XML document. As
such, it has a DTD named "rmi.dtd" which (alas) also fails to conceal this
attribute.
While we're pursuing what has now become a gratifyingly productive vein,
let's offer another possible approach to this issue, namely the use of
per-VM timeouts for socket calls. These were introduced in the Sun JDK 1.4.1
and so are available to people running WebLogic 8.1 today. Again these
suggestions have been tested for basic socket I/O (I haven't tried RMI). To
my knowledge these are not endorsed by BEA:
For client connections, timeouts can be set globally by passing the
following startup parameters to the JVM:
sun.net.client.defaultConnectTimeout (default: -1)
sun.net.client.defaultReadTimeout (default: -1)
sun.net.client.defaultConnectTimeout specifies the timeout (in milliseconds)
to establish the connection to the host.
sun.net.client.defaultReadTimeout specifies the timeout (in milliseconds)
when reading from input stream when a connection is established to a
resource.
(reference: Sun's Networking Properties documentation at
http://java.sun.com/j2se/1.4.1/docs/guide/net/properties.html )
Supplying non-default values for these properties should allow a server to
recover from a blocked downstream system if socket I/O is being used.
Regards
Alex
On Fri, 30 Jan 2004 09:02:23 -0800, Andy Piper <[email protected]>
wrote:
Alex Thomas <[email protected]> writes:
Hardly a constructive response to what is clearly a significant problem.
First, it should be obvious that this issue will be encountered in any
distributed system using T3 or RMI, not because of any shortcoming in
WebLogic services but because of blocking behaviour in downstream systems.That's why it is a new feature in 9.0. I'm not arguing that the
use-case is invalid, simply that your suggestion should not be tried.
Second, the solution I proposed has been tested (on 7.0SP1) and does improve
overall stability, as other people here have apparently discovered.Tested by who? Not by us because we know there are significant,
dangerous, problems with the approach you describe. Support are
under strict instructions not to offer this as a solution to
anyone. If anyone in BEA told you, they shouldn't have. If you
discovered it by decompiling the code then you are in violation of
your license agreement.
There is a supported mechanism for timing out connect attempts in 7.0
and later which you can get through support, but for timing out RMI
calls you will have to wait until 9.0 (or cook up a business case for
a backport).
Hope this clears things up.
andy

Timeout setting in RMI

Hello,
Does RMI provide a method like setSoTimeout in Socket?
Thank you.

I don't think that's possible, but do you mean on a really low level, and to reprogram an own version of RMI?
see for a similar question with some answers:
http://forum.java.sun.com/thread.jsp?thread=334856&forum=58&message=1370065
/ linda

RMI Connect timeout

hello,
need help with this problem :
i have a server web (tomcat) and many clients witch implements a rmi server.
on intranet, all works fine, that is every minute, the server instanciates every client, gives him the adresses of all others so that every client can communicate with another client using RMI.
when putting the server on Internet and a public line (adsl in france), the server can't access the clients and from the client point of vue, the error is
RemoteException Connection refused to host: 80.9.149.71; nested exception is:
java.net.ConnectException: Connection timed out: connect
thnks for your help and sorry for my english

no firewall on the web server witch is located at home, but i don't know about public network adsl
no ping because ip adress is configured at connect time, but all the clients access the server via http, and the server can't access the clients via rmi, nor each client can access other client via rmi.
thank's for your replyy

RMI Timeout

Similar Messages

Maybe you are looking for