How can a http response be delayed?

Hi,
I have a "simple" problem to solve but I am new to J2EE so I hope to find someone who has already faced with this problem.
I know a servlet can receive an http request from a client and prepare the response. The response will be sent as soon as the servlet returns. The same servlet serves the requests of many users, so if there are 100 users requesting for a servlet I guess the requests are queued and there will be a response outcoming the web-server for each request in the same order.
Please correct me if I say wrong things.
Now I would like to write a java code that receives the request but waits for a trigger event to send the response, without blocking other users.
I mean, a user sends a request, a servlet should forward the request to a thread (what else?) and exit without sending any response. The thread (I thought about a thread for each user) will wait for a trigger event and ONLY if the event happens the response will be sent to the requesting client.
In the meantime other users can send their requests and receive their response when their trigger events has happened.
1. Can J2EE technology allow me to do that?
2. How can I let the thread send the response to the right client?
3. Http packets do not contain the info about the client, how can I trigger the sending of the response?
I had a look to pushlet and comet but the problem I have seems to me different from let the server start a communication with a client. In my case the paradigm request-response is kept but the response is delayed. Of course the time-out should be increased (I have read that time out is settable in the client).
To be more clear I'm adding two examples:
1. one client (an applet) sends its request, the server receives the http packet and it will start a timer of 5 sec. Only when the timer expires the response will be sent. In the meantime other clients perform their requests and they all will be answered after 5 secs from their request (because there is a timer for each request).
2. the trigger is the number of requests received: if the server receives 10 requests from 10 clients then all the 10 clients will receive the response. If the server receives less of 10 requests in the time out period there will be issued a time-out error.
Please could you help me?
Thanks in advance.

maxqua72 wrote:
Hi,
I have a "simple" problem to solve but I am new to J2EE so I hope to find someone who has already faced with this problem.
I know a servlet can receive an http request from a client and prepare the response. The response will be sent as soon as the servlet returns. Not exactly. The request is sent with the contents of the output stream are flushed to the client. This may be right away, or it may not be. This usually does happen before the 'servlet returns' but you have control of when that happens.
The same servlet serves the requests of many users, so if there are 100 users requesting for a servlet I guess the requests are queued and there will be a response outcoming the web-server for each request in the same order.Again, not really true. requests are handled in the order in which they are received, but that does not mean that the first in is the first out. It just means the first in will begin before the second in. The second in could very well finish before the first one.
Please correct me if I say wrong things.
Now I would like to write a java code that receives the request but waits for a trigger event to send the response, without blocking other users.
I mean, a user sends a request, a servlet should forward the request to a thread (what else?) and exit without sending any response. The thread (I thought about a thread for each user) will wait for a trigger event and ONLY if the event happens the response will be sent to the requesting client.
In the meantime other users can send their requests and receive their response when their trigger events has happened.Each request will have its own thread. There is no need to make your own. You could delay the response as long as you like without making your own thread.
This assumes you do not implement the 'SingleThreadModel' deprecated interface on your servlet.
>
1. Can J2EE technology allow me to do that?
2. How can I let the thread send the response to the right client?
3. Http packets do not contain the info about the client, how can I trigger the sending of the response?Yes, JEE can do that. Because each request has its own thread, you can do your delay right in the servlet's service method (doXXX). You would then have the same access to the request/response as always. Alternatively, you could pass the request/response to whatever method/object that will do your work.
Note that using a new thread for this will actually be a detriment, since that will allow the thread the servlet is called in to return, which will end the request cycle and destroy the response object. So you should do it in the same thread that your request is made in.
>
I had a look to pushlet and comet but the problem I have seems to me different from let the server start a communication with a client. In my case the paradigm request-response is kept but the response is delayed.This is actually the same thing as what pushlets do. Pushlets require the user to make an initial request, then the Pushlets hold on to the initial request's thread by never closing the response stream or returning from the service method. It lets Pushlets then continuously send data to the client. Your system would be simpler, you would receive the request, open the response stream, maybe send initial data to client, then delay. When the delay finishes send the rest of the content and close the response stream and let the request die.
Of course the time-out should be increased (I have read that time out is settable in the client).That is true, but is a nightmare if you expect your client base to be more than a few people whom you know. A better option would be to regularly send small bits of 'keep alive' packets so the client knows the page is still active (which may also help prevent the client from navigating away from your site because it would appear frozen). What you could do is send small bits of javascript that update a specific portion of the page (like a count of current request, or a countdown until activation... whatever seems appropriate). Or just send nonsense javascript that doesn't affect the client at all (but which doesn't break your page either).
To be more clear I'm adding two examples:
1. one client (an applet) sends its request, the server receives the http packet and it will start a timer of 5 sec. Only when the timer expires the response will be sent. In the meantime other clients perform their requests and they all will be answered after 5 secs from their request (because there is a timer for each request).No problem. Adding a Thread.sleep(5000) will do this nicely.
>
2. the trigger is the number of requests received: if the server receives 10 requests from 10 clients then all the 10 clients will receive the response. If the server receives less of 10 requests in the time out period there will be issued a time-out error.Again, possible, but you would need to device a means of locking and unlocking threads based on client count. Two quickie methods would be to keep a member variable count of current request. Each request would use synchronized blocks to first increment the count, then in a loop, check the count, wait if it is < 10, or end the synchronized block if it is >= 10.
    public void doGet(...) {
        int localCount = 0;
        //Count this request, then recall how many total requests are already made
        synchronized (countLock) {
            this.count++;
            localCount = this.count;
        //Loop until this thread knows there are 10 requests
        while (localCount <10) {
            //wait for a while, and let other threads work
            try { Thread.sleep(100); } catch (InterruptedException ie) { /* don't care */ }
            //get an updated count of requests
            synchronized (countLock) {
                localcount = this.count;
        } //end while
        //do the rest of the work
        //Remove this request from the count
        synchronized(countLock) {
            this.count--;
    }A second, better approach would use a 'request registrar' which each request adds itself to. The registrar blocks each request on the condition that 10 requests come in, then releases them all. This would use the tools available in java.util.concurrent to work (java 5+). Untested example:
public class RequestRegistrar {
    private final ReentrantLock requestLock = new ReentrantLock();
    private final Condition requestCountMet = requestLock.newCondition();
    private int count;
    private static final int TARGET_COUNT  = 10;
    private static final RequestRegistrar instance = new RequestRegistrar();
    public static RequestRegistrar getInstance() { return instance; }
    private RequestRegistrar() { }
    public void registerRequest() {
        this.requestLock.lock();
        try {
            this.count++;
            if (this.count < TARGET_COUNT) {
                while (this.count < TARGET_COUNT) {
                    try {
                        this.requestCountMet.await();
                    } catch (InterruptedException ie) {
                        /* don't care, will just loop back and await again */
                } //end loop of waiting
            } else { //the target count has been met or exceeded
                this.count = 0; //reset count
                this.requestCountMet.signalAll();
        } finally {
            this.requestLock.unlock();
//Then your requests would all do:
    public void doGet(...) {
        RequestRegistrar.getInstance().registerRequest(); //waits for 10 requests to be registered
        //The rest of your work
    }Undoubtedly there is a lot of work to fix this up, but if you understand Locks and conditions it won't be hard (and if you don't perhaps now is a good time to learn...)
>
Please could you help me?
Thanks in advance.

Similar Messages

  • How can I put a time delay between specific events in a while loop?

    How can I put a time delay between specific events within the same while loop? I'm already using the "wait" command to control the overall loop iteration speed. But I want to time the individual events as well.

    Hi Jesse,
    You can use a flat sequence. In each box you can put your individual events and attached wait.
    Don't forget to reduce your total loop time of the time you added in the individual sequences.
    Doc-Doc
    Doc-Doc
    http://www.machinevision.ch
    http://visionindustrielle.ch
    Please take time to rate this answer

  • How can I use the Variable Delay VI (From DSP Module) to Deploy Variable Delay on Speddy-33

    Dears,
    I dont Understand How Can I Use the Variable Delay to Deploy Variable Delay on a Speedy-33 Kit

    Hello,
    Thank you for posting to the NI Forums!! You should be able to incorporate the Variable Delay.vi into your vi and then build as normal and the functionality of the vi should translate to the build. What type of build are you doing? Are you receiving any specific errors? Thanks!
    Regards,
    Margaret Barrett
    National Instruments
    Applications Engineer
    Digital Multimeters and LCR Meters

  • How can I fix the long delay (15 min  ) between the time I select a song to the time it begins playing?

    How can I fix the lond delay (15 min.) between the time I select a song to the time it begins playing?

    Hi Ben,
    This is a common issue that NI LabVIEW does not address.
    The good news is that OpenG provides a function to do just what you need.
    See this article: Resize Front Panel to largest decoration
    The way to get the OpenG functions is to use the JKI VI Package Manager (VIPM).
    In addition to the OpenG functions there are several other community provided tools avaliable through VIPM.
    For a quick test, I'm attaching these files (LV version 2010 sp1):
    Fit VI window to Largest Dec__ogtk.vi and the required sub vi Current VIs Parents Ref__ogtk.vi.
    You should really use VIPM to get the full OpenG library - there are many gems in there.
    steve
    Help the forum when you get help. Click the "Solution?" icon on the reply that answers your
    question. Give "Kudos" to replies that help.
    Attachments:
    Fit VI window to Largest Dec__ogtk.vi ‏29 KB
    Current VIs Parents Ref__ogtk.vi ‏9 KB

  • How can I set up a delayed analog trigger on PCI 6115 DAQ

    I have an S-Series PCI 6115 DAQ which I’m running with Labview. I’m using it to measure signals from an acoustic emission sensor and two force transducers. I’d like to set up a delayed analog trigger which will start acquisition on all three channels a period of time after a selected channel’s voltage exceeds a threshold.
    Currently I’m using the AI Config VI in line with the AI Start VI and AI Read VI to capture data after a analog hardware trigger occurs. A software trigger probably wouldn't work because I have to sample my data at 10MS/sec. My setup works fine for triggering without any delay or skip counts. However, if I set the delay or skip count in the additional trigger parameter field of the AI start VI, there is no effect, and the device still starts capturing data immediately after the trigger is received. What is the cause of this, and how can I get around it?
    Also, is it possible to sample the channels of a PCI-6115 DAQ at different rates? Right now, I’m sampling all my channels at 10MS/sec and throwing away data on all channels except one. However, this seems relatively slow and eventually I would like to attempt pseudo-real time control using my data.

    rpursley8 is right about needing to get the counters involved if you want a hardware timed delay in your application.
    Concerning whether or not you can sample at different rate, check this document out.
    Sampling Different Channels at Different Rates with NI-DAQmx
    Otis
    Training and Certification
    Product Support Engineer
    National Instruments

  • How to see http response from an axis web service call (Eclipse)

    Hello,
    I would like to see the raw http response which is returned from web service calls. I have a dynamic web project in Eclipse which uses a local instance of Tomcat 6. I'm using all of the default setting for a top-down web service generated from a WSDL file. I've generated the client using the built-in "generate client" using default settings.
    I've tried using the Eclipse plugin TCP/IP monitor and apache's TCPMON, but I am only able to see the http request, not the http response returned from the web server I am querying.
    I've seen some sparse documentation outlining how to use logging handlers and a client-config.wsdd file, but I haven't been able to get that working.
    So to recap, I'm looking for a way to view raw http responses using a web service client and server generated from a WSDL file in eclipse. I don't mind creating a new project using different code-generating libraries if someone has an easy way to do this using a different configuration.
    Thanks very much,
    Craig

    908794 wrote:
    Hello,
    I would like to see the raw http response which is returned from web service calls.Why the HTTP response? Isn't the soap message body enough? If it is, you probably want to check out SoapUI.
    A simple google query for "apache axis2 http response" also return this article:
    http://blogs.cocoondev.org/dims/archives/004668.html
    And finally, you did go through the Axis2 website right? It has a wiki with a rather staggering amount of articles in there.
    http://wiki.apache.org/ws/FrontPage/Axis2/

  • How can I keep responses for each question of quiz.

    For auditing purposes, I need to be able to record the user's answers for each of the 35 quiz questions.
    How can I accomplish this?
    Perhaps I can have the answers generated to a final answer page.

    Normally this is done by the LMS...
    If you need to do it in CP: there is a quizzing system variable cpQuizInfoAnswerChoice (Secrets of cpQuizInfoAnswerChoice and.... - Captivate blog)
    Because this variable is reused for each question slide, you'll have to store its value in user variables, one for each question slide. But beware: depending on the type of question, this system variable doesn't have the full answer, but maybe only the number or character identifying the answer (A,B,...).

  • How can prevent http chunks receiving in a web service response?

    Hi guys,
    I'm developing a WS client mobile application, which consumes a Web Service (java based).
    Usually it works fine, but sometimes in the clinet I get: java.rmi.MarshalException: Unexpected Exception : unexpected end of stream
    This error occur when some big data has to be received from the server.
    I've done some research on this problem and find that this is caused because of sending of th response via HTTP chunks. Somehow the server and client sides does not
    communicates as expected.
    I have no access to the Web service implementation.
    Is there a workaround to solve this problem on the client side?
    Can I force WS client to not receive the response in http chunks?
    Thanks

    I managed to fix it by using the well known pattern:
    Web Service on the server(1) <--> Web Application(2) <--> Mobile Client(3).
    For communication between (1) and (2) was by using WS, and between (2) and (3) was made by plain http connection. Now the application is portable and don't have any limitations.
    radarada, thanks for the advice. It may work, may be I can try it when I have some time.
    Vladdy

  • How to check HTTP response code

    Hi,
    we have the following scenario:
    HTTP -> PI -> HTTP
    So it's easy to use HTTP or SOAP-adapter here.
    But we have to inform the sender of the message immediately about the successfull delivered message.
    So we have to check if response code 200 was sent from the receiver of the message and inform the sender about this.
    I can't find any possibility how I can check this in PI 7.1 (neither by using HTTP adapter nor by using SOAP-adapter). There is no acknowledgement mechanism (I can't find any).
    Does anybody know how we can solve this?
    Thanks for your replies.
    Regards
    Thorsten

    If the response has to travel from Receiver to Sender then you have to make use of SYNC mode of processing or use a BPM.
    In Async flow the Sender will get a HTTP 200 OK response from the sender adapter of SAP XI/ PI.
    Regards,
    Abhishek.

  • How can I get responses to return to a specific email address?

    How do I get all my for responses to return to a specific email address? Does the Acrobat user need to be logged in with this email or is there an option for this?
    Any help would be appreciated!

    Looks to me like the Distribute Forms feature in Acrobat taps into the email address you have specified in Acrobat as your identity; that's where it gets the mailto: address associated with the submit button in the document message bar. 
    Furthermore, if you create a form in LiveCycle the distribute process still seems to tap into Acrobat.  Just for grins you can open the distributed form in LiveCycle (better make and open a copy so you don't break the Reader extended usage rights), go & look in the object palette at the submit address property of your email submit button and it will be blank.  Surprise!  This data has migrated to one or more than one script at the document level.  And if your user then opens your LiveCycle-made form in Reader to complete and return, the submit button in the document bar corresponds to your identity email address in Acrobat.
    So I managed a quick but ugly workaround for an email submitted form by:
    Going into Acrobat:Edit/Preferences/Identity
    changing the email address in my identity to the desired return address
    distributing my form (can do either in LiveCycle or in Acrobat)
    putting my Identity email address back where it was afterwards.
    Good luck with all that.  And if anyone knows a more elegant method, do share!

  • How can an OSB response be synchronized?

    Hello experts,
    I have a problem with synchronizing an OSB response.The scenario is as follows: I make a request to OSB which in turn, after processing the request replies with an answer.But the problem is what happens if the same scenario is executed by multiple users, in the way that i cannot match the responses with the requests.
    Can a request be matched in some way with it's response in multiple user(multithreaded) environment?
    What needs to be synchronized there?
    I user 11g R2.
    Thanks in advance and best regards!
    Carol

    Hi Carol,
    You may be looking for something like this...
    http://docs.oracle.com/cd/E23943_01/dev.1111/e15866/jms.htm#OSBDV1051
    Cheers,
    Vlad

  • How can make Jpanel response click event?(help!)

    in a program i want to make jpanel response mouseclicked event, but jpanel class does not provide addActionListener(). through which way? i am able to arrive at the purpose! (create a new class extends jpanel, how to do?)
    thank a lot!!!

    Hello,
    have a look:import java.awt.Color;
    import java.awt.event.*;
    import javax.swing.*;
    public class ClickPanel extends JPanel implements MouseListener
         public static void main(String[]args)
              JFrame frame = new JFrame("Click-Panel");
              frame.setDefaultCloseOperation(JFrame.EXIT_ON_CLOSE);
              frame.getContentPane().setLayout(null);
              frame.getContentPane().add(new ClickPanel());
              frame.setBounds(200,200,400,400);
              frame.setVisible(true);
         public ClickPanel()
              addMouseListener(this);
              setBounds(200,200,100,100);
              setBackground(Color.CYAN);
         public void mouseClicked(MouseEvent e)
              System.out.println("clicked");
         public void mouseEntered(MouseEvent e)
              System.out.println("entered");
              setBackground(Color.ORANGE);
         public void mouseExited(MouseEvent e)
              System.out.println("exited");
              setBackground(Color.CYAN);
         public void mousePressed(MouseEvent e)
              System.out.println("pressed");
         public void mouseReleased(MouseEvent e)
              System.out.println("released");
    }regards,
    Tim

  • How can I set Response Queue for standalone client using JMS Jax-RPC ?

    Hi,
    I am developing a standalone client for a web service running on a Weblogic 10.3 server using JMS transport. The response seems to be placed in JMSServer!JMSServer.TemporaryQueue0 when I try to test the application.
    The request xml created:
    <?xml version="1.0" encoding="UTF-8" standalone="yes"?>
    <content>
        <entry type="1">
            <textMessage fromQueue="true" JMSTimestamp="1285343150328" JMSReplyToDomain="1" JMSReplyTo="JMSServer!JMSServer.TemporaryQueue1" JMSRedelivered="false" JMSPriority="4" JMSMessageID="ID:&lt;520181.1285343150328.0&gt;" JMSExpiration="0" JMSDestination="SystemModule!demoQueue" JMSDeliveryMode="2">
                <headerProperty type="java.lang.String" value="text/xml; charset=utf-8" name="_wls_mimehdrContent_Type"/>
                <headerProperty type="java.lang.String" value="&quot;http://oracle/communications/platform/demo/webservices/rpc/secure/gen/getPhoneList&quot;" name="_wls_mimehdrSOAPAction"/>
                <headerProperty type="java.lang.String" value="/DemoWebServices/DemoWebServicesJMS" name="URI"/>
                <text>&lt;env:Envelope xmlns:env=&quot;http://schemas.xmlsoap.org/soap/envelope/&quot;&gt;&lt;env:Header/&gt;&lt;env:Body&gt;&lt;gen:lastName xmlns:gen=&quot;http://oracle/communications/platform/demo/webservices/rpc/gen&quot;&gt;8&lt;/gen:lastName&gt;&lt;/env:Body&gt;&lt;/env:Envelope&gt;</text>
            </textMessage>
        </entry>
    </content>On the application server I have configured a Response Queue that I would like my application to use. The client stubs where generated using the ant Task: weblogic.wsee.tools.anttasks.ClientGenTask.
    Standalone client code snippet:
          DemoWebServices_Impl service = null;
          try {
              service = new DemoWebServices_Impl(wsdl);
          } catch (ServiceException e) {
              e.printStackTrace();
          DemoWebServicesPortType port = null;
          //create credential provider and set it to the Stub
          try {
              port = service.getDemoWebServicesJMSPort();
          } catch (ServiceException e) {
            e.printStackTrace();
          PhoneList response = null;
          try {
              response = port.getPhoneList("8");
          } catch (RemoteException e) {
              e.printStackTrace();
          if(response != null) {
            for(Phone aPhone : response.getPhone()) {
                System.out.println("aReturned Phone number is:  " + aPhone.getPhoneNumber());
          }Would you please help me with modifying the client code to define which Queue the web service response should be placed in? setQueue() of the JMSTransportInfo class did not do the trick. Please help....
    Thanks,
    Sajitha
    Edited by: Sajitha on Sep 24, 2010 12:29 PM
    Edited by: Sajitha on Sep 27, 2010 8:44 AM

    Hi,
    I am developing a standalone client for a web service running on a Weblogic 10.3 server using JMS transport. The response seems to be placed in JMSServer!JMSServer.TemporaryQueue0 when I try to test the application.
    The request xml created:
    <?xml version="1.0" encoding="UTF-8" standalone="yes"?>
    <content>
        <entry type="1">
            <textMessage fromQueue="true" JMSTimestamp="1285343150328" JMSReplyToDomain="1" JMSReplyTo="JMSServer!JMSServer.TemporaryQueue1" JMSRedelivered="false" JMSPriority="4" JMSMessageID="ID:&lt;520181.1285343150328.0&gt;" JMSExpiration="0" JMSDestination="SystemModule!demoQueue" JMSDeliveryMode="2">
                <headerProperty type="java.lang.String" value="text/xml; charset=utf-8" name="_wls_mimehdrContent_Type"/>
                <headerProperty type="java.lang.String" value="&quot;http://oracle/communications/platform/demo/webservices/rpc/secure/gen/getPhoneList&quot;" name="_wls_mimehdrSOAPAction"/>
                <headerProperty type="java.lang.String" value="/DemoWebServices/DemoWebServicesJMS" name="URI"/>
                <text>&lt;env:Envelope xmlns:env=&quot;http://schemas.xmlsoap.org/soap/envelope/&quot;&gt;&lt;env:Header/&gt;&lt;env:Body&gt;&lt;gen:lastName xmlns:gen=&quot;http://oracle/communications/platform/demo/webservices/rpc/gen&quot;&gt;8&lt;/gen:lastName&gt;&lt;/env:Body&gt;&lt;/env:Envelope&gt;</text>
            </textMessage>
        </entry>
    </content>On the application server I have configured a Response Queue that I would like my application to use. The client stubs where generated using the ant Task: weblogic.wsee.tools.anttasks.ClientGenTask.
    Standalone client code snippet:
          DemoWebServices_Impl service = null;
          try {
              service = new DemoWebServices_Impl(wsdl);
          } catch (ServiceException e) {
              e.printStackTrace();
          DemoWebServicesPortType port = null;
          //create credential provider and set it to the Stub
          try {
              port = service.getDemoWebServicesJMSPort();
          } catch (ServiceException e) {
            e.printStackTrace();
          PhoneList response = null;
          try {
              response = port.getPhoneList("8");
          } catch (RemoteException e) {
              e.printStackTrace();
          if(response != null) {
            for(Phone aPhone : response.getPhone()) {
                System.out.println("aReturned Phone number is:  " + aPhone.getPhoneNumber());
          }Would you please help me with modifying the client code to define which Queue the web service response should be placed in? setQueue() of the JMSTransportInfo class did not do the trick. Please help....
    Thanks,
    Sajitha
    Edited by: Sajitha on Sep 24, 2010 12:29 PM
    Edited by: Sajitha on Sep 27, 2010 8:44 AM

  • Performance Analysis: java classes run fast, but http response is delayed

    Hello,
    I´m analysing a performance issue on our application and I have the following symptons:
    - A profile analysis with YourKit Java Profiler 7.5 told us that our java classes are running "fast" ( mostly less than 10s ).
    - The access_log of the Oracle HTTP Server (httpd.conf LogFormat %T parameter) show us that the time taken to process the entire request is 130s
    - The users told us that the application runs fast for a while, but after some minutes, the entire application runs slowly. After some minutes again, the entire application runs fast again.
    - There are no overload of CPU, MEM, IO, Networking. All have been checked and are ok.
    - At the same time the application A runs slowly, the application B on the same server ( but another instance ) runs normally.
    - There is no database botleneck. All the application queries have been profiled and optimized for best run.
    - A huge number of application threads exists on the java virtual machine. The peak was 187 threads ( mostly sleeping. Just one or two were running ). The threads average in the vm is 150.
    - Our application has few EJB ( 4 ), few ejb-calls. All the EJB calls are cached with EhCache. The cache timeout is 2hours. When time-out occurs, the cache is invalidated, cleared and reloaded.
    - We have only 10 users.
    My hypothesis are:
    1. Application server thread blocking. Some threads take until 13 minutes to finish, even when the java processing took only 23 seconds.
    2. Some problem with the size of the AJP packages between HTTPServer and Application server.
    I´m really desperate for help. Any help will be widely appreciated.
    Oracle Application Server 9.0.3.3 running on
    Sun Solaris 5.9
    Sun Java Development Kit 1.4.1_05-b01
    Is there any way to configure the max number of threads in the application server? Where?
    Thanks in advance,
    Murilo

    You can tweak the application server threads as described here:
    http://download.oracle.com/docs/cd/B31017_01/web.1013/b28950/threadpool.htm#BHBDGJBI
    Oh just noticed you are on 9.0.3.3 -- that is a very, very, very* old release. I don't think we had a tunable thread-pool back then. The oldest doc I can find is for 10.1.2, which I think is where we first introduced the thread-pool manager:
    http://download.oracle.com/docs/cd/B15904_01/web.1012/b14011/advanced.htm#i1014357
    Is there any garbage collection occurring that could account for the slow-down in ART? I guess not if APP-B continues to run well over the same period.
    -steve-

  • How can I create responsive motion path animation

    I have an animate file that contains many amimations and one motion path animation.  All animation elements of the animate file are responsive with the exception of the motion path animation which is not for some reason.  Why does the motion path animation not repsonsive even though it is L, T, W, H coordinates are set as percentages?
    Thanks

    Hi, evetsrelfel-
    There's no way to make motion paths responsive at the moment - we limit them to pixel-based animations.
    Thanks,
    -Elaine

Maybe you are looking for

  • How unlock my iphone 4s

    Hi! A week ago I swap with some girl for the phones! But now I have a problem because no one can unlock my iphone :/ A girl who had this phone previously had him on contract in the Orange. And now is the question.. can I unlock my iphone 4s somehow!

  • List of Desirable/Missing Components

    Obviously, the DataGrid is missing, as are any video controls... but what are some of your fave Flex/Flash components that aren't in Flash Catalyst?

  • No more image but the media is supposedly their????

    Hi Well I have and FCP5 project that I opened in FCP6, the clip is on the timeline but it just black. I have tried to reconnect through media manager but it is connected. Has anybody seen this before??? Thanks Guy

  • OTL Retrieval Process

    In OTL, if there is requirement like we need to the state field in the Payroll timecard, and it should be transferred as an Input value (along with hours) once Transfer to BEE is run. Could you please confirm if these are the steps to be followed a)

  • BLOB causing FRM-40505

    Any clue why a blob field in a datablock would cause an FRM-40505 (query failed). This only happens only after I upload to the server (Forms Servlet and 9iDB). Works fine in the developer (client/server mode). HELP!