Skip to main content

Improving JBoss 5 slow startup


We upgraded from JBoss 4 (and JDK 5) to JBoss 5 (and JDK 6). The problem is that the start time has gone from 1.5 minutes (on JBoss 4) to more than 4 minutes.




18:53:35,444 INFO [ServerImpl] JBoss (Microcontainer) [5.1.0.GA (build: SVNTag=JBoss_5_1_0_GA date=200905221053)] Started in 3m:9s:262ms



It seems like the component that is taking JBoss the longest time to initialize is the JMX




18:50:41,926 INFO [LogNotificationListener] Adding notification listener for logging mbean "jboss.system:service=Logging,type=Log4jService" to server org.jboss.mx.server.MBeanServerImpl@1adc122[ defaultDomain='jboss' ]
18:52:38,797 INFO [JMXConnectorServerService] JMX Connector server: service:jmx:rmi://lharel2/jndi/rmi://lharel2:1090/jmxconnector



From the DEBUG server log, I get these lines at the problematic time:




2009-12-18 18:51:00,886 DEBUG [org.jboss.deployment.MappedReferenceMetaDataResolverDeployer] (main) vfsfile:/C:/QC/Views/QCDev/jboss-5.1.0.GA/server/default/deploy/jmx-console.war/ endpoint mappings:
2009-12-18 18:51:00,886 DEBUG [org.jboss.deployment.MappedReferenceMetaDataResolverDeployer] (main) Processing unit=jmx-console.war, structure: jmx-console.war
2009-12-18 18:52:35,209 DEBUG [org.jboss.deployment.OptAnnotationMetaDataDeployer] (main) Deployment is metadata-complete, skipping annotation processing, ejbJarMetaData=null, jbossWebMetaData=org.jboss.metadata.web.spec.Web23MetaData@1f, jbossClientMetaData=null, metaDataCompleteIsDefault=false



There is no EJB in the project.



The memory settings are:




-Xms128m -Xmx512m -XX:MaxPermSize=256m



Do you have any idea how JBoss start time can be improved?



Update: so far no luck, I tried shreeni's suggestion (changed the scanning xmls). The server is not running in debug mode so MicSim's suggestion is not relevant


Source: Tips4all

Comments

  1. A shot into the blue sky without more information


    Network timeouts: 1,5 minutes of delay when deploying jmx-console.war may indicate a network timeout (e.g. 3 x 30 seconds). Try to start JBoss and bind it to a specific IP address using the -b command line argument or the jboss.bind.address system property. Also, try to make sure your host and DNS resolution settings on your system are correct.
    JMX is also using RMI and you may want to set the RMI server host name as system property. On some Linux distributions, RMI has problems with looking up the correct hostname and jmx-console.war may try to connect to the 'wrong localhost'. The system property is java.rmi.server.hostname
    System tracing: If that does not help, you may want to use strace to start the java process, so you can see the point where the system hangs (if it really does hang due to a network timeout or similar).

    ReplyDelete
  2. That is an awfully big gap in the logs. I suggest changing the log configuration to log everything at DEBUG level, rather than INFO. This will generate an awful lot more log entries, but hopefully will help you narrow it down.

    The easiest way to do this is to set the -Djboss.server.log.threshold=DEBUG system property when you start JBoss

    ReplyDelete
  3. Your suspicion about the jmx-console can be misleading. There may be other components doing work in the background unrelated to the jmx-console. In my experience, we had an issue where a small war file appeared to take 3 minutes to load! It was innocent. The culprit was partly due to an EAR file with many wars and jars.

    While I'm no expert, I would suggest the following:
    1) Try turning up the logging to TRACE. By doing this, I witnessed one of the deployers (EJBDeployer, I think) unnecessarily scanning WARs in one of our EARs. I then manually disabled the scanning of those WAR files.
    2) Run wireshark during the startup. I discovered some war files were hanging while waiting for a response from an external DTD request. Those websites were either now non-existent or would not properly serve the DTD files to java-based programs. I could speed it up by either having the programmers use local DTD files or mirroring those DTDs locally and having /etc/host loopback locally.

    ReplyDelete
  4. I had an issue like this but I found a good improvement by setting the initial and max heap size to same values, I mean:

    -Xms512m -Xmx512m


    With this, I improved from 4 to 2 minutes the starting time.

    ReplyDelete
  5. You could refer to this link to avoid unnecessary annotation scanning which could speedup your server start

    ReplyDelete
  6. I suppose you are starting in debug mode. This mode can be up to 3 times slower than normal mode.
    But there might be also a problem when switching from JDK5 to JDK6. I found this solution here on the net:


    I've solved that. It's a debugging issue. I've changed my debug settings from:

    wrapper.java.additional.26=-Xdebug
    wrapper.java.additional.27=-Xnoagent
    wrapper.java.additional.28=-Djava.compiler=NONE
    wrapper.java.additional.27=-Xrunjdwp:transport=dt_socket,address=7199,server=y,suspend=n


    to:

    wrapper.java.additional.26=-Xdebug
    wrapper.java.additional.27=-Xrunjdwp:transport=dt_socket,address=7199,server=y,suspend=n


    and JBoss becomes fast again.


    Hope this helps.

    ReplyDelete
  7. Do you need the JMX console application? Pragmatic thing would be to un-deploy it from the server, you could still use the jconsole or jvisualvm for basically the same thing.

    ReplyDelete
  8. See http://community.jboss.org/wiki/jboss5xtuningslimming especially the "Tuning" part.

    ReplyDelete
  9. I'm using JBoss 5.1.0 with a Macbook pro (2.26ghz 4gb) without applications it start in 54s

    15:00:26,449 INFO [ServerImpl] JBoss (Microcontainer) [5.1.0.GA (build: SVNTag=JBoss_5_1_0_GA date=200905221634)] Started in 54s:720ms


    I made a new configuration based on the "default". The JMS dataosurce points to a Postgres database instead of the "Hypersonic Database" (in memory database)

    ReplyDelete

Post a Comment

Popular posts from this blog

[韓日関係] 首相含む大幅な内閣改造の可能性…早ければ来月10日ごろ=韓国

div not scrolling properly with slimScroll plugin

I am using the slimScroll plugin for jQuery by Piotr Rochala Which is a great plugin for nice scrollbars on most browsers but I am stuck because I am using it for a chat box and whenever the user appends new text to the boxit does scroll using the .scrollTop() method however the plugin's scrollbar doesnt scroll with it and when the user wants to look though the chat history it will start scrolling from near the top. I have made a quick demo of my situation http://jsfiddle.net/DY9CT/2/ Does anyone know how to solve this problem?

Why does this javascript based printing cause Safari to refresh the page?

The page I am working on has a javascript function executed to print parts of the page. For some reason, printing in Safari, causes the window to somehow update. I say somehow, because it does not really refresh as in reload the page, but rather it starts the "rendering" of the page from start, i.e. scroll to top, flash animations start from 0, and so forth. The effect is reproduced by this fiddle: http://jsfiddle.net/fYmnB/ Clicking the print button and finishing or cancelling a print in Safari causes the screen to "go white" for a sec, which in my real website manifests itself as something "like" a reload. While running print button with, let's say, Firefox, just opens and closes the print dialogue without affecting the fiddle page in any way. Is there something with my way of calling the browsers print method that causes this, or how can it be explained - and preferably, avoided? P.S.: On my real site the same occurs with Chrome. In the ex