tag:old.nabble.com,2006:forum-11766Nabble - codehaus - ActiveCluster2007-04-28T19:47:43ZActiveCluster is a framework for writing cluster-based applications, monitoring the available nodes in a cluster and communicating with them. ActiveCluster consists of an API together with a default implementation. codehaus - ActiveCluster home is <a href="http://activecluster.codehaus.org/" target="_top" rel="nofollow">here</a>.tag:old.nabble.com,2006:post-10238305i have a problem, who can help me?2007-04-28T19:47:43Z2007-04-28T19:47:43Zcsnowfox<DIV>i am a green hand on xfire</DIV>
<DIV>i catch a error like this</DIV>
<DIV>i think this is caused by the version,but i don't know how to deal with it.</DIV>
<DIV>help me~~</DIV>
<DIV>thanks</DIV>
<DIV> </DIV>
<DIV>Exception in thread "main" javax.xml.stream.FactoryConfigurationError: Provider null could not be instantiated: java.lang.NullPointerException<BR> at javax.xml.stream.FactoryFinder.newInstance(FactoryFinder.java:75)<BR> at javax.xml.stream.FactoryFinder.find(FactoryFinder.java:136)<BR> at javax.xml.stream.FactoryFinder.find(FactoryFinder.java:92)<BR> at javax.xml.stream.XMLInputFactory.newInstance(XMLInputFactory.java:136)<BR> at at org.codehaus.xfire.util.STAXUtils.<clinit>(STAXUtils.java:48).null(Unknown Source)<BR> at org.codehaus.xfire.transport.http.HttpChannel.writeWithoutAttachments(HttpChannel.java:54)<BR> at org.codehaus.xfire.transport.http.CommonsHttpMessageSender.getByteArrayRequestEntity(CommonsHttpMessageSender.java:417)<BR> at org.codehaus.xfire.transport.http.CommonsHttpMessageSender.send(CommonsHttpMessageSender.java:355)<BR> at org.codehaus.xfire.transport.http.HttpChannel.sendViaClient(HttpChannel.java:123)<BR> at org.codehaus.xfire.transport.http.HttpChannel.send(HttpChannel.java:48)<BR> at org.codehaus.xfire.handler.OutMessageSender.invoke(OutMessageSender.java:26)<BR> at org.codehaus.xfire.handler.HandlerPipeline.invoke(HandlerPipeline.java:131)<BR> at org.codehaus.xfire.client.Invocation.invoke(Invocation.java:75)<BR> at org.codehaus.xfire.client.Client.invoke(Client.java:335)<BR> at org.codehaus.xfire.client.XFireProxy.handleRequest(XFireProxy.java:77)<BR> at org.codehaus.xfire.client.XFireProxy.invoke(XFireProxy.java:57)<BR> at $Proxy0.echo(Unknown Source)<BR> at com.c2.xfire.example.CallService.runSimpleService(CallService.java:49)<BR> at com.c2.xfire.example.CallService.main(CallService.java:24)</DIV>
<DIV> </DIV>
<DIV></DIV><br><!-- footer --><br><hr>
<a style="font-size:14px;line-height:15px; color:#000; text-decoration:none" href="http://event.mail.163.com/chanel/xyq.htm?from=126" target="_blank" rel="nofollow"><span style="text-decoration:underline; color:blue">免费试玩2006中国最佳网络游戏--梦幻西游 </span> </a><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-9709466Cannot get example working2007-03-28T02:15:49Z2007-03-28T02:15:49ZgarethuHi,
<br><br>I have followed the "Getting Started" tutorial.
<br>Step 1 is complete and compiled successfully.
<br>Step 2 does not work. I have all my environment setup correctly.
<br>I have the following setup :
<br>1. My project resides in the directory :
<br>D:\Apps\EclipseWorkspace\Enunciate with the source code in
<br>D:\Apps\EclipseWorkspace\Enunciate\src
<br>2. I go to the directory D:\Apps\EclipseWorkspace\Enunciate and type in
<br>enunciate -Exfire.war ifyouwannabecool.war\
<br>src/com/ifyouwannabecool/api/ExclusiveGroupException.java\
<br>src/com/ifyouwannabecool/api/PermissionDeniedException.java\
<br>src/com/ifyouwannabecool/api/LinkageService.java\
<br>src/com/ifyouwannabecool/api/PersonaService.java\
<br>src/com/ifyouwannabecool/impl/PersonaServiceImpl.java\
<br>src/com/ifyouwannabecool/impl/LinkageServiceImpl.java\
<br>src/com/ifyouwannabecool/domain/persona/Name.java\
<br>src/com/ifyouwannabecool/domain/persona/Persona.java\
<br>src/com/ifyouwannabecool/domain/link/Link.java\
<br>src/com/ifyouwannabecool/domain/link/SocialGroup.java
<br><br>I receive the following message from the console :
<br><br><br>D:\Apps\EclipseWorkspace\Enunciate>enunciate -Exfire.war
<br>ifyouwannabecool.war\ s
<br>rc/com/ifyouwannabecool/api/ExclusiveGroupException.java\
<br>src/com/ifyouwannabeco
<br>ol/api/PermissionDeniedException.java\
<br>src/com/ifyouwannabecool/api/LinkageServi
<br>ce.java\ src/com/ifyouwannabecool/api/PersonaService.java\
<br>src/com/ifyouwannabec
<br>ool/impl/PersonaServiceImpl.java\
<br>src/com/ifyouwannabecool/impl/LinkageServiceIm
<br>pl.java\ src/com/ifyouwannabecool/domain/persona/Name.java\
<br>src/com/ifyouwannabe
<br>cool/domain/persona/Persona.java\
<br>src/com/ifyouwannabecool/domain/link/Link.java
<br>\ src/com/ifyouwannabecool/domain/link/SocialGroup.java
<br>Usage: java [-options] class [args...]
<br> (to execute a class)
<br> or java [-options] -jar jarfile [args...]
<br> (to execute a jar file)
<br><br>where options include:
<br> -client to select the "client" VM
<br> -server to select the "server" VM
<br> -hotspot is a synonym for the "client" VM [deprecated]
<br> The default VM is client.
<br><br> -cp <class search path of directories and zip/jar files>
<br> -classpath <class search path of directories and zip/jar files>
<br> A ; separated list of directories, JAR archives,
<br> and ZIP archives to search for class files.
<br> -D<name>=<value>
<br> set a system property
<br> -verbose[:class|gc|jni]
<br> enable verbose output
<br> -version print product version and exit
<br> -version:<value>
<br> require the specified version to run
<br> -showversion print product version and continue
<br> -jre-restrict-search | -jre-no-restrict-search
<br> include/exclude user private JREs in the version
<br>search
<br> -? -help print this help message
<br> -X print help on non-standard options
<br> -ea[:<packagename>...|:<classname>]
<br> -enableassertions[:<packagename>...|:<classname>]
<br> enable assertions
<br> -da[:<packagename>...|:<classname>]
<br> -disableassertions[:<packagename>...|:<classname>]
<br> disable assertions
<br> -esa | -enablesystemassertions
<br> enable system assertions
<br> -dsa | -disablesystemassertions
<br> disable system assertions
<br> -agentlib:<libname>[=<options>]
<br> load native agent library <libname>, e.g.
<br>-agentlib:hprof
<br> see also, -agentlib:jdwp=help and
<br>-agentlib:hprof=help
<br> -agentpath:<pathname>[=<options>]
<br> load native agent library by full pathname
<br> -javaagent:<jarpath>[=<options>]
<br> load Java programming language agent, see
<br>java.lang.instrument
<br><br>I am using Windows XP (unfortunately).
<br><br>Will appreciate any help.
<br><br>Gareth Uren
<br>Software Developer
<br>Hyphen Technology (Pty) Ltd
<br>Direct Line : 011 303 0333
<br>Cell No : 082 379 0154
<br>Fax No : 0860 634 8185
<br><br>The information contained in this email is confidential and may contain proprietary information. It is meant solely for the intended recipient. Access to this email by anyone else is unauthorised. If you are not the intended recipient, any disclosure, copying, distribution or any action taken or omitted in reliance on this, is prohibited and may be unlawful. No liability or responsibility is accepted if information or data is, for whatever reason corrupted or does not reach its intended recipient. No warranty is given that this email is free of viruses. The views expressed in this email are, unless otherwise stated, those of the author and not those of HYPHEN Technology (Pty) Ltd or its management. HYPHEN Technology (Pty) Ltd reserves the right to monitor, intercept and block emails addressed to its users or take any other action in accordance with its email use policy.
<br><br>---------------------------------------------------------------------
<br>To unsubscribe from this list please visit:
<br><br> <a href="http://xircles.codehaus.org/manage_email" target="_top" rel="nofollow">http://xircles.codehaus.org/manage_email</a><br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-3193380Re: ActiveCluster move to ActiveMQ2006-03-01T14:48:44Z2006-03-01T14:48:44ZrajdaviesHi Ben,
<br><br>that's great! I didn't see the attachments - can you send them again ?
<br><br>cheers,
<br><br>Rob
<br><br>On 1 Mar 2006, at 21:59, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3193380&i=0" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<br><div class='shrinkable-quote'><br>> I've modified the DestinationMarshaller,
<br>> DefaultDestinationMarshaller, and
<br>> DefaultClusterFactory to use "Session.createTopic(String
<br>> destinationName)"
<br>> instead of "new ActiveMQTopic(String destinationName)". This will
<br>> work
<br>> with ActiveMQ, SonicMQ, Tibco, and probably most other vendors.
<br>>
<br>> Since DestinationMarshaller.getDestination(String name) is called
<br>> on every
<br>> heartbeat message, and most JMS implementations (ActiveMQ excluded)
<br>> perform
<br>> a remote call to create a new Destination, I've also added a lookup
<br>> cache
<br>> of name->destination mappings in DefaultDestinationMarshaller, and
<br>> changed
<br>> the interface definition to throw a JMSException.
<br>>
<br>> Since Weblogic (and perhaps others) may use JNDI lookups or other
<br>> mechanism
<br>> to do name->Destination mappings, I've modified the
<br>> DefaultClusterFactory,
<br>> making the protected createCluster method that takes a Connection and
<br>> Session object to be public. This way you can create the
<br>> Connection and
<br>> Session objects prior to creating your DestinationMarshaller object.
<br>>
<br>>
<br>> (See attached file: DestinationMarshaller.java)(See attached file:
<br>> DefaultClusterFactory.java)(See attached file:
<br>> DefaultDestinationMarshaller.java)
<br>>
<br>> I had to modify the unit tests slightly to throw the new
<br>> JMSException, and
<br>> to provide a Session object to the DefaultDestinationMarshaller
<br>> constructor. Once modified, all the unit tests ran successfully.
<br>>
<br>> Does this sound like a workable solution? Should I give up on a
<br>> cross-vendor implementation, and write my own ClusterFactory and
<br>> Marshaller
<br>> implementations? (in hindsight, it would have been simpler ;-)
<br>>
<br>> Thanks,
<br>> Ben
<br>>
<br>>
<br>>
<br>>
<br>> |---------+---------------------------->
<br>> | | James Strachan |
<br>> | | <james.strachan@g|
<br>> | | mail.com> |
<br>> | | |
<br>> | | 03/01/2006 01:55 |
<br>> | | PM |
<br>> | | Please respond to|
<br>> | | user |
<br>> | | |
<br>> |---------+---------------------------->
<br>>> ---------------------------------------------------------------------
<br>>> ---------------------------------------------------------|
<br>>
<br>> |
<br>> |
<br>> | To:
<br>> <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3193380&i=1" target="_top" rel="nofollow">user@...</a>
<br>> |
<br>> | cc: activemq-
<br>> <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3193380&i=2" target="_top" rel="nofollow">dev@...</a>
<br>> |
<br>> | Subject: Re: [activecluster-user] ActiveCluster move to
<br>> ActiveMQ |
<br>>> ---------------------------------------------------------------------
<br>>> ---------------------------------------------------------|
<br>>
<br>>
<br>>
<br>>
<br>> On 1 Mar 2006, at 18:16, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3193380&i=3" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<br>>> Unfortunately, activecluster-4.0-M4 is dependent on activemq-core,
<br>>> so it
<br>>> can't be used easily with another JMS provider.
<br>>
<br>> Depending on classes inside activemq-core.jar != not working with
<br>> another JMS provider :)
<br>>
<br>>
<br>>> The DefaultClusterFactory class depends on
<br>>> org.apache.activemq.util.IdGenerator. That seems like a simple
<br>>> dependency
<br>>> to remove (maybe use Jakarta Commons Id package?).
<br>>
<br>> Maybe we should just copy that one class across.
<br>>
<br>>
<br>>> Not so obvious is an issue where the DestinationMarshaller
<br>>> interface is
<br>>> required to create a Destination from a String. Creating a
<br>>> destination for
<br>>> most JMS vendors requires access to the Session object (e.g.
<br>>> Session.createTopic(String name)), which is encapsulated in the
<br>>> DefaultClusterFactory class. How would I create a destination in
<br>>> my own
<br>>> marshaller implementation, with no access to the Session object? The
<br>>> DefaultDestinationMarshaller implementation creates a new
<br>>> ActiveMQTopic
<br>>> object, which obviously only works in ActiveMQ....
<br>>
<br>> We could just use a dummy Destination object for now that gets
<br>> serialised?
<br>>
<br>>
<br>>> So, if I tweak the DefaultClusterFactory class, changing the
<br>>> protected
<br>>> createCluster method to public, then I can easily write my own
<br>>> DestinationMarshaller, and the code is truly independent of ActiveMQ.
<br>>>
<br>>> Is anyone else interested in using ActiveCluster with a JMS
<br>>> provider other
<br>>> than ActiveMQ?
<br>>
<br>> Sure - supporting any JMS is a good thing.
<br>>
<br>>
<br>>> Any plans to make ActiveCluster a top-level Geronimo
<br>>> project, and not just a subcomponent of ActiveMQ?
<br>>
<br>> We'll see what happens with Geronimo / WADI / ActiveMQ. It might make
<br>> sense as a module in Geronimo or in ActiveMQ. Am not sure if its big
<br>> enough to be a top level project
<br>>
<br>> James
<br>> -------
<br>> <a href="http://radio.weblogs.com/0112098/" target="_top" rel="nofollow">http://radio.weblogs.com/0112098/</a><br>>
<br>>
<br>>
<br>>
<br>> This communication is for informational purposes only. It is not
<br>> intended
<br>> as an offer or solicitation for the purchase or sale of any financial
<br>> instrument or as an official confirmation of any transaction. All
<br>> market prices,
<br>> data and other information are not warranted as to completeness or
<br>> accuracy and
<br>> are subject to change without notice. Any comments or statements
<br>> made herein
<br>> do not necessarily reflect those of JPMorgan Chase & Co., its
<br>> subsidiaries
<br>> and affiliates.
</div><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-3192506Re: ActiveCluster move to ActiveMQ2006-03-01T13:59:20Z2006-03-01T13:59:20Zben.kiblerI've modified the DestinationMarshaller, DefaultDestinationMarshaller, and
<br>DefaultClusterFactory to use "Session.createTopic(String destinationName)"
<br>instead of "new ActiveMQTopic(String destinationName)". This will work
<br>with ActiveMQ, SonicMQ, Tibco, and probably most other vendors.
<br><br>Since DestinationMarshaller.getDestination(String name) is called on every
<br>heartbeat message, and most JMS implementations (ActiveMQ excluded) perform
<br>a remote call to create a new Destination, I've also added a lookup cache
<br>of name->destination mappings in DefaultDestinationMarshaller, and changed
<br>the interface definition to throw a JMSException.
<br><br>Since Weblogic (and perhaps others) may use JNDI lookups or other mechanism
<br>to do name->Destination mappings, I've modified the DefaultClusterFactory,
<br>making the protected createCluster method that takes a Connection and
<br>Session object to be public. This way you can create the Connection and
<br>Session objects prior to creating your DestinationMarshaller object.
<br><br><br>(See attached file: DestinationMarshaller.java)(See attached file:
<br>DefaultClusterFactory.java)(See attached file:
<br>DefaultDestinationMarshaller.java)
<br><br>I had to modify the unit tests slightly to throw the new JMSException, and
<br>to provide a Session object to the DefaultDestinationMarshaller
<br>constructor. Once modified, all the unit tests ran successfully.
<br><br>Does this sound like a workable solution? Should I give up on a
<br>cross-vendor implementation, and write my own ClusterFactory and Marshaller
<br>implementations? (in hindsight, it would have been simpler ;-)
<br><br>Thanks,
<br>Ben
<br><br><br><br><br>|---------+---------------------------->
<br>| | James Strachan |
<br>| | <james.strachan@g|
<br>| | mail.com> |
<br>| | |
<br>| | 03/01/2006 01:55 |
<br>| | PM |
<br>| | Please respond to|
<br>| | user |
<br>| | |
<br>|---------+---------------------------->
<br> >------------------------------------------------------------------------------------------------------------------------------|
<br> | |
<br> | To: <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3192506&i=0" target="_top" rel="nofollow">user@...</a> |
<br> | cc: <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3192506&i=1" target="_top" rel="nofollow">activemq-dev@...</a> |
<br> | Subject: Re: [activecluster-user] ActiveCluster move to ActiveMQ |
<br> >------------------------------------------------------------------------------------------------------------------------------|
<br><br><br><br><br>On 1 Mar 2006, at 18:16, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3192506&i=2" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<br>> Unfortunately, activecluster-4.0-M4 is dependent on activemq-core,
<br>> so it
<br>> can't be used easily with another JMS provider.
<br><br>Depending on classes inside activemq-core.jar != not working with
<br>another JMS provider :)
<br><br><br>> The DefaultClusterFactory class depends on
<br>> org.apache.activemq.util.IdGenerator. That seems like a simple
<br>> dependency
<br>> to remove (maybe use Jakarta Commons Id package?).
<br><br>Maybe we should just copy that one class across.
<br><br><div class='shrinkable-quote'><div class='shrinkable-quote'><br>> Not so obvious is an issue where the DestinationMarshaller
<br>> interface is
<br>> required to create a Destination from a String. Creating a
<br>> destination for
<br>> most JMS vendors requires access to the Session object (e.g.
<br>> Session.createTopic(String name)), which is encapsulated in the
<br>> DefaultClusterFactory class. How would I create a destination in
<br>> my own
<br>> marshaller implementation, with no access to the Session object? The
<br>> DefaultDestinationMarshaller implementation creates a new
<br>> ActiveMQTopic
<br>> object, which obviously only works in ActiveMQ....
</div></div>We could just use a dummy Destination object for now that gets
<br>serialised?
<br><br><br>> So, if I tweak the DefaultClusterFactory class, changing the protected
<br>> createCluster method to public, then I can easily write my own
<br>> DestinationMarshaller, and the code is truly independent of ActiveMQ.
<br>>
<br>> Is anyone else interested in using ActiveCluster with a JMS
<br>> provider other
<br>> than ActiveMQ?
<br><br>Sure - supporting any JMS is a good thing.
<br><br><br>> Any plans to make ActiveCluster a top-level Geronimo
<br>> project, and not just a subcomponent of ActiveMQ?
<br><br>We'll see what happens with Geronimo / WADI / ActiveMQ. It might make
<br>sense as a module in Geronimo or in ActiveMQ. Am not sure if its big
<br>enough to be a top level project
<br><br>James
<br>-------
<br><a href="http://radio.weblogs.com/0112098/" target="_top" rel="nofollow">http://radio.weblogs.com/0112098/</a><br><br><br><br><br>This communication is for informational purposes only. It is not intended
<br>as an offer or solicitation for the purchase or sale of any financial
<br>instrument or as an official confirmation of any transaction. All market prices,
<br>data and other information are not warranted as to completeness or accuracy and
<br>are subject to change without notice. Any comments or statements made herein
<br>do not necessarily reflect those of JPMorgan Chase & Co., its subsidiaries
<br>and affiliates.<p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-3189598Re: ActiveCluster move to ActiveMQ2006-03-01T11:25:21Z2006-03-01T11:25:21Zben.kiblerHi James,
<br><br>Thanks for your feedback. I'll submit a patch for the default
<br>implementation which will work for both ActiveMQ and other JMS
<br>implementations.
<br><br>> Anyone know how to make a dependency marked
<br>> as for testing only?
<br><br>Maven 2.0 provides the <scope/> tag, which can be placed inside a
<br><dependency> element, to indicate the type of dependency.
<br><br><scope>test</scope> is used for dependencies only needed for unit tests,
<br>like junit, or in this case ActiveMQ.
<br><br><scope>provided</scope> is normally used when writing to a Java spec, in
<br>this case JMS. So, you can include ActiveMQ or j2ee-jms-spec as a
<br>dependency, but when using Maven to construct your build package, it won't
<br>include the ActiveMQ libraries, and will expect the implementation to
<br>provide a suitable jar file (i.e. Weblogic or some other provider).
<br><br>For more details see
<br><a href="http://maven.apache.org/guides/introduction/introduction-to-dependency-mechanism.html" target="_top" rel="nofollow">http://maven.apache.org/guides/introduction/introduction-to-dependency-mechanism.html</a><br><br>Regards,
<br>Ben
<br><br><br><br>|---------+---------------------------->
<br>| | James Strachan |
<br>| | <james.strachan@g|
<br>| | mail.com> |
<br>| | |
<br>| | 03/01/2006 01:51 |
<br>| | PM |
<br>| | Please respond to|
<br>| | user |
<br>| | |
<br>|---------+---------------------------->
<br> >------------------------------------------------------------------------------------------------------------------------------|
<br> | |
<br> | To: <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3189598&i=0" target="_top" rel="nofollow">user@...</a> |
<br> | cc: |
<br> | Subject: Re: [activecluster-user] ActiveCluster move to ActiveMQ |
<br> >------------------------------------------------------------------------------------------------------------------------------|
<br><br><br><br><br>On 28 Feb 2006, at 20:10, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3189598&i=1" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<div class='shrinkable-quote'><br>> It appears that new development in ActiveCluster has moved to the
<br>> ActiveMQ
<br>> project, and there are no new standalone releases of ActiveCluster
<br>> since
<br>> rel 1.0.
<br>>
<br>> Looking at the ActiveCluster build file in Apache SVN
<br>> (<a href="https://svn.apache.org/repos/asf/incubator/activemq/trunk/" target="_top" rel="nofollow">https://svn.apache.org/repos/asf/incubator/activemq/trunk/</a><br>> activecluster/pom.xml)
<br>> it appears that ActiveCluster is completely dependent on ActiveMQ
<br>> as well.
<br>>
<br>> I've been using ActiveCluster for some time now with other JMS
<br>> implementations, and would like to continue doing so.
</div><br>Great, glad to hear it.
<br><br>Its purely a gremlin in the M2 POM - there's no hard and fast
<br>dependency, ActiveMQ is purely used for testing. If I knew how to do
<br>it, I'd fix the pom.xml. Anyone know how to make a dependency marked
<br>as for testing only?
<br><br>FWIW we're only using Maven 1 on ActiveMQ/ActiveCluster/ActiveIO
<br>right now - we backed off using m2
<br><br>> Is there intent to maintain
<br>> ActiveCluster as a standalone cluster API, not dependent on ActiveMQ?
<br><br>Yes. Its just given that ActiveMQ/WADI/Geronimo uses it, we felt it
<br>was easier to move the code to Apache with ActiveMQ
<br><br>James
<br>-------
<br><a href="http://radio.weblogs.com/0112098/" target="_top" rel="nofollow">http://radio.weblogs.com/0112098/</a><br><br><br><br><br><br><br>This communication is for informational purposes only. It is not intended
<br>as an offer or solicitation for the purchase or sale of any financial
<br>instrument or as an official confirmation of any transaction. All market prices,
<br>data and other information are not warranted as to completeness or accuracy and
<br>are subject to change without notice. Any comments or statements made herein
<br>do not necessarily reflect those of JPMorgan Chase & Co., its subsidiaries
<br>and affiliates.
<br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-3188984Re: ActiveCluster move to ActiveMQ2006-03-01T10:55:18Z2006-03-01T10:55:18ZJames.StrachanOn 1 Mar 2006, at 18:16, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3188984&i=0" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<br>> Unfortunately, activecluster-4.0-M4 is dependent on activemq-core,
<br>> so it
<br>> can't be used easily with another JMS provider.
<br><br>Depending on classes inside activemq-core.jar != not working with
<br>another JMS provider :)
<br><br><br>> The DefaultClusterFactory class depends on
<br>> org.apache.activemq.util.IdGenerator. That seems like a simple
<br>> dependency
<br>> to remove (maybe use Jakarta Commons Id package?).
<br><br>Maybe we should just copy that one class across.
<br><br><div class='shrinkable-quote'><br>> Not so obvious is an issue where the DestinationMarshaller
<br>> interface is
<br>> required to create a Destination from a String. Creating a
<br>> destination for
<br>> most JMS vendors requires access to the Session object (e.g.
<br>> Session.createTopic(String name)), which is encapsulated in the
<br>> DefaultClusterFactory class. How would I create a destination in
<br>> my own
<br>> marshaller implementation, with no access to the Session object? The
<br>> DefaultDestinationMarshaller implementation creates a new
<br>> ActiveMQTopic
<br>> object, which obviously only works in ActiveMQ....
</div><br>We could just use a dummy Destination object for now that gets
<br>serialised?
<br><br><br>> So, if I tweak the DefaultClusterFactory class, changing the protected
<br>> createCluster method to public, then I can easily write my own
<br>> DestinationMarshaller, and the code is truly independent of ActiveMQ.
<br>>
<br>> Is anyone else interested in using ActiveCluster with a JMS
<br>> provider other
<br>> than ActiveMQ?
<br><br>Sure - supporting any JMS is a good thing.
<br><br><br>> Any plans to make ActiveCluster a top-level Geronimo
<br>> project, and not just a subcomponent of ActiveMQ?
<br><br>We'll see what happens with Geronimo / WADI / ActiveMQ. It might make
<br>sense as a module in Geronimo or in ActiveMQ. Am not sure if its big
<br>enough to be a top level project
<br><br>James
<br>-------
<br><a href="http://radio.weblogs.com/0112098/" target="_top" rel="nofollow">http://radio.weblogs.com/0112098/</a><br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-3188926Re: ActiveCluster move to ActiveMQ2006-03-01T10:51:45Z2006-03-01T10:51:45ZJames.StrachanOn 28 Feb 2006, at 20:10, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3188926&i=0" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<div class='shrinkable-quote'><br>> It appears that new development in ActiveCluster has moved to the
<br>> ActiveMQ
<br>> project, and there are no new standalone releases of ActiveCluster
<br>> since
<br>> rel 1.0.
<br>>
<br>> Looking at the ActiveCluster build file in Apache SVN
<br>> (<a href="https://svn.apache.org/repos/asf/incubator/activemq/trunk/" target="_top" rel="nofollow">https://svn.apache.org/repos/asf/incubator/activemq/trunk/</a>
<br>> activecluster/pom.xml)
<br>> it appears that ActiveCluster is completely dependent on ActiveMQ
<br>> as well.
<br>>
<br>> I've been using ActiveCluster for some time now with other JMS
<br>> implementations, and would like to continue doing so.
</div><br>Great, glad to hear it.
<br><br>Its purely a gremlin in the M2 POM - there's no hard and fast
<br>dependency, ActiveMQ is purely used for testing. If I knew how to do
<br>it, I'd fix the pom.xml. Anyone know how to make a dependency marked
<br>as for testing only?
<br><br>FWIW we're only using Maven 1 on ActiveMQ/ActiveCluster/ActiveIO
<br>right now - we backed off using m2
<br><br>> Is there intent to maintain
<br>> ActiveCluster as a standalone cluster API, not dependent on ActiveMQ?
<br><br>Yes. Its just given that ActiveMQ/WADI/Geronimo uses it, we felt it
<br>was easier to move the code to Apache with ActiveMQ
<br><br>James
<br>-------
<br><a href="http://radio.weblogs.com/0112098/" target="_top" rel="nofollow">http://radio.weblogs.com/0112098/</a><br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-3188571Re: ActiveCluster move to ActiveMQ2006-03-01T10:16:10Z2006-03-01T10:16:10Zben.kiblerUnfortunately, activecluster-4.0-M4 is dependent on activemq-core, so it
<br>can't be used easily with another JMS provider.
<br><br>The DefaultClusterFactory class depends on
<br>org.apache.activemq.util.IdGenerator. That seems like a simple dependency
<br>to remove (maybe use Jakarta Commons Id package?).
<br><br>Not so obvious is an issue where the DestinationMarshaller interface is
<br>required to create a Destination from a String. Creating a destination for
<br>most JMS vendors requires access to the Session object (e.g.
<br>Session.createTopic(String name)), which is encapsulated in the
<br>DefaultClusterFactory class. How would I create a destination in my own
<br>marshaller implementation, with no access to the Session object? The
<br>DefaultDestinationMarshaller implementation creates a new ActiveMQTopic
<br>object, which obviously only works in ActiveMQ....
<br><br>So, if I tweak the DefaultClusterFactory class, changing the protected
<br>createCluster method to public, then I can easily write my own
<br>DestinationMarshaller, and the code is truly independent of ActiveMQ.
<br><br>Is anyone else interested in using ActiveCluster with a JMS provider other
<br>than ActiveMQ? Any plans to make ActiveCluster a top-level Geronimo
<br>project, and not just a subcomponent of ActiveMQ?
<br><br><br><br>IB Technology GCCG - Tel +1-212-622-4653
<br>|---------+---------------------------->
<br>| | Ben X Kibler |
<br>| | |
<br>| | 02/28/2006 06:15 |
<br>| | PM |
<br>| | |
<br>|---------+---------------------------->
<br> >-----------------------------------------------------------------------------------------------------------------------------|
<br> | |
<br> | To: <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3188571&i=0" target="_top" rel="nofollow">user@...</a> |
<br> | cc: |
<br> | Subject: Re: [activecluster-user] ActiveCluster move to ActiveMQ(Document link: Ben X Kibler) |
<br> >-----------------------------------------------------------------------------------------------------------------------------|
<br><br><br><br>Hi Rob,
<br><br>Thanks for the feedback. I found an activecluster-4.0-M4 jar file in the
<br>latest ActiveMQ distribution, containing the DestinationMarshaller
<br>interface and default implementation.
<br><br>It appears it should work without the ActiveMQ jar files. I'll start
<br>testing tomorrow and let you know how it goes.
<br><br>Thanks,
<br>Ben
<br><br><br><br><br>|---------+---------------------------->
<br>| | Rob Davies |
<br>| | <rajdavies@gmail.|
<br>| | com> |
<br>| | |
<br>| | 02/28/2006 05:39 |
<br>| | PM |
<br>| | Please respond to|
<br>| | user |
<br>| | |
<br>|---------+---------------------------->
<br> >------------------------------------------------------------------------------------------------------------------------------|
<br> | |
<br> | To: <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3188571&i=1" target="_top" rel="nofollow">user@...</a> |
<br> | cc: |
<br> | Subject: Re: [activecluster-user] ActiveCluster move to ActiveMQ |
<br> >------------------------------------------------------------------------------------------------------------------------------|
<br><br><br><br><br>Hi Ben,
<br><br>long time no talk!
<br>We had to move the api back to use javax.jms.Destination to support
<br>WADI for Apache Geronimo 1.0. If I remember correctly, you've been
<br>using SonicMQ that did not support serialization of Destinations.
<br>We've tried to cater for that by adding the DestinationMarshaller
<br>interface - that can be an optional argument for creating a cluster.
<br>The default is to use an ActiveMQDestination, but it should be
<br>trivial to pass in an implementation that uses Sonic.
<br><br>Please let us know if this is not the case.
<br><br>cheers,
<br><br>Rob
<br><br>On 28 Feb 2006, at 20:10, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3188571&i=2" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<br><div class='shrinkable-quote'><br>> It appears that new development in ActiveCluster has moved to the
<br>> ActiveMQ
<br>> project, and there are no new standalone releases of ActiveCluster
<br>> since
<br>> rel 1.0.
<br>>
<br>> Looking at the ActiveCluster build file in Apache SVN
<br>> (<a href="https://svn.apache.org/repos/asf/incubator/activemq/trunk/" target="_top" rel="nofollow">https://svn.apache.org/repos/asf/incubator/activemq/trunk/</a><br>> activecluster/pom.xml)
<br>> it appears that ActiveCluster is completely dependent on ActiveMQ
<br>> as well.
<br>>
<br>> I've been using ActiveCluster for some time now with other JMS
<br>> implementations, and would like to continue doing so.
<br>>
<br>> Is there a version 1.1 or greater distribution of ActiveCluster
<br>> (including
<br>> Destination serialisation fix, <a href="http://jira.activemq.org/jira/browse/" target="_top" rel="nofollow">http://jira.activemq.org/jira/browse/</a><br>> ACL-6)
<br>> available as a standalone JAR file somewhere? Is there intent to
<br>> maintain
<br>> ActiveCluster as a standalone cluster API, not dependent on ActiveMQ?
<br>>
<br>> Thanks,
<br>> Ben
<br>>
<br>>
<br>>
<br>> This communication is for informational purposes only. It is not
<br>> intended
<br>> as an offer or solicitation for the purchase or sale of any financial
<br>> instrument or as an official confirmation of any transaction. All
<br>> market prices,
<br>> data and other information are not warranted as to completeness or
<br>> accuracy and
<br>> are subject to change without notice. Any comments or statements
<br>> made herein
<br>> do not necessarily reflect those of JPMorgan Chase & Co., its
<br>> subsidiaries
<br>> and affiliates.
</div><br><br><br><br><br><br><br><br>This communication is for informational purposes only. It is not intended
<br>as an offer or solicitation for the purchase or sale of any financial
<br>instrument or as an official confirmation of any transaction. All market prices,
<br>data and other information are not warranted as to completeness or accuracy and
<br>are subject to change without notice. Any comments or statements made herein
<br>do not necessarily reflect those of JPMorgan Chase & Co., its subsidiaries
<br>and affiliates.
<br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-3174050Re: ActiveCluster move to ActiveMQ2006-02-28T15:15:58Z2006-02-28T15:15:58Zben.kiblerHi Rob,
<br><br>Thanks for the feedback. I found an activecluster-4.0-M4 jar file in the
<br>latest ActiveMQ distribution, containing the DestinationMarshaller
<br>interface and default implementation.
<br><br>It appears it should work without the ActiveMQ jar files. I'll start
<br>testing tomorrow and let you know how it goes.
<br><br>Thanks,
<br>Ben
<br><br><br><br><br>|---------+---------------------------->
<br>| | Rob Davies |
<br>| | <rajdavies@gmail.|
<br>| | com> |
<br>| | |
<br>| | 02/28/2006 05:39 |
<br>| | PM |
<br>| | Please respond to|
<br>| | user |
<br>| | |
<br>|---------+---------------------------->
<br> >------------------------------------------------------------------------------------------------------------------------------|
<br> | |
<br> | To: <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3174050&i=0" target="_top" rel="nofollow">user@...</a> |
<br> | cc: |
<br> | Subject: Re: [activecluster-user] ActiveCluster move to ActiveMQ |
<br> >------------------------------------------------------------------------------------------------------------------------------|
<br><br><br><br><br>Hi Ben,
<br><br>long time no talk!
<br>We had to move the api back to use javax.jms.Destination to support
<br>WADI for Apache Geronimo 1.0. If I remember correctly, you've been
<br>using SonicMQ that did not support serialization of Destinations.
<br>We've tried to cater for that by adding the DestinationMarshaller
<br>interface - that can be an optional argument for creating a cluster.
<br>The default is to use an ActiveMQDestination, but it should be
<br>trivial to pass in an implementation that uses Sonic.
<br><br>Please let us know if this is not the case.
<br><br>cheers,
<br><br>Rob
<br><br>On 28 Feb 2006, at 20:10, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3174050&i=1" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<br><div class='shrinkable-quote'><br>> It appears that new development in ActiveCluster has moved to the
<br>> ActiveMQ
<br>> project, and there are no new standalone releases of ActiveCluster
<br>> since
<br>> rel 1.0.
<br>>
<br>> Looking at the ActiveCluster build file in Apache SVN
<br>> (<a href="https://svn.apache.org/repos/asf/incubator/activemq/trunk/" target="_top" rel="nofollow">https://svn.apache.org/repos/asf/incubator/activemq/trunk/</a><br>> activecluster/pom.xml)
<br>> it appears that ActiveCluster is completely dependent on ActiveMQ
<br>> as well.
<br>>
<br>> I've been using ActiveCluster for some time now with other JMS
<br>> implementations, and would like to continue doing so.
<br>>
<br>> Is there a version 1.1 or greater distribution of ActiveCluster
<br>> (including
<br>> Destination serialisation fix, <a href="http://jira.activemq.org/jira/browse/" target="_top" rel="nofollow">http://jira.activemq.org/jira/browse/</a><br>> ACL-6)
<br>> available as a standalone JAR file somewhere? Is there intent to
<br>> maintain
<br>> ActiveCluster as a standalone cluster API, not dependent on ActiveMQ?
<br>>
<br>> Thanks,
<br>> Ben
<br>>
<br>>
<br>>
<br>> This communication is for informational purposes only. It is not
<br>> intended
<br>> as an offer or solicitation for the purchase or sale of any financial
<br>> instrument or as an official confirmation of any transaction. All
<br>> market prices,
<br>> data and other information are not warranted as to completeness or
<br>> accuracy and
<br>> are subject to change without notice. Any comments or statements
<br>> made herein
<br>> do not necessarily reflect those of JPMorgan Chase & Co., its
<br>> subsidiaries
<br>> and affiliates.
</div><br><br><br><br><br><br>This communication is for informational purposes only. It is not intended
<br>as an offer or solicitation for the purchase or sale of any financial
<br>instrument or as an official confirmation of any transaction. All market prices,
<br>data and other information are not warranted as to completeness or accuracy and
<br>are subject to change without notice. Any comments or statements made herein
<br>do not necessarily reflect those of JPMorgan Chase & Co., its subsidiaries
<br>and affiliates.
<br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-3173604Re: ActiveCluster move to ActiveMQ2006-02-28T14:50:50Z2006-02-28T14:50:50ZJonathan DutyGreat. Glad to know it is still going on. You may want to put a
<br>note on the activecluster website. The talk around town is that the
<br>project is dead (since no activity on the website in over a year).
<br><br>~Jonathan
<br><br>On Feb 28, 2006, at 5:41 PM, Rob Davies wrote:
<br><div class='shrinkable-quote'><br>> ActiveCluster is currently using <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3173604&i=0" target="_top" rel="nofollow">activemq-dev@...</a>
<br>>
<br>> So it's part of the ActiveMQ project in the Apache incubator at the
<br>> moment.
<br>>
<br>> cheers,
<br>>
<br>> Rob
<br>>
<br>> On 28 Feb 2006, at 22:31, Jonathan Duty wrote:
<br>>
<br>>> Ahh, cool. I wasn't complaining. Just trying to help Ben get his
<br>>> question answered. I'm very happy with ActiveCluster and I'm glad
<br>>> to know it is still alive and well.
<br>>>
<br>>> Just curious, what mailing list? I don't see an activecluster
<br>>> project in the incubator.
<br>>>
<br>>> ~Jonathan
<br>>>
<br>>> On Feb 28, 2006, at 5:26 PM, Rob Davies wrote:
<br>>>
<br>>>> Hi Jonathan,
<br>>>>
<br>>>> apologies if we've missed you're questions - but as we've moved
<br>>>> to apache - we've been concentrating on the apache mailing lists
<br>>>> - as we really want the community to grow there.
<br>>>>
<br>>>> cheers,
<br>>>>
<br>>>> Rob
<br>>>>
<br>>>> On 28 Feb 2006, at 20:27, Jonathan Duty wrote:
<br>>>>
<br>>>>> Ben,
<br>>>>> I have been listening of this list for a few months and never
<br>>>>> given a response to any of my questions, and I have seen 0
<br>>>>> activity. I finally had to go post to the activemq list. I'm
<br>>>>> guessing if you want a response from the core developers you
<br>>>>> will need to do that also.
<br>>>>>
<br>>>>> ~Jonathan
<br>>>>>
<br>>>>> On Feb 28, 2006, at 3:10 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3173604&i=1" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<br>>>>>
<br>>>>>> It appears that new development in ActiveCluster has moved to
<br>>>>>> the ActiveMQ
<br>>>>>> project, and there are no new standalone releases of
<br>>>>>> ActiveCluster since
<br>>>>>> rel 1.0.
<br>>>>>>
<br>>>>>> Looking at the ActiveCluster build file in Apache SVN
<br>>>>>> (<a href="https://svn.apache.org/repos/asf/incubator/activemq/trunk/" target="_top" rel="nofollow">https://svn.apache.org/repos/asf/incubator/activemq/trunk/</a>
<br>>>>>> activecluster/pom.xml)
<br>>>>>> it appears that ActiveCluster is completely dependent on
<br>>>>>> ActiveMQ as well.
<br>>>>>>
<br>>>>>> I've been using ActiveCluster for some time now with other JMS
<br>>>>>> implementations, and would like to continue doing so.
<br>>>>>>
<br>>>>>> Is there a version 1.1 or greater distribution of ActiveCluster
<br>>>>>> (including
<br>>>>>> Destination serialisation fix, <a href="http://jira.activemq.org/jira/" target="_top" rel="nofollow">http://jira.activemq.org/jira/</a>
<br>>>>>> browse/ACL-6)
<br>>>>>> available as a standalone JAR file somewhere? Is there intent
<br>>>>>> to maintain
<br>>>>>> ActiveCluster as a standalone cluster API, not dependent on
<br>>>>>> ActiveMQ?
<br>>>>>>
<br>>>>>> Thanks,
<br>>>>>> Ben
<br>>>>>>
<br>>>>>>
<br>>>>>>
<br>>>>>> This communication is for informational purposes only. It is
<br>>>>>> not intended
<br>>>>>> as an offer or solicitation for the purchase or sale of any
<br>>>>>> financial
<br>>>>>> instrument or as an official confirmation of any transaction.
<br>>>>>> All market prices,
<br>>>>>> data and other information are not warranted as to completeness
<br>>>>>> or accuracy and
<br>>>>>> are subject to change without notice. Any comments or
<br>>>>>> statements made herein
<br>>>>>> do not necessarily reflect those of JPMorgan Chase & Co., its
<br>>>>>> subsidiaries
<br>>>>>> and affiliates.
<br>>>>>
<br>>>>
<br>>>
<br>>
</div><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-3173507Re: ActiveCluster move to ActiveMQ2006-02-28T14:41:02Z2006-02-28T14:41:02ZrajdaviesActiveCluster is currently using <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3173507&i=0" target="_top" rel="nofollow">activemq-dev@...</a>
<br><br>So it's part of the ActiveMQ project in the Apache incubator at the
<br>moment.
<br><br>cheers,
<br><br>Rob
<br><br>On 28 Feb 2006, at 22:31, Jonathan Duty wrote:
<br><div class='shrinkable-quote'><br>> Ahh, cool. I wasn't complaining. Just trying to help Ben get his
<br>> question answered. I'm very happy with ActiveCluster and I'm glad
<br>> to know it is still alive and well.
<br>>
<br>> Just curious, what mailing list? I don't see an activecluster
<br>> project in the incubator.
<br>>
<br>> ~Jonathan
<br>>
<br>> On Feb 28, 2006, at 5:26 PM, Rob Davies wrote:
<br>>
<br>>> Hi Jonathan,
<br>>>
<br>>> apologies if we've missed you're questions - but as we've moved to
<br>>> apache - we've been concentrating on the apache mailing lists - as
<br>>> we really want the community to grow there.
<br>>>
<br>>> cheers,
<br>>>
<br>>> Rob
<br>>>
<br>>> On 28 Feb 2006, at 20:27, Jonathan Duty wrote:
<br>>>
<br>>>> Ben,
<br>>>> I have been listening of this list for a few months and never
<br>>>> given a response to any of my questions, and I have seen 0
<br>>>> activity. I finally had to go post to the activemq list. I'm
<br>>>> guessing if you want a response from the core developers you will
<br>>>> need to do that also.
<br>>>>
<br>>>> ~Jonathan
<br>>>>
<br>>>> On Feb 28, 2006, at 3:10 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3173507&i=1" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<br>>>>
<br>>>>> It appears that new development in ActiveCluster has moved to
<br>>>>> the ActiveMQ
<br>>>>> project, and there are no new standalone releases of
<br>>>>> ActiveCluster since
<br>>>>> rel 1.0.
<br>>>>>
<br>>>>> Looking at the ActiveCluster build file in Apache SVN
<br>>>>> (<a href="https://svn.apache.org/repos/asf/incubator/activemq/trunk/" target="_top" rel="nofollow">https://svn.apache.org/repos/asf/incubator/activemq/trunk/</a>
<br>>>>> activecluster/pom.xml)
<br>>>>> it appears that ActiveCluster is completely dependent on
<br>>>>> ActiveMQ as well.
<br>>>>>
<br>>>>> I've been using ActiveCluster for some time now with other JMS
<br>>>>> implementations, and would like to continue doing so.
<br>>>>>
<br>>>>> Is there a version 1.1 or greater distribution of ActiveCluster
<br>>>>> (including
<br>>>>> Destination serialisation fix, <a href="http://jira.activemq.org/jira/" target="_top" rel="nofollow">http://jira.activemq.org/jira/</a>
<br>>>>> browse/ACL-6)
<br>>>>> available as a standalone JAR file somewhere? Is there intent
<br>>>>> to maintain
<br>>>>> ActiveCluster as a standalone cluster API, not dependent on
<br>>>>> ActiveMQ?
<br>>>>>
<br>>>>> Thanks,
<br>>>>> Ben
<br>>>>>
<br>>>>>
<br>>>>>
<br>>>>> This communication is for informational purposes only. It is not
<br>>>>> intended
<br>>>>> as an offer or solicitation for the purchase or sale of any
<br>>>>> financial
<br>>>>> instrument or as an official confirmation of any transaction.
<br>>>>> All market prices,
<br>>>>> data and other information are not warranted as to completeness
<br>>>>> or accuracy and
<br>>>>> are subject to change without notice. Any comments or statements
<br>>>>> made herein
<br>>>>> do not necessarily reflect those of JPMorgan Chase & Co., its
<br>>>>> subsidiaries
<br>>>>> and affiliates.
<br>>>>
<br>>>
<br>>
</div><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-3173442Re: ActiveCluster move to ActiveMQ2006-02-28T14:39:22Z2006-02-28T14:39:22ZrajdaviesHi Ben,
<br><br>long time no talk!
<br>We had to move the api back to use javax.jms.Destination to support
<br>WADI for Apache Geronimo 1.0. If I remember correctly, you've been
<br>using SonicMQ that did not support serialization of Destinations.
<br>We've tried to cater for that by adding the DestinationMarshaller
<br>interface - that can be an optional argument for creating a cluster.
<br>The default is to use an ActiveMQDestination, but it should be
<br>trivial to pass in an implementation that uses Sonic.
<br><br>Please let us know if this is not the case.
<br><br>cheers,
<br><br>Rob
<br><br>On 28 Feb 2006, at 20:10, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3173442&i=0" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<br><div class='shrinkable-quote'><br>> It appears that new development in ActiveCluster has moved to the
<br>> ActiveMQ
<br>> project, and there are no new standalone releases of ActiveCluster
<br>> since
<br>> rel 1.0.
<br>>
<br>> Looking at the ActiveCluster build file in Apache SVN
<br>> (<a href="https://svn.apache.org/repos/asf/incubator/activemq/trunk/" target="_top" rel="nofollow">https://svn.apache.org/repos/asf/incubator/activemq/trunk/</a>
<br>> activecluster/pom.xml)
<br>> it appears that ActiveCluster is completely dependent on ActiveMQ
<br>> as well.
<br>>
<br>> I've been using ActiveCluster for some time now with other JMS
<br>> implementations, and would like to continue doing so.
<br>>
<br>> Is there a version 1.1 or greater distribution of ActiveCluster
<br>> (including
<br>> Destination serialisation fix, <a href="http://jira.activemq.org/jira/browse/" target="_top" rel="nofollow">http://jira.activemq.org/jira/browse/</a>
<br>> ACL-6)
<br>> available as a standalone JAR file somewhere? Is there intent to
<br>> maintain
<br>> ActiveCluster as a standalone cluster API, not dependent on ActiveMQ?
<br>>
<br>> Thanks,
<br>> Ben
<br>>
<br>>
<br>>
<br>> This communication is for informational purposes only. It is not
<br>> intended
<br>> as an offer or solicitation for the purchase or sale of any financial
<br>> instrument or as an official confirmation of any transaction. All
<br>> market prices,
<br>> data and other information are not warranted as to completeness or
<br>> accuracy and
<br>> are subject to change without notice. Any comments or statements
<br>> made herein
<br>> do not necessarily reflect those of JPMorgan Chase & Co., its
<br>> subsidiaries
<br>> and affiliates.
</div><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-3173322Re: ActiveCluster move to ActiveMQ2006-02-28T14:31:10Z2006-02-28T14:31:10ZJonathan DutyAhh, cool. I wasn't complaining. Just trying to help Ben get his
<br>question answered. I'm very happy with ActiveCluster and I'm glad to
<br>know it is still alive and well.
<br><br>Just curious, what mailing list? I don't see an activecluster
<br>project in the incubator.
<br><br>~Jonathan
<br><br>On Feb 28, 2006, at 5:26 PM, Rob Davies wrote:
<br><div class='shrinkable-quote'><br>> Hi Jonathan,
<br>>
<br>> apologies if we've missed you're questions - but as we've moved to
<br>> apache - we've been concentrating on the apache mailing lists - as
<br>> we really want the community to grow there.
<br>>
<br>> cheers,
<br>>
<br>> Rob
<br>>
<br>> On 28 Feb 2006, at 20:27, Jonathan Duty wrote:
<br>>
<br>>> Ben,
<br>>> I have been listening of this list for a few months and never
<br>>> given a response to any of my questions, and I have seen 0
<br>>> activity. I finally had to go post to the activemq list. I'm
<br>>> guessing if you want a response from the core developers you will
<br>>> need to do that also.
<br>>>
<br>>> ~Jonathan
<br>>>
<br>>> On Feb 28, 2006, at 3:10 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3173322&i=0" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<br>>>
<br>>>> It appears that new development in ActiveCluster has moved to the
<br>>>> ActiveMQ
<br>>>> project, and there are no new standalone releases of
<br>>>> ActiveCluster since
<br>>>> rel 1.0.
<br>>>>
<br>>>> Looking at the ActiveCluster build file in Apache SVN
<br>>>> (<a href="https://svn.apache.org/repos/asf/incubator/activemq/trunk/" target="_top" rel="nofollow">https://svn.apache.org/repos/asf/incubator/activemq/trunk/</a>
<br>>>> activecluster/pom.xml)
<br>>>> it appears that ActiveCluster is completely dependent on
<br>>>> ActiveMQ as well.
<br>>>>
<br>>>> I've been using ActiveCluster for some time now with other JMS
<br>>>> implementations, and would like to continue doing so.
<br>>>>
<br>>>> Is there a version 1.1 or greater distribution of ActiveCluster
<br>>>> (including
<br>>>> Destination serialisation fix, <a href="http://jira.activemq.org/jira/" target="_top" rel="nofollow">http://jira.activemq.org/jira/</a>
<br>>>> browse/ACL-6)
<br>>>> available as a standalone JAR file somewhere? Is there intent to
<br>>>> maintain
<br>>>> ActiveCluster as a standalone cluster API, not dependent on
<br>>>> ActiveMQ?
<br>>>>
<br>>>> Thanks,
<br>>>> Ben
<br>>>>
<br>>>>
<br>>>>
<br>>>> This communication is for informational purposes only. It is not
<br>>>> intended
<br>>>> as an offer or solicitation for the purchase or sale of any
<br>>>> financial
<br>>>> instrument or as an official confirmation of any transaction. All
<br>>>> market prices,
<br>>>> data and other information are not warranted as to completeness
<br>>>> or accuracy and
<br>>>> are subject to change without notice. Any comments or statements
<br>>>> made herein
<br>>>> do not necessarily reflect those of JPMorgan Chase & Co., its
<br>>>> subsidiaries
<br>>>> and affiliates.
<br>>>
<br>>
</div><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-3173254Re: ActiveCluster move to ActiveMQ2006-02-28T14:26:46Z2006-02-28T14:26:46ZrajdaviesHi Jonathan,
<br><br>apologies if we've missed you're questions - but as we've moved to
<br>apache - we've been concentrating on the apache mailing lists - as we
<br>really want the community to grow there.
<br><br>cheers,
<br><br>Rob
<br><br>On 28 Feb 2006, at 20:27, Jonathan Duty wrote:
<br><div class='shrinkable-quote'><br>> Ben,
<br>> I have been listening of this list for a few months and never given
<br>> a response to any of my questions, and I have seen 0 activity. I
<br>> finally had to go post to the activemq list. I'm guessing if you
<br>> want a response from the core developers you will need to do that
<br>> also.
<br>>
<br>> ~Jonathan
<br>>
<br>> On Feb 28, 2006, at 3:10 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3173254&i=0" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<br>>
<br>>> It appears that new development in ActiveCluster has moved to the
<br>>> ActiveMQ
<br>>> project, and there are no new standalone releases of ActiveCluster
<br>>> since
<br>>> rel 1.0.
<br>>>
<br>>> Looking at the ActiveCluster build file in Apache SVN
<br>>> (<a href="https://svn.apache.org/repos/asf/incubator/activemq/trunk/" target="_top" rel="nofollow">https://svn.apache.org/repos/asf/incubator/activemq/trunk/</a>
<br>>> activecluster/pom.xml)
<br>>> it appears that ActiveCluster is completely dependent on ActiveMQ
<br>>> as well.
<br>>>
<br>>> I've been using ActiveCluster for some time now with other JMS
<br>>> implementations, and would like to continue doing so.
<br>>>
<br>>> Is there a version 1.1 or greater distribution of ActiveCluster
<br>>> (including
<br>>> Destination serialisation fix, <a href="http://jira.activemq.org/jira/" target="_top" rel="nofollow">http://jira.activemq.org/jira/</a>
<br>>> browse/ACL-6)
<br>>> available as a standalone JAR file somewhere? Is there intent to
<br>>> maintain
<br>>> ActiveCluster as a standalone cluster API, not dependent on ActiveMQ?
<br>>>
<br>>> Thanks,
<br>>> Ben
<br>>>
<br>>>
<br>>>
<br>>> This communication is for informational purposes only. It is not
<br>>> intended
<br>>> as an offer or solicitation for the purchase or sale of any financial
<br>>> instrument or as an official confirmation of any transaction. All
<br>>> market prices,
<br>>> data and other information are not warranted as to completeness or
<br>>> accuracy and
<br>>> are subject to change without notice. Any comments or statements
<br>>> made herein
<br>>> do not necessarily reflect those of JPMorgan Chase & Co., its
<br>>> subsidiaries
<br>>> and affiliates.
<br>>
</div><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-3171149Re: ActiveCluster move to ActiveMQ2006-02-28T12:27:55Z2006-02-28T12:27:55ZJonathan DutyBen,
<br>I have been listening of this list for a few months and never given a
<br>response to any of my questions, and I have seen 0 activity. I
<br>finally had to go post to the activemq list. I'm guessing if you
<br>want a response from the core developers you will need to do that also.
<br><br>~Jonathan
<br><br>On Feb 28, 2006, at 3:10 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=3171149&i=0" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<br><div class='shrinkable-quote'><br>> It appears that new development in ActiveCluster has moved to the
<br>> ActiveMQ
<br>> project, and there are no new standalone releases of ActiveCluster
<br>> since
<br>> rel 1.0.
<br>>
<br>> Looking at the ActiveCluster build file in Apache SVN
<br>> (<a href="https://svn.apache.org/repos/asf/incubator/activemq/trunk/" target="_top" rel="nofollow">https://svn.apache.org/repos/asf/incubator/activemq/trunk/</a>
<br>> activecluster/pom.xml)
<br>> it appears that ActiveCluster is completely dependent on ActiveMQ
<br>> as well.
<br>>
<br>> I've been using ActiveCluster for some time now with other JMS
<br>> implementations, and would like to continue doing so.
<br>>
<br>> Is there a version 1.1 or greater distribution of ActiveCluster
<br>> (including
<br>> Destination serialisation fix, <a href="http://jira.activemq.org/jira/browse/" target="_top" rel="nofollow">http://jira.activemq.org/jira/browse/</a>
<br>> ACL-6)
<br>> available as a standalone JAR file somewhere? Is there intent to
<br>> maintain
<br>> ActiveCluster as a standalone cluster API, not dependent on ActiveMQ?
<br>>
<br>> Thanks,
<br>> Ben
<br>>
<br>>
<br>>
<br>> This communication is for informational purposes only. It is not
<br>> intended
<br>> as an offer or solicitation for the purchase or sale of any financial
<br>> instrument or as an official confirmation of any transaction. All
<br>> market prices,
<br>> data and other information are not warranted as to completeness or
<br>> accuracy and
<br>> are subject to change without notice. Any comments or statements
<br>> made herein
<br>> do not necessarily reflect those of JPMorgan Chase & Co., its
<br>> subsidiaries
<br>> and affiliates.
</div><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-3170851ActiveCluster move to ActiveMQ2006-02-28T12:10:09Z2006-02-28T12:10:09Zben.kiblerIt appears that new development in ActiveCluster has moved to the ActiveMQ
<br>project, and there are no new standalone releases of ActiveCluster since
<br>rel 1.0.
<br><br>Looking at the ActiveCluster build file in Apache SVN
<br>(<a href="https://svn.apache.org/repos/asf/incubator/activemq/trunk/activecluster/pom.xml" target="_top" rel="nofollow">https://svn.apache.org/repos/asf/incubator/activemq/trunk/activecluster/pom.xml</a>)
<br> it appears that ActiveCluster is completely dependent on ActiveMQ as well.
<br><br>I've been using ActiveCluster for some time now with other JMS
<br>implementations, and would like to continue doing so.
<br><br>Is there a version 1.1 or greater distribution of ActiveCluster (including
<br>Destination serialisation fix, <a href="http://jira.activemq.org/jira/browse/ACL-6" target="_top" rel="nofollow">http://jira.activemq.org/jira/browse/ACL-6</a>)
<br>available as a standalone JAR file somewhere? Is there intent to maintain
<br>ActiveCluster as a standalone cluster API, not dependent on ActiveMQ?
<br><br>Thanks,
<br>Ben
<br><br><br><br>This communication is for informational purposes only. It is not intended
<br>as an offer or solicitation for the purchase or sale of any financial
<br>instrument or as an official confirmation of any transaction. All market prices,
<br>data and other information are not warranted as to completeness or accuracy and
<br>are subject to change without notice. Any comments or statements made herein
<br>do not necessarily reflect those of JPMorgan Chase & Co., its subsidiaries
<br>and affiliates.
<br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1882212detect cluster failure?2005-12-10T06:38:28Z2005-12-10T06:38:28ZJonathan DutyHello,
<br>I would like to use active cluster to manage a cluster of distributed
<br>servers. I will run a broker in one location that these servers
<br>will connect to and run active cluster on top of. After playing
<br>around with this I have it working. Except, if I bring down the
<br>broker, the servers never get an event to let me know that the
<br>cluster is down. Is there a way I can tap into the cluster or
<br>activemq connection to notice when it goes down?
<br><br>Thanks in advance,
<br>~Jonathan
<br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1530197[wadi-dev] Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-11-17T05:49:33Z2005-11-17T05:49:33ZJules Gosnell-4James Strachan wrote:
<br><div class='shrinkable-quote'><br>>
<br>> On 17 Nov 2005, at 12:53, Jules Gosnell wrote:
<br>>
<br>>> Rob Davies wrote:
<br>>>
<br>>>> Hi Jules,
<br>>>>
<br>>>> although there isn't a test case for this yet - I've moved
<br>>>> ActiveCluster to activeMQ4.0 - which has an improved discovery
<br>>>> mechansim. I've also updated the api - so that Strings are used in
<br>>>> the nodes instead of JMS destinations - as one user wanted to use
<br>>>> ActiveCluster with SonicMQ - which doesn't support serialized
<br>>>> Destinations.
<br>>>>
<br>>>> I'm hoping activeMQ4.0 will solve the issue for you :)
<br>>>
<br>>>
<br>>> hmmm....
<br>>>
<br>>> on activemq-user, the other day, Hiram said of AMQ4.0's release date :
<br>>>
<br>>> "Months :( We like to cook major releases for a while before
<br>>> declaring them production ready. It should be ready around 2nd Qrt
<br>>> next year."
<br>>>
<br>>> So, I doubt that this will be the release that ships with
<br>>> Geronimo-1.0 :-(
<br>>
<br>>
<br>> 4.0 will go GA before the end of the year. Its pretty solid & stable
<br>> right now.
<br>>
</div>excellent - apologies for any misunderstanding.
<br><div class='shrinkable-quote'><br>>
<br>>> My problem is that I would like a working activecluster on AMQ-
<br>>> peer:// for Geronimo-1.0 - otherwise it is going to be hard to
<br>>> cluster Geronimo.
<br>>>
<br>>> A 1.0 release needs deps on hard releases of its subcomponents - not
<br>>> snapshots.
<br>>
<br>>
<br>> We've already got 4.0-M1 out and 4.0-M2 is a few days away.
</div><br>I will keep an eye out for it and give it a whirl as soon as it shows up.
<br><br>Thanks, James.
<br><br>Jules
<br><br>>
<br>> James
<br>> -------
<br>> <a href="http://radio.weblogs.com/0112098/" target="_top" rel="nofollow">http://radio.weblogs.com/0112098/</a><br><br><br><br>--
<br>"Open Source is a self-assembling organism. You dangle a piece of
<br>string into a super-saturated solution and a whole operating-system
<br>crystallises out around it."
<br><br>/**********************************
<br> * Jules Gosnell
<br> * Partner
<br> * Core Developers Network (Europe)
<br> *
<br> * www.coredevelopers.net
<br> *
<br> * Open Source Training & Support.
<br> **********************************/
<br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1529695Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-11-17T05:15:30Z2005-11-17T05:15:30ZJames.Strachan<br>On 17 Nov 2005, at 12:53, Jules Gosnell wrote:
<br><div class='shrinkable-quote'><br>> Rob Davies wrote:
<br>>
<br>>> Hi Jules,
<br>>>
<br>>> although there isn't a test case for this yet - I've moved
<br>>> ActiveCluster to activeMQ4.0 - which has an improved discovery
<br>>> mechansim. I've also updated the api - so that Strings are used
<br>>> in the nodes instead of JMS destinations - as one user wanted to
<br>>> use ActiveCluster with SonicMQ - which doesn't support
<br>>> serialized Destinations.
<br>>>
<br>>> I'm hoping activeMQ4.0 will solve the issue for you :)
<br>>
<br>> hmmm....
<br>>
<br>> on activemq-user, the other day, Hiram said of AMQ4.0's release date :
<br>>
<br>> "Months :( We like to cook major releases for a while before
<br>> declaring them production ready. It should be ready around 2nd
<br>> Qrt next year."
<br>>
<br>> So, I doubt that this will be the release that ships with
<br>> Geronimo-1.0 :-(
</div><br>4.0 will go GA before the end of the year. Its pretty solid & stable
<br>right now.
<br><br><br>> My problem is that I would like a working activecluster on AMQ-
<br>> peer:// for Geronimo-1.0 - otherwise it is going to be hard to
<br>> cluster Geronimo.
<br>>
<br>> A 1.0 release needs deps on hard releases of its subcomponents -
<br>> not snapshots.
<br><br>We've already got 4.0-M1 out and 4.0-M2 is a few days away.
<br><br>James
<br>-------
<br><a href="http://radio.weblogs.com/0112098/" target="_top" rel="nofollow">http://radio.weblogs.com/0112098/</a><br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1529356[wadi-dev] Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-11-17T04:53:42Z2005-11-17T04:53:42ZJules Gosnell-4Rob Davies wrote:
<br><div class='shrinkable-quote'><br>> Hi Jules,
<br>>
<br>> although there isn't a test case for this yet - I've moved
<br>> ActiveCluster to activeMQ4.0 - which has an improved discovery
<br>> mechansim. I've also updated the api - so that Strings are used in
<br>> the nodes instead of JMS destinations - as one user wanted to use
<br>> ActiveCluster with SonicMQ - which doesn't support serialized
<br>> Destinations.
<br>>
<br>> I'm hoping activeMQ4.0 will solve the issue for you :)
</div><br>hmmm....
<br><br>on activemq-user, the other day, Hiram said of AMQ4.0's release date :
<br><br>"Months :( We like to cook major releases for a while before
<br>declaring them production ready. It should be ready around 2nd Qrt
<br>next year."
<br><br>So, I doubt that this will be the release that ships with Geronimo-1.0 :-(
<br><br>My problem is that I would like a working activecluster on AMQ-peer://
<br>for Geronimo-1.0 - otherwise it is going to be hard to cluster Geronimo.
<br><br>A 1.0 release needs deps on hard releases of its subcomponents - not
<br>snapshots.
<br><br>Can you see your way to a resolution to this ? - Does Geronimo-1.0
<br>feature anywhere on your roadmap ?
<br><br>thank-you,
<br><br><br>Jules
<br><br><br><div class='shrinkable-quote'><br>>
<br>> cheers,
<br>>
<br>> Rob
<br>> On 17 Nov 2005, at 11:34, Jules Gosnell wrote:
<br>>
<br>>> Rob,
<br>>>
<br>>> any movement on this issue ?
<br>>>
<br>>> cheers,
<br>>>
<br>>>
<br>>> Jules
<br>>>
<br>>>
<br>>> Rob Davies wrote:
<br>>>
<br>>>> Thanks for testing this Jules, will have a look - though I think
<br>>>> you're right that this is a discovery problem.
<br>>>>
<br>>>> cheers,
<br>>>>
<br>>>> Rob
<br>>>> On 7 Nov 2005, at 12:15, Jules Gosnell wrote:
<br>>>>
<br>>>>> Jules Gosnell wrote:
<br>>>>>
<br>>>>>> Jules Gosnell wrote:
<br>>>>>>
<br>>>>>>> Rob,
<br>>>>>>>
<br>>>>>>> Thanks for putting some time into this...
<br>>>>>>>
<br>>>>>>> I've just rebuilt AC and run the test that I included again.
<br>>>>>>> I'm afraid that it doesn't fix the problem. Once the second
<br>>>>>>> node leaves the cluster the first node becomes unable to see
<br>>>>>>> any subsequent node that wishes to join. It looks as if it is
<br>>>>>>> some logic problem associated with cluster's dropping to one
<br>>>>>>> node, but, curiously it works fine with the tcp:// protocol...
<br>>>>>>> Maybe it is something to do with autodiscovery ? Do tcp:// and
<br>>>>>>> peer;// use common or separate autodiscovery mechanisms ?
<br>>>>>>>
<br>>>>>>> Is there some test that I can run to help pinpoint the problem ?
<br>>>>>>>
<br>>>>>>> Sorry to be the bearer of bad news :-(
<br>>>>>>>
<br>>>>>>>
<br>>>>>>> Jules
<br>>>>>>
<br>>>>>>
<br>>>>>>
<br>>>>>>
<br>>>>>>
<br>>>>>> P.S.
<br>>>>>>
<br>>>>>> I ran my tests on AMQ 3.2-M1 - I guess that I need to run them
<br>>>>>> on AMQ3-2 to be sure - only problem is that 3.2 breaks our
<br>>>>>> mavens build with deps that cannot be downloaded - maybe our
<br>>>>>> repo list is not what it should be - I will see what I can do
<br>>>>>> and get back to you in a little while.
<br>>>>>
<br>>>>>
<br>>>>>
<br>>>>> Just ran the test on HEAD of AC and AMQ-3.2 - same result...
<br>>>>>
<br>>>>> Everything is fine until a cluster drops down to one node. That
<br>>>>> node then becomes somehow out of reach of subsequent nodes. Any
<br>>>>> nodes that are started after this point, only see each other and
<br>>>>> not the original node. Taking the newly formed cluster down to
<br>>>>> one member has the same effect, so you can end up with more than
<br>>>>> one isolated node.
<br>>>>>
<br>>>>> Jules
<br>>>>>
<br>>>>>>
<br>>>>>>>
<br>>>>>>> Rob Davies wrote:
<br>>>>>>>
<br>>>>>>>> hi Jules,
<br>>>>>>>>
<br>>>>>>>> the latest ActiveCluster in the repo works with ActiveMQ 3.2
<br>>>>>>>> - amq4 will hopefully be ready soon. I'd appreciate it if
<br>>>>>>>> you could build activecluster from src - and see if your
<br>>>>>>>> problems have disappeared - there was a a nasty deadlock I
<br>>>>>>>> ended up fixing.
<br>>>>>>>>
<br>>>>>>>> cheers,
<br>>>>>>>>
<br>>>>>>>> Rob
<br>>>>>>>>
<br>>>>>>>> On 6 Nov 2005, at 18:41, Jules Gosnell wrote:
<br>>>>>>>>
<br>>>>>>>>> Rob Davies wrote:
<br>>>>>>>>>
<br>>>>>>>>>> Hi Jules,
<br>>>>>>>>>>
<br>>>>>>>>>> I'm in the process of porting activecluster to amq4. There's
<br>>>>>>>>>> an issue with discovery in amq4 at the moment that needs
<br>>>>>>>>>> to be fixed.
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>> Ok - no problem - if you let me know when the fix goes in, I
<br>>>>>>>>> will be happy to retest and report.
<br>>>>>>>>>
<br>>>>>>>>> BTW - do you know if AMQ4 is the release that is intended to
<br>>>>>>>>> be bundled with Geronimo-1.0 ? I am targetting a WADI
<br>>>>>>>>> release at Geronimo-1.0 and I need to make sure that we run
<br>>>>>>>>> on the same version of AMQ as is packaged with G.
<br>>>>>>>>>
<br>>>>>>>>> Cheers, Rob,
<br>>>>>>>>>
<br>>>>>>>>> Jules
<br>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>> cheers,
<br>>>>>>>>>>
<br>>>>>>>>>> Rob
<br>>>>>>>>>> On 29 Oct 2005, at 02:10, Jules Gosnell wrote:
<br>>>>>>>>>>
<br>>>>>>>>>>> Jules Gosnell wrote:
<br>>>>>>>>>>>
<br>>>>>>>>>>>
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> tested against activecluster-1.1-SNAPSHOT, activemq-4.0-
<br>>>>>>>>>>>> SNAPSHOT and Sun JDK 1.4.2_08 and 1.5.0 - two node
<br>>>>>>>>>>>> cluster, both on same machine.
<br>>>>>>>>>>>>
<br>>>>>>>>>>>
<br>>>>>>>>>>> oops - AMQ-4.0 was not on the machine that I ran this on -
<br>>>>>>>>>>> so this issue is actually with AMQ-3.2-M1.
<br>>>>>>>>>>>
<br>>>>>>>>>>> I am trying to run the test with AMQ-4.0, but the second
<br>>>>>>>>>>> node won't start because "java.io.IOException: Journal is
<br>>>>>>>>>>> allready opened by another application", So i will have
<br>>>>>>>>>>> to figure out how to tell it not to journal - I'll be
<br>>>>>>>>>>> back ;-)
<br>>>>>>>>>>>
<br>>>>>>>>>>> Jules
<br>>>>>>>>>>>
<br>>>>>>>>>>>
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> start your first node (red)
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster red
<br>>>>>>>>>>>> 2005/10/29 00:04:18:317 BST [INFO] ACCluster - starting...
<br>>>>>>>>>>>> 2005/10/29 00:04:18:321 BST [INFO] ACCluster - ...started
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> start your second node (green)
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>>>>>>>> 2005/10/29 00:04:31:874 BST [INFO] ACCluster - starting...
<br>>>>>>>>>>>> 2005/10/29 00:04:31:885 BST [INFO] ACCluster - ...started
<br>>>>>>>>>>>> 2005/10/29 00:04:33:323 BST [INFO] ACCluster - onNodeAdd: red
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> red says:
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> 2005/10/29 00:04:33:536 BST [INFO] ACCluster - onNodeAdd:
<br>>>>>>>>>>>> green
<br>>>>>>>>>>>> 2005/10/29 00:04:33:537 BST [INFO] ACCluster -
<br>>>>>>>>>>>> onCoordinatorChanged: green
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> ctl-c green, after a few seconds red says:
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> 2005/10/29 00:04:46:624 BST [INFO] ACCluster -
<br>>>>>>>>>>>> onNodeFailed: green
<br>>>>>>>>>>>> 2005/10/29 00:04:46:626 BST [INFO] ACCluster -
<br>>>>>>>>>>>> onCoordinatorChanged: red
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> all fine so far - now restart green:
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>>>>>>>> 2005/10/29 00:04:51:962 BST [INFO] ACCluster - starting...
<br>>>>>>>>>>>> 2005/10/29 00:04:51:967 BST [INFO] ACCluster - ...started
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> red says......nothing.
<br>>>>>>>>>>>>
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> if you then start a third node (blue):
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster blue
<br>>>>>>>>>>>> 2005/10/29 00:08:47:990 BST [INFO] ACCluster - starting...
<br>>>>>>>>>>>> 2005/10/29 00:08:47:994 BST [INFO] ACCluster - ...started
<br>>>>>>>>>>>> 2005/10/29 00:08:50:198 BST [INFO] ACCluster - onNodeAdd:
<br>>>>>>>>>>>> green
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> it sees green, but not red and green says:
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> 2005/10/29 00:08:49:661 BST [INFO] ACCluster - onNodeAdd: blue
<br>>>>>>>>>>>> 2005/10/29 00:08:49:662 BST [INFO] ACCluster -
<br>>>>>>>>>>>> onCoordinatorChanged: blue
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> so green sees blue too.
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> but red says.....nothing.
<br>>>>>>>>>>>>
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> Using tcp:// and a broker, red will happily see blue and
<br>>>>>>>>>>>> green after they are restarted.
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> This has been the case with activecluster/activemq for a
<br>>>>>>>>>>>> long time, but I was hoping that it would be fixed by
<br>>>>>>>>>>>> improvements to peer:// in activemq-4.0 - but :-(
<br>>>>>>>>>>>>
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> Jules
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> -------------------------------------------------------------
<br>>>>>>>>>>>> -- ---- -- ---
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> package org.codehaus.wadi.sandbox.partition.impl;
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> import java.util.HashMap;
<br>>>>>>>>>>>> import java.util.Map;
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> import javax.jms.JMSException;
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> import org.activecluster.ClusterEvent;
<br>>>>>>>>>>>> import org.activecluster.ClusterFactory;
<br>>>>>>>>>>>> import org.activecluster.ClusterListener;
<br>>>>>>>>>>>> import org.activecluster.impl.DefaultClusterFactory;
<br>>>>>>>>>>>> import org.activemq.ActiveMQConnectionFactory;
<br>>>>>>>>>>>> import org.activemq.store.vm.VMPersistenceAdapterFactory;
<br>>>>>>>>>>>> import org.apache.commons.logging.Log;
<br>>>>>>>>>>>> import org.apache.commons.logging.LogFactory;
<br>>>>>>>>>>>> import
<br>>>>>>>>>>>> org.codehaus.wadi.dindex.impl.SeniorityElectionStrategy;
<br>>>>>>>>>>>> import org.codehaus.wadi.impl.CustomClusterFactory;
<br>>>>>>>>>>>> import org.codehaus.wadi.sandbox.partition.Cluster;
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> public class ACCluster {
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> protected final Log _log = LogFactory.getLog(getClass());
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> //protected final String _clusterUri="tcp://smilodon:
<br>>>>>>>>>>>> 61616";
<br>>>>>>>>>>>> protected final String _clusterUri="peer://
<br>>>>>>>>>>>> org.codehaus.wadi";
<br>>>>>>>>>>>> protected final String
<br>>>>>>>>>>>> _clusterName="ORG.CODEHAUS.WADI.TEST";
<br>>>>>>>>>>>> protected final ActiveMQConnectionFactory
<br>>>>>>>>>>>> _connectionFactory=new ActiveMQConnectionFactory (_clusterUri);
<br>>>>>>>>>>>> protected final ClusterFactory _clusterFactory=new
<br>>>>>>>>>>>> DefaultClusterFactory(_connectionFactory);
<br>>>>>>>>>>>> protected final org.activecluster.Cluster _cluster;
<br>>>>>>>>>>>> protected final long _timeout=30*1000L;
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> public ACCluster(String nodeName) throws Exception {
<br>>>>>>>>>>>> System.setProperty
<br>>>>>>>>>>>> ("activemq.persistenceAdapterFactory",
<br>>>>>>>>>>>> VMPersistenceAdapterFactory.class.getName());
<br>>>>>>>>>>>> _cluster=_clusterFactory.createCluster(_clusterName);
<br>>>>>>>>>>>> Map state=new HashMap();
<br>>>>>>>>>>>> state.put("nodeName", nodeName);
<br>>>>>>>>>>>> _cluster.getLocalNode().setState(state);
<br>>>>>>>>>>>> _cluster.addClusterListener(new ClusterListener() {
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> public void onNodeAdd(ClusterEvent arg0) {
<br>>>>>>>>>>>> _log.info("onNodeAdd: "+arg0.getNode
<br>>>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>>>> }
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> public void onNodeUpdate(ClusterEvent arg0) {
<br>>>>>>>>>>>> _log.info("onNodeUpdate: "+arg0.getNode
<br>>>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>>>> }
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> public void onNodeRemoved(ClusterEvent arg0) {
<br>>>>>>>>>>>> _log.info("onNodeRemoved: "+arg0.getNode
<br>>>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>>>> }
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> public void onNodeFailed(ClusterEvent arg0) {
<br>>>>>>>>>>>> _log.info("onNodeFailed: "+arg0.getNode
<br>>>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>>>> }
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> public void onCoordinatorChanged(ClusterEvent
<br>>>>>>>>>>>> arg0) {
<br>>>>>>>>>>>> _log.info("onCoordinatorChanged:
<br>>>>>>>>>>>> "+arg0.getNode ().getState().get("nodeName"));
<br>>>>>>>>>>>> }
<br>>>>>>>>>>>> });
<br>>>>>>>>>>>> }
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> public void start() throws JMSException {
<br>>>>>>>>>>>> _log.info("starting...");
<br>>>>>>>>>>>> _cluster.start();
<br>>>>>>>>>>>> _log.info("...started");
<br>>>>>>>>>>>> }
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> public void stop() {
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> }
<br>>>>>>>>>>>>
<br>>>>>>>>>>>> public static void main(String[] args) throws Exception {
<br>>>>>>>>>>>> ACCluster cluster=new ACCluster(args[0]);
<br>>>>>>>>>>>> cluster.start();
<br>>>>>>>>>>>> Thread.sleep(100*1000);
<br>>>>>>>>>>>> cluster.stop();
<br>>>>>>>>>>>> }
<br>>>>>>>>>>>> }
<br>>>>>>>>>>>>
<br>>>>>>>>>>>>
<br>>>>>>>>>>>
<br>>>>>>>>>>>
<br>>>>>>>>>>> --
<br>>>>>>>>>>> "Open Source is a self-assembling organism. You dangle a
<br>>>>>>>>>>> piece of
<br>>>>>>>>>>> string into a super-saturated solution and a whole
<br>>>>>>>>>>> operating- system
<br>>>>>>>>>>> crystallises out around it."
<br>>>>>>>>>>>
<br>>>>>>>>>>> /**********************************
<br>>>>>>>>>>> * Jules Gosnell
<br>>>>>>>>>>> * Partner
<br>>>>>>>>>>> * Core Developers Network (Europe)
<br>>>>>>>>>>> *
<br>>>>>>>>>>> * www.coredevelopers.net
<br>>>>>>>>>>> *
<br>>>>>>>>>>> * Open Source Training & Support.
<br>>>>>>>>>>> **********************************/
<br>>>>>>>>>>>
<br>>>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>> --
<br>>>>>>>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>>>>>>>> string into a super-saturated solution and a whole operating-
<br>>>>>>>>> system
<br>>>>>>>>> crystallises out around it."
<br>>>>>>>>>
<br>>>>>>>>> /**********************************
<br>>>>>>>>> * Jules Gosnell
<br>>>>>>>>> * Partner
<br>>>>>>>>> * Core Developers Network (Europe)
<br>>>>>>>>> *
<br>>>>>>>>> * www.coredevelopers.net
<br>>>>>>>>> *
<br>>>>>>>>> * Open Source Training & Support.
<br>>>>>>>>> **********************************/
<br>>>>>>>>>
<br>>>>>>>
<br>>>>>>>
<br>>>>>>
<br>>>>>>
<br>>>>>
<br>>>>>
<br>>>>> --
<br>>>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>>>> string into a super-saturated solution and a whole operating-system
<br>>>>> crystallises out around it."
<br>>>>>
<br>>>>> /**********************************
<br>>>>> * Jules Gosnell
<br>>>>> * Partner
<br>>>>> * Core Developers Network (Europe)
<br>>>>> *
<br>>>>> * www.coredevelopers.net
<br>>>>> *
<br>>>>> * Open Source Training & Support.
<br>>>>> **********************************/
<br>>>>>
<br>>>
<br>>>
<br>>> --
<br>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>> string into a super-saturated solution and a whole operating-system
<br>>> crystallises out around it."
<br>>>
<br>>> /**********************************
<br>>> * Jules Gosnell
<br>>> * Partner
<br>>> * Core Developers Network (Europe)
<br>>> *
<br>>> * www.coredevelopers.net
<br>>> *
<br>>> * Open Source Training & Support.
<br>>> **********************************/
<br>>>
</div><br><br>--
<br>"Open Source is a self-assembling organism. You dangle a piece of
<br>string into a super-saturated solution and a whole operating-system
<br>crystallises out around it."
<br><br>/**********************************
<br> * Jules Gosnell
<br> * Partner
<br> * Core Developers Network (Europe)
<br> *
<br> * www.coredevelopers.net
<br> *
<br> * Open Source Training & Support.
<br> **********************************/
<br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1528328Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-11-17T03:41:37Z2005-11-17T03:41:37ZrajdaviesHi Jules,
<br><br>although there isn't a test case for this yet - I've moved
<br>ActiveCluster to activeMQ4.0 - which has an improved discovery
<br>mechansim. I've also updated the api - so that Strings are used in
<br>the nodes instead of JMS destinations - as one user wanted to use
<br>ActiveCluster with SonicMQ - which doesn't support serialized
<br>Destinations.
<br><br>I'm hoping activeMQ4.0 will solve the issue for you :)
<br><br>cheers,
<br><br>Rob
<br>On 17 Nov 2005, at 11:34, Jules Gosnell wrote:
<br><div class='shrinkable-quote'><br>> Rob,
<br>>
<br>> any movement on this issue ?
<br>>
<br>> cheers,
<br>>
<br>>
<br>> Jules
<br>>
<br>>
<br>> Rob Davies wrote:
<br>>
<br>>> Thanks for testing this Jules, will have a look - though I think
<br>>> you're right that this is a discovery problem.
<br>>>
<br>>> cheers,
<br>>>
<br>>> Rob
<br>>> On 7 Nov 2005, at 12:15, Jules Gosnell wrote:
<br>>>
<br>>>> Jules Gosnell wrote:
<br>>>>
<br>>>>> Jules Gosnell wrote:
<br>>>>>
<br>>>>>> Rob,
<br>>>>>>
<br>>>>>> Thanks for putting some time into this...
<br>>>>>>
<br>>>>>> I've just rebuilt AC and run the test that I included again.
<br>>>>>> I'm afraid that it doesn't fix the problem. Once the second
<br>>>>>> node leaves the cluster the first node becomes unable to see
<br>>>>>> any subsequent node that wishes to join. It looks as if it is
<br>>>>>> some logic problem associated with cluster's dropping to one
<br>>>>>> node, but, curiously it works fine with the tcp:// protocol...
<br>>>>>> Maybe it is something to do with autodiscovery ? Do tcp:// and
<br>>>>>> peer;// use common or separate autodiscovery mechanisms ?
<br>>>>>>
<br>>>>>> Is there some test that I can run to help pinpoint the problem ?
<br>>>>>>
<br>>>>>> Sorry to be the bearer of bad news :-(
<br>>>>>>
<br>>>>>>
<br>>>>>> Jules
<br>>>>>
<br>>>>>
<br>>>>>
<br>>>>>
<br>>>>> P.S.
<br>>>>>
<br>>>>> I ran my tests on AMQ 3.2-M1 - I guess that I need to run them
<br>>>>> on AMQ3-2 to be sure - only problem is that 3.2 breaks our
<br>>>>> mavens build with deps that cannot be downloaded - maybe our
<br>>>>> repo list is not what it should be - I will see what I can do
<br>>>>> and get back to you in a little while.
<br>>>>
<br>>>>
<br>>>> Just ran the test on HEAD of AC and AMQ-3.2 - same result...
<br>>>>
<br>>>> Everything is fine until a cluster drops down to one node. That
<br>>>> node then becomes somehow out of reach of subsequent nodes. Any
<br>>>> nodes that are started after this point, only see each other and
<br>>>> not the original node. Taking the newly formed cluster down to
<br>>>> one member has the same effect, so you can end up with more than
<br>>>> one isolated node.
<br>>>>
<br>>>> Jules
<br>>>>
<br>>>>>
<br>>>>>>
<br>>>>>> Rob Davies wrote:
<br>>>>>>
<br>>>>>>> hi Jules,
<br>>>>>>>
<br>>>>>>> the latest ActiveCluster in the repo works with ActiveMQ 3.2
<br>>>>>>> - amq4 will hopefully be ready soon. I'd appreciate it if
<br>>>>>>> you could build activecluster from src - and see if your
<br>>>>>>> problems have disappeared - there was a a nasty deadlock I
<br>>>>>>> ended up fixing.
<br>>>>>>>
<br>>>>>>> cheers,
<br>>>>>>>
<br>>>>>>> Rob
<br>>>>>>>
<br>>>>>>> On 6 Nov 2005, at 18:41, Jules Gosnell wrote:
<br>>>>>>>
<br>>>>>>>> Rob Davies wrote:
<br>>>>>>>>
<br>>>>>>>>> Hi Jules,
<br>>>>>>>>>
<br>>>>>>>>> I'm in the process of porting activecluster to amq4.
<br>>>>>>>>> There's an issue with discovery in amq4 at the moment
<br>>>>>>>>> that needs to be fixed.
<br>>>>>>>>
<br>>>>>>>>
<br>>>>>>>>
<br>>>>>>>>
<br>>>>>>>>
<br>>>>>>>> Ok - no problem - if you let me know when the fix goes in, I
<br>>>>>>>> will be happy to retest and report.
<br>>>>>>>>
<br>>>>>>>> BTW - do you know if AMQ4 is the release that is intended to
<br>>>>>>>> be bundled with Geronimo-1.0 ? I am targetting a WADI
<br>>>>>>>> release at Geronimo-1.0 and I need to make sure that we run
<br>>>>>>>> on the same version of AMQ as is packaged with G.
<br>>>>>>>>
<br>>>>>>>> Cheers, Rob,
<br>>>>>>>>
<br>>>>>>>> Jules
<br>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>> cheers,
<br>>>>>>>>>
<br>>>>>>>>> Rob
<br>>>>>>>>> On 29 Oct 2005, at 02:10, Jules Gosnell wrote:
<br>>>>>>>>>
<br>>>>>>>>>> Jules Gosnell wrote:
<br>>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>>>
<br>>>>>>>>>>> tested against activecluster-1.1-SNAPSHOT, activemq-4.0-
<br>>>>>>>>>>> SNAPSHOT and Sun JDK 1.4.2_08 and 1.5.0 - two node
<br>>>>>>>>>>> cluster, both on same machine.
<br>>>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>> oops - AMQ-4.0 was not on the machine that I ran this on -
<br>>>>>>>>>> so this issue is actually with AMQ-3.2-M1.
<br>>>>>>>>>>
<br>>>>>>>>>> I am trying to run the test with AMQ-4.0, but the second
<br>>>>>>>>>> node won't start because "java.io.IOException: Journal
<br>>>>>>>>>> is allready opened by another application", So i will
<br>>>>>>>>>> have to figure out how to tell it not to journal - I'll
<br>>>>>>>>>> be back ;-)
<br>>>>>>>>>>
<br>>>>>>>>>> Jules
<br>>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>>>
<br>>>>>>>>>>> start your first node (red)
<br>>>>>>>>>>>
<br>>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster red
<br>>>>>>>>>>> 2005/10/29 00:04:18:317 BST [INFO] ACCluster - starting...
<br>>>>>>>>>>> 2005/10/29 00:04:18:321 BST [INFO] ACCluster - ...started
<br>>>>>>>>>>>
<br>>>>>>>>>>> start your second node (green)
<br>>>>>>>>>>>
<br>>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>>>>>>> 2005/10/29 00:04:31:874 BST [INFO] ACCluster - starting...
<br>>>>>>>>>>> 2005/10/29 00:04:31:885 BST [INFO] ACCluster - ...started
<br>>>>>>>>>>> 2005/10/29 00:04:33:323 BST [INFO] ACCluster - onNodeAdd: red
<br>>>>>>>>>>>
<br>>>>>>>>>>> red says:
<br>>>>>>>>>>>
<br>>>>>>>>>>> 2005/10/29 00:04:33:536 BST [INFO] ACCluster - onNodeAdd:
<br>>>>>>>>>>> green
<br>>>>>>>>>>> 2005/10/29 00:04:33:537 BST [INFO] ACCluster -
<br>>>>>>>>>>> onCoordinatorChanged: green
<br>>>>>>>>>>>
<br>>>>>>>>>>> ctl-c green, after a few seconds red says:
<br>>>>>>>>>>>
<br>>>>>>>>>>> 2005/10/29 00:04:46:624 BST [INFO] ACCluster -
<br>>>>>>>>>>> onNodeFailed: green
<br>>>>>>>>>>> 2005/10/29 00:04:46:626 BST [INFO] ACCluster -
<br>>>>>>>>>>> onCoordinatorChanged: red
<br>>>>>>>>>>>
<br>>>>>>>>>>> all fine so far - now restart green:
<br>>>>>>>>>>>
<br>>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>>>>>>> 2005/10/29 00:04:51:962 BST [INFO] ACCluster - starting...
<br>>>>>>>>>>> 2005/10/29 00:04:51:967 BST [INFO] ACCluster - ...started
<br>>>>>>>>>>>
<br>>>>>>>>>>> red says......nothing.
<br>>>>>>>>>>>
<br>>>>>>>>>>>
<br>>>>>>>>>>> if you then start a third node (blue):
<br>>>>>>>>>>>
<br>>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster blue
<br>>>>>>>>>>> 2005/10/29 00:08:47:990 BST [INFO] ACCluster - starting...
<br>>>>>>>>>>> 2005/10/29 00:08:47:994 BST [INFO] ACCluster - ...started
<br>>>>>>>>>>> 2005/10/29 00:08:50:198 BST [INFO] ACCluster - onNodeAdd:
<br>>>>>>>>>>> green
<br>>>>>>>>>>>
<br>>>>>>>>>>> it sees green, but not red and green says:
<br>>>>>>>>>>>
<br>>>>>>>>>>> 2005/10/29 00:08:49:661 BST [INFO] ACCluster - onNodeAdd:
<br>>>>>>>>>>> blue
<br>>>>>>>>>>> 2005/10/29 00:08:49:662 BST [INFO] ACCluster -
<br>>>>>>>>>>> onCoordinatorChanged: blue
<br>>>>>>>>>>>
<br>>>>>>>>>>> so green sees blue too.
<br>>>>>>>>>>>
<br>>>>>>>>>>> but red says.....nothing.
<br>>>>>>>>>>>
<br>>>>>>>>>>>
<br>>>>>>>>>>> Using tcp:// and a broker, red will happily see blue and
<br>>>>>>>>>>> green after they are restarted.
<br>>>>>>>>>>>
<br>>>>>>>>>>> This has been the case with activecluster/activemq for a
<br>>>>>>>>>>> long time, but I was hoping that it would be fixed by
<br>>>>>>>>>>> improvements to peer:// in activemq-4.0 - but :-(
<br>>>>>>>>>>>
<br>>>>>>>>>>>
<br>>>>>>>>>>> Jules
<br>>>>>>>>>>>
<br>>>>>>>>>>> -------------------------------------------------------------
<br>>>>>>>>>>> -- ---- -- ---
<br>>>>>>>>>>>
<br>>>>>>>>>>> package org.codehaus.wadi.sandbox.partition.impl;
<br>>>>>>>>>>>
<br>>>>>>>>>>> import java.util.HashMap;
<br>>>>>>>>>>> import java.util.Map;
<br>>>>>>>>>>>
<br>>>>>>>>>>> import javax.jms.JMSException;
<br>>>>>>>>>>>
<br>>>>>>>>>>> import org.activecluster.ClusterEvent;
<br>>>>>>>>>>> import org.activecluster.ClusterFactory;
<br>>>>>>>>>>> import org.activecluster.ClusterListener;
<br>>>>>>>>>>> import org.activecluster.impl.DefaultClusterFactory;
<br>>>>>>>>>>> import org.activemq.ActiveMQConnectionFactory;
<br>>>>>>>>>>> import org.activemq.store.vm.VMPersistenceAdapterFactory;
<br>>>>>>>>>>> import org.apache.commons.logging.Log;
<br>>>>>>>>>>> import org.apache.commons.logging.LogFactory;
<br>>>>>>>>>>> import
<br>>>>>>>>>>> org.codehaus.wadi.dindex.impl.SeniorityElectionStrategy;
<br>>>>>>>>>>> import org.codehaus.wadi.impl.CustomClusterFactory;
<br>>>>>>>>>>> import org.codehaus.wadi.sandbox.partition.Cluster;
<br>>>>>>>>>>>
<br>>>>>>>>>>> public class ACCluster {
<br>>>>>>>>>>>
<br>>>>>>>>>>> protected final Log _log = LogFactory.getLog(getClass());
<br>>>>>>>>>>>
<br>>>>>>>>>>> //protected final String _clusterUri="tcp://smilodon:
<br>>>>>>>>>>> 61616";
<br>>>>>>>>>>> protected final String _clusterUri="peer://
<br>>>>>>>>>>> org.codehaus.wadi";
<br>>>>>>>>>>> protected final String
<br>>>>>>>>>>> _clusterName="ORG.CODEHAUS.WADI.TEST";
<br>>>>>>>>>>> protected final ActiveMQConnectionFactory
<br>>>>>>>>>>> _connectionFactory=new ActiveMQConnectionFactory
<br>>>>>>>>>>> (_clusterUri);
<br>>>>>>>>>>> protected final ClusterFactory _clusterFactory=new
<br>>>>>>>>>>> DefaultClusterFactory(_connectionFactory);
<br>>>>>>>>>>> protected final org.activecluster.Cluster _cluster;
<br>>>>>>>>>>> protected final long _timeout=30*1000L;
<br>>>>>>>>>>>
<br>>>>>>>>>>> public ACCluster(String nodeName) throws Exception {
<br>>>>>>>>>>> System.setProperty
<br>>>>>>>>>>> ("activemq.persistenceAdapterFactory",
<br>>>>>>>>>>> VMPersistenceAdapterFactory.class.getName());
<br>>>>>>>>>>> _cluster=_clusterFactory.createCluster(_clusterName);
<br>>>>>>>>>>> Map state=new HashMap();
<br>>>>>>>>>>> state.put("nodeName", nodeName);
<br>>>>>>>>>>> _cluster.getLocalNode().setState(state);
<br>>>>>>>>>>> _cluster.addClusterListener(new ClusterListener() {
<br>>>>>>>>>>>
<br>>>>>>>>>>> public void onNodeAdd(ClusterEvent arg0) {
<br>>>>>>>>>>> _log.info("onNodeAdd: "+arg0.getNode
<br>>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>>> }
<br>>>>>>>>>>>
<br>>>>>>>>>>> public void onNodeUpdate(ClusterEvent arg0) {
<br>>>>>>>>>>> _log.info("onNodeUpdate: "+arg0.getNode
<br>>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>>> }
<br>>>>>>>>>>>
<br>>>>>>>>>>> public void onNodeRemoved(ClusterEvent arg0) {
<br>>>>>>>>>>> _log.info("onNodeRemoved: "+arg0.getNode
<br>>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>>> }
<br>>>>>>>>>>>
<br>>>>>>>>>>> public void onNodeFailed(ClusterEvent arg0) {
<br>>>>>>>>>>> _log.info("onNodeFailed: "+arg0.getNode
<br>>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>>> }
<br>>>>>>>>>>>
<br>>>>>>>>>>> public void onCoordinatorChanged(ClusterEvent
<br>>>>>>>>>>> arg0) {
<br>>>>>>>>>>> _log.info("onCoordinatorChanged:
<br>>>>>>>>>>> "+arg0.getNode ().getState().get("nodeName"));
<br>>>>>>>>>>> }
<br>>>>>>>>>>> });
<br>>>>>>>>>>> }
<br>>>>>>>>>>>
<br>>>>>>>>>>> public void start() throws JMSException {
<br>>>>>>>>>>> _log.info("starting...");
<br>>>>>>>>>>> _cluster.start();
<br>>>>>>>>>>> _log.info("...started");
<br>>>>>>>>>>> }
<br>>>>>>>>>>>
<br>>>>>>>>>>> public void stop() {
<br>>>>>>>>>>>
<br>>>>>>>>>>> }
<br>>>>>>>>>>>
<br>>>>>>>>>>> public static void main(String[] args) throws Exception {
<br>>>>>>>>>>> ACCluster cluster=new ACCluster(args[0]);
<br>>>>>>>>>>> cluster.start();
<br>>>>>>>>>>> Thread.sleep(100*1000);
<br>>>>>>>>>>> cluster.stop();
<br>>>>>>>>>>> }
<br>>>>>>>>>>> }
<br>>>>>>>>>>>
<br>>>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>> --
<br>>>>>>>>>> "Open Source is a self-assembling organism. You dangle a
<br>>>>>>>>>> piece of
<br>>>>>>>>>> string into a super-saturated solution and a whole
<br>>>>>>>>>> operating- system
<br>>>>>>>>>> crystallises out around it."
<br>>>>>>>>>>
<br>>>>>>>>>> /**********************************
<br>>>>>>>>>> * Jules Gosnell
<br>>>>>>>>>> * Partner
<br>>>>>>>>>> * Core Developers Network (Europe)
<br>>>>>>>>>> *
<br>>>>>>>>>> * www.coredevelopers.net
<br>>>>>>>>>> *
<br>>>>>>>>>> * Open Source Training & Support.
<br>>>>>>>>>> **********************************/
<br>>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>
<br>>>>>>>>
<br>>>>>>>> --
<br>>>>>>>> "Open Source is a self-assembling organism. You dangle a
<br>>>>>>>> piece of
<br>>>>>>>> string into a super-saturated solution and a whole operating-
<br>>>>>>>> system
<br>>>>>>>> crystallises out around it."
<br>>>>>>>>
<br>>>>>>>> /**********************************
<br>>>>>>>> * Jules Gosnell
<br>>>>>>>> * Partner
<br>>>>>>>> * Core Developers Network (Europe)
<br>>>>>>>> *
<br>>>>>>>> * www.coredevelopers.net
<br>>>>>>>> *
<br>>>>>>>> * Open Source Training & Support.
<br>>>>>>>> **********************************/
<br>>>>>>>>
<br>>>>>>
<br>>>>>>
<br>>>>>
<br>>>>>
<br>>>>
<br>>>>
<br>>>> --
<br>>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>>> string into a super-saturated solution and a whole operating-system
<br>>>> crystallises out around it."
<br>>>>
<br>>>> /**********************************
<br>>>> * Jules Gosnell
<br>>>> * Partner
<br>>>> * Core Developers Network (Europe)
<br>>>> *
<br>>>> * www.coredevelopers.net
<br>>>> *
<br>>>> * Open Source Training & Support.
<br>>>> **********************************/
<br>>>>
<br>>
<br>>
<br>> --
<br>> "Open Source is a self-assembling organism. You dangle a piece of
<br>> string into a super-saturated solution and a whole operating-system
<br>> crystallises out around it."
<br>>
<br>> /**********************************
<br>> * Jules Gosnell
<br>> * Partner
<br>> * Core Developers Network (Europe)
<br>> *
<br>> * www.coredevelopers.net
<br>> *
<br>> * Open Source Training & Support.
<br>> **********************************/
<br>>
</div><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1528253Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-11-17T03:34:36Z2005-11-17T03:34:36ZJules Gosnell-4Rob,
<br><br>any movement on this issue ?
<br><br>cheers,
<br><br><br>Jules
<br><br><br>Rob Davies wrote:
<br><div class='shrinkable-quote'><br>> Thanks for testing this Jules, will have a look - though I think
<br>> you're right that this is a discovery problem.
<br>>
<br>> cheers,
<br>>
<br>> Rob
<br>> On 7 Nov 2005, at 12:15, Jules Gosnell wrote:
<br>>
<br>>> Jules Gosnell wrote:
<br>>>
<br>>>> Jules Gosnell wrote:
<br>>>>
<br>>>>> Rob,
<br>>>>>
<br>>>>> Thanks for putting some time into this...
<br>>>>>
<br>>>>> I've just rebuilt AC and run the test that I included again. I'm
<br>>>>> afraid that it doesn't fix the problem. Once the second node
<br>>>>> leaves the cluster the first node becomes unable to see any
<br>>>>> subsequent node that wishes to join. It looks as if it is some
<br>>>>> logic problem associated with cluster's dropping to one node, but,
<br>>>>> curiously it works fine with the tcp:// protocol... Maybe it is
<br>>>>> something to do with autodiscovery ? Do tcp:// and peer;// use
<br>>>>> common or separate autodiscovery mechanisms ?
<br>>>>>
<br>>>>> Is there some test that I can run to help pinpoint the problem ?
<br>>>>>
<br>>>>> Sorry to be the bearer of bad news :-(
<br>>>>>
<br>>>>>
<br>>>>> Jules
<br>>>>
<br>>>>
<br>>>>
<br>>>>
<br>>>> P.S.
<br>>>>
<br>>>> I ran my tests on AMQ 3.2-M1 - I guess that I need to run them on
<br>>>> AMQ3-2 to be sure - only problem is that 3.2 breaks our mavens
<br>>>> build with deps that cannot be downloaded - maybe our repo list is
<br>>>> not what it should be - I will see what I can do and get back to
<br>>>> you in a little while.
<br>>>
<br>>>
<br>>> Just ran the test on HEAD of AC and AMQ-3.2 - same result...
<br>>>
<br>>> Everything is fine until a cluster drops down to one node. That node
<br>>> then becomes somehow out of reach of subsequent nodes. Any nodes
<br>>> that are started after this point, only see each other and not the
<br>>> original node. Taking the newly formed cluster down to one member
<br>>> has the same effect, so you can end up with more than one isolated
<br>>> node.
<br>>>
<br>>> Jules
<br>>>
<br>>>>
<br>>>>>
<br>>>>> Rob Davies wrote:
<br>>>>>
<br>>>>>> hi Jules,
<br>>>>>>
<br>>>>>> the latest ActiveCluster in the repo works with ActiveMQ 3.2 -
<br>>>>>> amq4 will hopefully be ready soon. I'd appreciate it if you
<br>>>>>> could build activecluster from src - and see if your problems
<br>>>>>> have disappeared - there was a a nasty deadlock I ended up fixing.
<br>>>>>>
<br>>>>>> cheers,
<br>>>>>>
<br>>>>>> Rob
<br>>>>>>
<br>>>>>> On 6 Nov 2005, at 18:41, Jules Gosnell wrote:
<br>>>>>>
<br>>>>>>> Rob Davies wrote:
<br>>>>>>>
<br>>>>>>>> Hi Jules,
<br>>>>>>>>
<br>>>>>>>> I'm in the process of porting activecluster to amq4. There's
<br>>>>>>>> an issue with discovery in amq4 at the moment that needs to
<br>>>>>>>> be fixed.
<br>>>>>>>
<br>>>>>>>
<br>>>>>>>
<br>>>>>>>
<br>>>>>>>
<br>>>>>>> Ok - no problem - if you let me know when the fix goes in, I
<br>>>>>>> will be happy to retest and report.
<br>>>>>>>
<br>>>>>>> BTW - do you know if AMQ4 is the release that is intended to be
<br>>>>>>> bundled with Geronimo-1.0 ? I am targetting a WADI release at
<br>>>>>>> Geronimo-1.0 and I need to make sure that we run on the same
<br>>>>>>> version of AMQ as is packaged with G.
<br>>>>>>>
<br>>>>>>> Cheers, Rob,
<br>>>>>>>
<br>>>>>>> Jules
<br>>>>>>>
<br>>>>>>>>
<br>>>>>>>> cheers,
<br>>>>>>>>
<br>>>>>>>> Rob
<br>>>>>>>> On 29 Oct 2005, at 02:10, Jules Gosnell wrote:
<br>>>>>>>>
<br>>>>>>>>> Jules Gosnell wrote:
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>> tested against activecluster-1.1-SNAPSHOT, activemq-4.0-
<br>>>>>>>>>> SNAPSHOT and Sun JDK 1.4.2_08 and 1.5.0 - two node cluster,
<br>>>>>>>>>> both on same machine.
<br>>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>> oops - AMQ-4.0 was not on the machine that I ran this on - so
<br>>>>>>>>> this issue is actually with AMQ-3.2-M1.
<br>>>>>>>>>
<br>>>>>>>>> I am trying to run the test with AMQ-4.0, but the second node
<br>>>>>>>>> won't start because "java.io.IOException: Journal is
<br>>>>>>>>> allready opened by another application", So i will have to
<br>>>>>>>>> figure out how to tell it not to journal - I'll be back ;-)
<br>>>>>>>>>
<br>>>>>>>>> Jules
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>> start your first node (red)
<br>>>>>>>>>>
<br>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster red
<br>>>>>>>>>> 2005/10/29 00:04:18:317 BST [INFO] ACCluster - starting...
<br>>>>>>>>>> 2005/10/29 00:04:18:321 BST [INFO] ACCluster - ...started
<br>>>>>>>>>>
<br>>>>>>>>>> start your second node (green)
<br>>>>>>>>>>
<br>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>>>>>> 2005/10/29 00:04:31:874 BST [INFO] ACCluster - starting...
<br>>>>>>>>>> 2005/10/29 00:04:31:885 BST [INFO] ACCluster - ...started
<br>>>>>>>>>> 2005/10/29 00:04:33:323 BST [INFO] ACCluster - onNodeAdd: red
<br>>>>>>>>>>
<br>>>>>>>>>> red says:
<br>>>>>>>>>>
<br>>>>>>>>>> 2005/10/29 00:04:33:536 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>>>>>>> 2005/10/29 00:04:33:537 BST [INFO] ACCluster -
<br>>>>>>>>>> onCoordinatorChanged: green
<br>>>>>>>>>>
<br>>>>>>>>>> ctl-c green, after a few seconds red says:
<br>>>>>>>>>>
<br>>>>>>>>>> 2005/10/29 00:04:46:624 BST [INFO] ACCluster - onNodeFailed:
<br>>>>>>>>>> green
<br>>>>>>>>>> 2005/10/29 00:04:46:626 BST [INFO] ACCluster -
<br>>>>>>>>>> onCoordinatorChanged: red
<br>>>>>>>>>>
<br>>>>>>>>>> all fine so far - now restart green:
<br>>>>>>>>>>
<br>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>>>>>> 2005/10/29 00:04:51:962 BST [INFO] ACCluster - starting...
<br>>>>>>>>>> 2005/10/29 00:04:51:967 BST [INFO] ACCluster - ...started
<br>>>>>>>>>>
<br>>>>>>>>>> red says......nothing.
<br>>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>> if you then start a third node (blue):
<br>>>>>>>>>>
<br>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster blue
<br>>>>>>>>>> 2005/10/29 00:08:47:990 BST [INFO] ACCluster - starting...
<br>>>>>>>>>> 2005/10/29 00:08:47:994 BST [INFO] ACCluster - ...started
<br>>>>>>>>>> 2005/10/29 00:08:50:198 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>>>>>>>
<br>>>>>>>>>> it sees green, but not red and green says:
<br>>>>>>>>>>
<br>>>>>>>>>> 2005/10/29 00:08:49:661 BST [INFO] ACCluster - onNodeAdd: blue
<br>>>>>>>>>> 2005/10/29 00:08:49:662 BST [INFO] ACCluster -
<br>>>>>>>>>> onCoordinatorChanged: blue
<br>>>>>>>>>>
<br>>>>>>>>>> so green sees blue too.
<br>>>>>>>>>>
<br>>>>>>>>>> but red says.....nothing.
<br>>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>> Using tcp:// and a broker, red will happily see blue and
<br>>>>>>>>>> green after they are restarted.
<br>>>>>>>>>>
<br>>>>>>>>>> This has been the case with activecluster/activemq for a
<br>>>>>>>>>> long time, but I was hoping that it would be fixed by
<br>>>>>>>>>> improvements to peer:// in activemq-4.0 - but :-(
<br>>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>> Jules
<br>>>>>>>>>>
<br>>>>>>>>>> ---------------------------------------------------------------
<br>>>>>>>>>> ---- -- ---
<br>>>>>>>>>>
<br>>>>>>>>>> package org.codehaus.wadi.sandbox.partition.impl;
<br>>>>>>>>>>
<br>>>>>>>>>> import java.util.HashMap;
<br>>>>>>>>>> import java.util.Map;
<br>>>>>>>>>>
<br>>>>>>>>>> import javax.jms.JMSException;
<br>>>>>>>>>>
<br>>>>>>>>>> import org.activecluster.ClusterEvent;
<br>>>>>>>>>> import org.activecluster.ClusterFactory;
<br>>>>>>>>>> import org.activecluster.ClusterListener;
<br>>>>>>>>>> import org.activecluster.impl.DefaultClusterFactory;
<br>>>>>>>>>> import org.activemq.ActiveMQConnectionFactory;
<br>>>>>>>>>> import org.activemq.store.vm.VMPersistenceAdapterFactory;
<br>>>>>>>>>> import org.apache.commons.logging.Log;
<br>>>>>>>>>> import org.apache.commons.logging.LogFactory;
<br>>>>>>>>>> import org.codehaus.wadi.dindex.impl.SeniorityElectionStrategy;
<br>>>>>>>>>> import org.codehaus.wadi.impl.CustomClusterFactory;
<br>>>>>>>>>> import org.codehaus.wadi.sandbox.partition.Cluster;
<br>>>>>>>>>>
<br>>>>>>>>>> public class ACCluster {
<br>>>>>>>>>>
<br>>>>>>>>>> protected final Log _log = LogFactory.getLog(getClass());
<br>>>>>>>>>>
<br>>>>>>>>>> //protected final String _clusterUri="tcp://smilodon: 61616";
<br>>>>>>>>>> protected final String _clusterUri="peer://
<br>>>>>>>>>> org.codehaus.wadi";
<br>>>>>>>>>> protected final String _clusterName="ORG.CODEHAUS.WADI.TEST";
<br>>>>>>>>>> protected final ActiveMQConnectionFactory
<br>>>>>>>>>> _connectionFactory=new ActiveMQConnectionFactory(_clusterUri);
<br>>>>>>>>>> protected final ClusterFactory _clusterFactory=new
<br>>>>>>>>>> DefaultClusterFactory(_connectionFactory);
<br>>>>>>>>>> protected final org.activecluster.Cluster _cluster;
<br>>>>>>>>>> protected final long _timeout=30*1000L;
<br>>>>>>>>>>
<br>>>>>>>>>> public ACCluster(String nodeName) throws Exception {
<br>>>>>>>>>> System.setProperty
<br>>>>>>>>>> ("activemq.persistenceAdapterFactory",
<br>>>>>>>>>> VMPersistenceAdapterFactory.class.getName());
<br>>>>>>>>>> _cluster=_clusterFactory.createCluster(_clusterName);
<br>>>>>>>>>> Map state=new HashMap();
<br>>>>>>>>>> state.put("nodeName", nodeName);
<br>>>>>>>>>> _cluster.getLocalNode().setState(state);
<br>>>>>>>>>> _cluster.addClusterListener(new ClusterListener() {
<br>>>>>>>>>>
<br>>>>>>>>>> public void onNodeAdd(ClusterEvent arg0) {
<br>>>>>>>>>> _log.info("onNodeAdd: "+arg0.getNode
<br>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>> public void onNodeUpdate(ClusterEvent arg0) {
<br>>>>>>>>>> _log.info("onNodeUpdate: "+arg0.getNode
<br>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>> public void onNodeRemoved(ClusterEvent arg0) {
<br>>>>>>>>>> _log.info("onNodeRemoved: "+arg0.getNode
<br>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>> public void onNodeFailed(ClusterEvent arg0) {
<br>>>>>>>>>> _log.info("onNodeFailed: "+arg0.getNode
<br>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>> public void onCoordinatorChanged(ClusterEvent
<br>>>>>>>>>> arg0) {
<br>>>>>>>>>> _log.info("onCoordinatorChanged:
<br>>>>>>>>>> "+arg0.getNode ().getState().get("nodeName"));
<br>>>>>>>>>> }
<br>>>>>>>>>> });
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>> public void start() throws JMSException {
<br>>>>>>>>>> _log.info("starting...");
<br>>>>>>>>>> _cluster.start();
<br>>>>>>>>>> _log.info("...started");
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>> public void stop() {
<br>>>>>>>>>>
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>> public static void main(String[] args) throws Exception {
<br>>>>>>>>>> ACCluster cluster=new ACCluster(args[0]);
<br>>>>>>>>>> cluster.start();
<br>>>>>>>>>> Thread.sleep(100*1000);
<br>>>>>>>>>> cluster.stop();
<br>>>>>>>>>> }
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>> --
<br>>>>>>>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>>>>>>>> string into a super-saturated solution and a whole operating-
<br>>>>>>>>> system
<br>>>>>>>>> crystallises out around it."
<br>>>>>>>>>
<br>>>>>>>>> /**********************************
<br>>>>>>>>> * Jules Gosnell
<br>>>>>>>>> * Partner
<br>>>>>>>>> * Core Developers Network (Europe)
<br>>>>>>>>> *
<br>>>>>>>>> * www.coredevelopers.net
<br>>>>>>>>> *
<br>>>>>>>>> * Open Source Training & Support.
<br>>>>>>>>> **********************************/
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>
<br>>>>>>>
<br>>>>>>> --
<br>>>>>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>>>>>> string into a super-saturated solution and a whole operating- system
<br>>>>>>> crystallises out around it."
<br>>>>>>>
<br>>>>>>> /**********************************
<br>>>>>>> * Jules Gosnell
<br>>>>>>> * Partner
<br>>>>>>> * Core Developers Network (Europe)
<br>>>>>>> *
<br>>>>>>> * www.coredevelopers.net
<br>>>>>>> *
<br>>>>>>> * Open Source Training & Support.
<br>>>>>>> **********************************/
<br>>>>>>>
<br>>>>>
<br>>>>>
<br>>>>
<br>>>>
<br>>>
<br>>>
<br>>> --
<br>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>> string into a super-saturated solution and a whole operating-system
<br>>> crystallises out around it."
<br>>>
<br>>> /**********************************
<br>>> * Jules Gosnell
<br>>> * Partner
<br>>> * Core Developers Network (Europe)
<br>>> *
<br>>> * www.coredevelopers.net
<br>>> *
<br>>> * Open Source Training & Support.
<br>>> **********************************/
<br>>>
</div><br><br>--
<br>"Open Source is a self-assembling organism. You dangle a piece of
<br>string into a super-saturated solution and a whole operating-system
<br>crystallises out around it."
<br><br>/**********************************
<br> * Jules Gosnell
<br> * Partner
<br> * Core Developers Network (Europe)
<br> *
<br> * www.coredevelopers.net
<br> *
<br> * Open Source Training & Support.
<br> **********************************/
<br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1501408Re: [activecluster-user] ActiveCluster 1.1 and SonicMQ2005-11-15T13:14:45Z2005-11-15T13:14:45Zben.kibler<br>Thanks Rob,
<br><br>I made the same change (switching to String) locally, and that resolves the
<br>serialization problems with SonicMQ 6.1.
<br><br>Regards,
<br>Ben
<br><br><br><br>|---------+---------------------------->
<br>| | Rob Davies |
<br>| | <rajdavies@gmail.|
<br>| | com> |
<br>| | |
<br>| | 11/15/2005 12:51 |
<br>| | PM |
<br>| | Please respond to|
<br>| | user |
<br>| | |
<br>|---------+---------------------------->
<br> >------------------------------------------------------------------------------------------------------------------------------|
<br> | |
<br> | To: <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=1501408&i=0" target="_top" rel="nofollow">user@...</a> |
<br> | cc: |
<br> | Subject: Re: [activecluster-user] ActiveCluster 1.1 and SonicMQ |
<br> >------------------------------------------------------------------------------------------------------------------------------|
<br><br><br><br><br>Hi Ben,
<br><br>the latest in CVS head of ActiveCluster has changed to use Strings in
<br>the serialized nodes instead of jms Destinations.
<br>I haven't tried it with SonicMQ - so I don't know if it will fix
<br>everything - but fingers crossed :)
<br><br>cheers,
<br><br>Rob
<br>On 9 Nov 2005, at 17:03, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=1501408&i=1" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<br><div class='shrinkable-quote'><br>>
<br>> I'm trying to use ActiveCluster 1.1 with SonicMQ 6.1, with no
<br>> success. The
<br>> following error occurs at startup:
<br>>
<br>> 7389 [Timer-0] ERROR impl.StateServiceStub - Could not send JMS
<br>> message:
<br>> javax.jms.MessageFormatException: progress.message.jimpl.Connection
<br>> javax.jms.MessageFormatException: progress.message.jimpl.Connection
<br>> at progress.message.jimpl.ObjectMessage.setObject(Unknown
<br>> Source)
<br>> at progress.message.jimpl.Session.createObjectMessage(Unknown
<br>> Source)
<br>> at org.activecluster.impl.StateServiceStub.keepAlive(
<br>> StateServiceStub.java:54)
<br>> at org.activecluster.impl.ReplicatedLocalNode.pingRemoteNodes(
<br>> ReplicatedLocalNode.java:49)
<br>> at org.activecluster.impl.DefaultCluster$1.run
<br>> (DefaultCluster.java:74
<br>> )
<br>> at org.activecluster.impl.StateServiceImpl.checkForTimeouts(
<br>> StateServiceImpl.java:150)
<br>> at org.activecluster.impl.StateServiceImpl$1.run(
<br>> StateServiceImpl.java:165)
<br>> at java.util.TimerThread.mainLoop(Unknown Source)
<br>> at java.util.TimerThread.run(Unknown Source)
<br>>
<br>> I'm fairly sure the problem is due to the fact that SonicMQ JMS
<br>> objects
<br>> don't implement Serializable, while ActiveMQ JMS objects do. The
<br>> ActiveCluster NodeImpl class assumes it can be serialized via
<br>> Session.createObjectMessage(), but contains a member variable with a
<br>> Destination object, implemented by the JMS provider.
<br>>
<br>> Has anyone ever tried to use this with Sonic MQ before?
<br>>
<br>> Thanks,
<br>> Ben
<br>>
<br>> ---------------------------------------------
<br>> <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=1501408&i=2" target="_top" rel="nofollow">ben.kibler@...</a>
<br>> +1-212-622-4653
<br>>
<br>>
<br>>
<br>> This communication is for informational purposes only. It is not
<br>> intended
<br>> as an offer or solicitation for the purchase or sale of any financial
<br>> instrument or as an official confirmation of any transaction. All
<br>> market prices,
<br>> data and other information are not warranted as to completeness or
<br>> accuracy and
<br>> are subject to change without notice. Any comments or statements
<br>> made herein
<br>> do not necessarily reflect those of JPMorgan Chase & Co., its
<br>> subsidiaries
<br>> and affiliates.
<br>>
</div><br><br><br><br><br>This communication is for informational purposes only. It is not intended
<br>as an offer or solicitation for the purchase or sale of any financial
<br>instrument or as an official confirmation of any transaction. All market prices,
<br>data and other information are not warranted as to completeness or accuracy and
<br>are subject to change without notice. Any comments or statements made herein
<br>do not necessarily reflect those of JPMorgan Chase & Co., its subsidiaries
<br>and affiliates
<br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1497928Re: [activecluster-user] ActiveCluster 1.1 and SonicMQ2005-11-15T09:51:03Z2005-11-15T09:51:03ZrajdaviesHi Ben,
<br><br>the latest in CVS head of ActiveCluster has changed to use Strings in
<br>the serialized nodes instead of jms Destinations.
<br>I haven't tried it with SonicMQ - so I don't know if it will fix
<br>everything - but fingers crossed :)
<br><br>cheers,
<br><br>Rob
<br>On 9 Nov 2005, at 17:03, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=1497928&i=0" target="_top" rel="nofollow">ben.kibler@...</a> wrote:
<br><div class='shrinkable-quote'><br>>
<br>> I'm trying to use ActiveCluster 1.1 with SonicMQ 6.1, with no
<br>> success. The
<br>> following error occurs at startup:
<br>>
<br>> 7389 [Timer-0] ERROR impl.StateServiceStub - Could not send JMS
<br>> message:
<br>> javax.jms.MessageFormatException: progress.message.jimpl.Connection
<br>> javax.jms.MessageFormatException: progress.message.jimpl.Connection
<br>> at progress.message.jimpl.ObjectMessage.setObject(Unknown
<br>> Source)
<br>> at progress.message.jimpl.Session.createObjectMessage(Unknown
<br>> Source)
<br>> at org.activecluster.impl.StateServiceStub.keepAlive(
<br>> StateServiceStub.java:54)
<br>> at org.activecluster.impl.ReplicatedLocalNode.pingRemoteNodes(
<br>> ReplicatedLocalNode.java:49)
<br>> at org.activecluster.impl.DefaultCluster$1.run
<br>> (DefaultCluster.java:74
<br>> )
<br>> at org.activecluster.impl.StateServiceImpl.checkForTimeouts(
<br>> StateServiceImpl.java:150)
<br>> at org.activecluster.impl.StateServiceImpl$1.run(
<br>> StateServiceImpl.java:165)
<br>> at java.util.TimerThread.mainLoop(Unknown Source)
<br>> at java.util.TimerThread.run(Unknown Source)
<br>>
<br>> I'm fairly sure the problem is due to the fact that SonicMQ JMS
<br>> objects
<br>> don't implement Serializable, while ActiveMQ JMS objects do. The
<br>> ActiveCluster NodeImpl class assumes it can be serialized via
<br>> Session.createObjectMessage(), but contains a member variable with a
<br>> Destination object, implemented by the JMS provider.
<br>>
<br>> Has anyone ever tried to use this with Sonic MQ before?
<br>>
<br>> Thanks,
<br>> Ben
<br>>
<br>> ---------------------------------------------
<br>> <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=1497928&i=1" target="_top" rel="nofollow">ben.kibler@...</a>
<br>> +1-212-622-4653
<br>>
<br>>
<br>>
<br>> This communication is for informational purposes only. It is not
<br>> intended
<br>> as an offer or solicitation for the purchase or sale of any financial
<br>> instrument or as an official confirmation of any transaction. All
<br>> market prices,
<br>> data and other information are not warranted as to completeness or
<br>> accuracy and
<br>> are subject to change without notice. Any comments or statements
<br>> made herein
<br>> do not necessarily reflect those of JPMorgan Chase & Co., its
<br>> subsidiaries
<br>> and affiliates.
<br>>
</div><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1407542[activecluster-user] ActiveCluster 1.1 and SonicMQ2005-11-09T09:03:09Z2005-11-09T09:03:09Zben.kibler<br>I'm trying to use ActiveCluster 1.1 with SonicMQ 6.1, with no success. The
<br>following error occurs at startup:
<br><br>7389 [Timer-0] ERROR impl.StateServiceStub - Could not send JMS message:
<br>javax.jms.MessageFormatException: progress.message.jimpl.Connection
<br>javax.jms.MessageFormatException: progress.message.jimpl.Connection
<br> at progress.message.jimpl.ObjectMessage.setObject(Unknown Source)
<br> at progress.message.jimpl.Session.createObjectMessage(Unknown Source)
<br> at org.activecluster.impl.StateServiceStub.keepAlive(
<br>StateServiceStub.java:54)
<br> at org.activecluster.impl.ReplicatedLocalNode.pingRemoteNodes(
<br>ReplicatedLocalNode.java:49)
<br> at org.activecluster.impl.DefaultCluster$1.run(DefaultCluster.java:74
<br>)
<br> at org.activecluster.impl.StateServiceImpl.checkForTimeouts(
<br>StateServiceImpl.java:150)
<br> at org.activecluster.impl.StateServiceImpl$1.run(
<br>StateServiceImpl.java:165)
<br> at java.util.TimerThread.mainLoop(Unknown Source)
<br> at java.util.TimerThread.run(Unknown Source)
<br><br>I'm fairly sure the problem is due to the fact that SonicMQ JMS objects
<br>don't implement Serializable, while ActiveMQ JMS objects do. The
<br>ActiveCluster NodeImpl class assumes it can be serialized via
<br>Session.createObjectMessage(), but contains a member variable with a
<br>Destination object, implemented by the JMS provider.
<br><br>Has anyone ever tried to use this with Sonic MQ before?
<br><br>Thanks,
<br>Ben
<br><br>---------------------------------------------
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=1407542&i=0" target="_top" rel="nofollow">ben.kibler@...</a>
<br>+1-212-622-4653
<br><br><br><br>This communication is for informational purposes only. It is not intended
<br>as an offer or solicitation for the purchase or sale of any financial
<br>instrument or as an official confirmation of any transaction. All market prices,
<br>data and other information are not warranted as to completeness or accuracy and
<br>are subject to change without notice. Any comments or statements made herein
<br>do not necessarily reflect those of JPMorgan Chase & Co., its subsidiaries
<br>and affiliates.
<br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1366683Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-11-07T04:25:46Z2005-11-07T04:25:46ZJules Gosnell-4Rob Davies wrote:
<br><br>> Thanks for testing this Jules, will have a look - though I think
<br>> you're right that this is a discovery problem.
<br><br>cheers, Rob
<br><div class='shrinkable-quote'><br>>
<br>> cheers,
<br>>
<br>> Rob
<br>> On 7 Nov 2005, at 12:15, Jules Gosnell wrote:
<br>>
<br>>> Jules Gosnell wrote:
<br>>>
<br>>>> Jules Gosnell wrote:
<br>>>>
<br>>>>> Rob,
<br>>>>>
<br>>>>> Thanks for putting some time into this...
<br>>>>>
<br>>>>> I've just rebuilt AC and run the test that I included again. I'm
<br>>>>> afraid that it doesn't fix the problem. Once the second node
<br>>>>> leaves the cluster the first node becomes unable to see any
<br>>>>> subsequent node that wishes to join. It looks as if it is some
<br>>>>> logic problem associated with cluster's dropping to one node, but,
<br>>>>> curiously it works fine with the tcp:// protocol... Maybe it is
<br>>>>> something to do with autodiscovery ? Do tcp:// and peer;// use
<br>>>>> common or separate autodiscovery mechanisms ?
<br>>>>>
<br>>>>> Is there some test that I can run to help pinpoint the problem ?
<br>>>>>
<br>>>>> Sorry to be the bearer of bad news :-(
<br>>>>>
<br>>>>>
<br>>>>> Jules
<br>>>>
<br>>>>
<br>>>>
<br>>>>
<br>>>> P.S.
<br>>>>
<br>>>> I ran my tests on AMQ 3.2-M1 - I guess that I need to run them on
<br>>>> AMQ3-2 to be sure - only problem is that 3.2 breaks our mavens
<br>>>> build with deps that cannot be downloaded - maybe our repo list is
<br>>>> not what it should be - I will see what I can do and get back to
<br>>>> you in a little while.
<br>>>
<br>>>
<br>>> Just ran the test on HEAD of AC and AMQ-3.2 - same result...
<br>>>
<br>>> Everything is fine until a cluster drops down to one node. That node
<br>>> then becomes somehow out of reach of subsequent nodes. Any nodes
<br>>> that are started after this point, only see each other and not the
<br>>> original node. Taking the newly formed cluster down to one member
<br>>> has the same effect, so you can end up with more than one isolated
<br>>> node.
<br>>>
<br>>> Jules
<br>>>
<br>>>>
<br>>>>>
<br>>>>> Rob Davies wrote:
<br>>>>>
<br>>>>>> hi Jules,
<br>>>>>>
<br>>>>>> the latest ActiveCluster in the repo works with ActiveMQ 3.2 -
<br>>>>>> amq4 will hopefully be ready soon. I'd appreciate it if you
<br>>>>>> could build activecluster from src - and see if your problems
<br>>>>>> have disappeared - there was a a nasty deadlock I ended up fixing.
<br>>>>>>
<br>>>>>> cheers,
<br>>>>>>
<br>>>>>> Rob
<br>>>>>>
<br>>>>>> On 6 Nov 2005, at 18:41, Jules Gosnell wrote:
<br>>>>>>
<br>>>>>>> Rob Davies wrote:
<br>>>>>>>
<br>>>>>>>> Hi Jules,
<br>>>>>>>>
<br>>>>>>>> I'm in the process of porting activecluster to amq4. There's
<br>>>>>>>> an issue with discovery in amq4 at the moment that needs to
<br>>>>>>>> be fixed.
<br>>>>>>>
<br>>>>>>>
<br>>>>>>>
<br>>>>>>>
<br>>>>>>>
<br>>>>>>> Ok - no problem - if you let me know when the fix goes in, I
<br>>>>>>> will be happy to retest and report.
<br>>>>>>>
<br>>>>>>> BTW - do you know if AMQ4 is the release that is intended to be
<br>>>>>>> bundled with Geronimo-1.0 ? I am targetting a WADI release at
<br>>>>>>> Geronimo-1.0 and I need to make sure that we run on the same
<br>>>>>>> version of AMQ as is packaged with G.
<br>>>>>>>
<br>>>>>>> Cheers, Rob,
<br>>>>>>>
<br>>>>>>> Jules
<br>>>>>>>
<br>>>>>>>>
<br>>>>>>>> cheers,
<br>>>>>>>>
<br>>>>>>>> Rob
<br>>>>>>>> On 29 Oct 2005, at 02:10, Jules Gosnell wrote:
<br>>>>>>>>
<br>>>>>>>>> Jules Gosnell wrote:
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>> tested against activecluster-1.1-SNAPSHOT, activemq-4.0-
<br>>>>>>>>>> SNAPSHOT and Sun JDK 1.4.2_08 and 1.5.0 - two node cluster,
<br>>>>>>>>>> both on same machine.
<br>>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>> oops - AMQ-4.0 was not on the machine that I ran this on - so
<br>>>>>>>>> this issue is actually with AMQ-3.2-M1.
<br>>>>>>>>>
<br>>>>>>>>> I am trying to run the test with AMQ-4.0, but the second node
<br>>>>>>>>> won't start because "java.io.IOException: Journal is
<br>>>>>>>>> allready opened by another application", So i will have to
<br>>>>>>>>> figure out how to tell it not to journal - I'll be back ;-)
<br>>>>>>>>>
<br>>>>>>>>> Jules
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>> start your first node (red)
<br>>>>>>>>>>
<br>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster red
<br>>>>>>>>>> 2005/10/29 00:04:18:317 BST [INFO] ACCluster - starting...
<br>>>>>>>>>> 2005/10/29 00:04:18:321 BST [INFO] ACCluster - ...started
<br>>>>>>>>>>
<br>>>>>>>>>> start your second node (green)
<br>>>>>>>>>>
<br>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>>>>>> 2005/10/29 00:04:31:874 BST [INFO] ACCluster - starting...
<br>>>>>>>>>> 2005/10/29 00:04:31:885 BST [INFO] ACCluster - ...started
<br>>>>>>>>>> 2005/10/29 00:04:33:323 BST [INFO] ACCluster - onNodeAdd: red
<br>>>>>>>>>>
<br>>>>>>>>>> red says:
<br>>>>>>>>>>
<br>>>>>>>>>> 2005/10/29 00:04:33:536 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>>>>>>> 2005/10/29 00:04:33:537 BST [INFO] ACCluster -
<br>>>>>>>>>> onCoordinatorChanged: green
<br>>>>>>>>>>
<br>>>>>>>>>> ctl-c green, after a few seconds red says:
<br>>>>>>>>>>
<br>>>>>>>>>> 2005/10/29 00:04:46:624 BST [INFO] ACCluster - onNodeFailed:
<br>>>>>>>>>> green
<br>>>>>>>>>> 2005/10/29 00:04:46:626 BST [INFO] ACCluster -
<br>>>>>>>>>> onCoordinatorChanged: red
<br>>>>>>>>>>
<br>>>>>>>>>> all fine so far - now restart green:
<br>>>>>>>>>>
<br>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>>>>>> 2005/10/29 00:04:51:962 BST [INFO] ACCluster - starting...
<br>>>>>>>>>> 2005/10/29 00:04:51:967 BST [INFO] ACCluster - ...started
<br>>>>>>>>>>
<br>>>>>>>>>> red says......nothing.
<br>>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>> if you then start a third node (blue):
<br>>>>>>>>>>
<br>>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster blue
<br>>>>>>>>>> 2005/10/29 00:08:47:990 BST [INFO] ACCluster - starting...
<br>>>>>>>>>> 2005/10/29 00:08:47:994 BST [INFO] ACCluster - ...started
<br>>>>>>>>>> 2005/10/29 00:08:50:198 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>>>>>>>
<br>>>>>>>>>> it sees green, but not red and green says:
<br>>>>>>>>>>
<br>>>>>>>>>> 2005/10/29 00:08:49:661 BST [INFO] ACCluster - onNodeAdd: blue
<br>>>>>>>>>> 2005/10/29 00:08:49:662 BST [INFO] ACCluster -
<br>>>>>>>>>> onCoordinatorChanged: blue
<br>>>>>>>>>>
<br>>>>>>>>>> so green sees blue too.
<br>>>>>>>>>>
<br>>>>>>>>>> but red says.....nothing.
<br>>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>> Using tcp:// and a broker, red will happily see blue and
<br>>>>>>>>>> green after they are restarted.
<br>>>>>>>>>>
<br>>>>>>>>>> This has been the case with activecluster/activemq for a
<br>>>>>>>>>> long time, but I was hoping that it would be fixed by
<br>>>>>>>>>> improvements to peer:// in activemq-4.0 - but :-(
<br>>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>> Jules
<br>>>>>>>>>>
<br>>>>>>>>>> ---------------------------------------------------------------
<br>>>>>>>>>> ---- -- ---
<br>>>>>>>>>>
<br>>>>>>>>>> package org.codehaus.wadi.sandbox.partition.impl;
<br>>>>>>>>>>
<br>>>>>>>>>> import java.util.HashMap;
<br>>>>>>>>>> import java.util.Map;
<br>>>>>>>>>>
<br>>>>>>>>>> import javax.jms.JMSException;
<br>>>>>>>>>>
<br>>>>>>>>>> import org.activecluster.ClusterEvent;
<br>>>>>>>>>> import org.activecluster.ClusterFactory;
<br>>>>>>>>>> import org.activecluster.ClusterListener;
<br>>>>>>>>>> import org.activecluster.impl.DefaultClusterFactory;
<br>>>>>>>>>> import org.activemq.ActiveMQConnectionFactory;
<br>>>>>>>>>> import org.activemq.store.vm.VMPersistenceAdapterFactory;
<br>>>>>>>>>> import org.apache.commons.logging.Log;
<br>>>>>>>>>> import org.apache.commons.logging.LogFactory;
<br>>>>>>>>>> import org.codehaus.wadi.dindex.impl.SeniorityElectionStrategy;
<br>>>>>>>>>> import org.codehaus.wadi.impl.CustomClusterFactory;
<br>>>>>>>>>> import org.codehaus.wadi.sandbox.partition.Cluster;
<br>>>>>>>>>>
<br>>>>>>>>>> public class ACCluster {
<br>>>>>>>>>>
<br>>>>>>>>>> protected final Log _log = LogFactory.getLog(getClass());
<br>>>>>>>>>>
<br>>>>>>>>>> //protected final String _clusterUri="tcp://smilodon: 61616";
<br>>>>>>>>>> protected final String _clusterUri="peer://
<br>>>>>>>>>> org.codehaus.wadi";
<br>>>>>>>>>> protected final String _clusterName="ORG.CODEHAUS.WADI.TEST";
<br>>>>>>>>>> protected final ActiveMQConnectionFactory
<br>>>>>>>>>> _connectionFactory=new ActiveMQConnectionFactory(_clusterUri);
<br>>>>>>>>>> protected final ClusterFactory _clusterFactory=new
<br>>>>>>>>>> DefaultClusterFactory(_connectionFactory);
<br>>>>>>>>>> protected final org.activecluster.Cluster _cluster;
<br>>>>>>>>>> protected final long _timeout=30*1000L;
<br>>>>>>>>>>
<br>>>>>>>>>> public ACCluster(String nodeName) throws Exception {
<br>>>>>>>>>> System.setProperty
<br>>>>>>>>>> ("activemq.persistenceAdapterFactory",
<br>>>>>>>>>> VMPersistenceAdapterFactory.class.getName());
<br>>>>>>>>>> _cluster=_clusterFactory.createCluster(_clusterName);
<br>>>>>>>>>> Map state=new HashMap();
<br>>>>>>>>>> state.put("nodeName", nodeName);
<br>>>>>>>>>> _cluster.getLocalNode().setState(state);
<br>>>>>>>>>> _cluster.addClusterListener(new ClusterListener() {
<br>>>>>>>>>>
<br>>>>>>>>>> public void onNodeAdd(ClusterEvent arg0) {
<br>>>>>>>>>> _log.info("onNodeAdd: "+arg0.getNode
<br>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>> public void onNodeUpdate(ClusterEvent arg0) {
<br>>>>>>>>>> _log.info("onNodeUpdate: "+arg0.getNode
<br>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>> public void onNodeRemoved(ClusterEvent arg0) {
<br>>>>>>>>>> _log.info("onNodeRemoved: "+arg0.getNode
<br>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>> public void onNodeFailed(ClusterEvent arg0) {
<br>>>>>>>>>> _log.info("onNodeFailed: "+arg0.getNode
<br>>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>> public void onCoordinatorChanged(ClusterEvent
<br>>>>>>>>>> arg0) {
<br>>>>>>>>>> _log.info("onCoordinatorChanged:
<br>>>>>>>>>> "+arg0.getNode ().getState().get("nodeName"));
<br>>>>>>>>>> }
<br>>>>>>>>>> });
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>> public void start() throws JMSException {
<br>>>>>>>>>> _log.info("starting...");
<br>>>>>>>>>> _cluster.start();
<br>>>>>>>>>> _log.info("...started");
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>> public void stop() {
<br>>>>>>>>>>
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>> public static void main(String[] args) throws Exception {
<br>>>>>>>>>> ACCluster cluster=new ACCluster(args[0]);
<br>>>>>>>>>> cluster.start();
<br>>>>>>>>>> Thread.sleep(100*1000);
<br>>>>>>>>>> cluster.stop();
<br>>>>>>>>>> }
<br>>>>>>>>>> }
<br>>>>>>>>>>
<br>>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>> --
<br>>>>>>>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>>>>>>>> string into a super-saturated solution and a whole operating-
<br>>>>>>>>> system
<br>>>>>>>>> crystallises out around it."
<br>>>>>>>>>
<br>>>>>>>>> /**********************************
<br>>>>>>>>> * Jules Gosnell
<br>>>>>>>>> * Partner
<br>>>>>>>>> * Core Developers Network (Europe)
<br>>>>>>>>> *
<br>>>>>>>>> * www.coredevelopers.net
<br>>>>>>>>> *
<br>>>>>>>>> * Open Source Training & Support.
<br>>>>>>>>> **********************************/
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>
<br>>>>>>>
<br>>>>>>> --
<br>>>>>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>>>>>> string into a super-saturated solution and a whole operating- system
<br>>>>>>> crystallises out around it."
<br>>>>>>>
<br>>>>>>> /**********************************
<br>>>>>>> * Jules Gosnell
<br>>>>>>> * Partner
<br>>>>>>> * Core Developers Network (Europe)
<br>>>>>>> *
<br>>>>>>> * www.coredevelopers.net
<br>>>>>>> *
<br>>>>>>> * Open Source Training & Support.
<br>>>>>>> **********************************/
<br>>>>>>>
<br>>>>>
<br>>>>>
<br>>>>
<br>>>>
<br>>>
<br>>>
<br>>> --
<br>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>> string into a super-saturated solution and a whole operating-system
<br>>> crystallises out around it."
<br>>>
<br>>> /**********************************
<br>>> * Jules Gosnell
<br>>> * Partner
<br>>> * Core Developers Network (Europe)
<br>>> *
<br>>> * www.coredevelopers.net
<br>>> *
<br>>> * Open Source Training & Support.
<br>>> **********************************/
<br>>>
</div><br><br>--
<br>"Open Source is a self-assembling organism. You dangle a piece of
<br>string into a super-saturated solution and a whole operating-system
<br>crystallises out around it."
<br><br>/**********************************
<br> * Jules Gosnell
<br> * Partner
<br> * Core Developers Network (Europe)
<br> *
<br> * www.coredevelopers.net
<br> *
<br> * Open Source Training & Support.
<br> **********************************/
<br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1366650Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-11-07T04:23:07Z2005-11-07T04:23:07ZrajdaviesThanks for testing this Jules, will have a look - though I think
<br>you're right that this is a discovery problem.
<br><br>cheers,
<br><br>Rob
<br>On 7 Nov 2005, at 12:15, Jules Gosnell wrote:
<br><div class='shrinkable-quote'><br>> Jules Gosnell wrote:
<br>>
<br>>> Jules Gosnell wrote:
<br>>>
<br>>>> Rob,
<br>>>>
<br>>>> Thanks for putting some time into this...
<br>>>>
<br>>>> I've just rebuilt AC and run the test that I included again. I'm
<br>>>> afraid that it doesn't fix the problem. Once the second node
<br>>>> leaves the cluster the first node becomes unable to see any
<br>>>> subsequent node that wishes to join. It looks as if it is some
<br>>>> logic problem associated with cluster's dropping to one node,
<br>>>> but, curiously it works fine with the tcp:// protocol... Maybe it
<br>>>> is something to do with autodiscovery ? Do tcp:// and peer;// use
<br>>>> common or separate autodiscovery mechanisms ?
<br>>>>
<br>>>> Is there some test that I can run to help pinpoint the problem ?
<br>>>>
<br>>>> Sorry to be the bearer of bad news :-(
<br>>>>
<br>>>>
<br>>>> Jules
<br>>>
<br>>>
<br>>>
<br>>> P.S.
<br>>>
<br>>> I ran my tests on AMQ 3.2-M1 - I guess that I need to run them on
<br>>> AMQ3-2 to be sure - only problem is that 3.2 breaks our mavens
<br>>> build with deps that cannot be downloaded - maybe our repo list is
<br>>> not what it should be - I will see what I can do and get back to
<br>>> you in a little while.
<br>>
<br>> Just ran the test on HEAD of AC and AMQ-3.2 - same result...
<br>>
<br>> Everything is fine until a cluster drops down to one node. That
<br>> node then becomes somehow out of reach of subsequent nodes. Any
<br>> nodes that are started after this point, only see each other and
<br>> not the original node. Taking the newly formed cluster down to one
<br>> member has the same effect, so you can end up with more than one
<br>> isolated node.
<br>>
<br>> Jules
<br>>
<br>>>
<br>>>>
<br>>>> Rob Davies wrote:
<br>>>>
<br>>>>> hi Jules,
<br>>>>>
<br>>>>> the latest ActiveCluster in the repo works with ActiveMQ 3.2 -
<br>>>>> amq4 will hopefully be ready soon. I'd appreciate it if you
<br>>>>> could build activecluster from src - and see if your problems
<br>>>>> have disappeared - there was a a nasty deadlock I ended up
<br>>>>> fixing.
<br>>>>>
<br>>>>> cheers,
<br>>>>>
<br>>>>> Rob
<br>>>>>
<br>>>>> On 6 Nov 2005, at 18:41, Jules Gosnell wrote:
<br>>>>>
<br>>>>>> Rob Davies wrote:
<br>>>>>>
<br>>>>>>> Hi Jules,
<br>>>>>>>
<br>>>>>>> I'm in the process of porting activecluster to amq4. There's
<br>>>>>>> an issue with discovery in amq4 at the moment that needs to
<br>>>>>>> be fixed.
<br>>>>>>
<br>>>>>>
<br>>>>>>
<br>>>>>>
<br>>>>>> Ok - no problem - if you let me know when the fix goes in, I
<br>>>>>> will be happy to retest and report.
<br>>>>>>
<br>>>>>> BTW - do you know if AMQ4 is the release that is intended to
<br>>>>>> be bundled with Geronimo-1.0 ? I am targetting a WADI release
<br>>>>>> at Geronimo-1.0 and I need to make sure that we run on the
<br>>>>>> same version of AMQ as is packaged with G.
<br>>>>>>
<br>>>>>> Cheers, Rob,
<br>>>>>>
<br>>>>>> Jules
<br>>>>>>
<br>>>>>>>
<br>>>>>>> cheers,
<br>>>>>>>
<br>>>>>>> Rob
<br>>>>>>> On 29 Oct 2005, at 02:10, Jules Gosnell wrote:
<br>>>>>>>
<br>>>>>>>> Jules Gosnell wrote:
<br>>>>>>>>
<br>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>> tested against activecluster-1.1-SNAPSHOT, activemq-4.0-
<br>>>>>>>>> SNAPSHOT and Sun JDK 1.4.2_08 and 1.5.0 - two node
<br>>>>>>>>> cluster, both on same machine.
<br>>>>>>>>>
<br>>>>>>>>
<br>>>>>>>> oops - AMQ-4.0 was not on the machine that I ran this on -
<br>>>>>>>> so this issue is actually with AMQ-3.2-M1.
<br>>>>>>>>
<br>>>>>>>> I am trying to run the test with AMQ-4.0, but the second
<br>>>>>>>> node won't start because "java.io.IOException: Journal is
<br>>>>>>>> allready opened by another application", So i will have to
<br>>>>>>>> figure out how to tell it not to journal - I'll be back ;-)
<br>>>>>>>>
<br>>>>>>>> Jules
<br>>>>>>>>
<br>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>> start your first node (red)
<br>>>>>>>>>
<br>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster red
<br>>>>>>>>> 2005/10/29 00:04:18:317 BST [INFO] ACCluster - starting...
<br>>>>>>>>> 2005/10/29 00:04:18:321 BST [INFO] ACCluster - ...started
<br>>>>>>>>>
<br>>>>>>>>> start your second node (green)
<br>>>>>>>>>
<br>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>>>>> 2005/10/29 00:04:31:874 BST [INFO] ACCluster - starting...
<br>>>>>>>>> 2005/10/29 00:04:31:885 BST [INFO] ACCluster - ...started
<br>>>>>>>>> 2005/10/29 00:04:33:323 BST [INFO] ACCluster - onNodeAdd: red
<br>>>>>>>>>
<br>>>>>>>>> red says:
<br>>>>>>>>>
<br>>>>>>>>> 2005/10/29 00:04:33:536 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>>>>>> 2005/10/29 00:04:33:537 BST [INFO] ACCluster -
<br>>>>>>>>> onCoordinatorChanged: green
<br>>>>>>>>>
<br>>>>>>>>> ctl-c green, after a few seconds red says:
<br>>>>>>>>>
<br>>>>>>>>> 2005/10/29 00:04:46:624 BST [INFO] ACCluster - onNodeFailed:
<br>>>>>>>>> green
<br>>>>>>>>> 2005/10/29 00:04:46:626 BST [INFO] ACCluster -
<br>>>>>>>>> onCoordinatorChanged: red
<br>>>>>>>>>
<br>>>>>>>>> all fine so far - now restart green:
<br>>>>>>>>>
<br>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>>>>> 2005/10/29 00:04:51:962 BST [INFO] ACCluster - starting...
<br>>>>>>>>> 2005/10/29 00:04:51:967 BST [INFO] ACCluster - ...started
<br>>>>>>>>>
<br>>>>>>>>> red says......nothing.
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>> if you then start a third node (blue):
<br>>>>>>>>>
<br>>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster blue
<br>>>>>>>>> 2005/10/29 00:08:47:990 BST [INFO] ACCluster - starting...
<br>>>>>>>>> 2005/10/29 00:08:47:994 BST [INFO] ACCluster - ...started
<br>>>>>>>>> 2005/10/29 00:08:50:198 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>>>>>>
<br>>>>>>>>> it sees green, but not red and green says:
<br>>>>>>>>>
<br>>>>>>>>> 2005/10/29 00:08:49:661 BST [INFO] ACCluster - onNodeAdd: blue
<br>>>>>>>>> 2005/10/29 00:08:49:662 BST [INFO] ACCluster -
<br>>>>>>>>> onCoordinatorChanged: blue
<br>>>>>>>>>
<br>>>>>>>>> so green sees blue too.
<br>>>>>>>>>
<br>>>>>>>>> but red says.....nothing.
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>> Using tcp:// and a broker, red will happily see blue and
<br>>>>>>>>> green after they are restarted.
<br>>>>>>>>>
<br>>>>>>>>> This has been the case with activecluster/activemq for a
<br>>>>>>>>> long time, but I was hoping that it would be fixed by
<br>>>>>>>>> improvements to peer:// in activemq-4.0 - but :-(
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>> Jules
<br>>>>>>>>>
<br>>>>>>>>> ---------------------------------------------------------------
<br>>>>>>>>> ---- -- ---
<br>>>>>>>>>
<br>>>>>>>>> package org.codehaus.wadi.sandbox.partition.impl;
<br>>>>>>>>>
<br>>>>>>>>> import java.util.HashMap;
<br>>>>>>>>> import java.util.Map;
<br>>>>>>>>>
<br>>>>>>>>> import javax.jms.JMSException;
<br>>>>>>>>>
<br>>>>>>>>> import org.activecluster.ClusterEvent;
<br>>>>>>>>> import org.activecluster.ClusterFactory;
<br>>>>>>>>> import org.activecluster.ClusterListener;
<br>>>>>>>>> import org.activecluster.impl.DefaultClusterFactory;
<br>>>>>>>>> import org.activemq.ActiveMQConnectionFactory;
<br>>>>>>>>> import org.activemq.store.vm.VMPersistenceAdapterFactory;
<br>>>>>>>>> import org.apache.commons.logging.Log;
<br>>>>>>>>> import org.apache.commons.logging.LogFactory;
<br>>>>>>>>> import org.codehaus.wadi.dindex.impl.SeniorityElectionStrategy;
<br>>>>>>>>> import org.codehaus.wadi.impl.CustomClusterFactory;
<br>>>>>>>>> import org.codehaus.wadi.sandbox.partition.Cluster;
<br>>>>>>>>>
<br>>>>>>>>> public class ACCluster {
<br>>>>>>>>>
<br>>>>>>>>> protected final Log _log = LogFactory.getLog(getClass());
<br>>>>>>>>>
<br>>>>>>>>> //protected final String _clusterUri="tcp://smilodon:
<br>>>>>>>>> 61616";
<br>>>>>>>>> protected final String _clusterUri="peer://
<br>>>>>>>>> org.codehaus.wadi";
<br>>>>>>>>> protected final String
<br>>>>>>>>> _clusterName="ORG.CODEHAUS.WADI.TEST";
<br>>>>>>>>> protected final ActiveMQConnectionFactory
<br>>>>>>>>> _connectionFactory=new ActiveMQConnectionFactory(_clusterUri);
<br>>>>>>>>> protected final ClusterFactory _clusterFactory=new
<br>>>>>>>>> DefaultClusterFactory(_connectionFactory);
<br>>>>>>>>> protected final org.activecluster.Cluster _cluster;
<br>>>>>>>>> protected final long _timeout=30*1000L;
<br>>>>>>>>>
<br>>>>>>>>> public ACCluster(String nodeName) throws Exception {
<br>>>>>>>>> System.setProperty
<br>>>>>>>>> ("activemq.persistenceAdapterFactory",
<br>>>>>>>>> VMPersistenceAdapterFactory.class.getName());
<br>>>>>>>>> _cluster=_clusterFactory.createCluster(_clusterName);
<br>>>>>>>>> Map state=new HashMap();
<br>>>>>>>>> state.put("nodeName", nodeName);
<br>>>>>>>>> _cluster.getLocalNode().setState(state);
<br>>>>>>>>> _cluster.addClusterListener(new ClusterListener() {
<br>>>>>>>>>
<br>>>>>>>>> public void onNodeAdd(ClusterEvent arg0) {
<br>>>>>>>>> _log.info("onNodeAdd: "+arg0.getNode
<br>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>> }
<br>>>>>>>>>
<br>>>>>>>>> public void onNodeUpdate(ClusterEvent arg0) {
<br>>>>>>>>> _log.info("onNodeUpdate: "+arg0.getNode
<br>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>> }
<br>>>>>>>>>
<br>>>>>>>>> public void onNodeRemoved(ClusterEvent arg0) {
<br>>>>>>>>> _log.info("onNodeRemoved: "+arg0.getNode
<br>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>> }
<br>>>>>>>>>
<br>>>>>>>>> public void onNodeFailed(ClusterEvent arg0) {
<br>>>>>>>>> _log.info("onNodeFailed: "+arg0.getNode
<br>>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>>> }
<br>>>>>>>>>
<br>>>>>>>>> public void onCoordinatorChanged(ClusterEvent
<br>>>>>>>>> arg0) {
<br>>>>>>>>> _log.info("onCoordinatorChanged:
<br>>>>>>>>> "+arg0.getNode ().getState().get("nodeName"));
<br>>>>>>>>> }
<br>>>>>>>>> });
<br>>>>>>>>> }
<br>>>>>>>>>
<br>>>>>>>>> public void start() throws JMSException {
<br>>>>>>>>> _log.info("starting...");
<br>>>>>>>>> _cluster.start();
<br>>>>>>>>> _log.info("...started");
<br>>>>>>>>> }
<br>>>>>>>>>
<br>>>>>>>>> public void stop() {
<br>>>>>>>>>
<br>>>>>>>>> }
<br>>>>>>>>>
<br>>>>>>>>> public static void main(String[] args) throws Exception {
<br>>>>>>>>> ACCluster cluster=new ACCluster(args[0]);
<br>>>>>>>>> cluster.start();
<br>>>>>>>>> Thread.sleep(100*1000);
<br>>>>>>>>> cluster.stop();
<br>>>>>>>>> }
<br>>>>>>>>> }
<br>>>>>>>>>
<br>>>>>>>>>
<br>>>>>>>>
<br>>>>>>>>
<br>>>>>>>> --
<br>>>>>>>> "Open Source is a self-assembling organism. You dangle a
<br>>>>>>>> piece of
<br>>>>>>>> string into a super-saturated solution and a whole operating-
<br>>>>>>>> system
<br>>>>>>>> crystallises out around it."
<br>>>>>>>>
<br>>>>>>>> /**********************************
<br>>>>>>>> * Jules Gosnell
<br>>>>>>>> * Partner
<br>>>>>>>> * Core Developers Network (Europe)
<br>>>>>>>> *
<br>>>>>>>> * www.coredevelopers.net
<br>>>>>>>> *
<br>>>>>>>> * Open Source Training & Support.
<br>>>>>>>> **********************************/
<br>>>>>>>>
<br>>>>>>>>
<br>>>>>>
<br>>>>>>
<br>>>>>> --
<br>>>>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>>>>> string into a super-saturated solution and a whole operating-
<br>>>>>> system
<br>>>>>> crystallises out around it."
<br>>>>>>
<br>>>>>> /**********************************
<br>>>>>> * Jules Gosnell
<br>>>>>> * Partner
<br>>>>>> * Core Developers Network (Europe)
<br>>>>>> *
<br>>>>>> * www.coredevelopers.net
<br>>>>>> *
<br>>>>>> * Open Source Training & Support.
<br>>>>>> **********************************/
<br>>>>>>
<br>>>>
<br>>>>
<br>>>
<br>>>
<br>>
<br>>
<br>> --
<br>> "Open Source is a self-assembling organism. You dangle a piece of
<br>> string into a super-saturated solution and a whole operating-system
<br>> crystallises out around it."
<br>>
<br>> /**********************************
<br>> * Jules Gosnell
<br>> * Partner
<br>> * Core Developers Network (Europe)
<br>> *
<br>> * www.coredevelopers.net
<br>> *
<br>> * Open Source Training & Support.
<br>> **********************************/
<br>>
</div><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1366605Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-11-07T04:15:56Z2005-11-07T04:15:56ZJules Gosnell-4Jules Gosnell wrote:
<br><div class='shrinkable-quote'><br>> Jules Gosnell wrote:
<br>>
<br>>> Rob,
<br>>>
<br>>> Thanks for putting some time into this...
<br>>>
<br>>> I've just rebuilt AC and run the test that I included again. I'm
<br>>> afraid that it doesn't fix the problem. Once the second node leaves
<br>>> the cluster the first node becomes unable to see any subsequent node
<br>>> that wishes to join. It looks as if it is some logic problem
<br>>> associated with cluster's dropping to one node, but, curiously it
<br>>> works fine with the tcp:// protocol... Maybe it is something to do
<br>>> with autodiscovery ? Do tcp:// and peer;// use common or separate
<br>>> autodiscovery mechanisms ?
<br>>>
<br>>> Is there some test that I can run to help pinpoint the problem ?
<br>>>
<br>>> Sorry to be the bearer of bad news :-(
<br>>>
<br>>>
<br>>> Jules
<br>>
<br>>
<br>>
<br>> P.S.
<br>>
<br>> I ran my tests on AMQ 3.2-M1 - I guess that I need to run them on
<br>> AMQ3-2 to be sure - only problem is that 3.2 breaks our mavens build
<br>> with deps that cannot be downloaded - maybe our repo list is not what
<br>> it should be - I will see what I can do and get back to you in a
<br>> little while.
</div><br>Just ran the test on HEAD of AC and AMQ-3.2 - same result...
<br><br>Everything is fine until a cluster drops down to one node. That node
<br>then becomes somehow out of reach of subsequent nodes. Any nodes that
<br>are started after this point, only see each other and not the original
<br>node. Taking the newly formed cluster down to one member has the same
<br>effect, so you can end up with more than one isolated node.
<br><br>Jules
<br><div class='shrinkable-quote'><br>>
<br>>>
<br>>> Rob Davies wrote:
<br>>>
<br>>>> hi Jules,
<br>>>>
<br>>>> the latest ActiveCluster in the repo works with ActiveMQ 3.2 -
<br>>>> amq4 will hopefully be ready soon. I'd appreciate it if you could
<br>>>> build activecluster from src - and see if your problems have
<br>>>> disappeared - there was a a nasty deadlock I ended up fixing.
<br>>>>
<br>>>> cheers,
<br>>>>
<br>>>> Rob
<br>>>>
<br>>>> On 6 Nov 2005, at 18:41, Jules Gosnell wrote:
<br>>>>
<br>>>>> Rob Davies wrote:
<br>>>>>
<br>>>>>> Hi Jules,
<br>>>>>>
<br>>>>>> I'm in the process of porting activecluster to amq4. There's an
<br>>>>>> issue with discovery in amq4 at the moment that needs to be fixed.
<br>>>>>
<br>>>>>
<br>>>>>
<br>>>>>
<br>>>>> Ok - no problem - if you let me know when the fix goes in, I will
<br>>>>> be happy to retest and report.
<br>>>>>
<br>>>>> BTW - do you know if AMQ4 is the release that is intended to be
<br>>>>> bundled with Geronimo-1.0 ? I am targetting a WADI release at
<br>>>>> Geronimo-1.0 and I need to make sure that we run on the same
<br>>>>> version of AMQ as is packaged with G.
<br>>>>>
<br>>>>> Cheers, Rob,
<br>>>>>
<br>>>>> Jules
<br>>>>>
<br>>>>>>
<br>>>>>> cheers,
<br>>>>>>
<br>>>>>> Rob
<br>>>>>> On 29 Oct 2005, at 02:10, Jules Gosnell wrote:
<br>>>>>>
<br>>>>>>> Jules Gosnell wrote:
<br>>>>>>>
<br>>>>>>>
<br>>>>>>>>
<br>>>>>>>> tested against activecluster-1.1-SNAPSHOT, activemq-4.0-
<br>>>>>>>> SNAPSHOT and Sun JDK 1.4.2_08 and 1.5.0 - two node cluster,
<br>>>>>>>> both on same machine.
<br>>>>>>>>
<br>>>>>>>
<br>>>>>>> oops - AMQ-4.0 was not on the machine that I ran this on - so
<br>>>>>>> this issue is actually with AMQ-3.2-M1.
<br>>>>>>>
<br>>>>>>> I am trying to run the test with AMQ-4.0, but the second node
<br>>>>>>> won't start because "java.io.IOException: Journal is allready
<br>>>>>>> opened by another application", So i will have to figure out
<br>>>>>>> how to tell it not to journal - I'll be back ;-)
<br>>>>>>>
<br>>>>>>> Jules
<br>>>>>>>
<br>>>>>>>
<br>>>>>>>>
<br>>>>>>>> start your first node (red)
<br>>>>>>>>
<br>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster red
<br>>>>>>>> 2005/10/29 00:04:18:317 BST [INFO] ACCluster - starting...
<br>>>>>>>> 2005/10/29 00:04:18:321 BST [INFO] ACCluster - ...started
<br>>>>>>>>
<br>>>>>>>> start your second node (green)
<br>>>>>>>>
<br>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>>>> 2005/10/29 00:04:31:874 BST [INFO] ACCluster - starting...
<br>>>>>>>> 2005/10/29 00:04:31:885 BST [INFO] ACCluster - ...started
<br>>>>>>>> 2005/10/29 00:04:33:323 BST [INFO] ACCluster - onNodeAdd: red
<br>>>>>>>>
<br>>>>>>>> red says:
<br>>>>>>>>
<br>>>>>>>> 2005/10/29 00:04:33:536 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>>>>> 2005/10/29 00:04:33:537 BST [INFO] ACCluster -
<br>>>>>>>> onCoordinatorChanged: green
<br>>>>>>>>
<br>>>>>>>> ctl-c green, after a few seconds red says:
<br>>>>>>>>
<br>>>>>>>> 2005/10/29 00:04:46:624 BST [INFO] ACCluster - onNodeFailed: green
<br>>>>>>>> 2005/10/29 00:04:46:626 BST [INFO] ACCluster -
<br>>>>>>>> onCoordinatorChanged: red
<br>>>>>>>>
<br>>>>>>>> all fine so far - now restart green:
<br>>>>>>>>
<br>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>>>> 2005/10/29 00:04:51:962 BST [INFO] ACCluster - starting...
<br>>>>>>>> 2005/10/29 00:04:51:967 BST [INFO] ACCluster - ...started
<br>>>>>>>>
<br>>>>>>>> red says......nothing.
<br>>>>>>>>
<br>>>>>>>>
<br>>>>>>>> if you then start a third node (blue):
<br>>>>>>>>
<br>>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster blue
<br>>>>>>>> 2005/10/29 00:08:47:990 BST [INFO] ACCluster - starting...
<br>>>>>>>> 2005/10/29 00:08:47:994 BST [INFO] ACCluster - ...started
<br>>>>>>>> 2005/10/29 00:08:50:198 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>>>>>
<br>>>>>>>> it sees green, but not red and green says:
<br>>>>>>>>
<br>>>>>>>> 2005/10/29 00:08:49:661 BST [INFO] ACCluster - onNodeAdd: blue
<br>>>>>>>> 2005/10/29 00:08:49:662 BST [INFO] ACCluster -
<br>>>>>>>> onCoordinatorChanged: blue
<br>>>>>>>>
<br>>>>>>>> so green sees blue too.
<br>>>>>>>>
<br>>>>>>>> but red says.....nothing.
<br>>>>>>>>
<br>>>>>>>>
<br>>>>>>>> Using tcp:// and a broker, red will happily see blue and green
<br>>>>>>>> after they are restarted.
<br>>>>>>>>
<br>>>>>>>> This has been the case with activecluster/activemq for a long
<br>>>>>>>> time, but I was hoping that it would be fixed by improvements
<br>>>>>>>> to peer:// in activemq-4.0 - but :-(
<br>>>>>>>>
<br>>>>>>>>
<br>>>>>>>> Jules
<br>>>>>>>>
<br>>>>>>>> -------------------------------------------------------------------
<br>>>>>>>> -- ---
<br>>>>>>>>
<br>>>>>>>> package org.codehaus.wadi.sandbox.partition.impl;
<br>>>>>>>>
<br>>>>>>>> import java.util.HashMap;
<br>>>>>>>> import java.util.Map;
<br>>>>>>>>
<br>>>>>>>> import javax.jms.JMSException;
<br>>>>>>>>
<br>>>>>>>> import org.activecluster.ClusterEvent;
<br>>>>>>>> import org.activecluster.ClusterFactory;
<br>>>>>>>> import org.activecluster.ClusterListener;
<br>>>>>>>> import org.activecluster.impl.DefaultClusterFactory;
<br>>>>>>>> import org.activemq.ActiveMQConnectionFactory;
<br>>>>>>>> import org.activemq.store.vm.VMPersistenceAdapterFactory;
<br>>>>>>>> import org.apache.commons.logging.Log;
<br>>>>>>>> import org.apache.commons.logging.LogFactory;
<br>>>>>>>> import org.codehaus.wadi.dindex.impl.SeniorityElectionStrategy;
<br>>>>>>>> import org.codehaus.wadi.impl.CustomClusterFactory;
<br>>>>>>>> import org.codehaus.wadi.sandbox.partition.Cluster;
<br>>>>>>>>
<br>>>>>>>> public class ACCluster {
<br>>>>>>>>
<br>>>>>>>> protected final Log _log = LogFactory.getLog(getClass());
<br>>>>>>>>
<br>>>>>>>> //protected final String _clusterUri="tcp://smilodon:61616";
<br>>>>>>>> protected final String _clusterUri="peer://org.codehaus.wadi";
<br>>>>>>>> protected final String _clusterName="ORG.CODEHAUS.WADI.TEST";
<br>>>>>>>> protected final ActiveMQConnectionFactory
<br>>>>>>>> _connectionFactory=new ActiveMQConnectionFactory(_clusterUri);
<br>>>>>>>> protected final ClusterFactory _clusterFactory=new
<br>>>>>>>> DefaultClusterFactory(_connectionFactory);
<br>>>>>>>> protected final org.activecluster.Cluster _cluster;
<br>>>>>>>> protected final long _timeout=30*1000L;
<br>>>>>>>>
<br>>>>>>>> public ACCluster(String nodeName) throws Exception {
<br>>>>>>>> System.setProperty
<br>>>>>>>> ("activemq.persistenceAdapterFactory",
<br>>>>>>>> VMPersistenceAdapterFactory.class.getName());
<br>>>>>>>> _cluster=_clusterFactory.createCluster(_clusterName);
<br>>>>>>>> Map state=new HashMap();
<br>>>>>>>> state.put("nodeName", nodeName);
<br>>>>>>>> _cluster.getLocalNode().setState(state);
<br>>>>>>>> _cluster.addClusterListener(new ClusterListener() {
<br>>>>>>>>
<br>>>>>>>> public void onNodeAdd(ClusterEvent arg0) {
<br>>>>>>>> _log.info("onNodeAdd: "+arg0.getNode().getState
<br>>>>>>>> ().get("nodeName"));
<br>>>>>>>> }
<br>>>>>>>>
<br>>>>>>>> public void onNodeUpdate(ClusterEvent arg0) {
<br>>>>>>>> _log.info("onNodeUpdate: "+arg0.getNode
<br>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>> }
<br>>>>>>>>
<br>>>>>>>> public void onNodeRemoved(ClusterEvent arg0) {
<br>>>>>>>> _log.info("onNodeRemoved: "+arg0.getNode
<br>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>> }
<br>>>>>>>>
<br>>>>>>>> public void onNodeFailed(ClusterEvent arg0) {
<br>>>>>>>> _log.info("onNodeFailed: "+arg0.getNode
<br>>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>>> }
<br>>>>>>>>
<br>>>>>>>> public void onCoordinatorChanged(ClusterEvent arg0) {
<br>>>>>>>> _log.info("onCoordinatorChanged: "+arg0.getNode
<br>>>>>>>> ().getState().get("nodeName"));
<br>>>>>>>> }
<br>>>>>>>> });
<br>>>>>>>> }
<br>>>>>>>>
<br>>>>>>>> public void start() throws JMSException {
<br>>>>>>>> _log.info("starting...");
<br>>>>>>>> _cluster.start();
<br>>>>>>>> _log.info("...started");
<br>>>>>>>> }
<br>>>>>>>>
<br>>>>>>>> public void stop() {
<br>>>>>>>>
<br>>>>>>>> }
<br>>>>>>>>
<br>>>>>>>> public static void main(String[] args) throws Exception {
<br>>>>>>>> ACCluster cluster=new ACCluster(args[0]);
<br>>>>>>>> cluster.start();
<br>>>>>>>> Thread.sleep(100*1000);
<br>>>>>>>> cluster.stop();
<br>>>>>>>> }
<br>>>>>>>> }
<br>>>>>>>>
<br>>>>>>>>
<br>>>>>>>
<br>>>>>>>
<br>>>>>>> --
<br>>>>>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>>>>>> string into a super-saturated solution and a whole operating-system
<br>>>>>>> crystallises out around it."
<br>>>>>>>
<br>>>>>>> /**********************************
<br>>>>>>> * Jules Gosnell
<br>>>>>>> * Partner
<br>>>>>>> * Core Developers Network (Europe)
<br>>>>>>> *
<br>>>>>>> * www.coredevelopers.net
<br>>>>>>> *
<br>>>>>>> * Open Source Training & Support.
<br>>>>>>> **********************************/
<br>>>>>>>
<br>>>>>>>
<br>>>>>
<br>>>>>
<br>>>>> --
<br>>>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>>>> string into a super-saturated solution and a whole operating-system
<br>>>>> crystallises out around it."
<br>>>>>
<br>>>>> /**********************************
<br>>>>> * Jules Gosnell
<br>>>>> * Partner
<br>>>>> * Core Developers Network (Europe)
<br>>>>> *
<br>>>>> * www.coredevelopers.net
<br>>>>> *
<br>>>>> * Open Source Training & Support.
<br>>>>> **********************************/
<br>>>>>
<br>>>
<br>>>
<br>>
<br>>
</div><br><br>--
<br>"Open Source is a self-assembling organism. You dangle a piece of
<br>string into a super-saturated solution and a whole operating-system
<br>crystallises out around it."
<br><br>/**********************************
<br> * Jules Gosnell
<br> * Partner
<br> * Core Developers Network (Europe)
<br> *
<br> * www.coredevelopers.net
<br> *
<br> * Open Source Training & Support.
<br> **********************************/
<br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1366432Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-11-07T04:01:49Z2005-11-07T04:01:49ZJules Gosnell-4Jules Gosnell wrote:
<br><div class='shrinkable-quote'><br>> Rob,
<br>>
<br>> Thanks for putting some time into this...
<br>>
<br>> I've just rebuilt AC and run the test that I included again. I'm
<br>> afraid that it doesn't fix the problem. Once the second node leaves
<br>> the cluster the first node becomes unable to see any subsequent node
<br>> that wishes to join. It looks as if it is some logic problem
<br>> associated with cluster's dropping to one node, but, curiously it
<br>> works fine with the tcp:// protocol... Maybe it is something to do
<br>> with autodiscovery ? Do tcp:// and peer;// use common or separate
<br>> autodiscovery mechanisms ?
<br>>
<br>> Is there some test that I can run to help pinpoint the problem ?
<br>>
<br>> Sorry to be the bearer of bad news :-(
<br>>
<br>>
<br>> Jules
</div><br><br>P.S.
<br><br>I ran my tests on AMQ 3.2-M1 - I guess that I need to run them on AMQ3-2
<br>to be sure - only problem is that 3.2 breaks our mavens build with deps
<br>that cannot be downloaded - maybe our repo list is not what it should be
<br>- I will see what I can do and get back to you in a little while.
<br><div class='shrinkable-quote'><br>>
<br>> Rob Davies wrote:
<br>>
<br>>> hi Jules,
<br>>>
<br>>> the latest ActiveCluster in the repo works with ActiveMQ 3.2 - amq4
<br>>> will hopefully be ready soon. I'd appreciate it if you could build
<br>>> activecluster from src - and see if your problems have disappeared
<br>>> - there was a a nasty deadlock I ended up fixing.
<br>>>
<br>>> cheers,
<br>>>
<br>>> Rob
<br>>>
<br>>> On 6 Nov 2005, at 18:41, Jules Gosnell wrote:
<br>>>
<br>>>> Rob Davies wrote:
<br>>>>
<br>>>>> Hi Jules,
<br>>>>>
<br>>>>> I'm in the process of porting activecluster to amq4. There's an
<br>>>>> issue with discovery in amq4 at the moment that needs to be fixed.
<br>>>>
<br>>>>
<br>>>>
<br>>>> Ok - no problem - if you let me know when the fix goes in, I will
<br>>>> be happy to retest and report.
<br>>>>
<br>>>> BTW - do you know if AMQ4 is the release that is intended to be
<br>>>> bundled with Geronimo-1.0 ? I am targetting a WADI release at
<br>>>> Geronimo-1.0 and I need to make sure that we run on the same
<br>>>> version of AMQ as is packaged with G.
<br>>>>
<br>>>> Cheers, Rob,
<br>>>>
<br>>>> Jules
<br>>>>
<br>>>>>
<br>>>>> cheers,
<br>>>>>
<br>>>>> Rob
<br>>>>> On 29 Oct 2005, at 02:10, Jules Gosnell wrote:
<br>>>>>
<br>>>>>> Jules Gosnell wrote:
<br>>>>>>
<br>>>>>>
<br>>>>>>>
<br>>>>>>> tested against activecluster-1.1-SNAPSHOT, activemq-4.0-
<br>>>>>>> SNAPSHOT and Sun JDK 1.4.2_08 and 1.5.0 - two node cluster,
<br>>>>>>> both on same machine.
<br>>>>>>>
<br>>>>>>
<br>>>>>> oops - AMQ-4.0 was not on the machine that I ran this on - so
<br>>>>>> this issue is actually with AMQ-3.2-M1.
<br>>>>>>
<br>>>>>> I am trying to run the test with AMQ-4.0, but the second node
<br>>>>>> won't start because "java.io.IOException: Journal is allready
<br>>>>>> opened by another application", So i will have to figure out how
<br>>>>>> to tell it not to journal - I'll be back ;-)
<br>>>>>>
<br>>>>>> Jules
<br>>>>>>
<br>>>>>>
<br>>>>>>>
<br>>>>>>> start your first node (red)
<br>>>>>>>
<br>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster red
<br>>>>>>> 2005/10/29 00:04:18:317 BST [INFO] ACCluster - starting...
<br>>>>>>> 2005/10/29 00:04:18:321 BST [INFO] ACCluster - ...started
<br>>>>>>>
<br>>>>>>> start your second node (green)
<br>>>>>>>
<br>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>>> 2005/10/29 00:04:31:874 BST [INFO] ACCluster - starting...
<br>>>>>>> 2005/10/29 00:04:31:885 BST [INFO] ACCluster - ...started
<br>>>>>>> 2005/10/29 00:04:33:323 BST [INFO] ACCluster - onNodeAdd: red
<br>>>>>>>
<br>>>>>>> red says:
<br>>>>>>>
<br>>>>>>> 2005/10/29 00:04:33:536 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>>>> 2005/10/29 00:04:33:537 BST [INFO] ACCluster -
<br>>>>>>> onCoordinatorChanged: green
<br>>>>>>>
<br>>>>>>> ctl-c green, after a few seconds red says:
<br>>>>>>>
<br>>>>>>> 2005/10/29 00:04:46:624 BST [INFO] ACCluster - onNodeFailed: green
<br>>>>>>> 2005/10/29 00:04:46:626 BST [INFO] ACCluster -
<br>>>>>>> onCoordinatorChanged: red
<br>>>>>>>
<br>>>>>>> all fine so far - now restart green:
<br>>>>>>>
<br>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>>> 2005/10/29 00:04:51:962 BST [INFO] ACCluster - starting...
<br>>>>>>> 2005/10/29 00:04:51:967 BST [INFO] ACCluster - ...started
<br>>>>>>>
<br>>>>>>> red says......nothing.
<br>>>>>>>
<br>>>>>>>
<br>>>>>>> if you then start a third node (blue):
<br>>>>>>>
<br>>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster blue
<br>>>>>>> 2005/10/29 00:08:47:990 BST [INFO] ACCluster - starting...
<br>>>>>>> 2005/10/29 00:08:47:994 BST [INFO] ACCluster - ...started
<br>>>>>>> 2005/10/29 00:08:50:198 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>>>>
<br>>>>>>> it sees green, but not red and green says:
<br>>>>>>>
<br>>>>>>> 2005/10/29 00:08:49:661 BST [INFO] ACCluster - onNodeAdd: blue
<br>>>>>>> 2005/10/29 00:08:49:662 BST [INFO] ACCluster -
<br>>>>>>> onCoordinatorChanged: blue
<br>>>>>>>
<br>>>>>>> so green sees blue too.
<br>>>>>>>
<br>>>>>>> but red says.....nothing.
<br>>>>>>>
<br>>>>>>>
<br>>>>>>> Using tcp:// and a broker, red will happily see blue and green
<br>>>>>>> after they are restarted.
<br>>>>>>>
<br>>>>>>> This has been the case with activecluster/activemq for a long
<br>>>>>>> time, but I was hoping that it would be fixed by improvements
<br>>>>>>> to peer:// in activemq-4.0 - but :-(
<br>>>>>>>
<br>>>>>>>
<br>>>>>>> Jules
<br>>>>>>>
<br>>>>>>> -------------------------------------------------------------------
<br>>>>>>> -- ---
<br>>>>>>>
<br>>>>>>> package org.codehaus.wadi.sandbox.partition.impl;
<br>>>>>>>
<br>>>>>>> import java.util.HashMap;
<br>>>>>>> import java.util.Map;
<br>>>>>>>
<br>>>>>>> import javax.jms.JMSException;
<br>>>>>>>
<br>>>>>>> import org.activecluster.ClusterEvent;
<br>>>>>>> import org.activecluster.ClusterFactory;
<br>>>>>>> import org.activecluster.ClusterListener;
<br>>>>>>> import org.activecluster.impl.DefaultClusterFactory;
<br>>>>>>> import org.activemq.ActiveMQConnectionFactory;
<br>>>>>>> import org.activemq.store.vm.VMPersistenceAdapterFactory;
<br>>>>>>> import org.apache.commons.logging.Log;
<br>>>>>>> import org.apache.commons.logging.LogFactory;
<br>>>>>>> import org.codehaus.wadi.dindex.impl.SeniorityElectionStrategy;
<br>>>>>>> import org.codehaus.wadi.impl.CustomClusterFactory;
<br>>>>>>> import org.codehaus.wadi.sandbox.partition.Cluster;
<br>>>>>>>
<br>>>>>>> public class ACCluster {
<br>>>>>>>
<br>>>>>>> protected final Log _log = LogFactory.getLog(getClass());
<br>>>>>>>
<br>>>>>>> //protected final String _clusterUri="tcp://smilodon:61616";
<br>>>>>>> protected final String _clusterUri="peer://org.codehaus.wadi";
<br>>>>>>> protected final String _clusterName="ORG.CODEHAUS.WADI.TEST";
<br>>>>>>> protected final ActiveMQConnectionFactory
<br>>>>>>> _connectionFactory=new ActiveMQConnectionFactory(_clusterUri);
<br>>>>>>> protected final ClusterFactory _clusterFactory=new
<br>>>>>>> DefaultClusterFactory(_connectionFactory);
<br>>>>>>> protected final org.activecluster.Cluster _cluster;
<br>>>>>>> protected final long _timeout=30*1000L;
<br>>>>>>>
<br>>>>>>> public ACCluster(String nodeName) throws Exception {
<br>>>>>>> System.setProperty
<br>>>>>>> ("activemq.persistenceAdapterFactory",
<br>>>>>>> VMPersistenceAdapterFactory.class.getName());
<br>>>>>>> _cluster=_clusterFactory.createCluster(_clusterName);
<br>>>>>>> Map state=new HashMap();
<br>>>>>>> state.put("nodeName", nodeName);
<br>>>>>>> _cluster.getLocalNode().setState(state);
<br>>>>>>> _cluster.addClusterListener(new ClusterListener() {
<br>>>>>>>
<br>>>>>>> public void onNodeAdd(ClusterEvent arg0) {
<br>>>>>>> _log.info("onNodeAdd: "+arg0.getNode().getState
<br>>>>>>> ().get("nodeName"));
<br>>>>>>> }
<br>>>>>>>
<br>>>>>>> public void onNodeUpdate(ClusterEvent arg0) {
<br>>>>>>> _log.info("onNodeUpdate: "+arg0.getNode
<br>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>> }
<br>>>>>>>
<br>>>>>>> public void onNodeRemoved(ClusterEvent arg0) {
<br>>>>>>> _log.info("onNodeRemoved: "+arg0.getNode
<br>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>> }
<br>>>>>>>
<br>>>>>>> public void onNodeFailed(ClusterEvent arg0) {
<br>>>>>>> _log.info("onNodeFailed: "+arg0.getNode
<br>>>>>>> ().getState ().get("nodeName"));
<br>>>>>>> }
<br>>>>>>>
<br>>>>>>> public void onCoordinatorChanged(ClusterEvent arg0) {
<br>>>>>>> _log.info("onCoordinatorChanged: "+arg0.getNode
<br>>>>>>> ().getState().get("nodeName"));
<br>>>>>>> }
<br>>>>>>> });
<br>>>>>>> }
<br>>>>>>>
<br>>>>>>> public void start() throws JMSException {
<br>>>>>>> _log.info("starting...");
<br>>>>>>> _cluster.start();
<br>>>>>>> _log.info("...started");
<br>>>>>>> }
<br>>>>>>>
<br>>>>>>> public void stop() {
<br>>>>>>>
<br>>>>>>> }
<br>>>>>>>
<br>>>>>>> public static void main(String[] args) throws Exception {
<br>>>>>>> ACCluster cluster=new ACCluster(args[0]);
<br>>>>>>> cluster.start();
<br>>>>>>> Thread.sleep(100*1000);
<br>>>>>>> cluster.stop();
<br>>>>>>> }
<br>>>>>>> }
<br>>>>>>>
<br>>>>>>>
<br>>>>>>
<br>>>>>>
<br>>>>>> --
<br>>>>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>>>>> string into a super-saturated solution and a whole operating-system
<br>>>>>> crystallises out around it."
<br>>>>>>
<br>>>>>> /**********************************
<br>>>>>> * Jules Gosnell
<br>>>>>> * Partner
<br>>>>>> * Core Developers Network (Europe)
<br>>>>>> *
<br>>>>>> * www.coredevelopers.net
<br>>>>>> *
<br>>>>>> * Open Source Training & Support.
<br>>>>>> **********************************/
<br>>>>>>
<br>>>>>>
<br>>>>
<br>>>>
<br>>>> --
<br>>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>>> string into a super-saturated solution and a whole operating-system
<br>>>> crystallises out around it."
<br>>>>
<br>>>> /**********************************
<br>>>> * Jules Gosnell
<br>>>> * Partner
<br>>>> * Core Developers Network (Europe)
<br>>>> *
<br>>>> * www.coredevelopers.net
<br>>>> *
<br>>>> * Open Source Training & Support.
<br>>>> **********************************/
<br>>>>
<br>>
<br>>
</div><br><br>--
<br>"Open Source is a self-assembling organism. You dangle a piece of
<br>string into a super-saturated solution and a whole operating-system
<br>crystallises out around it."
<br><br>/**********************************
<br> * Jules Gosnell
<br> * Partner
<br> * Core Developers Network (Europe)
<br> *
<br> * www.coredevelopers.net
<br> *
<br> * Open Source Training & Support.
<br> **********************************/
<br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1366180Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-11-07T03:36:53Z2005-11-07T03:36:53ZJules Gosnell-4Rob,
<br><br>Thanks for putting some time into this...
<br><br>I've just rebuilt AC and run the test that I included again. I'm afraid
<br>that it doesn't fix the problem. Once the second node leaves the cluster
<br>the first node becomes unable to see any subsequent node that wishes to
<br>join. It looks as if it is some logic problem associated with cluster's
<br>dropping to one node, but, curiously it works fine with the tcp://
<br>protocol... Maybe it is something to do with autodiscovery ? Do tcp://
<br>and peer;// use common or separate autodiscovery mechanisms ?
<br><br>Is there some test that I can run to help pinpoint the problem ?
<br><br>Sorry to be the bearer of bad news :-(
<br><br><br>Jules
<br><br>Rob Davies wrote:
<br><div class='shrinkable-quote'><br>> hi Jules,
<br>>
<br>> the latest ActiveCluster in the repo works with ActiveMQ 3.2 - amq4
<br>> will hopefully be ready soon. I'd appreciate it if you could build
<br>> activecluster from src - and see if your problems have disappeared -
<br>> there was a a nasty deadlock I ended up fixing.
<br>>
<br>> cheers,
<br>>
<br>> Rob
<br>>
<br>> On 6 Nov 2005, at 18:41, Jules Gosnell wrote:
<br>>
<br>>> Rob Davies wrote:
<br>>>
<br>>>> Hi Jules,
<br>>>>
<br>>>> I'm in the process of porting activecluster to amq4. There's an
<br>>>> issue with discovery in amq4 at the moment that needs to be fixed.
<br>>>
<br>>>
<br>>> Ok - no problem - if you let me know when the fix goes in, I will be
<br>>> happy to retest and report.
<br>>>
<br>>> BTW - do you know if AMQ4 is the release that is intended to be
<br>>> bundled with Geronimo-1.0 ? I am targetting a WADI release at
<br>>> Geronimo-1.0 and I need to make sure that we run on the same version
<br>>> of AMQ as is packaged with G.
<br>>>
<br>>> Cheers, Rob,
<br>>>
<br>>> Jules
<br>>>
<br>>>>
<br>>>> cheers,
<br>>>>
<br>>>> Rob
<br>>>> On 29 Oct 2005, at 02:10, Jules Gosnell wrote:
<br>>>>
<br>>>>> Jules Gosnell wrote:
<br>>>>>
<br>>>>>
<br>>>>>>
<br>>>>>> tested against activecluster-1.1-SNAPSHOT, activemq-4.0- SNAPSHOT
<br>>>>>> and Sun JDK 1.4.2_08 and 1.5.0 - two node cluster, both on same
<br>>>>>> machine.
<br>>>>>>
<br>>>>>
<br>>>>> oops - AMQ-4.0 was not on the machine that I ran this on - so
<br>>>>> this issue is actually with AMQ-3.2-M1.
<br>>>>>
<br>>>>> I am trying to run the test with AMQ-4.0, but the second node
<br>>>>> won't start because "java.io.IOException: Journal is allready
<br>>>>> opened by another application", So i will have to figure out how
<br>>>>> to tell it not to journal - I'll be back ;-)
<br>>>>>
<br>>>>> Jules
<br>>>>>
<br>>>>>
<br>>>>>>
<br>>>>>> start your first node (red)
<br>>>>>>
<br>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster red
<br>>>>>> 2005/10/29 00:04:18:317 BST [INFO] ACCluster - starting...
<br>>>>>> 2005/10/29 00:04:18:321 BST [INFO] ACCluster - ...started
<br>>>>>>
<br>>>>>> start your second node (green)
<br>>>>>>
<br>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>> 2005/10/29 00:04:31:874 BST [INFO] ACCluster - starting...
<br>>>>>> 2005/10/29 00:04:31:885 BST [INFO] ACCluster - ...started
<br>>>>>> 2005/10/29 00:04:33:323 BST [INFO] ACCluster - onNodeAdd: red
<br>>>>>>
<br>>>>>> red says:
<br>>>>>>
<br>>>>>> 2005/10/29 00:04:33:536 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>>> 2005/10/29 00:04:33:537 BST [INFO] ACCluster -
<br>>>>>> onCoordinatorChanged: green
<br>>>>>>
<br>>>>>> ctl-c green, after a few seconds red says:
<br>>>>>>
<br>>>>>> 2005/10/29 00:04:46:624 BST [INFO] ACCluster - onNodeFailed: green
<br>>>>>> 2005/10/29 00:04:46:626 BST [INFO] ACCluster -
<br>>>>>> onCoordinatorChanged: red
<br>>>>>>
<br>>>>>> all fine so far - now restart green:
<br>>>>>>
<br>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>>> 2005/10/29 00:04:51:962 BST [INFO] ACCluster - starting...
<br>>>>>> 2005/10/29 00:04:51:967 BST [INFO] ACCluster - ...started
<br>>>>>>
<br>>>>>> red says......nothing.
<br>>>>>>
<br>>>>>>
<br>>>>>> if you then start a third node (blue):
<br>>>>>>
<br>>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster blue
<br>>>>>> 2005/10/29 00:08:47:990 BST [INFO] ACCluster - starting...
<br>>>>>> 2005/10/29 00:08:47:994 BST [INFO] ACCluster - ...started
<br>>>>>> 2005/10/29 00:08:50:198 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>>>
<br>>>>>> it sees green, but not red and green says:
<br>>>>>>
<br>>>>>> 2005/10/29 00:08:49:661 BST [INFO] ACCluster - onNodeAdd: blue
<br>>>>>> 2005/10/29 00:08:49:662 BST [INFO] ACCluster -
<br>>>>>> onCoordinatorChanged: blue
<br>>>>>>
<br>>>>>> so green sees blue too.
<br>>>>>>
<br>>>>>> but red says.....nothing.
<br>>>>>>
<br>>>>>>
<br>>>>>> Using tcp:// and a broker, red will happily see blue and green
<br>>>>>> after they are restarted.
<br>>>>>>
<br>>>>>> This has been the case with activecluster/activemq for a long
<br>>>>>> time, but I was hoping that it would be fixed by improvements to
<br>>>>>> peer:// in activemq-4.0 - but :-(
<br>>>>>>
<br>>>>>>
<br>>>>>> Jules
<br>>>>>>
<br>>>>>> -------------------------------------------------------------------
<br>>>>>> -- ---
<br>>>>>>
<br>>>>>> package org.codehaus.wadi.sandbox.partition.impl;
<br>>>>>>
<br>>>>>> import java.util.HashMap;
<br>>>>>> import java.util.Map;
<br>>>>>>
<br>>>>>> import javax.jms.JMSException;
<br>>>>>>
<br>>>>>> import org.activecluster.ClusterEvent;
<br>>>>>> import org.activecluster.ClusterFactory;
<br>>>>>> import org.activecluster.ClusterListener;
<br>>>>>> import org.activecluster.impl.DefaultClusterFactory;
<br>>>>>> import org.activemq.ActiveMQConnectionFactory;
<br>>>>>> import org.activemq.store.vm.VMPersistenceAdapterFactory;
<br>>>>>> import org.apache.commons.logging.Log;
<br>>>>>> import org.apache.commons.logging.LogFactory;
<br>>>>>> import org.codehaus.wadi.dindex.impl.SeniorityElectionStrategy;
<br>>>>>> import org.codehaus.wadi.impl.CustomClusterFactory;
<br>>>>>> import org.codehaus.wadi.sandbox.partition.Cluster;
<br>>>>>>
<br>>>>>> public class ACCluster {
<br>>>>>>
<br>>>>>> protected final Log _log = LogFactory.getLog(getClass());
<br>>>>>>
<br>>>>>> //protected final String _clusterUri="tcp://smilodon:61616";
<br>>>>>> protected final String _clusterUri="peer://org.codehaus.wadi";
<br>>>>>> protected final String _clusterName="ORG.CODEHAUS.WADI.TEST";
<br>>>>>> protected final ActiveMQConnectionFactory
<br>>>>>> _connectionFactory=new ActiveMQConnectionFactory(_clusterUri);
<br>>>>>> protected final ClusterFactory _clusterFactory=new
<br>>>>>> DefaultClusterFactory(_connectionFactory);
<br>>>>>> protected final org.activecluster.Cluster _cluster;
<br>>>>>> protected final long _timeout=30*1000L;
<br>>>>>>
<br>>>>>> public ACCluster(String nodeName) throws Exception {
<br>>>>>> System.setProperty
<br>>>>>> ("activemq.persistenceAdapterFactory",
<br>>>>>> VMPersistenceAdapterFactory.class.getName());
<br>>>>>> _cluster=_clusterFactory.createCluster(_clusterName);
<br>>>>>> Map state=new HashMap();
<br>>>>>> state.put("nodeName", nodeName);
<br>>>>>> _cluster.getLocalNode().setState(state);
<br>>>>>> _cluster.addClusterListener(new ClusterListener() {
<br>>>>>>
<br>>>>>> public void onNodeAdd(ClusterEvent arg0) {
<br>>>>>> _log.info("onNodeAdd: "+arg0.getNode().getState
<br>>>>>> ().get("nodeName"));
<br>>>>>> }
<br>>>>>>
<br>>>>>> public void onNodeUpdate(ClusterEvent arg0) {
<br>>>>>> _log.info("onNodeUpdate: "+arg0.getNode
<br>>>>>> ().getState ().get("nodeName"));
<br>>>>>> }
<br>>>>>>
<br>>>>>> public void onNodeRemoved(ClusterEvent arg0) {
<br>>>>>> _log.info("onNodeRemoved: "+arg0.getNode
<br>>>>>> ().getState ().get("nodeName"));
<br>>>>>> }
<br>>>>>>
<br>>>>>> public void onNodeFailed(ClusterEvent arg0) {
<br>>>>>> _log.info("onNodeFailed: "+arg0.getNode
<br>>>>>> ().getState ().get("nodeName"));
<br>>>>>> }
<br>>>>>>
<br>>>>>> public void onCoordinatorChanged(ClusterEvent arg0) {
<br>>>>>> _log.info("onCoordinatorChanged: "+arg0.getNode
<br>>>>>> ().getState().get("nodeName"));
<br>>>>>> }
<br>>>>>> });
<br>>>>>> }
<br>>>>>>
<br>>>>>> public void start() throws JMSException {
<br>>>>>> _log.info("starting...");
<br>>>>>> _cluster.start();
<br>>>>>> _log.info("...started");
<br>>>>>> }
<br>>>>>>
<br>>>>>> public void stop() {
<br>>>>>>
<br>>>>>> }
<br>>>>>>
<br>>>>>> public static void main(String[] args) throws Exception {
<br>>>>>> ACCluster cluster=new ACCluster(args[0]);
<br>>>>>> cluster.start();
<br>>>>>> Thread.sleep(100*1000);
<br>>>>>> cluster.stop();
<br>>>>>> }
<br>>>>>> }
<br>>>>>>
<br>>>>>>
<br>>>>>
<br>>>>>
<br>>>>> --
<br>>>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>>>> string into a super-saturated solution and a whole operating-system
<br>>>>> crystallises out around it."
<br>>>>>
<br>>>>> /**********************************
<br>>>>> * Jules Gosnell
<br>>>>> * Partner
<br>>>>> * Core Developers Network (Europe)
<br>>>>> *
<br>>>>> * www.coredevelopers.net
<br>>>>> *
<br>>>>> * Open Source Training & Support.
<br>>>>> **********************************/
<br>>>>>
<br>>>>>
<br>>>
<br>>>
<br>>> --
<br>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>> string into a super-saturated solution and a whole operating-system
<br>>> crystallises out around it."
<br>>>
<br>>> /**********************************
<br>>> * Jules Gosnell
<br>>> * Partner
<br>>> * Core Developers Network (Europe)
<br>>> *
<br>>> * www.coredevelopers.net
<br>>> *
<br>>> * Open Source Training & Support.
<br>>> **********************************/
<br>>>
</div><br><br>--
<br>"Open Source is a self-assembling organism. You dangle a piece of
<br>string into a super-saturated solution and a whole operating-system
<br>crystallises out around it."
<br><br>/**********************************
<br> * Jules Gosnell
<br> * Partner
<br> * Core Developers Network (Europe)
<br> *
<br> * www.coredevelopers.net
<br> *
<br> * Open Source Training & Support.
<br> **********************************/
<br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1363949Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-11-06T23:17:11Z2005-11-06T23:17:11Zrajdavieshi Jules,
<br><br>the latest ActiveCluster in the repo works with ActiveMQ 3.2 - amq4
<br>will hopefully be ready soon. I'd appreciate it if you could build
<br>activecluster from src - and see if your problems have disappeared -
<br>there was a a nasty deadlock I ended up fixing.
<br><br>cheers,
<br><br>Rob
<br><br>On 6 Nov 2005, at 18:41, Jules Gosnell wrote:
<br><div class='shrinkable-quote'><br>> Rob Davies wrote:
<br>>
<br>>> Hi Jules,
<br>>>
<br>>> I'm in the process of porting activecluster to amq4. There's an
<br>>> issue with discovery in amq4 at the moment that needs to be fixed.
<br>>
<br>> Ok - no problem - if you let me know when the fix goes in, I will
<br>> be happy to retest and report.
<br>>
<br>> BTW - do you know if AMQ4 is the release that is intended to be
<br>> bundled with Geronimo-1.0 ? I am targetting a WADI release at
<br>> Geronimo-1.0 and I need to make sure that we run on the same
<br>> version of AMQ as is packaged with G.
<br>>
<br>> Cheers, Rob,
<br>>
<br>> Jules
<br>>
<br>>>
<br>>> cheers,
<br>>>
<br>>> Rob
<br>>> On 29 Oct 2005, at 02:10, Jules Gosnell wrote:
<br>>>
<br>>>> Jules Gosnell wrote:
<br>>>>
<br>>>>
<br>>>>>
<br>>>>> tested against activecluster-1.1-SNAPSHOT, activemq-4.0-
<br>>>>> SNAPSHOT and Sun JDK 1.4.2_08 and 1.5.0 - two node cluster,
<br>>>>> both on same machine.
<br>>>>>
<br>>>>
<br>>>> oops - AMQ-4.0 was not on the machine that I ran this on - so
<br>>>> this issue is actually with AMQ-3.2-M1.
<br>>>>
<br>>>> I am trying to run the test with AMQ-4.0, but the second node
<br>>>> won't start because "java.io.IOException: Journal is allready
<br>>>> opened by another application", So i will have to figure out how
<br>>>> to tell it not to journal - I'll be back ;-)
<br>>>>
<br>>>> Jules
<br>>>>
<br>>>>
<br>>>>>
<br>>>>> start your first node (red)
<br>>>>>
<br>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster red
<br>>>>> 2005/10/29 00:04:18:317 BST [INFO] ACCluster - starting...
<br>>>>> 2005/10/29 00:04:18:321 BST [INFO] ACCluster - ...started
<br>>>>>
<br>>>>> start your second node (green)
<br>>>>>
<br>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>> 2005/10/29 00:04:31:874 BST [INFO] ACCluster - starting...
<br>>>>> 2005/10/29 00:04:31:885 BST [INFO] ACCluster - ...started
<br>>>>> 2005/10/29 00:04:33:323 BST [INFO] ACCluster - onNodeAdd: red
<br>>>>>
<br>>>>> red says:
<br>>>>>
<br>>>>> 2005/10/29 00:04:33:536 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>> 2005/10/29 00:04:33:537 BST [INFO] ACCluster -
<br>>>>> onCoordinatorChanged: green
<br>>>>>
<br>>>>> ctl-c green, after a few seconds red says:
<br>>>>>
<br>>>>> 2005/10/29 00:04:46:624 BST [INFO] ACCluster - onNodeFailed: green
<br>>>>> 2005/10/29 00:04:46:626 BST [INFO] ACCluster -
<br>>>>> onCoordinatorChanged: red
<br>>>>>
<br>>>>> all fine so far - now restart green:
<br>>>>>
<br>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>>> 2005/10/29 00:04:51:962 BST [INFO] ACCluster - starting...
<br>>>>> 2005/10/29 00:04:51:967 BST [INFO] ACCluster - ...started
<br>>>>>
<br>>>>> red says......nothing.
<br>>>>>
<br>>>>>
<br>>>>> if you then start a third node (blue):
<br>>>>>
<br>>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster blue
<br>>>>> 2005/10/29 00:08:47:990 BST [INFO] ACCluster - starting...
<br>>>>> 2005/10/29 00:08:47:994 BST [INFO] ACCluster - ...started
<br>>>>> 2005/10/29 00:08:50:198 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>>
<br>>>>> it sees green, but not red and green says:
<br>>>>>
<br>>>>> 2005/10/29 00:08:49:661 BST [INFO] ACCluster - onNodeAdd: blue
<br>>>>> 2005/10/29 00:08:49:662 BST [INFO] ACCluster -
<br>>>>> onCoordinatorChanged: blue
<br>>>>>
<br>>>>> so green sees blue too.
<br>>>>>
<br>>>>> but red says.....nothing.
<br>>>>>
<br>>>>>
<br>>>>> Using tcp:// and a broker, red will happily see blue and green
<br>>>>> after they are restarted.
<br>>>>>
<br>>>>> This has been the case with activecluster/activemq for a long
<br>>>>> time, but I was hoping that it would be fixed by improvements
<br>>>>> to peer:// in activemq-4.0 - but :-(
<br>>>>>
<br>>>>>
<br>>>>> Jules
<br>>>>>
<br>>>>> -------------------------------------------------------------------
<br>>>>> -- ---
<br>>>>>
<br>>>>> package org.codehaus.wadi.sandbox.partition.impl;
<br>>>>>
<br>>>>> import java.util.HashMap;
<br>>>>> import java.util.Map;
<br>>>>>
<br>>>>> import javax.jms.JMSException;
<br>>>>>
<br>>>>> import org.activecluster.ClusterEvent;
<br>>>>> import org.activecluster.ClusterFactory;
<br>>>>> import org.activecluster.ClusterListener;
<br>>>>> import org.activecluster.impl.DefaultClusterFactory;
<br>>>>> import org.activemq.ActiveMQConnectionFactory;
<br>>>>> import org.activemq.store.vm.VMPersistenceAdapterFactory;
<br>>>>> import org.apache.commons.logging.Log;
<br>>>>> import org.apache.commons.logging.LogFactory;
<br>>>>> import org.codehaus.wadi.dindex.impl.SeniorityElectionStrategy;
<br>>>>> import org.codehaus.wadi.impl.CustomClusterFactory;
<br>>>>> import org.codehaus.wadi.sandbox.partition.Cluster;
<br>>>>>
<br>>>>> public class ACCluster {
<br>>>>>
<br>>>>> protected final Log _log = LogFactory.getLog(getClass());
<br>>>>>
<br>>>>> //protected final String _clusterUri="tcp://smilodon:61616";
<br>>>>> protected final String _clusterUri="peer://org.codehaus.wadi";
<br>>>>> protected final String _clusterName="ORG.CODEHAUS.WADI.TEST";
<br>>>>> protected final ActiveMQConnectionFactory
<br>>>>> _connectionFactory=new ActiveMQConnectionFactory(_clusterUri);
<br>>>>> protected final ClusterFactory _clusterFactory=new
<br>>>>> DefaultClusterFactory(_connectionFactory);
<br>>>>> protected final org.activecluster.Cluster _cluster;
<br>>>>> protected final long _timeout=30*1000L;
<br>>>>>
<br>>>>> public ACCluster(String nodeName) throws Exception {
<br>>>>> System.setProperty
<br>>>>> ("activemq.persistenceAdapterFactory",
<br>>>>> VMPersistenceAdapterFactory.class.getName());
<br>>>>> _cluster=_clusterFactory.createCluster(_clusterName);
<br>>>>> Map state=new HashMap();
<br>>>>> state.put("nodeName", nodeName);
<br>>>>> _cluster.getLocalNode().setState(state);
<br>>>>> _cluster.addClusterListener(new ClusterListener() {
<br>>>>>
<br>>>>> public void onNodeAdd(ClusterEvent arg0) {
<br>>>>> _log.info("onNodeAdd: "+arg0.getNode().getState
<br>>>>> ().get("nodeName"));
<br>>>>> }
<br>>>>>
<br>>>>> public void onNodeUpdate(ClusterEvent arg0) {
<br>>>>> _log.info("onNodeUpdate: "+arg0.getNode
<br>>>>> ().getState ().get("nodeName"));
<br>>>>> }
<br>>>>>
<br>>>>> public void onNodeRemoved(ClusterEvent arg0) {
<br>>>>> _log.info("onNodeRemoved: "+arg0.getNode
<br>>>>> ().getState ().get("nodeName"));
<br>>>>> }
<br>>>>>
<br>>>>> public void onNodeFailed(ClusterEvent arg0) {
<br>>>>> _log.info("onNodeFailed: "+arg0.getNode
<br>>>>> ().getState ().get("nodeName"));
<br>>>>> }
<br>>>>>
<br>>>>> public void onCoordinatorChanged(ClusterEvent arg0) {
<br>>>>> _log.info("onCoordinatorChanged: "+arg0.getNode
<br>>>>> ().getState().get("nodeName"));
<br>>>>> }
<br>>>>> });
<br>>>>> }
<br>>>>>
<br>>>>> public void start() throws JMSException {
<br>>>>> _log.info("starting...");
<br>>>>> _cluster.start();
<br>>>>> _log.info("...started");
<br>>>>> }
<br>>>>>
<br>>>>> public void stop() {
<br>>>>>
<br>>>>> }
<br>>>>>
<br>>>>> public static void main(String[] args) throws Exception {
<br>>>>> ACCluster cluster=new ACCluster(args[0]);
<br>>>>> cluster.start();
<br>>>>> Thread.sleep(100*1000);
<br>>>>> cluster.stop();
<br>>>>> }
<br>>>>> }
<br>>>>>
<br>>>>>
<br>>>>
<br>>>>
<br>>>> --
<br>>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>>> string into a super-saturated solution and a whole operating-system
<br>>>> crystallises out around it."
<br>>>>
<br>>>> /**********************************
<br>>>> * Jules Gosnell
<br>>>> * Partner
<br>>>> * Core Developers Network (Europe)
<br>>>> *
<br>>>> * www.coredevelopers.net
<br>>>> *
<br>>>> * Open Source Training & Support.
<br>>>> **********************************/
<br>>>>
<br>>>>
<br>>
<br>>
<br>> --
<br>> "Open Source is a self-assembling organism. You dangle a piece of
<br>> string into a super-saturated solution and a whole operating-system
<br>> crystallises out around it."
<br>>
<br>> /**********************************
<br>> * Jules Gosnell
<br>> * Partner
<br>> * Core Developers Network (Europe)
<br>> *
<br>> * www.coredevelopers.net
<br>> *
<br>> * Open Source Training & Support.
<br>> **********************************/
<br>>
</div><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1358686Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-11-06T10:41:13Z2005-11-06T10:41:13ZJules Gosnell-4Rob Davies wrote:
<br><br>> Hi Jules,
<br>>
<br>> I'm in the process of porting activecluster to amq4. There's an issue
<br>> with discovery in amq4 at the moment that needs to be fixed.
<br><br>Ok - no problem - if you let me know when the fix goes in, I will be
<br>happy to retest and report.
<br><br>BTW - do you know if AMQ4 is the release that is intended to be bundled
<br>with Geronimo-1.0 ? I am targetting a WADI release at Geronimo-1.0 and I
<br>need to make sure that we run on the same version of AMQ as is packaged
<br>with G.
<br><br>Cheers, Rob,
<br><br>Jules
<br><div class='shrinkable-quote'><br>>
<br>> cheers,
<br>>
<br>> Rob
<br>> On 29 Oct 2005, at 02:10, Jules Gosnell wrote:
<br>>
<br>>> Jules Gosnell wrote:
<br>>>
<br>>>
<br>>>>
<br>>>> tested against activecluster-1.1-SNAPSHOT, activemq-4.0-SNAPSHOT
<br>>>> and Sun JDK 1.4.2_08 and 1.5.0 - two node cluster, both on same
<br>>>> machine.
<br>>>>
<br>>>
<br>>> oops - AMQ-4.0 was not on the machine that I ran this on - so this
<br>>> issue is actually with AMQ-3.2-M1.
<br>>>
<br>>> I am trying to run the test with AMQ-4.0, but the second node won't
<br>>> start because "java.io.IOException: Journal is allready opened by
<br>>> another application", So i will have to figure out how to tell it
<br>>> not to journal - I'll be back ;-)
<br>>>
<br>>> Jules
<br>>>
<br>>>
<br>>>>
<br>>>> start your first node (red)
<br>>>>
<br>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster red
<br>>>> 2005/10/29 00:04:18:317 BST [INFO] ACCluster - starting...
<br>>>> 2005/10/29 00:04:18:321 BST [INFO] ACCluster - ...started
<br>>>>
<br>>>> start your second node (green)
<br>>>>
<br>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>> 2005/10/29 00:04:31:874 BST [INFO] ACCluster - starting...
<br>>>> 2005/10/29 00:04:31:885 BST [INFO] ACCluster - ...started
<br>>>> 2005/10/29 00:04:33:323 BST [INFO] ACCluster - onNodeAdd: red
<br>>>>
<br>>>> red says:
<br>>>>
<br>>>> 2005/10/29 00:04:33:536 BST [INFO] ACCluster - onNodeAdd: green
<br>>>> 2005/10/29 00:04:33:537 BST [INFO] ACCluster -
<br>>>> onCoordinatorChanged: green
<br>>>>
<br>>>> ctl-c green, after a few seconds red says:
<br>>>>
<br>>>> 2005/10/29 00:04:46:624 BST [INFO] ACCluster - onNodeFailed: green
<br>>>> 2005/10/29 00:04:46:626 BST [INFO] ACCluster -
<br>>>> onCoordinatorChanged: red
<br>>>>
<br>>>> all fine so far - now restart green:
<br>>>>
<br>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>>> 2005/10/29 00:04:51:962 BST [INFO] ACCluster - starting...
<br>>>> 2005/10/29 00:04:51:967 BST [INFO] ACCluster - ...started
<br>>>>
<br>>>> red says......nothing.
<br>>>>
<br>>>>
<br>>>> if you then start a third node (blue):
<br>>>>
<br>>>> [jules@zeuglodon core]$ ./cluster.sh activecluster blue
<br>>>> 2005/10/29 00:08:47:990 BST [INFO] ACCluster - starting...
<br>>>> 2005/10/29 00:08:47:994 BST [INFO] ACCluster - ...started
<br>>>> 2005/10/29 00:08:50:198 BST [INFO] ACCluster - onNodeAdd: green
<br>>>>
<br>>>> it sees green, but not red and green says:
<br>>>>
<br>>>> 2005/10/29 00:08:49:661 BST [INFO] ACCluster - onNodeAdd: blue
<br>>>> 2005/10/29 00:08:49:662 BST [INFO] ACCluster -
<br>>>> onCoordinatorChanged: blue
<br>>>>
<br>>>> so green sees blue too.
<br>>>>
<br>>>> but red says.....nothing.
<br>>>>
<br>>>>
<br>>>> Using tcp:// and a broker, red will happily see blue and green
<br>>>> after they are restarted.
<br>>>>
<br>>>> This has been the case with activecluster/activemq for a long time,
<br>>>> but I was hoping that it would be fixed by improvements to peer://
<br>>>> in activemq-4.0 - but :-(
<br>>>>
<br>>>>
<br>>>> Jules
<br>>>>
<br>>>> ---------------------------------------------------------------------
<br>>>> ---
<br>>>>
<br>>>> package org.codehaus.wadi.sandbox.partition.impl;
<br>>>>
<br>>>> import java.util.HashMap;
<br>>>> import java.util.Map;
<br>>>>
<br>>>> import javax.jms.JMSException;
<br>>>>
<br>>>> import org.activecluster.ClusterEvent;
<br>>>> import org.activecluster.ClusterFactory;
<br>>>> import org.activecluster.ClusterListener;
<br>>>> import org.activecluster.impl.DefaultClusterFactory;
<br>>>> import org.activemq.ActiveMQConnectionFactory;
<br>>>> import org.activemq.store.vm.VMPersistenceAdapterFactory;
<br>>>> import org.apache.commons.logging.Log;
<br>>>> import org.apache.commons.logging.LogFactory;
<br>>>> import org.codehaus.wadi.dindex.impl.SeniorityElectionStrategy;
<br>>>> import org.codehaus.wadi.impl.CustomClusterFactory;
<br>>>> import org.codehaus.wadi.sandbox.partition.Cluster;
<br>>>>
<br>>>> public class ACCluster {
<br>>>>
<br>>>> protected final Log _log = LogFactory.getLog(getClass());
<br>>>>
<br>>>> //protected final String _clusterUri="tcp://smilodon:61616";
<br>>>> protected final String _clusterUri="peer://org.codehaus.wadi";
<br>>>> protected final String _clusterName="ORG.CODEHAUS.WADI.TEST";
<br>>>> protected final ActiveMQConnectionFactory _connectionFactory=new
<br>>>> ActiveMQConnectionFactory(_clusterUri);
<br>>>> protected final ClusterFactory _clusterFactory=new
<br>>>> DefaultClusterFactory(_connectionFactory);
<br>>>> protected final org.activecluster.Cluster _cluster;
<br>>>> protected final long _timeout=30*1000L;
<br>>>>
<br>>>> public ACCluster(String nodeName) throws Exception {
<br>>>> System.setProperty("activemq.persistenceAdapterFactory",
<br>>>> VMPersistenceAdapterFactory.class.getName());
<br>>>> _cluster=_clusterFactory.createCluster(_clusterName);
<br>>>> Map state=new HashMap();
<br>>>> state.put("nodeName", nodeName);
<br>>>> _cluster.getLocalNode().setState(state);
<br>>>> _cluster.addClusterListener(new ClusterListener() {
<br>>>>
<br>>>> public void onNodeAdd(ClusterEvent arg0) {
<br>>>> _log.info("onNodeAdd: "+arg0.getNode().getState
<br>>>> ().get("nodeName"));
<br>>>> }
<br>>>>
<br>>>> public void onNodeUpdate(ClusterEvent arg0) {
<br>>>> _log.info("onNodeUpdate: "+arg0.getNode().getState
<br>>>> ().get("nodeName"));
<br>>>> }
<br>>>>
<br>>>> public void onNodeRemoved(ClusterEvent arg0) {
<br>>>> _log.info("onNodeRemoved: "+arg0.getNode().getState
<br>>>> ().get("nodeName"));
<br>>>> }
<br>>>>
<br>>>> public void onNodeFailed(ClusterEvent arg0) {
<br>>>> _log.info("onNodeFailed: "+arg0.getNode().getState
<br>>>> ().get("nodeName"));
<br>>>> }
<br>>>>
<br>>>> public void onCoordinatorChanged(ClusterEvent arg0) {
<br>>>> _log.info("onCoordinatorChanged: "+arg0.getNode
<br>>>> ().getState().get("nodeName"));
<br>>>> }
<br>>>> });
<br>>>> }
<br>>>>
<br>>>> public void start() throws JMSException {
<br>>>> _log.info("starting...");
<br>>>> _cluster.start();
<br>>>> _log.info("...started");
<br>>>> }
<br>>>>
<br>>>> public void stop() {
<br>>>>
<br>>>> }
<br>>>>
<br>>>> public static void main(String[] args) throws Exception {
<br>>>> ACCluster cluster=new ACCluster(args[0]);
<br>>>> cluster.start();
<br>>>> Thread.sleep(100*1000);
<br>>>> cluster.stop();
<br>>>> }
<br>>>> }
<br>>>>
<br>>>>
<br>>>
<br>>>
<br>>> --
<br>>> "Open Source is a self-assembling organism. You dangle a piece of
<br>>> string into a super-saturated solution and a whole operating-system
<br>>> crystallises out around it."
<br>>>
<br>>> /**********************************
<br>>> * Jules Gosnell
<br>>> * Partner
<br>>> * Core Developers Network (Europe)
<br>>> *
<br>>> * www.coredevelopers.net
<br>>> *
<br>>> * Open Source Training & Support.
<br>>> **********************************/
<br>>>
<br>>>
</div><br><br>--
<br>"Open Source is a self-assembling organism. You dangle a piece of
<br>string into a super-saturated solution and a whole operating-system
<br>crystallises out around it."
<br><br>/**********************************
<br> * Jules Gosnell
<br> * Partner
<br> * Core Developers Network (Europe)
<br> *
<br> * www.coredevelopers.net
<br> *
<br> * Open Source Training & Support.
<br> **********************************/
<br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1282444Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-10-31T02:46:36Z2005-10-31T02:46:36ZJules Gosnell-4Jules Gosnell wrote:
<br><br>> Jules Gosnell wrote:
<br>>
<br>>>
<br>>> tested against activecluster-1.1-SNAPSHOT, activemq-4.0-SNAPSHOT and
<br>>> Sun JDK 1.4.2_08 and 1.5.0 - two node cluster, both on same machine.
<br>>
<br>>
<br>> oops - AMQ-4.0 was not on the machine that I ran this on - so this
<br>> issue is actually with AMQ-3.2-M1.
<br><br>I just ran this on AMQ-4.0
<br><br>The red and green nodes do not find each other in the first place. :-(
<br><br>So, in conclusion, this seems to be an AMQ issue (there does not appear
<br>to be any special case code for peer:// in AC).
<br><br>using AMQ-3.2, nodes fail to rejoin a cluster
<br>using AMQ-4.0, nodes fail to join a cluster in the first place
<br><br>thanks for you time,
<br><br><br>Jules
<br><div class='shrinkable-quote'><br>>
<br>> I am trying to run the test with AMQ-4.0, but the second node won't
<br>> start because "java.io.IOException: Journal is allready opened by
<br>> another application", So i will have to figure out how to tell it not
<br>> to journal - I'll be back ;-)
<br>>
<br>> Jules
<br>>
<br>>>
<br>>> start your first node (red)
<br>>>
<br>>> [jules@zeuglodon core]$ ./cluster.sh activecluster red
<br>>> 2005/10/29 00:04:18:317 BST [INFO] ACCluster - starting...
<br>>> 2005/10/29 00:04:18:321 BST [INFO] ACCluster - ...started
<br>>>
<br>>> start your second node (green)
<br>>>
<br>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>> 2005/10/29 00:04:31:874 BST [INFO] ACCluster - starting...
<br>>> 2005/10/29 00:04:31:885 BST [INFO] ACCluster - ...started
<br>>> 2005/10/29 00:04:33:323 BST [INFO] ACCluster - onNodeAdd: red
<br>>>
<br>>> red says:
<br>>>
<br>>> 2005/10/29 00:04:33:536 BST [INFO] ACCluster - onNodeAdd: green
<br>>> 2005/10/29 00:04:33:537 BST [INFO] ACCluster - onCoordinatorChanged:
<br>>> green
<br>>>
<br>>> ctl-c green, after a few seconds red says:
<br>>>
<br>>> 2005/10/29 00:04:46:624 BST [INFO] ACCluster - onNodeFailed: green
<br>>> 2005/10/29 00:04:46:626 BST [INFO] ACCluster - onCoordinatorChanged: red
<br>>>
<br>>> all fine so far - now restart green:
<br>>>
<br>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>> 2005/10/29 00:04:51:962 BST [INFO] ACCluster - starting...
<br>>> 2005/10/29 00:04:51:967 BST [INFO] ACCluster - ...started
<br>>>
<br>>> red says......nothing.
<br>>>
<br>>>
<br>>> if you then start a third node (blue):
<br>>>
<br>>> [jules@zeuglodon core]$ ./cluster.sh activecluster blue
<br>>> 2005/10/29 00:08:47:990 BST [INFO] ACCluster - starting...
<br>>> 2005/10/29 00:08:47:994 BST [INFO] ACCluster - ...started
<br>>> 2005/10/29 00:08:50:198 BST [INFO] ACCluster - onNodeAdd: green
<br>>>
<br>>> it sees green, but not red and green says:
<br>>>
<br>>> 2005/10/29 00:08:49:661 BST [INFO] ACCluster - onNodeAdd: blue
<br>>> 2005/10/29 00:08:49:662 BST [INFO] ACCluster - onCoordinatorChanged:
<br>>> blue
<br>>>
<br>>> so green sees blue too.
<br>>>
<br>>> but red says.....nothing.
<br>>>
<br>>>
<br>>> Using tcp:// and a broker, red will happily see blue and green after
<br>>> they are restarted.
<br>>>
<br>>> This has been the case with activecluster/activemq for a long time,
<br>>> but I was hoping that it would be fixed by improvements to peer:// in
<br>>> activemq-4.0 - but :-(
<br>>>
<br>>>
<br>>> Jules
<br>>>
<br>>> ------------------------------------------------------------------------
<br>>>
<br>>> package org.codehaus.wadi.sandbox.partition.impl;
<br>>>
<br>>> import java.util.HashMap;
<br>>> import java.util.Map;
<br>>>
<br>>> import javax.jms.JMSException;
<br>>>
<br>>> import org.activecluster.ClusterEvent;
<br>>> import org.activecluster.ClusterFactory;
<br>>> import org.activecluster.ClusterListener;
<br>>> import org.activecluster.impl.DefaultClusterFactory;
<br>>> import org.activemq.ActiveMQConnectionFactory;
<br>>> import org.activemq.store.vm.VMPersistenceAdapterFactory;
<br>>> import org.apache.commons.logging.Log;
<br>>> import org.apache.commons.logging.LogFactory;
<br>>> import org.codehaus.wadi.dindex.impl.SeniorityElectionStrategy;
<br>>> import org.codehaus.wadi.impl.CustomClusterFactory;
<br>>> import org.codehaus.wadi.sandbox.partition.Cluster;
<br>>>
<br>>> public class ACCluster {
<br>>>
<br>>> protected final Log _log = LogFactory.getLog(getClass());
<br>>>
<br>>> //protected final String _clusterUri="tcp://smilodon:61616";
<br>>> protected final String _clusterUri="peer://org.codehaus.wadi";
<br>>> protected final String _clusterName="ORG.CODEHAUS.WADI.TEST";
<br>>> protected final ActiveMQConnectionFactory _connectionFactory=new
<br>>> ActiveMQConnectionFactory(_clusterUri);
<br>>> protected final ClusterFactory _clusterFactory=new
<br>>> DefaultClusterFactory(_connectionFactory);
<br>>> protected final org.activecluster.Cluster _cluster;
<br>>> protected final long _timeout=30*1000L;
<br>>>
<br>>> public ACCluster(String nodeName) throws Exception {
<br>>> System.setProperty("activemq.persistenceAdapterFactory",
<br>>> VMPersistenceAdapterFactory.class.getName());
<br>>> _cluster=_clusterFactory.createCluster(_clusterName);
<br>>> Map state=new HashMap();
<br>>> state.put("nodeName", nodeName);
<br>>> _cluster.getLocalNode().setState(state);
<br>>> _cluster.addClusterListener(new ClusterListener() {
<br>>>
<br>>> public void onNodeAdd(ClusterEvent arg0) {
<br>>> _log.info("onNodeAdd:
<br>>> "+arg0.getNode().getState().get("nodeName"));
<br>>> }
<br>>>
<br>>> public void onNodeUpdate(ClusterEvent arg0) {
<br>>> _log.info("onNodeUpdate:
<br>>> "+arg0.getNode().getState().get("nodeName"));
<br>>> }
<br>>>
<br>>> public void onNodeRemoved(ClusterEvent arg0) {
<br>>> _log.info("onNodeRemoved:
<br>>> "+arg0.getNode().getState().get("nodeName"));
<br>>> }
<br>>>
<br>>> public void onNodeFailed(ClusterEvent arg0) {
<br>>> _log.info("onNodeFailed:
<br>>> "+arg0.getNode().getState().get("nodeName"));
<br>>> }
<br>>>
<br>>> public void onCoordinatorChanged(ClusterEvent arg0) {
<br>>> _log.info("onCoordinatorChanged:
<br>>> "+arg0.getNode().getState().get("nodeName"));
<br>>> }
<br>>> });
<br>>> }
<br>>>
<br>>> public void start() throws JMSException {
<br>>> _log.info("starting...");
<br>>> _cluster.start();
<br>>> _log.info("...started");
<br>>> }
<br>>>
<br>>> public void stop() {
<br>>>
<br>>> }
<br>>>
<br>>> public static void main(String[] args) throws Exception {
<br>>> ACCluster cluster=new ACCluster(args[0]);
<br>>> cluster.start();
<br>>> Thread.sleep(100*1000);
<br>>> cluster.stop();
<br>>> }
<br>>> }
<br>>>
<br>>>
<br>>
<br>>
</div><br><br>--
<br>"Open Source is a self-assembling organism. You dangle a piece of
<br>string into a super-saturated solution and a whole operating-system
<br>crystallises out around it."
<br><br>/**********************************
<br> * Jules Gosnell
<br> * Partner
<br> * Core Developers Network (Europe)
<br> *
<br> * www.coredevelopers.net
<br> *
<br> * Open Source Training & Support.
<br> **********************************/
<br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1335723Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-10-29T00:24:54Z2005-10-29T00:24:54ZrajdaviesHi Jules,
<br><br>I'm in the process of porting activecluster to amq4. There's an issue
<br>with discovery in amq4 at the moment that needs to be fixed.
<br><br>cheers,
<br><br>Rob
<br>On 29 Oct 2005, at 02:10, Jules Gosnell wrote:
<br><div class='shrinkable-quote'><br>> Jules Gosnell wrote:
<br>>
<br>>
<br>>>
<br>>> tested against activecluster-1.1-SNAPSHOT, activemq-4.0-SNAPSHOT
<br>>> and Sun JDK 1.4.2_08 and 1.5.0 - two node cluster, both on same
<br>>> machine.
<br>>>
<br>>
<br>> oops - AMQ-4.0 was not on the machine that I ran this on - so this
<br>> issue is actually with AMQ-3.2-M1.
<br>>
<br>> I am trying to run the test with AMQ-4.0, but the second node won't
<br>> start because "java.io.IOException: Journal is allready opened by
<br>> another application", So i will have to figure out how to tell it
<br>> not to journal - I'll be back ;-)
<br>>
<br>> Jules
<br>>
<br>>
<br>>>
<br>>> start your first node (red)
<br>>>
<br>>> [jules@zeuglodon core]$ ./cluster.sh activecluster red
<br>>> 2005/10/29 00:04:18:317 BST [INFO] ACCluster - starting...
<br>>> 2005/10/29 00:04:18:321 BST [INFO] ACCluster - ...started
<br>>>
<br>>> start your second node (green)
<br>>>
<br>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>> 2005/10/29 00:04:31:874 BST [INFO] ACCluster - starting...
<br>>> 2005/10/29 00:04:31:885 BST [INFO] ACCluster - ...started
<br>>> 2005/10/29 00:04:33:323 BST [INFO] ACCluster - onNodeAdd: red
<br>>>
<br>>> red says:
<br>>>
<br>>> 2005/10/29 00:04:33:536 BST [INFO] ACCluster - onNodeAdd: green
<br>>> 2005/10/29 00:04:33:537 BST [INFO] ACCluster -
<br>>> onCoordinatorChanged: green
<br>>>
<br>>> ctl-c green, after a few seconds red says:
<br>>>
<br>>> 2005/10/29 00:04:46:624 BST [INFO] ACCluster - onNodeFailed: green
<br>>> 2005/10/29 00:04:46:626 BST [INFO] ACCluster -
<br>>> onCoordinatorChanged: red
<br>>>
<br>>> all fine so far - now restart green:
<br>>>
<br>>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>>> 2005/10/29 00:04:51:962 BST [INFO] ACCluster - starting...
<br>>> 2005/10/29 00:04:51:967 BST [INFO] ACCluster - ...started
<br>>>
<br>>> red says......nothing.
<br>>>
<br>>>
<br>>> if you then start a third node (blue):
<br>>>
<br>>> [jules@zeuglodon core]$ ./cluster.sh activecluster blue
<br>>> 2005/10/29 00:08:47:990 BST [INFO] ACCluster - starting...
<br>>> 2005/10/29 00:08:47:994 BST [INFO] ACCluster - ...started
<br>>> 2005/10/29 00:08:50:198 BST [INFO] ACCluster - onNodeAdd: green
<br>>>
<br>>> it sees green, but not red and green says:
<br>>>
<br>>> 2005/10/29 00:08:49:661 BST [INFO] ACCluster - onNodeAdd: blue
<br>>> 2005/10/29 00:08:49:662 BST [INFO] ACCluster -
<br>>> onCoordinatorChanged: blue
<br>>>
<br>>> so green sees blue too.
<br>>>
<br>>> but red says.....nothing.
<br>>>
<br>>>
<br>>> Using tcp:// and a broker, red will happily see blue and green
<br>>> after they are restarted.
<br>>>
<br>>> This has been the case with activecluster/activemq for a long
<br>>> time, but I was hoping that it would be fixed by improvements to
<br>>> peer:// in activemq-4.0 - but :-(
<br>>>
<br>>>
<br>>> Jules
<br>>>
<br>>> ---------------------------------------------------------------------
<br>>> ---
<br>>>
<br>>> package org.codehaus.wadi.sandbox.partition.impl;
<br>>>
<br>>> import java.util.HashMap;
<br>>> import java.util.Map;
<br>>>
<br>>> import javax.jms.JMSException;
<br>>>
<br>>> import org.activecluster.ClusterEvent;
<br>>> import org.activecluster.ClusterFactory;
<br>>> import org.activecluster.ClusterListener;
<br>>> import org.activecluster.impl.DefaultClusterFactory;
<br>>> import org.activemq.ActiveMQConnectionFactory;
<br>>> import org.activemq.store.vm.VMPersistenceAdapterFactory;
<br>>> import org.apache.commons.logging.Log;
<br>>> import org.apache.commons.logging.LogFactory;
<br>>> import org.codehaus.wadi.dindex.impl.SeniorityElectionStrategy;
<br>>> import org.codehaus.wadi.impl.CustomClusterFactory;
<br>>> import org.codehaus.wadi.sandbox.partition.Cluster;
<br>>>
<br>>> public class ACCluster {
<br>>>
<br>>> protected final Log _log = LogFactory.getLog(getClass());
<br>>>
<br>>> //protected final String _clusterUri="tcp://smilodon:61616";
<br>>> protected final String _clusterUri="peer://org.codehaus.wadi";
<br>>> protected final String _clusterName="ORG.CODEHAUS.WADI.TEST";
<br>>> protected final ActiveMQConnectionFactory
<br>>> _connectionFactory=new ActiveMQConnectionFactory(_clusterUri);
<br>>> protected final ClusterFactory _clusterFactory=new
<br>>> DefaultClusterFactory(_connectionFactory);
<br>>> protected final org.activecluster.Cluster _cluster;
<br>>> protected final long _timeout=30*1000L;
<br>>>
<br>>> public ACCluster(String nodeName) throws Exception {
<br>>> System.setProperty("activemq.persistenceAdapterFactory",
<br>>> VMPersistenceAdapterFactory.class.getName());
<br>>> _cluster=_clusterFactory.createCluster(_clusterName);
<br>>> Map state=new HashMap();
<br>>> state.put("nodeName", nodeName);
<br>>> _cluster.getLocalNode().setState(state);
<br>>> _cluster.addClusterListener(new ClusterListener() {
<br>>>
<br>>> public void onNodeAdd(ClusterEvent arg0) {
<br>>> _log.info("onNodeAdd: "+arg0.getNode().getState
<br>>> ().get("nodeName"));
<br>>> }
<br>>>
<br>>> public void onNodeUpdate(ClusterEvent arg0) {
<br>>> _log.info("onNodeUpdate: "+arg0.getNode().getState
<br>>> ().get("nodeName"));
<br>>> }
<br>>>
<br>>> public void onNodeRemoved(ClusterEvent arg0) {
<br>>> _log.info("onNodeRemoved: "+arg0.getNode().getState
<br>>> ().get("nodeName"));
<br>>> }
<br>>>
<br>>> public void onNodeFailed(ClusterEvent arg0) {
<br>>> _log.info("onNodeFailed: "+arg0.getNode().getState
<br>>> ().get("nodeName"));
<br>>> }
<br>>>
<br>>> public void onCoordinatorChanged(ClusterEvent arg0) {
<br>>> _log.info("onCoordinatorChanged: "+arg0.getNode
<br>>> ().getState().get("nodeName"));
<br>>> }
<br>>> });
<br>>> }
<br>>>
<br>>> public void start() throws JMSException {
<br>>> _log.info("starting...");
<br>>> _cluster.start();
<br>>> _log.info("...started");
<br>>> }
<br>>>
<br>>> public void stop() {
<br>>>
<br>>> }
<br>>>
<br>>> public static void main(String[] args) throws Exception {
<br>>> ACCluster cluster=new ACCluster(args[0]);
<br>>> cluster.start();
<br>>> Thread.sleep(100*1000);
<br>>> cluster.stop();
<br>>> }
<br>>> }
<br>>>
<br>>>
<br>>
<br>>
<br>> --
<br>> "Open Source is a self-assembling organism. You dangle a piece of
<br>> string into a super-saturated solution and a whole operating-system
<br>> crystallises out around it."
<br>>
<br>> /**********************************
<br>> * Jules Gosnell
<br>> * Partner
<br>> * Core Developers Network (Europe)
<br>> *
<br>> * www.coredevelopers.net
<br>> *
<br>> * Open Source Training & Support.
<br>> **********************************/
<br>>
<br>>
</div><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>tag:old.nabble.com,2006:post-1266842Re: [activecluster-user] peer:// - nodes do not rejoin cluster... (test enclosed)2005-10-28T19:10:28Z2005-10-28T19:10:28ZJules Gosnell-4Jules Gosnell wrote:
<br><br>>
<br>> tested against activecluster-1.1-SNAPSHOT, activemq-4.0-SNAPSHOT and
<br>> Sun JDK 1.4.2_08 and 1.5.0 - two node cluster, both on same machine.
<br><br>oops - AMQ-4.0 was not on the machine that I ran this on - so this issue
<br>is actually with AMQ-3.2-M1.
<br><br>I am trying to run the test with AMQ-4.0, but the second node won't
<br>start because "java.io.IOException: Journal is allready opened by
<br>another application", So i will have to figure out how to tell it not to
<br>journal - I'll be back ;-)
<br><br>Jules
<br><div class='shrinkable-quote'><br>>
<br>> start your first node (red)
<br>>
<br>> [jules@zeuglodon core]$ ./cluster.sh activecluster red
<br>> 2005/10/29 00:04:18:317 BST [INFO] ACCluster - starting...
<br>> 2005/10/29 00:04:18:321 BST [INFO] ACCluster - ...started
<br>>
<br>> start your second node (green)
<br>>
<br>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>> 2005/10/29 00:04:31:874 BST [INFO] ACCluster - starting...
<br>> 2005/10/29 00:04:31:885 BST [INFO] ACCluster - ...started
<br>> 2005/10/29 00:04:33:323 BST [INFO] ACCluster - onNodeAdd: red
<br>>
<br>> red says:
<br>>
<br>> 2005/10/29 00:04:33:536 BST [INFO] ACCluster - onNodeAdd: green
<br>> 2005/10/29 00:04:33:537 BST [INFO] ACCluster - onCoordinatorChanged:
<br>> green
<br>>
<br>> ctl-c green, after a few seconds red says:
<br>>
<br>> 2005/10/29 00:04:46:624 BST [INFO] ACCluster - onNodeFailed: green
<br>> 2005/10/29 00:04:46:626 BST [INFO] ACCluster - onCoordinatorChanged: red
<br>>
<br>> all fine so far - now restart green:
<br>>
<br>> [jules@zeuglodon core]$ ./cluster.sh activecluster green
<br>> 2005/10/29 00:04:51:962 BST [INFO] ACCluster - starting...
<br>> 2005/10/29 00:04:51:967 BST [INFO] ACCluster - ...started
<br>>
<br>> red says......nothing.
<br>>
<br>>
<br>> if you then start a third node (blue):
<br>>
<br>> [jules@zeuglodon core]$ ./cluster.sh activecluster blue
<br>> 2005/10/29 00:08:47:990 BST [INFO] ACCluster - starting...
<br>> 2005/10/29 00:08:47:994 BST [INFO] ACCluster - ...started
<br>> 2005/10/29 00:08:50:198 BST [INFO] ACCluster - onNodeAdd: green
<br>>
<br>> it sees green, but not red and green says:
<br>>
<br>> 2005/10/29 00:08:49:661 BST [INFO] ACCluster - onNodeAdd: blue
<br>> 2005/10/29 00:08:49:662 BST [INFO] ACCluster - onCoordinatorChanged: blue
<br>>
<br>> so green sees blue too.
<br>>
<br>> but red says.....nothing.
<br>>
<br>>
<br>> Using tcp:// and a broker, red will happily see blue and green after
<br>> they are restarted.
<br>>
<br>> This has been the case with activecluster/activemq for a long time,
<br>> but I was hoping that it would be fixed by improvements to peer:// in
<br>> activemq-4.0 - but :-(
<br>>
<br>>
<br>> Jules
<br>>
<br>>------------------------------------------------------------------------
<br>>
<br>>package org.codehaus.wadi.sandbox.partition.impl;
<br>>
<br>>import java.util.HashMap;
<br>>import java.util.Map;
<br>>
<br>>import javax.jms.JMSException;
<br>>
<br>>import org.activecluster.ClusterEvent;
<br>>import org.activecluster.ClusterFactory;
<br>>import org.activecluster.ClusterListener;
<br>>import org.activecluster.impl.DefaultClusterFactory;
<br>>import org.activemq.ActiveMQConnectionFactory;
<br>>import org.activemq.store.vm.VMPersistenceAdapterFactory;
<br>>import org.apache.commons.logging.Log;
<br>>import org.apache.commons.logging.LogFactory;
<br>>import org.codehaus.wadi.dindex.impl.SeniorityElectionStrategy;
<br>>import org.codehaus.wadi.impl.CustomClusterFactory;
<br>>import org.codehaus.wadi.sandbox.partition.Cluster;
<br>>
<br>>public class ACCluster {
<br>>
<br>> protected final Log _log = LogFactory.getLog(getClass());
<br>>
<br>> //protected final String _clusterUri="tcp://smilodon:61616";
<br>> protected final String _clusterUri="peer://org.codehaus.wadi";
<br>> protected final String _clusterName="ORG.CODEHAUS.WADI.TEST";
<br>> protected final ActiveMQConnectionFactory _connectionFactory=new ActiveMQConnectionFactory(_clusterUri);
<br>> protected final ClusterFactory _clusterFactory=new DefaultClusterFactory(_connectionFactory);
<br>> protected final org.activecluster.Cluster _cluster;
<br>> protected final long _timeout=30*1000L;
<br>>
<br>> public ACCluster(String nodeName) throws Exception {
<br>> System.setProperty("activemq.persistenceAdapterFactory", VMPersistenceAdapterFactory.class.getName());
<br>> _cluster=_clusterFactory.createCluster(_clusterName);
<br>> Map state=new HashMap();
<br>> state.put("nodeName", nodeName);
<br>> _cluster.getLocalNode().setState(state);
<br>> _cluster.addClusterListener(new ClusterListener() {
<br>>
<br>> public void onNodeAdd(ClusterEvent arg0) {
<br>> _log.info("onNodeAdd: "+arg0.getNode().getState().get("nodeName"));
<br>> }
<br>>
<br>> public void onNodeUpdate(ClusterEvent arg0) {
<br>> _log.info("onNodeUpdate: "+arg0.getNode().getState().get("nodeName"));
<br>> }
<br>>
<br>> public void onNodeRemoved(ClusterEvent arg0) {
<br>> _log.info("onNodeRemoved: "+arg0.getNode().getState().get("nodeName"));
<br>> }
<br>>
<br>> public void onNodeFailed(ClusterEvent arg0) {
<br>> _log.info("onNodeFailed: "+arg0.getNode().getState().get("nodeName"));
<br>> }
<br>>
<br>> public void onCoordinatorChanged(ClusterEvent arg0) {
<br>> _log.info("onCoordinatorChanged: "+arg0.getNode().getState().get("nodeName"));
<br>> }
<br>> });
<br>> }
<br>>
<br>> public void start() throws JMSException {
<br>> _log.info("starting...");
<br>> _cluster.start();
<br>> _log.info("...started");
<br>> }
<br>>
<br>> public void stop() {
<br>>
<br>> }
<br>>
<br>> public static void main(String[] args) throws Exception {
<br>> ACCluster cluster=new ACCluster(args[0]);
<br>> cluster.start();
<br>> Thread.sleep(100*1000);
<br>> cluster.stop();
<br>> }
<br>>}
<br>>
<br>>
</div><br><br>--
<br>"Open Source is a self-assembling organism. You dangle a piece of
<br>string into a super-saturated solution and a whole operating-system
<br>crystallises out around it."
<br><br>/**********************************
<br> * Jules Gosnell
<br> * Partner
<br> * Core Developers Network (Europe)
<br> *
<br> * www.coredevelopers.net
<br> *
<br> * Open Source Training & Support.
<br> **********************************/
<br><br><p>From forum: <a href="http://old.nabble.com/activecluster---user-f11767.html" embed="fixTarget[11767]" target="_top" >activecluster - user</a></p>