hmm, well, i'm only running one broker per machine actually. you mention that i might be initializing the jms before the broker, but if i am, i'm not aware of it. i'm simply starting the broker using cmd-line "activemq xbean:file:./foo.xml" then ssh'ing to the next box and doing the same. here's my full "foo.xml" startup file, if you don't mind taking a look to see if anything looks odd - thanks
<?xml version="1.0" encoding="UTF-8"?> <beans xmlns:amq="http://activemq.org/config/1.0";> <amq:broker brokerName="brokerA" useJmx="true"> <amq:destinationPolicy> <amq:policyMap> <amq:policyEntries> <amq:policyEntry topic="FOO.>"> <amq:dispatchPolicy> <amq:strictOrderDispatchPolicy/> </amq:dispatchPolicy> <amq:subscriptionRecoveryPolicy> <amq:lastImageSubscriptionRecoveryPolicy/> </amq:subscriptionRecoveryPolicy> </amq:policyEntry> <amq:policyEntry topic="Prices.>"> <amq:pendingMessageLimitStrategy> <amq:constantPendingMessageLimitStrategy limit="0"/> </amq:pendingMessageLimitStrategy> </amq:policyEntry> </amq:policyEntries> </amq:policyMap> </amq:destinationPolicy> <amq:persistenceAdapter> <amq:journalPersistenceAdapter> <constructor-arg index="0"> <ref bean="myJournalImpl"/> </constructor-arg> <constructor-arg index="1"> <amq:memoryPersistenceAdapter init-method="createTransactionStore"/> </constructor-arg> <constructor-arg index="2"> <bean id="myTaskRunnerFactory" class="org.apache.activemq.thread.TaskRunnerFactory"/> </constructor-arg> </amq:journalPersistenceAdapter> </amq:persistenceAdapter> <amq:transportConnectors> <amq:transportConnector uri="tcp://localhost:0" discoveryUri="multicast://default"/> </amq:transportConnectors> <amq:networkConnectors> <!-- by default just auto discover the other brokers --> <amq:networkConnector uri="multicast://default"/> </amq:networkConnectors> </amq:broker> <bean id="myJournalImpl" class="org.apache.activeio.journal.active.JournalImpl"> <constructor-arg index="0"> <bean id="myFile" class="java.io.File"> <constructor-arg index="0"> <value>/home/ptomsic/projects/mbus/journal</value> </constructor-arg> </bean> </constructor-arg> </bean> </beans> --- James Strachan <[EMAIL PROTECTED]> wrote: > On 5/6/06, Paul Tomsic <[EMAIL PROTECTED]> wrote: > > there's no replication at the OS level, if that's > what > > you're asking. > > no NFS magic at all, actually. i've even tried > > putting the journals in > > box1:/tmp/foo, box2:/tmp/bar, box3:/tmp/baz, etc. > > > > so, each journal is a different directory, in that > it > > resides on a different machine. > > it just seems weird that i can start box[1], then > > box[2], but box[3] fails due to the aforementioned > > issues. > > any other suggestions? > > Yes - see the other scenario in the FAQ... > http://goopen.org/confluence/display/ACTIVEMQ/Journal+is+already+opened+by+this+application > > basically the only way this exception occurs is if > another process > locks the file or you start two brokers. So I > suspect you are > actually running 2 brokers in a JVM - where you > start a JMS connection > using the vm:// connector before you have > initialised the broker you > intended to, thus causing 2 brokers to start. So any > chance you are > creating a JMS connection before your broker? > > -- > > James > ------- > http://radio.weblogs.com/0112098/ >
