Can anyone help me with this?
On Wed, Jun 5, 2013 at 6:24 PM, Jean Silga <jean.si...@net-vitesse.com>wrote:
> I tried in two places without success.:
>
> 1) In the destination policy of the broker
>
> amq.'broker'(id: 'broker', brokerName: 'localhost', useJmx: true,
> persistent: false, advisorySupport='true') {
> amq.'managementContext' {
> amq.'managementContext'(connectorPort: 2011,
> jmxDomainName: 'embeddedBroker')
> }
>
> amq.'systemUsage' {
> amq.'systemUsage' {
> amq.'memoryUsage' {
> amq.'memoryUsage'(limit:"200 mb")
> }
> }
> }
>
> * amq.'destinationPolicy' {*
> amq.'policyMap' {
> amq.'policyEntries' {
> amq.'policyEntry'(topic: "index.mixml",
> memoryLimit: "100mb")
> amq.'policyEntry'(topic:
> "index.onedirectory", memoryLimit: "150mb")
> * amq.'policyEntry'(topic:
> "index.onedirectory", durableTopicPrefetch: "2")
> amq.'policyEntry'(topic:
> "index.onedirectory", topicPrefetch: "2")*
>
> }
> }
> * }*
>
> amq.'plugins' {
> amq.'loggingBrokerPlugin'
> }
>
> if (Environment.current != Environment.TEST) {
> amq.'transportConnectors' {
> amq.'transportConnector'(name: 'connector1',
> uri:'nio://0.0.0.0:61616')
> amq.'transportConnector'(name: 'connector2',
> uri:'vm://localhost')
> if (nvconnect.clustering == true &&
> Environment.current == Environment.PRODUCTION)
> amq.'transportConnector'(uri: 'tcp://
> 0.0.0.0:0',
>
> discoveryUri:'multicast://default?group=nvconnect')
> }
> }
>
> if (nvconnect.clustering == true &&
> Environment.current == Environment.PRODUCTION) {
> amq.'networkConnectors' {
> amq.'networkConnector'(uri:
> 'multicast://default?group=nvconnect', duplex: "true", dynamicOnly: "true",
> prefetchSize:5) {
> amq.'dynamicallyIncludedDestinations'{
> amq.'topic'(physicalName:'index.mixml')
>
> amq.'topic'(physicalName:'index.mixml.response')
>
> amq.'topic'(physicalName:'index.onedirectory')
>
> amq.'topic'(physicalName:'index.onedirectory.response')
> }
> }
> }
> }
> }
>
>
> 2) In the connection factory
>
>
> jmsConnectionFactory(org.springframework.jms.connection.SingleConnectionFactory)
> { bean ->
> bean.dependsOn = "broker"
> clientId = "nv-jms-connection-" + UUID.randomUUID()
> targetConnectionFactory = {
> org.apache.activemq.ActiveMQConnectionFactory cf ->
> brokerURL ='vm://localhost'
> * prefetchPolicy = ref("customPrefetchPolicy")*
> }
> }
>
> *
> customPrefetchPolicy(org.apache.activemq.ActiveMQPrefetchPolicy){ bean ->
> topicPrefetch=2
> durableTopicPrefetch=2
> }*
>
>
> "OnedirectoryIndexJmsListenerContainer"(org.springframework.jms.listener.DefaultMessageListenerContainer)
> {
> it.'abstract' = true
> pubSubDomain = true
> autoStartup = false
> concurrentConsumers = 1
> subscriptionDurable = true
> * connectionFactory = ref("jmsConnectionFactory")*
> messageSelector = null
> cacheLevel = DefaultMessageListenerContainer.CACHE_SESSION
> durableSubscriptionName = "onedirectory.index-" +
> UUID.randomUUID()
> }
>
>
>
>
> On Wed, Jun 5, 2013 at 6:05 PM, Christian Posta <christian.po...@gmail.com
> > wrote:
>
>> yes, that's used in many places. where in the broker are you trying to
>> tune
>> it? the store?
>>
>>
>> On Wed, Jun 5, 2013 at 9:00 AM, Jean Silga <jean.si...@net-vitesse.com
>> >wrote:
>>
>> > I am talking about this LRUCache: org.apache.activemq.util.LRUCache.
>> >
>> >
>> > On Wed, Jun 5, 2013 at 5:45 PM, Christian Posta
>> > <christian.po...@gmail.com>wrote:
>> >
>> > > Which LRUCache are you talking about?
>> > >
>> > >
>> > >
>> > > On Wed, Jun 5, 2013 at 8:19 AM, Jean Silga <
>> jean.si...@net-vitesse.com
>> > > >wrote:
>> > >
>> > > > Hi guys
>> > > >
>> > > > We are currently using activemq for one of our project and we run
>> into
>> > > some
>> > > > out of memory trouble lately. Our application performs some caching
>> > every
>> > > > two hours: It reads the information to cache, encapsulates it in an
>> > > > ActiveMQObjectMessage object, sends the ActiveMQObjectMessage object
>> > to a
>> > > > topic so that the consumer can receive it and perform the caching.
>> The
>> > > > problem is that the LRUCache is keeping up to 100 messages in its
>> > cache
>> > > > (maxCacheSize=100), and this is causing an out of memory error
>> because
>> > > each
>> > > > message is quite big (10 MB of size). I tried to solve the issue by
>> > using
>> > > > the PrefetchPolicy feature, but that does not seem to have and
>> impact
>> > on
>> > > > how much messages are kept in LRUCache. Is it possible to specify
>> the
>> > > > amount of messages kept in the LRUCache? How ?
>> > > >
>> > > > Here are the versions of the plugins/frameworks I am using:
>> > > > Activemq 5.4.0
>> > > > JMS 0.5.1
>> > > > xbean-spring 3.6
>> > > > Grails 1.3.6
>> > > > Java 1.6
>> > > > I am using topics and non-persistent messaging.
>> > > >
>> > > > Thank you very much
>> > > >
>> > > > --
>> > > >
>> > > >
>> > > > *Jean Silga *
>> > > > Software Engineer at Netvitesse
>> > > > +33 (0) 9 72 22 44 10
>> > > >
>> > >
>> > >
>> > >
>> > > --
>> > > *Christian Posta*
>> > > http://www.christianposta.com/blog
>> > > twitter: @christianposta
>> > >
>> >
>> >
>> >
>> > --
>> >
>> >
>> > *Jean Silga *
>> > Software Engineer at Netvitesse
>> > +33 (0) 9 72 22 44 10
>> >
>>
>>
>>
>> --
>> *Christian Posta*
>> http://www.christianposta.com/blog
>> twitter: @christianposta
>>
>
>
>
> --
>
>
> *Jean Silga *
> Software Engineer at Netvitesse
> +33 (0) 9 72 22 44 10
>
--
*Jean Silga *
Software Engineer at Netvitesse
+33 (0) 9 72 22 44 10
