Re: ActiveMQ becomes slow for consumer

Also, if the consumer is embedded with the broker make sure your using the vm transport e.g. vm://localhost vs tcp://localhost:61616. You may also need to configure some concurrent consumers on your camel route. Regards /Dave On Sun, Jan 3, 2010 at 5:34 PM, Rob Davies wrote: > I'd try 5.3 - ther

Re: How to limit queue by number of messages

I think an EvictionStrategy might work. You can evict least recent or most recent messages once your queue gets to a certain size. http://activemq.apache.org/maven/activemq-core/apidocs/org/apache/activemq/broker/region/policy/MessageEvictionStrategy.html Cheer's, /Dave On Fri, Oct 23, 2009 at 6

Re: Possible causes of end-to-end packet loss in Activemq, and TCP protocol

You should dump the traffic using tcpdump or equivalent. Then analyze it with wireshark. That should help you pinpoint the problem further. Regardws /Dave On Thu, Oct 15, 2009 at 3:55 PM, praveen.sharma wrote: > > What are the possible causes of packet loss in activemq even with TCP > protocol.

Re: About Socket close questions,

The TIME_WAIT state is a normal part of connection teardown when the connection is closed. If you are seeing lots of connections in TIMED_WAIT state then you can combat that by caching the connection on the client side (i.e. opening a connection for every message you send would result in lots of so

Re: How to set

For a client failover connection the network: uri is unnecessary (i.e. network connectors are for broker to broker connections). I think something like the uri below should work. failover://(tcp://master1:0,tcp://slave1:0,tcp://master2:0,tcp://slave2:0))?randomize=false /Dave On Fri, May 8, 2009

Re: Connect to IBM MQ

If java is an option, you can use the IBM MQQueueConnectionFactory, something like.. Regards /Dave On Thu, Apr 23, 2009 at 2:20 PM, xaos wrote: > > Well, I figured as much. Is there any Open Source IBM MQ client out th

Re: Real time market data publisher

Hi Manu, Try message selectors. This should allow you to reduce the number of topics significantly. http://activemq.apache.org/selectors.html Regards /Dave On Tue, Apr 21, 2009 at 10:13 AM, manua wrote: > > Hi All, > > I am writing a real time market data publisher and publisher will publish >

Re: Network of brokers and listener with selectors

Hi Roland, How much traffic does the servlet client get - I'm wondering if perhaps your "dynamic" listener could be timing out. It might be worth trying to change the broker url you use to connect the dynamic client to the broker to something like: tcp://localhost:61616?wireFormat.maxInactivityDur

Re: Changing the reference store for the AMQ Message Store

> > > > > > > discoveryUri="multicast://default"/> > > > > > >http://mortbay.com/schemas/jetty/1.0";> > > >

Re: Changing the reference store for the AMQ Message Store

ser wrote: > > Bad news. After running fine as a standalone broker for two days, it died > again today with the same error. I've tried the syncOnWrite=true setting, > but it slows down the broker too much for what is acceptable in my system. > I think I'm pretty much out o

Re: Changing the reference store for the AMQ Message Store

internals of how ActiveMQ runs? It makes it quite > impractical if I can't use the service wrapper once I deploy my app. > > Cheers, > Stefan > > > Dave Stanley wrote: > > > > It seems as though your message store has been corrupted. You will > > probably > &g

Re: Changing the reference store for the AMQ Message Store

It seems as though your message store has been corrupted. You will probably need to wipe the /data directory to start from a clean In order to try and narrow down why this might be occurring: 1) Try test without the java service wrapper (standalone broker) and see if it makes any difference 2) Tr

Re: NFS locking issue for Master/Slave

Hi Jason, Thanks for posting the results and configuration to this one. I just noticed the reply :-) Glad you were able to verify nfsv4 does the job. /Dave On Mon, Mar 23, 2009 at 2:42 PM, JasonCzerak wrote: > > I have found a solution!!! > > NFSv4. In it's default configuration on NetApp (Da

Re: problem with activemq on freebsd 7

It seems like the embedded broker doesn't have the correct permissions to create its message store (Usually it will be created in the working directory for the process). Regards /Dave On Mon, Mar 9, 2009 at 4:49 AM, vlad2005 wrote: > > I try to deploy an application, weborb, to an fresh install

Re: NFS locking issue for Master/Slave

Hi Jason, Have you tried using nfsv4. It may be that you can use that (and tweak the lease period on the netapp if necessary) to give you the lock expiration behaviour your looking for. Hope this helps /Dave On Mon, Mar 9, 2009 at 2:16 PM, JasonCzerak wrote: > > > > JasonCzerak wrote: > > > > I

Re: Messages rollback

/Dave On Thu, Feb 12, 2009 at 12:07 PM, jzhang1 wrote: > > Thanks very much for the information. For the C/C++ sender and receiver, > does stomp provide the similar functionality as JMS session ? > > > > > Dave Stanley wrote: > > > > You can created a transacte

Re: Messages rollback

You can created a transacted jms session and either commit() or rollback() that session. See: http://svn.apache.org/viewvc/activemq/trunk/activemq-core/src/test/java/org/apache/activemq/JmsTransactionTestSupport.java?view=markup Regards /Dave On Wed, Feb 11, 2009 at 8:45 PM, jzhang1 wrote: > >

Re: Persistence DDL

I think this might be it .. http://activemq.apache.org/jdbc-support.html On Tue, Feb 10, 2009 at 2:28 PM, Aaron Mulder wrote: > I seem to remember a page on the ActiveMQ site that discussed the > specific tables/columns used or needed in a database used as a JDBC > persistent store. But I can'

Re: Stack Overflow

It looks like a recursive/nested invocation is overflowing the stack. The stack does look to be repeating itself. Is the problem easily reproducible? Also what version of the broker are you using? /Dave On Tue, Nov 11, 2008 at 11:32 AM, sub3 <[EMAIL PROTECTED]> wrote: > > Hi, > > Can someone tel

Re: activemq, session management?

Hi Eugene, >> . If the consumer dies for some reason - we need to resume it's job. Would a durable subscription [1] for your consumer work. Then if the consumer dies, you would just need to recover from where you left off. If you used CLIENT_ACKNOWLEDGE you should be able to control how many mess

Re: use AMQ to transfer, about Security problem

broker, > itself and remote broker. > > > how do u think this? > > > > > On Thu, Oct 16, 2008 at 11:18 PM, Dave Stanley <[EMAIL PROTECTED]> > wrote: > > > If the authorization is different for both ends of the network connector, > I > > think you

Re: use AMQ to transfer, about Security problem

If the authorization is different for both ends of the network connector, I think you would need to remove the duplex="true" On Wed, Oct 15, 2008 at 11:46 PM, gui pei <[EMAIL PROTECTED]> wrote: > Hi, all > Here is a problem, does anyone know how to resolve it? > > I have two ActiveMQ broker, A an

Re: QueueSize and InFlightCount on a Topic keeps growing

I'm not sure if it will help, but you could try disabling producer flowControl: It would also be interestin

Re: kr-store/index-queue-data grows indefinitely?

I think one of these two fixes should help. http://issues.apache.org/activemq/browse/AMQ-1797 http://issues.apache.org/activemq/browse/AMQ-1814 Hope this helps, /Dave On Mon, Aug 4, 2008 at 8:06 PM, Dylan Vanderhoof <[EMAIL PROTECTED]>wrote:

Re: Clients failover in HA mode

This might be too simple an explanation but maybe check you have a persistence adapter configured on both your brokers and also that your sending persistent messages e.g. message.setJMSDeliveryMode(DeliveryMode.PERSISTENT); HTH /Dave On Fri, Jun 27, 2008 at 12:52 AM, dan neo <[EMAIL PROTECTED]>

Re: How to configure keystore/trusted stores with Weblogic

https://issues.apache.org/activemq/browse/AMQ-1724 may be helpful for an example of the spring approach. You should also be able to create an sslContext programatically if thats preferred. Regards /Dave On Thu, Jun 19, 2008 at 8:58 AM, joemo <[EMAIL PROTECTED]> wrote: > > We have just started to

Re: Getting BindException on client side after restart of broker

It may be taking a while for the broker.stop() to complete. If you try and restart the broker before the transport connector from the previous broker instance has been fully shutdown, that might explain the bind exception (as the socket is already in use). HTH /Dave On Tue, Jun 17, 2008 at 1:23 P

Re: how many topics can I create in a broker before getting OutOfMemoryError: Java heap space

You can change the the maximum heap size for the JVM by tweaking the -Xmx setting in /bin/activemq.bat. There will be a memory and processing overhead for each topic you create. Is it possible to reduce the number of Topics and just send more messages to each Topic? Regards, /Dave On Tue, Jun 17,

Re: having problem with durable subscription

"); > > > > TopicSubscriber consumer = > > session.createDurableSubscriber(topic,"myDurable1"); > > //MessageConsumer consumer = session.createConsumer(topic); > > consumer.setMessageListener(this); > > > >

Re: having problem with durable subscription

Your not running this from within a thread by any chance? I don't believe connection.start() is going to block, so try and create the connection outside the thread and pass it in by reference to the thread instance. HTH /Dave On Thu, Jun 12, 2008 at 10:30 PM, eaglepointe <[EMAIL PROTECTED]> wrot

Re: Lost messages - not all messages sent to the queue are delivered to the message receiver

I think this may be a problem on the consumer side given you are using the Spring DMLC with no connection caching. The connection will be recycled each time, but this can cause problems if your prefetch policy is not aligned. Try using a pooled connection factory, also try and set the DMLC cacheLe