I haven't tried it, but you can take a look at
the ConditionalNetworkBridgeFilterFactory which can throttle messages to
subscriptions.
See bottom of this wiki:
http://activemq.apache.org/networks-of-brokers.html
On Wed, Jan 2, 2013 at 3:19 PM, billy wrote:
> I have two ActiveMQ-5.5.1 broker's c
So persistent=true is a broker setting. The setting in question for your
scenario is that for the client (producer) of the message.
See here:
http://activemq.apache.org/what-is-the-difference-between-persistent-and-non-persistent-delivery.html
On Wed, Jan 2, 2013 at 2:02 PM, Mohit Anchlia wrote
Thanks for the info. I'll look at those parameter.
It does look like that I am using persistence based on below
"persistent=true" parameter. So can I safely assume that only one producer
gets blocked?
http://activemq.apache.org/schema/core";
brokerName="static-broker1" persistent="true"
dataDirec
Yah, it's probably best to figure out how things are being sent right now.
If messages are sent async and without a producer window then you'll hit
the type of flow control that you did.
Messages are sent async by default if using non-persistent messages or
during a transaction. The only time any
Thanks! Sounds like there might be another way we can configure the flow
control that blocks only one producer? Do you have any recommendation on
what changes I should make to the config?
On Wed, Jan 2, 2013 at 12:33 PM, Christian Posta
wrote:
> Enqueue count = number of messages sent to the queu
Enqueue count = number of messages sent to the queue
Dequeue count = number of messages ack'd by the consumer
Inflight count = number of messages sent to consumer's session (and are not
ack'd by consumer)
Dispatch count = messages sent to the consumer (dequeue + inflight)
I asked about the async s
Like I said, One of Two RA is fail, Nothing in relation works, Queue Topic
etc...
In Glassfish, When I go in Administration Web Console and I try to ping
provider from poolFactory. its fail.
But, if I restart the ra, everything works fine. until next fail.
thx for reply
> What sort of stability
Thanks again! Is the dispatch queue size the one that relates to what's
held in memory?
We are using mule and the endpoint has no transactions configured. What did
you mean by forcing async?
On Wed, Jan 2, 2013 at 11:06 AM, Christian Posta
wrote:
> Correct. The memoryUsage is increased when mess
Correct. The memoryUsage is increased when messages are held in memory.
This happens when messages are paged in from the store. They are then
dispatched to consumers and held until the consumer acks.
In your case looks like a particular destination (eventsEndpoint) is
filling up somehow. Can you l
So if I understand correctly I shouldn't be seeing this behaviour unless
something is being held in memory. We ack immediately so I am kind of
wondering what might be going on here. Is there a way to find out what
might be happening here?
On Wed, Jan 2, 2013 at 10:03 AM, Christian Posta
wrote:
>
Anyway. Tested with 5.7.0 and 5.5.1:
if define vmcursor and lower memoryUsage (limit 64mb)
then memory fills up adding more and more messages and then producer is
stopped.
is here any setting to have not big memory usage and small bach size to
fetch messages from persistence DB (I use mysql) and p
What sort of stability issues are you seeing?
On Tue, Jan 1, 2013 at 11:48 AM, hhf wrote:
> Hi,
> I use Activemq with glassfish 3.01, I Deployed ra inside, and configure it.
> everything works.
> But, If I want use 2 differents activemq (not for fail over, for separate
> provider), I have to de
What info can you provide about how clients are consuming from the queues?
Any chance there are a bunch of messages "inflight" which haven't been
ack'd by the consumer?
On Wed, Jan 2, 2013 at 10:20 AM, Mohit Anchlia wrote:
> I see these messages being logged. Queues are setup to use persistent
>
Hi,
I use Activemq with glassfish 3.01, I Deployed ra inside, and configure it.
everything works.
But, If I want use 2 differents activemq (not for fail over, for separate
provider), I have to deploy a second ra inside.
But with two ra (or more), everything is very unstable.
often one ra is in err
Yah, Apollo has a slick REST API that can be used for that kind management.
See here:
http://activemq.apache.org/apollo/documentation/management-api.html
On Wed, Jan 2, 2013 at 8:31 AM, Fulko Hew wrote:
> On Wed, Jan 2, 2013 at 10:16 AM, Christian Posta
> wrote:
>
> > Hey All,
> >
> > I've add
that does look wrong. I wonder if you could wrangle up a testcase?
There are some existing unit tests that have most of the angles covered,
for example
org.apache.activemq.ActiveMQXAConnectionFactoryTest#testVanilaTransactionalProduceReceive
I tried a quick rollback variant but did not reproduce
On Wed, Jan 2, 2013 at 10:16 AM, Christian Posta
wrote:
> Hey All,
>
> I've added Part II to my introduction to Apache Apollo:
>
> http://www.christianposta.com/blog/?p=278
>
> Check Apollo here: http://activemq.apache.org/apollo/
>
Interesting!
Although I'm not an ActiveMQ user (directly) let a
Thank you very much,
I have created queue and created a junk destination to create the queue
physically (I would discard the junk message at consumer side), this way it
will create queue on the server.
Then my third party producer and consumer can refer the queue created on the
server.
Now the onl
You shoud not call the receive method of ActiveMQInputStream or it will
cause the data you recevied is one part of the whole.The lost part is the
payload of the ActiveMQMessage the call of the receive method returns.
Using the getJMSProperties method is the right way.
I found that every message c
A network connector looks like a normal consumer to the broker but the
consumerInfo (associated with the subscription) has a boolean
attribute networkSubscription=true.
>From an onSend, you can find the destinationName, get the corresponding
broker destination (from the regionBroker) and query its
Yes, I think we understand what you're trying to do, but it's not clear why
you need step 1. The queues don't need to be created ahead of time (like,
for example in IBM MQ series). They will be created automatically when a
producer tries to send to it or a consumer tries to receive from it.
JMS do
I think I am not clear in providing information here. Let me try one more
time.
So according to requirement, I need to create queue at run time in java code
(because I would come to know the number of queues to create at run time).
this would be independent class.
There are three part (you can say
Which version of ActiveMQ are you using? I just plugged your test case into
the trunk code and it worked fine.
On Tue, Jan 1, 2013 at 8:48 AM, gfshaikh wrote:
> Hi All
>
> I am trying to transfer a large file over ActiveMQ using JMS Streams - the
> transfer works fine however along with the fi
the reader thread (broker side) will be the one with:
org.apache.activemq.broker.TransportConnection#service
in its stack trace. That is the method that deals with each received
command.
In this case, it should be dealing with a ConnectionInfo command.
On 2 January 2013 05:11, arjun wrote:
> Th
Hi,
I agree with you however the only issue is I would be getting the number of
queue dynamically from some other class and hence according to requirement I
need to create queues dynamically in the program.
I have code something like this:
public static void main(String[] args) throws Exce
You could configure the broker to create those queues on startup. That way
you don't need any separate for-loop and the queues will already be visible
via the admin console. See:
http://activemq.apache.org/configure-startup-destinations.html
On Wed, Jan 2, 2013 at 3:02 AM, jainmanglesh wrote:
>
Hi
Thank you very much for your reply.
my requirement is like:
I want to create a list of activemq (which I am trying to create using for
loop as in code). These queues will be used by third party to send message
and on the other end used by third party to consume the messages on queue.
Third pa
27 matches
Mail list logo
