Our backend application is pumping a large number of messages through an
ActiveMQ
queue (upwards of 6,000 typically) to our frontend application (Flex based)
through Flex
Data Services.
We are using Jconsole to track the number of messages that go in to the queue,
and the
number of messages that get pulled off of the queue. In our testing, 100% of
the time Flex
Data Services consumes 100% of the messages off of the queue. Roughly 30% of
the time
however, the Flex application's Consumer receives only a small portion of the
total
messages from FDS (the number varies, but it is usually around 5-10% of the
total
messages).
We have looked through the FDS documentation extensively, and added all
appropriate
settings to both our messaging-config.xml and our services-config, here are
some
excerpts of those settings:
Our queue as defined in messaging-config.xml:
<destination id="MQ00001">
<adapter ref="jms" />
<properties>
<network>
<session-timeout>0</session-timeout>
</network>
<jms>
<destination-type>Queue</destination-type>
<message-type>javax.jms.TextMessage</message-type>
<connection-factory>ConnectionFactory</connection-factory>
<destination-jndi-name>dynamicQueues/MQ00001</destination-
jndi-name>
<delivery-mode>PERSISTENT</delivery-mode>
<message-priority>DEFAULT_PRIORITY</message-priority>
<acknowledge-mode>AUTO_ACKNOWLEDGE</acknowledge-mode>
<transacted-sessions>false</transacted-sessions>
<initial-context-environment>
<property>
<name>Context.INITIAL_CONTEXT_FACTORY</name>
<value>org.apache.activemq.jndi.ActiveMQInitialContextFactory</value>
</property>
<property>
<name>Context.PROVIDER_URL</name>
<value>tcp://192.168.2.187:61616</value>
</property>
</initial-context-environment>
</jms>
<server>
<max-cache-size>100000</max-cache-size>
<message-time-to-live>0</message-time-to-live>
<durable>true</durable>
</server>
</properties>
<channels>
<channel ref="my-rtmp" />
</channels>
</destination>
As you can see, we are setting generous values for our max-cache-size (number
of
messages to cache), and making sure that our messages live forever on the
queue, and we
are also ensuring delivery by setting their durable setting to true. We also
ensure that our
consumers never automatically disconnect from the queue (session-timeout = 0).
Our RTMP channel definition in services-config.xml:
<channel-definition id="my-rtmp"
class="mx.messaging.channels.RTMPChannel">
<endpoint uri="rtmp://{server.name}:443"
class="flex.messaging.endpoints.RTMPEndpoint"/>
<properties>
<idle-timeout-minutes>720</idle-timeout-minutes>
<client-to-server-maxbps>10000K</client-to-server-maxbps>
<server-to-client-maxbps>10000K</server-to-client-maxbps>
<accept-backlog>100000</accept-backlog>
</properties>
</channel-definition>
Again, we are setting generous values for all of our values here to ensure
delivery of
messages.
We had initially thought that perhaps this was based on the UDP protocol and
packets
were getting dropped (for whatever reason). After some research, we are led to
believe
that RTMP (the protocol of which these messages are transferred over) is based
on TCP,
which would mean that there shouldn't be any dropping of packets. We have
pretty much
ruled this out as a possible problem.
Does anybody have any experience with these types of issues? Or any thoughts on
what
else we might look at for settings, optimization, etc?
Help for any Adobe guys or general gurus in Flex Data Services would be much
appreciated.
Thanks for your time,
Brian
