I'm on a Corda V2 project (but will be migrating to V3 shortly). Per the docs
Artemis is hidden behind a thin interface...
But I'm exploring some business cases around the queue. Specifically is any of the following exposed for configuration? (Couldn't find anything specific in node config, but about to look at source.) Or would I need to do my own broker and specify with messagingServerAddress?
delivery retry rate
backoff rate
timeout, or when Artemis gives up on delivering the message
Sorry, might be separate question but can the internal queue be queried to see if a node has proposed tx's still waiting to be sent to a different node?
As of Corda 3, these settings are not configurable.
It is recommended not to try and interfere with these settings, as many Corda components do not have timeouts configured.
If your use case absolutely requires configuring these settings, please update to original question to explain why :)
Related
I feel like I am missing something very fundamental here.
I can bring up a RabbitMQ cluster with three nodes (rabbit1, rabbit2 and rabbit3) without an issue. Then when I start writing my microservices it seems like each client connects to only one rabbit instance. So let's say I have all my services connect to rabbit1.
If rabbit1 then goes down will my entire infrastructure blow up? Do the services have a way of switching to another rabbit node? It seems like they cannot, in which case, what is the point of having a cluster?
In case someone else runs into this and has trouble (like myself) finding this in the documentation, RabbitMQ does not manage client connection auto-recovery. From the docs:
Some client libraries provide a mechanism for automatic recovery from
network connection failures... Other clients may consider network
failure recovery to be a responsibility of the application.
So first check if you library offers auto-recovery, if not you'll have to implement it yourself.
There are multiple articles suggesting that load-balancer should be used in front of RabbitMQ cluster.
However, there are also multiple references that Spring AMQP is using some
failover implementation like connection reset when broker comes back to life.
I have several questions regarding this topic (given that those articles are more or less old and it's 2018 today)
When using Spring AMQP, is it load-balancing for still required?
If load-balancing is still suggested, how would I solve affinity of primary queue to its node? There would be much inter-connect between cluster nodes, because round-robin load-balancer would have 1-(1/n) success rate of hitting correct cluster node
Does Spring AMQP support some kind of topology awareness, which would allow it to consume from correct node?
There were some articles suggesting that clients should publish/consume to nodes respecting locality of queues. Does this still apply? How does this all fits together given load-balancing, Spring AMQP failover and CachingConnectionFactory?
Can anybody please provide answers to those topics and also provide relevant references, which would provide additional information for verification?
Thanks a lot
For each of your bullets:
a load balancer makes little sense with default configuration of Spring AMQP since it opens a single, long-lived, connection that is shared across all consumers. In, 2.0, you can configure the RabbitTemplate to use a separate connections; this is because it is a recommended configuration to use a different connection for publishers/consumers; this will be default in 2.1.
It might make sense to use a load balancer if you configure the connection factory to cache connections (instead of just channels) since, then, each component gets its own connection.
See next bullet.
See Queue Affinity and the LocalizedQueueConnectionFactory. It uses the management plugin to determine which node currently hosts the queue and connects to that. It will not work with a load balancer since it needs to connect to the actual node.
It is my understanding from several discussions that queue affinity is only needed in the most extreme environments and that, in most environments, the difference is immeasurable. However, environments/networks differ so much, YMMV so you may want to test. My general rule of thumb is to avoid premature optimization since the added complexity of the configuration may simply not be worth the benefit (and you may not have a problem in the first place).
I would like to know are there feature wise same or different? Could you also mention any pros and cons about both of these? Also please mention real-world use case for both Embedded BrokerService vs installed ActiveMQ broker. Thanks in advance!
ActiveMQ is just a Java application, and the embedded version offers essentially the same features as the stand-alone version. In fact, you can configure an embedded broker to take its configuration from an XML file, in which case it will look very similar to the stand-alone broker.
Embedding a broker is a reasonable thing to do if you need the benefit of programmatic configuration; that is, you want to configure things according to rules which are hard to implement in an XML file. It also makes sense if you want close-coupled operation between the broker and the application components, with message data being passed in memory. This might be the situation if you're using JMS as an inter-module communication mechanism within the application.
Embedding a broker has the disadvantage -- and it can be a profound one -- of making it difficult to disentangle problems in the broker from problems in your application. Figuring out the cause of, say, runaway memory consumption could be very difficult. You can get commercial support for ActiveMQ, should you need it, but it will be hard for any commercial organization to support a hybrid broker+application installation.
I'm looking to use a message queue system for an ongoing project, which now is relying on a custom (and brittle) message subsystem to interconnect multiple applications. Both the pub/sub and queue patterns are heavily used in my system.
Apache Apollo is one of the message queue systems I'm taking into account, but I don't find information about how can I handle (for instance) an Apollo server failure.
Is there a way to provide failover support in Apollo?
No, as of now this has not been resolved. Apollo is a very good broker, indeed, but lacks some production critical features like fail over. Apollo was an attempt to make a core for the next generation of ActiveMQ. However, the development is no loger active.
Have you considered other brokers like Apache Artemis? It's basically a new attempt to remake ActiveMQ with code from HornetQ, ActiveMQ and Apollo. Development is very active at the moment and there is support for fail over etc.
I have tried and tested the JMX API and it is pretty simple to use and provides a vast number of statistics required for monitoring ActiveMQ.
But the problem is, i dont want to monitor my ActiveMQ remotely and also i dont want to use another API.To be more precise, i want to use the JMS API itself to get statistics related to various destinations and the broker itself.
Advisory messages seem to be an alternative but they provide limited Amount of Administrative Messages to monitor.
Any input is highly appreciated...
There is no built-in support for this. But you can implement a JMS topic which publishes the monitoring data every few seconds. Make the connection non-persistent so that it doesn't pile up when there are no listeners or when they loose connection.
Now you can write a client that connects to this topic and it will receive updates.
AMQ-2379 resulted in a broker plugin for grabbing statistics from destinations by sending a simple JMS message. Check out the docs that show how to use it here:
http://activemq.apache.org/statisticsplugin.html
The statistics plugin is available in the 5.3 release.
You can checkout this http://issues.apache.org/activemq/browse/AMQ-2379, it will be avaiable in upcoming 5.3.0 release
There's a blog post queued up to go on http://issues.apache.org/activemq/browse/AMQ-2379 - will post it in a couple of days or so