CODEWITHSUNDEEP
activemq-classic
matt helliwell
matt helliwell

Reputation: 2678

What does failover transport do with a single URI

In my ActiveMQ client when I specify which broker to which to connect, what's the difference between specifying

tcp://host:port

and

failover:(tcp://host:port)

I've tried stopping and starting the broker and in both cases the client finds the broker again when it restarts. I haven't experimented with messages in flow (I'm using queues) or other network failures yet to see if there is a difference.

Hopefully someone has a definitive answer as the documentation I've found is a bit vague.

Thanks

Upvotes: 5

Views: 5556

Answers (3)

Ritesh Kumar
Ritesh Kumar

Reputation: 89

Failover is very useful to deal with the problem of broker or system failure and provides durability.

failover:(tcp://host1:port,tcp://host2:port,tcp://host3:port)?jms.useAsyncSend=true&randomize=false

Here if you make randomize = true, then it will use all 3 brokers simultaneously and randomize = false means use one broker at a time and when the running broker goes down, pick any broker from the sleeping brokers. And useAsyncSend = true will run the messaging system in asynchronous manner which is fast as compared to synchronous.

Upvotes: 1

anubhava
anubhava

Reputation: 785128

failover protocol also lets you provide some useful connection parameters. For ex:

failover:(tcp://host:port)?jms.useAsyncSend=true&timeout=5000

Here timeout=5000 will make sure to bail out initial connection attempt after 5 sec if broker isn't available or not responding.

Upvotes: 2

Hiram Chirino
Hiram Chirino

Reputation: 4091

If you use failover, then the client library go into a reconnect loop when you loose connection to the server. The network failure will be hidden from your app. If you don't use failover, then when you loose your connection then your app will get a JMS error telling you you lost the connection.

Upvotes: 5

Related Questions