High Availability

[clopezhe360]

I've successfully deployed the operator to my Kubernetes cluster along with the brokers. I noticed that in the Broker chart there's this line:

  deploymentPlan:
    size: 1

From my testing, if i change that to 2, it'll deploy two brokers. Does that mean High Availability gets configured automatically by the broker and my messages/queues will not get lost when one of my brokers goes down?

[brusdev]

The High Availability is provided with messageMigration and the persistence is provided persisting messages to persistent storage with persistenceEnabled, i.e.

apiVersion: broker.amq.io/v1beta1
kind: ActiveMQArtemis
metadata:
  name: artemis-ha
spec:
  deploymentPlan:
    size: 2
    persistenceEnabled: true
    messageMigration: true

[clopezhe360]

How is redundancy configured for the broker? I'm noticing queues that I'm creating are found on one broker and not the other.

[Jeroen0494]

From what I read here, messages are only distributed on node shutdown / failure:
https://artemiscloud.io/docs/getting-started/quick-start/#draining-messages-on-scale-down
Also, the default ON_DEMAND loadbalancer only loadbalances to nodes with queues with consumers. So you would need to balance your consumers.
At least, this is what my currently limited understanding of ActiveMQ tells me...

If this is set to ON_DEMAND then Apache ActiveMQ Artemis will only forward messages to other nodes of the cluster if the address to which they are being forwarded has queues which have consumers, [...]

[clopezhe360]

Thanks for the response! That's great to know!