The younger brother configured with a ActiveMQ cluster, each node configuration is as follows:

Namely: the connection between the node and node relationship can be determined by dynamic lookup algorithm, producers or consumers are then connected nodes when dynamic address list will be updated,
Encountered a problem now, is in some cases, all consumers are going to the same node, this leads to was not on the node, the data are transmitted to the node, this caused a certain pressure to the network,
Consult everybody, how to solve this problem, ActiveMQ not random choice node address? How can appear this kind of circumstance? If ActiveMQ cannot solve this problem, what other alternative middleware can be used to solve this problem? thank you

CodePudding user response:

Should be able to specify the target server (one or more) to send a message, if can't do this, then you re selection from the underlying framework,

CodePudding user response:

Well, what good underlying framework can be recommended, tried Kafka cannot meet the requirements

CodePudding user response:

Next according to the way of the building Lord, on two MQ cluster configuration, dynamic BK1 start first, and then BK2 started, I even BK1 producer, consumer BK2, but you can't get a message broke2, could you tell me the sending end and receiving end uri has any special configuration?

CodePudding user response:

1, make sure your cluster is not the master-slave architecture? Hang up of the Lord, will switch to the automatic from
2, the background of each mq administration page can open?

CodePudding user response:

Should be able to specify the target server (one or more) to send a message, if can't do this, then you re selection from the underlying framework,