We are planning to install and configure flume in our CDH 5.9 cluster.Wanted to know whether we can do single node installation or mulitple node installation is also possible.
And what will be advantage of multiple node installation over single node installation?
I am new to flume .
Please help me with this.
It depends on your application usecase. However its good to have multi-node flume agents for resiliency purposes.
Multinode/resiliency/loadbalancer, if you have more data coming in from the source, could scale up number of flume agents
Flume agent 1 - Node 1
Flume agent 2 - Node 2
Load balancer/virtual IP(VIP) connecting Node 1 and Node 2
Happy Fluming !