You can certainly use the Pulsar standalone to feed streaming topologies. The only constraints of the standalone are more around: * Throughput (only 1 broker) * Availability (if that machine/process goes down, the "service" is down) * Durability (data is stored in a single node, lose one disk, lose the data)
Deploying a multi-node clusters with brokers and bookies is a bit more involved, and probably not worth it to just do simple tests. You can find docs for deploying on bare metal or Kubernetes in https://pulsar.incubator.apache.org/docs/latest/deployment/InstanceSetup/ On Fri, Aug 18, 2017 at 9:38 AM J.R. Pauley <[email protected]> wrote: > is standalone suitable for use w/ Heron and PulsarSpout adaptor or does > pulsar need to be started in broker mode? If the latter how do you specify > the config file to use? > > -- Matteo Merli <[email protected]>
