Hi, You're right, you should always start a node with Ignition.start() and get Ignite instance with Ignition.ignite() after that.
You can start several nodes in one process, and in this case you will have to give unique grid name to each node within this process. But this is used mostly for unit testing, because it allows to start the whole cluster in one JVM and debug conveniently. Otherwise the most common deployment is one node per JVM with the default (null) name. You can start one or several nodes per host depending on your use case. Any cache is available only to one cluster and can't be shared. -Val On Jul 23, 2015 7:28 AM, "hueb1" <[email protected]> wrote: > I'm new to Ignite and trying to understand the API of Ignition.ignite() and > Ignition.start(). I believe we'd always need to call Ignition.start(..) > first to initialize the data grid(s). And then call Ignition.ignite(..) > giving it the data grid name or none for default data grid. That being > said, is there an example of a configuration file that specifies multiple > data grids? Is it not good practice to have multiple data grids running on > the same set of hosts? Should we always just use the default data grid? > Also I'm assuming the distributed caches are scoped by data grids yes? You > can't have multiple data grids access the same distributed cache? > > > > -- > View this message in context: > http://apache-ignite-users.70518.x6.nabble.com/Multiple-grids-in-a-cluster-tp692.html > Sent from the Apache Ignite Users mailing list archive at Nabble.com. >
