deepakmca05 wrote > I have couple of questions around ignite architecture for production set > up. > As the documentations says the cluster can be partitioned into cluster > groups of Data Grid/Compute grid etc. > So my question here is do we need to define the data grid and compute grid > on the same nodes of the cluster , or is it like separate nodes for data > grid and separate nodes for compute grids and then group them in cluster > groups? > If it has to be separate , then how is the data collocality achieved on > the compute nodes? > Please let me know if tere is any architecture document around production > deployment of Ignite.
Hi, All nodes in Ignite are equal and there is no separation between data nodes and compute nodes. By default, all server nodes participate in both storing the data and computations. But if required, you can create cluster groups and use them for different purposes. Obviously, to use collocation you will need to execute jobs on the nodes where corresponding data resides. BTW, this is done automatically by Ignite if you use IgniteCompute.affinityRun of IgniteCompute.affinityCall methods. -Val -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Ignite-architecture-for-production-set-up-tp1154p1169.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
