Distributed query/timeout


#1

We are evaluating aerospike community edition for our use case which basically fire multiple queries in distributed dataset in which data is indexed.

  1. aerospike will fire query in every node or on selective nodes depending on some primary key, like in cassandra ? or it will fire query in every node (like in ignite) ?

  2. what about the time out suppose from 10 node one node is slow , can i get 90% result or teh query will just fail ?

thanks


#2

I assume you are talking about secondary index queries.

If so,

  1. Query execution is scatter gather. It executes on every node at the same time, and then streams back the results.

  2. I’ve never seen a query time out, unless the application can’t consume the data coming back - but yes I believe the behavior may be similar to what you assume. Timeout is configurable though and it will throw an exception. Assuming the data modeling and cluster/app hw is good this shouldn’t occur.