Hi,
So, while creating a table how do we specify that it uses only certain set of nodes to store data?
How to tag certain nodes as hot and warm, so that we can create a table that stores data only in warm nodes?
Hi Sandy. Can you tell me what version are you using? Thanks
Sandy,
You don’t need to specify any of that as SingleStore takes care of it all for you. You just insert/update/delete your data as you would against any relational database. Behind the scenes SingleStore will distribute the data across the cluster, it will create two copies of each piece of data and put them on different nodes to insure redundancy and will mix the primary copies of data with the secondary copies so all the nodes get utilized. We use hash partitioning so the data gets randomly distributed across the cluster (assuming you pick a good shard key when you are creating your schema. This is one of the few things the user is responsible for doing. We have some good docs that teach you how to do this.). So when you run a query the system runs that query on all the nodes, against all the partitions that have the data you are looking for, which gives you the power of parallelism and then aggregates all the results and sends it back to you as one result set.
Rick