ignite-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From russ <russellwi...@gmail.com>
Subject Ignite user architecture question - Node Roles
Date Thu, 24 Nov 2016 10:07:57 GMT
Hi, I'm new to Ignite and I'm looking for some basic getting started advice
for an architecture on the scenario below. 

My use case is; I want to load data from 3 SQL Server databases into an
Ignite data grid and be able to do distributed SQL queries across the 3 sets
of data from client apps. Also I want to implement write-through updates on
the data so any client cache will get updated when the database does.

Question: How should I architect this? Would you create 4 "Worker Nodes"
containing zero custom code (say you wanted the grid to have 4 nodes) and
then create 3 "Loader Nodes", one for each database to load the data into
caches? My problem with this approach is what if a ''Loader Node" fails -
when i have the write-through updates working there will be no redundancy
having a single node for a given source of data.

Or would you put the data loading code into each of the nodes so all nodes
can load the data. But I'm thinking in this case all nodes would load the
data on start-up which is unnecessary in a standard start-up scenario. 

If you could advice on a node setup architecture for this that would be

Thanks, Russ

View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Ignite-user-architecture-question-Node-Roles-tp9181.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.

View raw message