0
votes

I have two Linux boxes, each of them hosts Ignite server instance which are linked into cluster of two nodes. Both servers have some caches with data ,caches are configured as REPLICATED. Also each Linux box has about 15 Java processes that have Ignite client and connect to Ignite cluster to get some data.

Is it possible to configure Ignite clients in such way that they will first try to get data from cache which is running in Ignite server which is on the same Linux box as Ignite client? If there is no such server then go to any other active Ignite server in cluster.

I've tried AffinityFunction but it doesn't have information about Ignite client.

1
Are they thick or thin clients? (Affinity function affects how data is distributed around your cluster. For a replicated cache, where all your data is on every node, that's not going to have much effect.)Stephen Darlington
They are thick clients.Ivan

1 Answers

0
votes

Thick client will try to go where data is. If data is on collocated node then it will be queried. Of course for SQL or Scan queries it is not always possible to know, in this case query will be broadcast.