Hi,

Please properly subscribe to the mailing list so that the community can
receive email notifications for your messages. To subscribe, send empty
email to [email protected] and follow simple instructions in
the reply.

Yes, for now, query will be send to all Item cache data nodes. There is a
ticket, that should resolve this problem IGNITE-4509 [1] and related
IGNITE-4510 [2] that should solve IN operator issue.


[1] https://issues.apache.org/jira/browse/IGNITE-4509
[2] https://issues.apache.org/jira/browse/IGNITE-4510

You wrote

> Hi,
> Let's say we have an ignite cache 'Item' where we have id, name, desc, and
> some other fields, and we have indexes on id and name.
> The query that we intend to hit is
> "select * from Item where _KEY in (?,?,?,?,?,?,?,?,?,?) and name = ?" and
> i always need to pass 10 values of key.
> As has been mentioned in other threads, this will not use the indexes, so
> we will be using query like
> "select I.* from Item I, table(_KEY integer = ?)N where T._KEY = I._KEY
> and name=?"
> The above works very well, but the question is that does this query
> execute on all the nodes of the cluster.
> Let's say i have 20 nodes in the cluster, and ignite executes this on all
> the nodes of the cluster, and returns back the data.
> Now, lets say that this query takes 1 ms when executing on every node, so
> that each node will be able to serve upto 1000 queries per second. But
> considering each query executes on every node, the cluster will be able to
> serve upto 1000 qps. Even though i double my cluster size to 40 nodes when
> my dataset increases, the number of queries per second will still be 1000.
> Here, i am assuming that each query runs on every node.Is it true? If yes,
> then doesn't it prevent the system from horizontal scaling while
> querying(and not while storing)?
> Thanks
> Tuco


-- 
Best regards,
Andrey V. Mashenkov

Reply via email to