Hi,

I'd say that it's normal to have bigger latency in the case when you
starting more queries simultaneously than you have threads in query pool(
https://apacheignite.readme.io/docs/thread-pools#section-queries-pool) and
cores) Some of the queries will wait till other will finish.

As for the thing that only one node was busy - you have replicated caches,
so, there is no need to search on a different node, it's enough just to
collect all data on the one. Instead of connecting to the one node only you
can connect drivers to the different nodes, not only to one.

By the way, how do you start 100 requests simultaneously? Do you run them
from one machine? You can have slowdown even on this step, since you
probably have fewer cores than the threads count.

Evgenii

чт, 6 дек. 2018 г. в 11:31, Skollur <[email protected]>:

> Currently I have only one node and other node is idle and brought it down.
> I have noticed that only one node is busy and I have code as below.
>
> IgniteConfiguration cfg = new IgniteConfiguration();
>        TcpDiscoverySpi discovery = new TcpDiscoverySpi();
>         discovery.setLocalAddress("192.0.1.2");
>         discovery.setLocalPort(10800);
>         cfg.setDiscoverySpi(discovery);
>         cfg.setPeerClassLoadingEnabled(true);
>
>
>
> --
> Sent from: http://apache-ignite-users.70518.x6.nabble.com/
>

Reply via email to