Re: Distributed join does not work

[Evgenii Zhuravlev <e....@gmail.com>]

Hey Vladyslav,

Most possible that your data wasn't collocated.

By default Ignite use joins only for collocated data and it's a preferable
option. To read more about data collocation please check this doc:
https://apacheignite.readme.io/docs/affinity-collocation

Distributed joins without collocation will always create overhead due to
the fact that there will be much more network round-trips and data movement
between the nodes to fulfill a query.

To read more about joins please refer to
https://apacheignite-sql.readme.io/docs/distributed-joins

Evgenii

2018-03-31 22:28 GMT+03:00 Melnychenko Vladyslav <vl...@outlook.com>:

> I created a simple project by Ignite web console.
> There are two caches were configured in the project.
> The caches are populated from Oracle database.
>
> I used a simple model like EMP, DEPT but there were more records ~ 1000000
> per table
>
> EMP
> ID NAME DEPNO
> 1 Smith 10
> 2 Alen 20
> 3 Ward 30
>
> DEPT
> ID NAME
> 10 ACCOUNTING
> 20 SALES
>
>
> I run Ignite server on the first environment and connected to the instance
> over JDBC thin driver
>
> The pretty simple query retrieved the result.
>
> SELECT e.id, e.name FROM emp e, dept d WHERE e.id=1 AND e.depno = d.id
> ------------------
> 1 Smith
> ------------------
>
> I run another Ignite server on the second environment. It said that I had
> a small cluster.
>
> Topology snapshot [ver=2, servers=2, clients=0, CPUs=8, offheap=11.0GB,
> heap=12.0GB]
>
> I run the query again and got the same result. Everything was OK.
>
> After a while, I repeated the query and got nothing.
> The separate queries from a single table retrieve the result but join does
> not work anymore.
>
> I guess it is not typical behavior.
> Is there was a misconfiguration or anything else?
>
>