Hi Bart,

With the diagnosis pack, the problem is much clearer.


   - Seems you are using a CDH quickstart sandbox which is very limited in
   resource. In that case, it is expected that running 2 or 3 table stats at
   the same time could hang due to resource competition (thus deadlock) in
   YARN. The same thing happens in my HDP sandbox btw.
   - Also the cube Ksiegarski_cube contains 30 dimensions, which is
   challenging even for some Kylin expert. Suggest start with a simpler cube
   with 20 dimensions or less. And use hierarchy / joint / mandatory more
   heavily to control the overall cuboid number.


Cheers

Yang



On Fri, Apr 28, 2017 at 2:32 AM, Bart <[email protected]> wrote:

> I can see that the Kylin has problems with counting cardinality with more
> than 3-4 dimensions. I just took the data source to my project, counting
> cardinality and... 3 out of 8 are staying at 5%. If I count one by one - it
> works...
>
>
>
> --
> View this message in context: http://apache-kylin.74782.x6.
> nabble.com/Creating-Snowflake-schema-Link-2-Cubes-tp7602p7803.html
> Sent from the Apache Kylin mailing list archive at Nabble.com.
>

Reply via email to