You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@atlas.apache.org by Bolke de Bruin <bd...@gmail.com> on 2019/08/10 20:55:50 UTC

Re: Performance tuning Atlas / Solr + HBase

And i should add Hbase to this question.

We run 1 hbase node at the moment.

Verstuurd vanaf mijn iPad

> Op 10 aug. 2019 om 22:01 heeft Bolke de Bruin <bd...@gmail.com> het volgende geschreven:
> 
> Hi,
> 
> We are facing some performance issues with our Atlas config deployed by Ambari. For example we see a Table search with a simple ‘like’ clause taking easily 20s. So I am wondering what people have been doing to tune Atlas and Solr.
> 
> Our current setup was configured with 24 solr nodes, 5 indexed shards. 2G max heap, 2G max direct memory. Caches kept at standard levels (Filter cache 2000, query cache 26000, document cache 26000).
> 
> Atlas is one server, default settings.
> 
> Any recommendations? I’m looking for recommendations around the amount of shards/nodes, heap memory, direct memory and caches obviously.
> 
> Cheers!
> Bolke
> 
> Verstuurd vanaf mijn iPad