| title | description | keywords | services | documentationcenter | author | manager | editor | ms.service | ms.workload | ms.tgt_pltfrm | ms.devlang | ms.topic | ms.date | ms.author |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Why are my LLAP queries running slow? | Microsoft Docs |
Use the Hive FAQ for answers to common questions on Hive on Azure HDInsight platform. |
Azure HDInsight, ambari, Tez, FAQ, troubleshooting guide, |
Azure HDInsight |
na |
multiple |
na |
na |
na |
article |
01/17/2018 |
msft-tacox |
The default cluster configurations are not sufficiently tuned for your workload
Queries in Hive LLAP are executing slower than expected. This can happen due to a variety of reasons.
LLAP is optimized for queries that involve joins and aggregates. Queries like the following don't perform well in an Interactive Hive cluster:
select * from table where column = "columnvalue"
To improve point query performance in Hive LLAP, set the following configurations:
hive.llap.io.enabled=false; (disable LLAP IO)
hive.optimize.index.filter=false; (disable ORC row index)
hive.exec.orc.split.strategy=BI; (to avoid recombining splits)
You can also increase usage the LLAP cache to improve performance with the following configuration change:
hive.fetch.task.conversion=none