hive在E-MapReduce集群的实践（二）集群hive参数优化提高hdfs性能hive参数优化

本文介绍一些常见的集群跑hive作业参数优化，可以根据业务需要来使用。

dfs.client.read.shortcircuit=true //直读

dfs.client.read.shortcircuit.streams.cache.size=4096  //直读缓存

dfs.datanode.balance.bandwidthPerSec=30048576 //提高balance带宽，一般扩容后调整

dfs.datanode.max.transfer.threads=16384 //提高线程数

dfs.namenode.checkpoint.period=21600 //延长checkpoint时间

dfs.namenode.handler.count=100  //并发数，大集群要提高

dfs.namenode.fslock.fair=false //降低写性能，但提高读锁性能

dfs.namenode.lifeline.handler.count=1 //ha集群优化，大集群使用

hive.metastore.server.max.threads=100000

hive.compactor.worker.threads=5

hive.metastore.client.socket.timeout=1800s

hive.metastore.failure.retries=5

hive.exec.max.dynamic.partitions=5000

hive.exec.max.dynamic.partitions.pernode=2000

set hive.execution.engine=tez;

SET hive.tez.auto.reducer.parallelism=true;

SET hive.tez.max.partition.factor=20;

STORED AS ORC tblproperties (“orc.compress" = “SNAPPY”)

hive.exec.orc.default.compress=SNAPPY

hive.exec.parallel=true

SET hive.exec.reducers.bytes.per.reducer=128000000;

set hive.vectorized.execution.enabled = true;

set hive.vectorized.execution.reduce.enabled = true;

hive.limit.optimize.enable=true

set hive.cbo.enable=true;

set hive.compute.query.using.stats=true;

set hive.stats.fetch.column.stats=true;

set hive.stats.fetch.partition.stats=true;

查询前先统计常用表的静态信息，常join的列

analyze table tweets compute statistics;

analyze table tweets compute statistics for columns sender, topic;

set hive.enforce.bucketing = <b>true</b>; 

set hive.optimize.bucketmapjoin = true;

set hive.optimize.bucketmapjoin.sortedmerge = true;

set hive.input.format = org.apache.hadoop.hive.ql.io.BucketizedHiveInputFormat;