【apache-hive-1.2.1】local mr

在hive中运行sql有很多是比较小的SQL，数据量小，计算量小。这些比较小的SQL 如果也采用分布式的方式来执行，那么就得不偿失，因为SQL真正执行的时间可能只有10s，但是分布式任务生成的其他过程执行可能要1min。这样小任务采用local mr方式执行，就是本地执行，通过把输入数据拉回到客户端来执行

三个参数来决定：

hive.exec.mode.local.auto=true 是否启动本地mr模式

hive.exec.mode.local.auto.input.files.max=4 input files的数量，默认是4个

hive.exec.mode.local.auto.inputbytes.max=134217728 input files的大小，默认是128M

注意：

hive.exec.mode.local.auto是大前提，只有设置为true，才可能会启用local mr模式

hive.exec.mode.local.auto.input.files.max 和 hive.exec.mode.local.auto.inputbytes.max是 与的关系，只有同时满足才会执行local mr

t_1==> 5个文件

t_2==> 2个文件

hive>set hive.exec.mode.local.auto=false
hive> select * from t_2 order by id;
Query ID = hadoop_20160125132157_d767beb0-f674-4962-ac3c-8fbdd2949d01
Total jobs = 1
Launching Job 1 out of 1
Number of reduce tasks determined at compile time: 1
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
set mapreduce.job.reduces=<number>
Starting Job = job_1453706740954_0006, Tracking URL = http://hftest0001.webex.com:8088/proxy/application_1453706740954_0006/ Kill Command = /home/hadoop/hadoop-2.7.1/bin/hadoop job  -kill job_1453706740954_0006
Hadoop job information for Stage-1: number of mappers: 1; number of reducers: 1
2016-01-25 13:22:19,210 Stage-1 map = 0%,  reduce = 0%
2016-01-25 13:22:26,497 Stage-1 map = 100%,  reduce = 0%, Cumulative CPU 1.47 sec
2016-01-25 13:22:40,207 Stage-1 map = 100%,  reduce = 100%, Cumulative CPU 3.68 sec
MapReduce Total cumulative CPU time: 3 seconds 680 msec
Ended Job = job_1453706740954_0006
MapReduce Jobs Launched:
Stage-Stage-1: Map: 1  Reduce: 1   Cumulative CPU: 3.68 sec   HDFS Read: 5465 HDFS Write: 32 SUCCESS
Total MapReduce CPU Time Spent: 3 seconds 680 msec
OK
...
...

hive>set hive.exec.mode.local.auto=true
hive> select * from t_2 order by id;
hive> select * from t_2 order by id;
Automatically selecting local only mode for query                ==> 启动用本地模式
Query ID = hadoop_20160125132322_9649b904-ad87-47fa-89ad-5e5f67315ac8
Total jobs = 1
Launching Job 1 out of 1
Number of reduce tasks determined at compile time: 1
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
set mapreduce.job.reduces=<number>
Job running in-process (local Hadoop)
2016-01-25 13:23:27,192 Stage-1 map = 100%,  reduce = 100%
Ended Job = job_local1850780899_0002
MapReduce Jobs Launched:
Stage-Stage-1:  HDFS Read: 1464 HDFS Write: 1618252652 SUCCESS
Total MapReduce CPU Time Spent: 0 msec
OK
...
...

hive>set hive.exec.mode.local.auto=true
hive> select * from t_1 order by id;
Query ID = hadoop_20160125132411_3ecd7ee9-8ccb-4bcc-8582-6d797c13babd
Total jobs = 1
Launching Job 1 out of 1
Number of reduce tasks determined at compile time: 1
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
set mapreduce.job.reduces=<number>
Cannot run job locally: Number of Input Files (= 5) is larger than hive.exec.mode.local.auto.input.files.max(= 4)                            ==>5 > 4 还是启用了分布式
Starting Job = job_1453706740954_0007, Tracking URL = http://hftest0001.webex.com:8088/proxy/application_1453706740954_0007/ Kill Command = /home/hadoop/hadoop-2.7.1/bin/hadoop job  -kill job_1453706740954_0007
Hadoop job information for Stage-1: number of mappers: 1; number of reducers: 1
2016-01-25 13:24:38,775 Stage-1 map = 0%,  reduce = 0%
2016-01-25 13:24:52,115 Stage-1 map = 100%,  reduce = 0%, Cumulative CPU 1.55 sec
2016-01-25 13:24:59,548 Stage-1 map = 100%,  reduce = 100%, Cumulative CPU 3.84 sec
MapReduce Total cumulative CPU time: 3 seconds 840 msec
Ended Job = job_1453706740954_0007
MapReduce Jobs Launched:
Stage-Stage-1: Map: 1  Reduce: 1   Cumulative CPU: 3.84 sec   HDFS Read: 5814 HDFS Write: 56 SUCCESS
Total MapReduce CPU Time Spent: 3 seconds 840 msec
OK
...
...

hive>set hive.exec.mode.local.auto=true
hive> set hive.exec.mode.local.auto.input.files.max=5;            ==> 设置输入文件数max=5
hive> select * from t_1 order by id;
Automatically selecting local only mode for query                 ==> 启用了本地模式
Query ID = hadoop_20160125132558_db2f4fca-f6bf-4b91-9569-c779a3b13386
Total jobs = 1
Launching Job 1 out of 1
Number of reduce tasks determined at compile time: 1
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
set mapreduce.job.reduces=<number>
Job running in-process (local Hadoop)
2016-01-25 13:26:03,232 Stage-1 map = 100%,  reduce = 100%
Ended Job = job_local264155444_0003
MapReduce Jobs Launched:
Stage-Stage-1:  HDFS Read: 1920 HDFS Write: 1887961792 SUCCESS
Total MapReduce CPU Time Spent: 0 msec
OK