sparksql与hive整合

hive配置

编辑 $HIVE_HOME/conf/hive-site.xml,增加如下内容:


 hive.metastore.uris
 thrift://master:9083
 Thrift uri for the remote metastore. Used by metastore client to connect to remote metastore.12345

启动hive metastore

启动 metastore: $hive --service metastore &
查看 metastore: $jobs[1]+  Running                 hive --service metastore &

关闭 metastore:$kill %1kill %jobid,1代表job id1234567891011

spark配置

将 $HIVE_HOME/conf/hive-site.xml copy或者软链 到 $SPARK_HOME/conf/将 $HIVE_HOME/lib/mysql-connector-java-5.1.12.jar copy或者软链到$SPARK_HOME/lib/copy或者软链$SPARK_HOME/lib/ 是方便spark standalone模式使用123

启动spark-sql

  1. standalone模式

    ./bin/spark-sql --master spark:master:7077 --jars /home/stark_summer/spark/spark-1.4/spark-1.4.1/lib/mysql-connector-java-5.1.12.jar
  • 1

yarn-client模式

$./bin/spark-sql --master yarn-client --jars /home/stark_summer/spark/spark-1.4/spark-1.4.1/lib/mysql-connector-java-5.1.12.jar执行 sql:
select count(*) from o2o_app;结果:302Time taken: 0.828 seconds, Fetched 1 row(s)2015-09-14 18:27:43,158 INFO  [main] CliDriver (SessionState.java:printInfo(536)) - Time taken: 0.828 seconds, Fetched 1 row(s)
spark-sql> 2015-09-14 18:27:43,160 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) - Finished stage: org.apache.spark.scheduler.StageInfo@5939ed302015-09-14 18:27:43,161 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) - task runtime:(count: 1, mean: 242.000000, stdev: 0.000000, max: 242.000000, min: 242.000000)2015-09-14 18:27:43,161 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) -    0%      5%      10%     25%     50%     75%     90%     95%     100%2015-09-14 18:27:43,161 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) -    242.0 ms        242.0 ms        242.0 ms        242.0 ms        242.0 ms        242.0 ms    242.0 ms 242.0 ms        242.0 ms2015-09-14 18:27:43,162 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) - fetch wait time:(count: 1, mean: 0.000000, stdev: 0.000000, max: 0.000000, min: 0.000000)2015-09-14 18:27:43,162 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) -    0%      5%      10%     25%     50%     75%     90%     95%     100%2015-09-14 18:27:43,162 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) -    0.0 ms  0.0 ms  0.0 ms  0.0 ms  0.0 ms  0.0 ms  0.0 ms  0.0 ms  0.0 ms2015-09-14 18:27:43,163 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) - remote bytes read:(count: 1, mean: 31.000000, stdev: 0.000000, max: 31.000000, min: 31.000000)2015-09-14 18:27:43,163 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) -    0%      5%      10%     25%     50%     75%     90%     95%     100%2015-09-14 18:27:43,163 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) -    31.0 B  31.0 B  31.0 B  31.0 B  31.0 B  31.0 B  31.0 B  31.0 B  31.0 B2015-09-14 18:27:43,163 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) - task result size:(count: 1, mean: 1228.000000, stdev: 0.000000, max: 1228.000000, min: 1228.000000)2015-09-14 18:27:43,163 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) -    0%      5%      10%     25%     50%     75%     90%     95%     100%2015-09-14 18:27:43,163 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) -    1228.0 B        1228.0 B        1228.0 B        1228.0 B        1228.0 B        1228.0 B    1228.0 B 1228.0 B        1228.0 B2015-09-14 18:27:43,164 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) - executor (non-fetch) time pct: (count: 1, mean: 69.834711, stdev: 0.000000, max: 69.834711, min: 69.834711)2015-09-14 18:27:43,164 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) -    0%      5%      10%     25%     50%     75%     90%     95%     100%2015-09-14 18:27:43,164 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) -    70 %    70 %    70 %    70 %    70 %    70 %    70 %    70 %    70 %2015-09-14 18:27:43,165 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) - fetch wait time pct: (count: 1, mean: 0.000000, stdev: 0.000000, max: 0.000000, min: 0.000000)2015-09-14 18:27:43,165 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) -    0%      5%      10%     25%     50%     75%     90%     95%     100%2015-09-14 18:27:43,165 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) -     0 %     0 %     0 %     0 %     0 %     0 %     0 %     0 %     0 %2015-09-14 18:27:43,166 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) - other time pct: (count: 1, mean: 30.165289, stdev: 0.000000, max: 30.165289, min: 30.165289)2015-09-14 18:27:43,166 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) -    0%      5%      10%     25%     50%     75%     90%     95%     100%2015-09-14 18:27:43,166 INFO  [SparkListenerBus] scheduler.StatsReportListener (Logging.scala:logInfo(59)) -    30 %    30 %    30 %    30 %    30 %    30 %    30 %    30 %    30 %12345678910111213141516171819202122232425262728293031
  1. yarn-cluster模式

./bin/spark-sql --master yarn-cluster  --jars /home/dp/spark/spark-1.4/spark-1.4.1/lib/mysql-connector-java-5.1.12.jarError: Cluster deploy mode is not applicable to Spark SQL shell.
Run with --help for usage help or --verbose for debug output2015-09-14 18:28:28,291 INFO  [Thread-0] util.Utils (Logging.scala:logInfo(59)) - Shutdown hook called

Cluster deploy mode 不支持的123456

启动 spark-shell

  1. standalone模式

./bin/spark-shell --master spark:master:7077 --jars /home/stark_summer/spark/spark-1.4/spark-1.4.1/lib/mysql-connector-java-5.1.12.jar1
  1. yarn-client模式

./bin/spark-shell --master yarn-client   --jars /home/dp/spark/spark-1.4/spark-1.4.1/lib/mysql-connector-java-5.1.12.jarsqlContext.sql("from o2o_app SELECT count(appkey,name1,name2)").collect().foreach(println)1234

尊重原创,拒绝转载,http://blog.csdn.net/stark_summer/article/details/48443147

你可能感兴趣的:(spark,hive)