FederatedHDFS+beeline+hiveserver2遇到的坑

遇到的坑:

创新互联建站长期为上1000家客户提供的网站建设服务,团队从业经验10年,关注不同地域、不同群体,并针对不同对象提供差异化的产品和服务;打造开放共赢平台,与合作伙伴共同营造健康的互联网生态环境。为利通企业提供专业的成都网站制作、成都网站设计、外贸营销网站建设利通网站改版等技术服务。拥有十余年丰富建站经验和众多成功案例,为您定制开发。

1、 Hive的任务会从临时目录移动数据到数据仓库目录,默认hive使用/tmp作为临时目录,用户通常使用/user/hive/warehouse/作为数据仓库目录。在Federated HDFS情况下,/tmp 和 /user视为两个不同的ViewFS mount table,所以hive任务在这两个目录之间移动数据。Federated HDFS不支持这样做,所以任务会失败。

报错信息:

ERROR : Failed with exception Unable to move sourceviewfs://cluster9/tmp/.hive-staging_hive_2015-07-29_12-34-11_306_6082682065011532871-5/-ext-10002to destinationviewfs://cluster9/user/hive/warehouse/tandem.db/cust_loss_alarm_unit

org.apache.hadoop.hive.ql.metadata.HiveException: Unable to movesourceviewfs://cluster9/tmp/warehouse/.hive-staging_hive_2015-07-29_12-34-11_306_6082682065011532871-5/-ext-10002to destinationviewfs://cluster9/user/hive/warehouse/tandem.db/cust_loss_alarm_unit

        atorg.apache.hadoop.hive.ql.metadata.Hive.moveFile(Hive.java:2521)

        atorg.apache.hadoop.hive.ql.exec.MoveTask.moveFile(MoveTask.java:105)

        atorg.apache.hadoop.hive.ql.exec.MoveTask.execute(MoveTask.java:222)

        at org.apache.hadoop.hive.ql.exec.Task.executeTask(Task.java:160)

        atorg.apache.hadoop.hive.ql.exec.TaskRunner.runSequential(TaskRunner.java:88)

        atorg.apache.hadoop.hive.ql.Driver.launchTask(Driver.java:1640)

        atorg.apache.hadoop.hive.ql.Driver.execute(Driver.java:1399)

        atorg.apache.hadoop.hive.ql.Driver.runInternal(Driver.java:1183)

        atorg.apache.hadoop.hive.ql.Driver.run(Driver.java:1049)

        atorg.apache.hadoop.hive.ql.Driver.run(Driver.java:1044)

        at org.apache.hive.service.cli.operation.SQLOperation.runQuery(SQLOperation.java:144)

        atorg.apache.hive.service.cli.operation.SQLOperation.access$100(SQLOperation.java:69)

        atorg.apache.hive.service.cli.operation.SQLOperation$1$1.run(SQLOperation.java:196)

        atjava.security.AccessController.doPrivileged(Native Method)

        atjavax.security.auth.Subject.doAs(Subject.java:415)

        atorg.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1671)

        at org.apache.hive.service.cli.operation.SQLOperation$1.run(SQLOperation.java:208)

        atjava.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471)

        atjava.util.concurrent.FutureTask.run(FutureTask.java:262)

        atjava.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)

        atjava.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)

        atjava.lang.Thread.run(Thread.java:745)

Caused by: java.io.IOException: Renames across Mount points notsupported

        atorg.apache.hadoop.fs.viewfs.ViewFileSystem.rename(ViewFileSystem.java:444)

        atorg.apache.hadoop.hive.ql.metadata.Hive.moveFile(Hive.java:2509)

        ... 21 more

相关代码:

org.apache.hadoop.fs.viewfs.ViewFileSystem

 /**

    // Alternate 1: renames within same file system -valid but we disallow

    // Alternate 2: (as described in next para - valid butwe have disallowed it

    //

    // Note we compare the URIs. the URIs include the linktargets.

    // hence we allow renames across mount links as longas the mount links

    // point to the same target.

    if (!re***c.targetFileSystem.getUri().equals(

             resDst.targetFileSystem.getUri())) {

      throw new IOException("Renames acrossMount points not supported");

    }

    */

   

    //

    // Alternate 3 : renames ONLY within the the samemount links.

    //

    if (re***c.targetFileSystem!=resDst.targetFileSystem) {

      throw new IOException("Renames acrossMount points not supported");

    }

 

Workaround:

a、在hdfs中 创建 /user/hive/warehouse/staging 目录,赋予777权限

然后添加配置:

    hive.exec.stagingdir

 /user/hive/warehouse/staging/.hive-staging

b、 只创建一个加载点如 /cluser   然后在此加载点下创建/tmp  /user等目录,最后修改hive相关目录的默认值。

2、 当查询返回结果集很大的时候,beeline客户端会卡住或out-of-memory

报错信息:

org.apache.thrift.TException: Error in calling method FetchResults

        atorg.apache.hive.jdbc.HiveConnection$SynchronizedHandler.invoke(HiveConnection.java:1271)

        atcom.sun.proxy.$Proxy0.FetchResults(Unknown Source)

        atorg.apache.hive.jdbc.HiveQueryResultSet.next(HiveQueryResultSet.java:363)

        at org.apache.hive.beeline.BufferedRows.(BufferedRows.java:42)

        atorg.apache.hive.beeline.BeeLine.print(BeeLine.java:1756)

        atorg.apache.hive.beeline.Commands.execute(Commands.java:806)

        atorg.apache.hive.beeline.Commands.sql(Commands.java:665)

        atorg.apache.hive.beeline.BeeLine.dispatch(BeeLine.java:974)

        atorg.apache.hive.beeline.BeeLine.execute(BeeLine.java:810)

        atorg.apache.hive.beeline.BeeLine.begin(BeeLine.java:767)

        at org.apache.hive.beeline.BeeLine.mainWithInputRedirection(BeeLine.java:480)

        atorg.apache.hive.beeline.BeeLine.main(BeeLine.java:463)

        atsun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)

        atsun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)

        atsun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)

        atjava.lang.reflect.Method.invoke(Method.java:606)

        atorg.apache.hadoop.util.RunJar.run(RunJar.java:221)

        at org.apache.hadoop.util.RunJar.main(RunJar.java:136)

Caused by: java.lang.OutOfMemoryError: Java heap space

        atjava.lang.Double.valueOf(Double.java:521)

       

Workaround:

查看源码发现:beeline获取结果集有两种模式一种增量模式,一种buffer模式

org.apache.hive.beeline.BeeLine

 

int print(ResultSet rs) throws SQLException {

    String format = getOpts().getOutputFormat();

    OutputFormat f = (OutputFormat)formats.get(format);

 

    if (f == null) {

     error(loc("unknown-format", new Object[] {

          format,formats.keySet()}));

      f = new TableOutputFormat(this);

    }

 

    Rows rows;

 

    if (getOpts().getIncremental()) {

      rows = new IncrementalRows(this,rs); //增量模式

    } else {

      rows = new BufferedRows(this, rs);buffer模式

    }

    return f.print(rows);

  }

org.apache.hive.beeline.BeeLineOpts

private boolean incremental = false; //默认为buffer模式

但是通过beeline –help没有发现相关设置

beeline --help

Usage: java org.apache.hive.cli.beeline.BeeLine

   -u               the JDBC URL to connect to

   -n                   the username to connect as

   -p                  the password to connect as

   -d               the driver class to use

   -i                  script file for initialization

   -e                     query that should be executed

   -f                  script file that should be executed

   -w (or) --password-file  the password file to read password from

   --hiveconfproperty=value       Use value for given property

   --hivevarname=value           hive variable name and value

                                  This is Hive specific settings in which variables

                                  can be set at session level and referenced in Hive

                                  commands or queries.

  --color=[true/false]           control whether color is used for display

  --showHeader=[true/false]       show column namesin query results

  --headerInterval=ROWS;         the interval between which heades are displayed

  --fastConnect=[true/false]      skip buildingtable/column list for tab-completion

  --autoCommit=[true/false]       enable/disableautomatic transaction commit

  --verbose=[true/false]         show verbose error messages and debug info

   --showWarnings=[true/false]    display connection warnings

   --showNestedErrs=[true/false]   displaynested errors

  --numberFormat=[pattern]        formatnumbers using DecimalFormat pattern

  --force=[true/false]           continue running script even after errors

  --maxWidth=MAXWIDTH            the maximum width of the terminal

   --maxColumnWidth=MAXCOLWIDTH    themaximum width to use when displaying columns

  --silent=[true/false]          be more silent

  --autosave=[true/false]        automatically save preferences

  --outputformat=[table/vertical/csv2/tsv2/dsv/csv/tsv]  format mode forresult display

                                  Note that csv, and tsv are deprecated - use csv2, tsv2 instead

  --truncateTable=[true/false]    truncatetable column when it exceeds length

   --delimiterForDSV=DELIMITER    specify the delimiter for delimiter-separated values output format (default: |)

  --isolation=LEVEL              set the transaction isolation level

   --nullemptystring=[true/false]  set to true toget historic behavior of printing null as empty string

  --help                         display this message

Beeline version 1.1.0-cdh6.4.3 by Apache Hive

但是没关系通过

beeline -u jdbc:hive2://10.17.28.173:10000–n xxxx -pxxxx --incremental=true 还是能进入增量模式


当前文章:FederatedHDFS+beeline+hiveserver2遇到的坑
文章转载:http://pwwzsj.com/article/jcpoip.html