最近在部署Hive上线，结果在线上线下同时出现了MoveTask报错的现象，虽然两者错误的日志以及错误信息一样，但是经过分析解决又发现两者的原因是不一样的。
首先线下的错误日志：
2015-05-18 18:53:09,679 ERROR [main]: exec.Task (SessionState.java:printError(833)) - Failed with exception Unable to rename: hdfs:
//
hadoop-master:9000/tmp/hive/hadoop/4d905c9f-ee65-4b1f-be96-93115b3aad61/hive_2015-05-18_18-51-42_401_2711668916550397051-1/-ext-10000 to: /user/hive/partitions/users_statis/dt_user_statis_behavior/event=play/period=0
org.apache.hadoop.hive.ql.metadata.HiveException: Unable to rename: hdfs:
//
hadoop-master:9000/tmp/hive/hadoop/4d905c9f-ee65-4b1f-be96-93115b3aad61/hive_2015-05-18_18-51-42_401_2711668916550397051-1/-ext-10000 to: /user/hive/partitions/users_statis/dt_user_statis_behavior/event=play/period=0
at org.apache.hadoop.hive.ql.exec.MoveTask.moveFile(MoveTask.java:111
)
at org.apache.hadoop.hive.ql.exec.MoveTask.execute(MoveTask.java:
213
)
at org.apache.hadoop.hive.ql.exec.Task.executeTask(Task.java:
160
)
at org.apache.hadoop.hive.ql.exec.TaskRunner.runSequential(TaskRunner.java:
85
)
at org.apache.hadoop.hive.ql.Driver.launchTask(Driver.java:
1604
)
at org.apache.hadoop.hive.ql.Driver.execute(Driver.java:
1364
)
at org.apache.hadoop.hive.ql.Driver.runInternal(Driver.java:
1177
)
at org.apache.hadoop.hive.ql.Driver.run(Driver.java:
1004
)
at org.apache.hadoop.hive.ql.Driver.run(Driver.java:
994
)
at org.apache.hadoop.hive.cli.CliDriver.processLocalCmd(CliDriver.java:
247
)
at org.apache.hadoop.hive.cli.CliDriver.processCmd(CliDriver.java:
199
)
at org.apache.hadoop.hive.cli.CliDriver.processLine(CliDriver.java:
410
)
at org.apache.hadoop.hive.cli.CliDriver.executeDriver(CliDriver.java:
783
)
at org.apache.hadoop.hive.cli.CliDriver.run(CliDriver.java:
677
)
at org.apache.hadoop.hive.cli.CliDriver.main(CliDriver.java:
616
)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:
57
)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:
43
)
at java.lang.reflect.Method.invoke(Method.java:
606
)
at org.apache.hadoop.util.RunJar.run(RunJar.java:
221
)
at org.apache.hadoop.util.RunJar.main(RunJar.java:
136)
这个问题在http://blog.csdn.net/lucien_zong/article/details/10198533中有详细的描述
执行SQL时，最后一个任务是MoveTask,它的作用是将运行SQL生成的Mapeduce任务结果文件放到SQL中指定的存储查询结果的路径中，具体方法就是重命名
下面是 org.apache.hadoop.hive.ql.exec.MoveTask 中对结果文件重命名的一段代码：//这个sourcePath参数就是存放Mapeduce结果文件的目录，所以它的值可能是 //hdfs://indigo:8020/tmp/hive-root/hive_2013-08-22_18-42-03_218_2856924886757165243/-ext-10000 if (fs.exists(sourcePath)) { Path deletePath = null ; // If it multiple level of folder are there fs.rename is failing so first // create the targetpath.getParent() if it not exist if (HiveConf.getBoolVar(conf, HiveConf.ConfVars.HIVE_INSERT_INTO_MULTILEVEL_DIRS)) { deletePath = createTargetPath(targetPath, fs); } //这里targetPath的值就是指定的放置结果文件的目录，值可能是 result/userName154122639/4e574b5d9f894a70b074ccd3981ca0f1 if (!fs.rename(sourcePath, targetPath)) {//上面产生的异常就是因为这里rename失败，进了if，throw了个异常 try { if (deletePath != null ) { fs.delete(deletePath, true ); } } catch (IOException e) { LOG.info("Unable to delete the path created for facilitating rename" + deletePath); } throw new HiveException("Unable to rename: " + sourcePath + " to: " + targetPath); } }rename的targetPath必须存在。
其实之前已经检查和创建targetPath了：
private Path createTargetPath(Path targetPath, FileSystem fs) throws IOException { Path deletePath = null ; Path mkDirPath = targetPath.getParent(); if (mkDirPath != null & ! fs.exists(mkDirPath)) { Path actualPath = mkDirPath; while (actualPath != null && ! fs.exists(actualPath)) { deletePath = actualPath; actualPath = actualPath.getParent(); } <property> <name>hive.insert.into.multilevel.dirs</name> <value>true</value> </property> fs.mkdirs(mkDirPath); } return deletePath;//返回新创建的最顶层的目录，万一失败用来删除用 }Apache出现过这个 问题 ，已经解决掉了
CDH 竟然加了个参数 hive.insert.into.multilevel.dirs，默认是false，意思是我还有这BUG呢哈。
当你被坑了，想打个patch时，会发现改个配置就可以了。
意思是我保留这个BUG，但你要是被坑了也不能说我有BUG，自己改配置好了.
目前还没发现其他地方用到了这个参数，在这里唯一作用就是限制SQL中指定存放结果文件不存在的目录的深度不能大于1.
不过也没发现这有什么好处。折腾半天，加个配置就可以了:
<property> <name>hive.insert.into.multilevel.dirs</name> <value> true </value> </property>
解决完线下的问题满心欢喜的以为可以解决线上的问题，结果发现了不行，仔细查看日志发现原来是因为缺包造成的
2015-05-18 19:22:03,799 ERROR [main]: exec.Task (SessionState.java:printError(861)) - Failed with exception Unable to move source hdfs:
//
hadoop1:9000/tmp/hive/statistics/dt_statistics_content_daily/.hive-staging_hive_2015-05-18_19-11-45_323_132664610162390564-1/-ext-10000 to destination /tmp/hive/statistics/dt_statistics_content_daily
org.apache.hadoop.hive.ql.metadata.HiveException: Unable to move source hdfs:
//
hadoop1:9000/tmp/hive/statistics/dt_statistics_content_daily/.hive-staging_hive_2015-05-18_19-11-45_323_132664610162390564-1/-ext-10000 to destination /tmp/hive/statistics/dt_statistics_content_daily
at org.apache.hadoop.hive.ql.metadata.Hive.moveFile(Hive.java:2483
)
at org.apache.hadoop.hive.ql.exec.MoveTask.moveFile(MoveTask.java:
105
)
at org.apache.hadoop.hive.ql.exec.MoveTask.execute(MoveTask.java:
222
)
at org.apache.hadoop.hive.ql.exec.Task.executeTask(Task.java:
160
)
at org.apache.hadoop.hive.ql.exec.TaskRunner.runSequential(TaskRunner.java:
88
)
at org.apache.hadoop.hive.ql.Driver.launchTask(Driver.java:
1638
)
at org.apache.hadoop.hive.ql.Driver.execute(Driver.java:
1397
)
at org.apache.hadoop.hive.ql.Driver.runInternal(Driver.java:
1183
)
at org.apache.hadoop.hive.ql.Driver.run(Driver.java:
1049
)
at org.apache.hadoop.hive.ql.Driver.run(Driver.java:
1039
)
at org.apache.hadoop.hive.cli.CliDriver.processLocalCmd(CliDriver.java:
207
)
at org.apache.hadoop.hive.cli.CliDriver.processCmd(CliDriver.java:
159
)
at org.apache.hadoop.hive.cli.CliDriver.processLine(CliDriver.java:
370
)
at org.apache.hadoop.hive.cli.CliDriver.executeDriver(CliDriver.java:
754
)
at org.apache.hadoop.hive.cli.CliDriver.run(CliDriver.java:
675
)
at org.apache.hadoop.hive.cli.CliDriver.main(CliDriver.java:
615
)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:
57
)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:
43
)
at java.lang.reflect.Method.invoke(Method.java:
606
)
at org.apache.hadoop.util.RunJar.run(RunJar.java:
221
)
at org.apache.hadoop.util.RunJar.main(RunJar.java:
136
)
Caused by: java.io.IOException: Cannot find DistCp
class
package
: org.apache.hadoop.tools.DistCp
at org.apache.hadoop.hive.shims.Hadoop23Shims.runDistCp(Hadoop23Shims.java:
1123
)
at org.apache.hadoop.hive.common.FileUtils.copy(FileUtils.java:
553
)
at org.apache.hadoop.hive.ql.metadata.Hive.moveFile(Hive.java:
2461
)
...
21 more
查看源码发现在调用moveFile时候会去调用Hadoop的distcp接口
@Override
public
boolean
runDistCp(Path src, Path dst, Configuration conf)
throws
IOException {
int
rc;
//
Creates the command-line parameters for distcp
String[] params = {"-update", "-skipcrccheck"
, src.toString(), dst.toString()};
try
{
Class clazzDistCp
= Class.forName("org.apache.hadoop.tools.DistCp"
);
Constructor c
=
clazzDistCp.getConstructor();
c.setAccessible(
true
);
Tool distcp
=
(Tool)c.newInstance();
distcp.setConf(conf);
rc
=
distcp.run(params);
}
catch
(ClassNotFoundException e) {
throw
new
IOException("Cannot find DistCp class package: " +
e.getMessage());
}
catch
(NoSuchMethodException e) {
throw
new
IOException("Cannot get DistCp constructor: " +
e.getMessage());
}
catch
(Exception e) {
throw
new
IOException("Cannot execute DistCp process: " +
e, e);
}
return
(0 ==
rc);
}
使用命令发现Hive环境变量中并没有Hadoop-Distcp-2.6.0.jar
hive -e 'set' | grep distcp
于是只要把hadoop/share/tools/下的Hadoop-Distcp-2.6.0.jar加载下来就行。
需要说明的是线上的Hive是1.1.0版本，线下的hive是0.14版本。同时需要注意的是线下的move文件的方式是rename，线上则用distcp方式，这跟Hive的版本有关了。
