Shell Options

All of the shell commands will accept a common set of options. For some commands, these options are ignored. For example, passing ---hostnames on a command that only executes on a single host will be ignored.

Generic Options

Many subcommands honor a common set of configuration options to alter their behavior:

Hadoop Common Commands

All of these commands are executed from the hadoop shell command. They have been broken up into User Commands and Administration Commands.

archive

Creates a hadoop archive. More information can be found at Hadoop Archives Guide.

checknative

Usage: hadoop checknative [-a] [-h]

This command checks the availability of the Hadoop native code. See Native Libaries for more information. By default, this command only checks the availability of libhadoop.

classpath

Usage: hadoop classpath [--glob |--jar <path> |-h |--help]

Prints the class path needed to get the Hadoop jar and the required libraries. If called without arguments, then prints the classpath set up by the command scripts, which is likely to contain wildcards in the classpath entries. Additional options print the classpath after wildcard expansion or write the classpath into the manifest of a jar file. The latter is useful in environments where wildcards cannot be used and the expanded classpath exceeds the maximum supported command line length.

conftest

Usage: hadoop conftest [-conffile <path>]...

Validates configuration XML files. If the -conffile option is not specified, the files in ${HADOOP_CONF_DIR} whose name end with .xml will be verified. If specified, that path will be verified. You can specify either a file or directory, and if a directory specified, the files in that directory whose name end with .xml will be verified. You can specify -conffile option multiple times.

The validation is fairly minimal: the XML is parsed and duplicate and empty property names are checked for. The command does not support XInclude; if you using that to pull in configuration items, it will declare the XML file invalid.

credential

Usage: hadoop credential <subcommand> [options]

Command to manage credentials, passwords and secrets within credential providers.

The CredentialProvider API in Hadoop allows for the separation of applications and how they store their required passwords/secrets. In order to indicate a particular provider type and location, the user must provide the hadoop.security.credential.provider.path configuration element in core-site.xml or use the command line option -provider on each of the following commands. This provider path is a comma-separated list of URLs that indicates the type and location of a list of providers that should be consulted. For example, the following path: user:///,jceks://file/tmp/test.jceks,jceks://hdfs@nn1.example.com/my/path/test.jceks

indicates that the current user’s credentials file should be consulted through the User Provider, that the local file located at /tmp/test.jceks is a Java Keystore Provider and that the file located within HDFS at nn1.example.com/my/path/test.jceks is also a store for a Java Keystore Provider.

When utilizing the credential command it will often be for provisioning a password or secret to a particular credential store provider. In order to explicitly indicate which provider store to use the -provider option should be used. Otherwise, given a path of multiple providers, the first non-transient provider will be used. This may or may not be the one that you intended.

Providers frequently require that a password or other secret is supplied. If the provider requires a password and is unable to find one, it will use a default password and emit a warning message that the default password is being used. If the -strict flag is supplied, the warning message becomes an error message and the command returns immediately with an error status.

Example: hadoop credential list -provider jceks://file/tmp/test.jceks

distch

Usage: hadoop distch [-f urilist_url] [-i] [-log logdir] path:owner:group:permissions

Change the ownership and permissions on many files at once.

distcp

Copy file or directories recursively. More information can be found at Hadoop DistCp Guide.

dtutil

Usage: hadoop dtutil [-keytab keytab_file -principal principal_name ] subcommand [-format (java|protobuf)] [-alias alias ] [-renewer renewer ] filename…

Utility to fetch and manage hadoop delegation tokens inside credentials files. It is intended to replace the simpler command fetchdt. There are multiple subcommands, each with their own flags and options.

For every subcommand that writes out a file, the -format option will specify the internal format to use. java is the legacy format that matches fetchdt. The default is protobuf.

For every subcommand that connects to a service, convenience flags are provided to specify the kerberos principal name and keytab file to use for auth.

fs

This command is documented in the File System Shell Guide. It is a synonym for hdfs dfs when HDFS is in use.

gridmix

Gridmix is a benchmark tool for Hadoop cluster. More information can be found in the Gridmix Guide.

jar

Usage: hadoop jar <jar> [mainClass] args...

Runs a jar file.

Use yarn jar to launch YARN applications instead.

jnipath

Usage: hadoop jnipath

Print the computed java.library.path.

kerbname

Usage: hadoop kerbname principal

Convert the named principal via the auth_to_local rules to the Hadoop user name.

Example: hadoop kerbname user@EXAMPLE.COM

kdiag

Usage: hadoop kdiag

Diagnose Kerberos Problems

key

Usage: hadoop key <subcommand> [options]

| -help | Prints usage of this command |

Manage keys via the KeyProvider. For details on KeyProviders, see the Transparent Encryption Guide.

Providers frequently require that a password or other secret is supplied. If the provider requires a password and is unable to find one, it will use a default password and emit a warning message that the default password is being used. If the -strict flag is supplied, the warning message becomes an error message and the command returns immediately with an error status.

NOTE: Some KeyProviders (e.g. org.apache.hadoop.crypto.key.JavaKeyStoreProvider) do not support uppercase key names.

NOTE: Some KeyProviders do not directly execute a key deletion (e.g. performs a soft-delete instead, or delay the actual deletion, to prevent mistake). In these cases, one may encounter errors when creating/deleting a key with the same name after deleting it. Please check the underlying KeyProvider for details.

kms

Usage: hadoop kms

Run KMS, the Key Management Server.

version

Usage: hadoop version

Prints the version.

CLASSNAME

Usage: hadoop CLASSNAME

Runs the class named CLASSNAME. The class must be part of a package.

envvars

Usage: hadoop envvars

Display computed Hadoop environment variables.

daemonlog

Usage:

hadoop daemonlog -getlevel <host:port> <classname> [-protocol (http|https)]
hadoop daemonlog -setlevel <host:port> <classname> <level> [-protocol (http|https)]

Get/Set the log level for a Log identified by a qualified class name in the daemon dynamically. By default, the command sends a HTTP request, but this can be overridden by using argument -protocol https to send a HTTPS request.

Example:

$ bin/hadoop daemonlog -setlevel 127.0.0.1:9870 org.apache.hadoop.hdfs.server.namenode.NameNode DEBUG
$ bin/hadoop daemonlog -getlevel 127.0.0.1:9871 org.apache.hadoop.hdfs.server.namenode.NameNode -protocol https

Note that the setting is not permanent and will be reset when the daemon is restarted. This command works by sending a HTTP/HTTPS request to the daemon’s internal Jetty servlet, so it supports the following daemons:

• Common
  • key management server
• HDFS
  • name node
  • secondary name node
  • data node
  • journal node
  • HttpFS server
• YARN
  • resource manager
  • node manager
  • Timeline server

etc/hadoop/hadoop-env.sh

This file stores the global settings used by all Hadoop shell commands.

etc/hadoop/hadoop-user-functions.sh

This file allows for advanced users to override some shell functionality.

~/.hadooprc

This stores the personal environment for an individual user. It is processed after the hadoop-env.sh and hadoop-user-functions.sh files and can contain the same settings.