2021年大数据Flink(七):​​​​​​​参数总结

参数总结

[root@node1 bin]# /export/server/flink/bin/flink --help

./flink <ACTION> [OPTIONS] [ARGUMENTS]



The following actions are available:



Action "run" compiles and runs a program.



  Syntax: run [OPTIONS] <jar-file> <arguments>

  "run" action options:

     -c,--class <classname>               Class with the program entry point

                                          ("main()" method). Only needed if the

                                          JAR file does not specify the class in

                                          its manifest.

     -C,--classpath <url>                 Adds a URL to each user code

                                          classloader  on all nodes in the

                                          cluster. The paths must specify a

                                          protocol (e.g. file://) and be

                                          accessible on all nodes (e.g. by means

                                          of a NFS share). You can use this

                                          option multiple times for specifying

                                          more than one URL. The protocol must

                                          be supported by the {@link

                                          java.net.URLClassLoader}.

     -d,--detached                        If present, runs the job in detached

                                          mode

     -n,--allowNonRestoredState           Allow to skip savepoint state that

                                          cannot be restored. You need to allow

                                          this if you removed an operator from

                                          your program that was part of the

                                          program when the savepoint was

                                          triggered.

     -p,--parallelism <parallelism>       The parallelism with which to run the

                                          program. Optional flag to override the

                                          default value specified in the

                                          configuration.

     -py,--python <pythonFile>            Python script with the program entry

                                          point. The dependent resources can be

                                          configured with the `--pyFiles`

                                          option.

     -pyarch,--pyArchives <arg>           Add python archive files for job. The

                                          archive files will be extracted to the

                                          working directory of python UDF

                                          worker. Currently only zip-format is

                                          supported. For each archive file, a

                                          target directory be specified. If the

                                          target directory name is specified,

                                          the archive file will be extracted to

                                          a name can directory with the

                                          specified name. Otherwise, the archive

                                          file will be extracted to a directory

                                          with the same name of the archive

                                          file. The files uploaded via this

                                          option are accessible via relative

                                          path. '#' could be used as the

                                          separator of the archive file path and

                                          the target directory name. Comma (',')

                                          could be used as the separator to

                                          specify multiple archive files. This

                                          option can be used to upload the

                                          virtual environment, the data files

                                          used in Python UDF (e.g.: --pyArchives

                                          file:///tmp/py37.zip,file:///tmp/data.

                                          zip#data --pyExecutable

                                          py37.zip/py37/bin/python). The data

                                          files could be accessed in Python UDF,

                                          e.g.: f = open('data/data.txt', 'r').

     -pyexec,--pyExecutable <arg>         Specify the path of the python

                                          interpreter used to execute the python

                                          UDF worker (e.g.: --pyExecutable

                                          /usr/local/bin/python3). The python

                                          UDF worker depends on Python 3.5+,

                                          Apache Beam (version == 2.23.0), Pip

                                          (version >= 7.1.0) and SetupTools

                                          (version >= 37.0.0). Please ensure

                                          that the specified environment meets

                                          the above requirements.

     -pyfs,--pyFiles <pythonFiles>        Attach custom python files for job.

                                          These files will be added to the

                                          PYTHONPATH of both the local client

                                          and the remote python UDF worker. The

                                          standard python resource file suffixes

                                          such as .py/.egg/.zip or directory are

                                          all supported. Comma (',') could be

                                          used as the separator to specify

                                          multiple files (e.g.: --pyFiles

                                          file:///tmp/myresource.zip,hdfs:///$na

                                          menode_address/myresource2.zip).

     -pym,--pyModule <pythonModule>       Python module with the program entry

                                          point. This option must be used in

                                          conjunction with `--pyFiles`.

     -pyreq,--pyRequirements <arg>        Specify a requirements.txt file which

                                          defines the third-party dependencies.

                                          These dependencies will be installed

                                          and added to the PYTHONPATH of the

                                          python UDF worker. A directory which

                                          contains the installation packages of

                                          these dependencies could be specified

                                          optionally. Use '#' as the separator

                                          if the optional parameter exists

                                          (e.g.: --pyRequirements

                                          file:///tmp/requirements.txt#file:///t

                                          mp/cached_dir).

     -s,--fromSavepoint <savepointPath>   Path to a savepoint to restore the job

                                          from (for example

                                          hdfs:///flink/savepoint-1537).

     -sae,--shutdownOnAttachedExit        If the job is submitted in attached

                                          mode, perform a best-effort cluster

                                          shutdown when the CLI is terminated

                                          abruptly, e.g., in response to a user

                                          interrupt, such as typing Ctrl + C.

  Options for Generic CLI mode:

     -D <property=value>   Allows specifying multiple generic configuration

                           options. The available options can be found at

                           https://ci.apache.org/projects/flink/flink-docs-stabl

                           e/ops/config.html

     -e,--executor <arg>   DEPRECATED: Please use the -t option instead which is

                           also available with the "Application Mode".

                           The name of the executor to be used for executing the

                           given job, which is equivalent to the

                           "execution.target" config option. The currently

                           available executors are: "remote", "local",

                           "kubernetes-session", "yarn-per-job", "yarn-session".

     -t,--target <arg>     The deployment target for the given application,

                           which is equivalent to the "execution.target" config

                           option. For the "run" action the currently available

                           targets are: "remote", "local", "kubernetes-session",

                           "yarn-per-job", "yarn-session". For the

                           "run-application" action the currently available

                           targets are: "kubernetes-application",

                           "yarn-application".



  Options for yarn-cluster mode:

     -d,--detached                        If present, runs the job in detached

                                          mode

     -m,--jobmanager <arg>                Set to yarn-cluster to use YARN

                                          execution mode.

     -yat,--yarnapplicationType <arg>     Set a custom application type for the

                                          application on YARN

     -yD <property=value>                 use value for given property

     -yd,--yarndetached                   If present, runs the job in detached

                                          mode (deprecated; use non-YARN

                                          specific option instead)

     -yh,--yarnhelp                       Help for the Yarn session CLI.

     -yid,--yarnapplicationId <arg>       Attach to running YARN session

     -yj,--yarnjar <arg>                  Path to Flink jar file

     -yjm,--yarnjobManagerMemory <arg>    Memory for JobManager Container with

                                          optional unit (default: MB)

     -ynl,--yarnnodeLabel <arg>           Specify YARN node label for the YARN

                                          application

     -ynm,--yarnname <arg>                Set a custom name for the application

                                          on YARN

     -yq,--yarnquery                      Display available YARN resources

                                          (memory, cores)

     -yqu,--yarnqueue <arg>               Specify YARN queue.

     -ys,--yarnslots <arg>                Number of slots per TaskManager

     -yt,--yarnship <arg>                 Ship files in the specified directory

                                          (t for transfer)

     -ytm,--yarntaskManagerMemory <arg>   Memory per TaskManager Container with

                                          optional unit (default: MB)

     -yz,--yarnzookeeperNamespace <arg>   Namespace to create the Zookeeper

                                          sub-paths for high availability mode

     -z,--zookeeperNamespace <arg>        Namespace to create the Zookeeper

                                          sub-paths for high availability mode



  Options for default mode:

     -D <property=value>             Allows specifying multiple generic

                                     configuration options. The available

                                     options can be found at

                                     https://ci.apache.org/projects/flink/flink-

                                     docs-stable/ops/config.html

     -m,--jobmanager <arg>           Address of the JobManager to which to

                                     connect. Use this flag to connect to a

                                     different JobManager than the one specified

                                     in the configuration. Attention: This

                                     option is respected only if the

                                     high-availability configuration is NONE.

     -z,--zookeeperNamespace <arg>   Namespace to create the Zookeeper sub-paths

                                     for high availability mode







Action "run-application" runs an application in Application Mode.



  Syntax: run-application [OPTIONS] <jar-file> <arguments>

  Options for Generic CLI mode:

     -D <property=value>   Allows specifying multiple generic configuration

                           options. The available options can be found at

                           https://ci.apache.org/projects/flink/flink-docs-stabl

                           e/ops/config.html

     -e,--executor <arg>   DEPRECATED: Please use the -t option instead which is

                           also available with the "Application Mode".

                           The name of the executor to be used for executing the

                           given job, which is equivalent to the

                           "execution.target" config option. The currently

                           available executors are: "remote", "local",

                           "kubernetes-session", "yarn-per-job", "yarn-session".

     -t,--target <arg>     The deployment target for the given application,

                           which is equivalent to the "execution.target" config

                           option. For the "run" action the currently available

                           targets are: "remote", "local", "kubernetes-session",

                           "yarn-per-job", "yarn-session". For the

                           "run-application" action the currently available

                           targets are: "kubernetes-application",

                           "yarn-application".







Action "info" shows the optimized execution plan of the program (JSON).



  Syntax: info [OPTIONS] <jar-file> <arguments>

  "info" action options:

     -c,--class <classname>           Class with the program entry point

                                      ("main()" method). Only needed if the JAR

                                      file does not specify the class in its

                                      manifest.

     -p,--parallelism <parallelism>   The parallelism with which to run the

                                      program. Optional flag to override the

                                      default value specified in the

                                      configuration.





Action "list" lists running and scheduled programs.



  Syntax: list [OPTIONS]

  "list" action options:

     -a,--all         Show all programs and their JobIDs

     -r,--running     Show only running programs and their JobIDs

     -s,--scheduled   Show only scheduled programs and their JobIDs

  Options for Generic CLI mode:

     -D <property=value>   Allows specifying multiple generic configuration

                           options. The available options can be found at

                           https://ci.apache.org/projects/flink/flink-docs-stabl

                           e/ops/config.html

     -e,--executor <arg>   DEPRECATED: Please use the -t option instead which is

                           also available with the "Application Mode".

                           The name of the executor to be used for executing the

                           given job, which is equivalent to the

                           "execution.target" config option. The currently

                           available executors are: "remote", "local",

                           "kubernetes-session", "yarn-per-job", "yarn-session".

     -t,--target <arg>     The deployment target for the given application,

                           which is equivalent to the "execution.target" config

                           option. For the "run" action the currently available

                           targets are: "remote", "local", "kubernetes-session",

                           "yarn-per-job", "yarn-session". For the

                           "run-application" action the currently available

                           targets are: "kubernetes-application",

                           "yarn-application".



  Options for yarn-cluster mode:

     -m,--jobmanager <arg>            Set to yarn-cluster to use YARN execution

                                      mode.

     -yid,--yarnapplicationId <arg>   Attach to running YARN session

     -z,--zookeeperNamespace <arg>    Namespace to create the Zookeeper

                                      sub-paths for high availability mode



  Options for default mode:

     -D <property=value>             Allows specifying multiple generic

                                     configuration options. The available

                                     options can be found at

                                     https://ci.apache.org/projects/flink/flink-

                                     docs-stable/ops/config.html

     -m,--jobmanager <arg>           Address of the JobManager to which to

                                     connect. Use this flag to connect to a

                                     different JobManager than the one specified

                                     in the configuration. Attention: This

                                     option is respected only if the

                                     high-availability configuration is NONE.

     -z,--zookeeperNamespace <arg>   Namespace to create the Zookeeper sub-paths

                                     for high availability mode







Action "stop" stops a running program with a savepoint (streaming jobs only).



  Syntax: stop [OPTIONS] <Job ID>

  "stop" action options:

     -d,--drain                           Send MAX_WATERMARK before taking the

                                          savepoint and stopping the pipelne.

     -p,--savepointPath <savepointPath>   Path to the savepoint (for example

                                          hdfs:///flink/savepoint-1537). If no

                                          directory is specified, the configured

                                          default will be used

                                          ("state.savepoints.dir").

  Options for Generic CLI mode:

     -D <property=value>   Allows specifying multiple generic configuration

                           options. The available options can be found at

                           https://ci.apache.org/projects/flink/flink-docs-stabl

                           e/ops/config.html

     -e,--executor <arg>   DEPRECATED: Please use the -t option instead which is

                           also available with the "Application Mode".

                           The name of the executor to be used for executing the

                           given job, which is equivalent to the

                           "execution.target" config option. The currently

                           available executors are: "remote", "local",

                           "kubernetes-session", "yarn-per-job", "yarn-session".

     -t,--target <arg>     The deployment target for the given application,

                           which is equivalent to the "execution.target" config

                           option. For the "run" action the currently available

                           targets are: "remote", "local", "kubernetes-session",

                           "yarn-per-job", "yarn-session". For the

                           "run-application" action the currently available

                           targets are: "kubernetes-application",

                           "yarn-application".



  Options for yarn-cluster mode:

     -m,--jobmanager <arg>            Set to yarn-cluster to use YARN execution

                                      mode.

     -yid,--yarnapplicationId <arg>   Attach to running YARN session

     -z,--zookeeperNamespace <arg>    Namespace to create the Zookeeper

                                      sub-paths for high availability mode



  Options for default mode:

     -D <property=value>             Allows specifying multiple generic

                                     configuration options. The available

                                     options can be found at

                                     https://ci.apache.org/projects/flink/flink-

                                     docs-stable/ops/config.html

     -m,--jobmanager <arg>           Address of the JobManager to which to

                                     connect. Use this flag to connect to a

                                     different JobManager than the one specified

                                     in the configuration. Attention: This

                                     option is respected only if the

                                     high-availability configuration is NONE.

     -z,--zookeeperNamespace <arg>   Namespace to create the Zookeeper sub-paths

                                     for high availability mode







Action "cancel" cancels a running program.



  Syntax: cancel [OPTIONS] <Job ID>

  "cancel" action options:

     -s,--withSavepoint <targetDirectory>   **DEPRECATION WARNING**: Cancelling

                                            a job with savepoint is deprecated.

                                            Use "stop" instead.

                                            Trigger savepoint and cancel job.

                                            The target directory is optional. If

                                            no directory is specified, the

                                            configured default directory

                                            (state.savepoints.dir) is used.

  Options for Generic CLI mode:

     -D <property=value>   Allows specifying multiple generic configuration

                           options. The available options can be found at

                           https://ci.apache.org/projects/flink/flink-docs-stabl

                           e/ops/config.html

     -e,--executor <arg>   DEPRECATED: Please use the -t option instead which is

                           also available with the "Application Mode".

                           The name of the executor to be used for executing the

                           given job, which is equivalent to the

                           "execution.target" config option. The currently

                           available executors are: "remote", "local",

                           "kubernetes-session", "yarn-per-job", "yarn-session".

     -t,--target <arg>     The deployment target for the given application,

                           which is equivalent to the "execution.target" config

                           option. For the "run" action the currently available

                           targets are: "remote", "local", "kubernetes-session",

                           "yarn-per-job", "yarn-session". For the

                           "run-application" action the currently available

                           targets are: "kubernetes-application",

                           "yarn-application".



  Options for yarn-cluster mode:

     -m,--jobmanager <arg>            Set to yarn-cluster to use YARN execution

                                      mode.

     -yid,--yarnapplicationId <arg>   Attach to running YARN session

     -z,--zookeeperNamespace <arg>    Namespace to create the Zookeeper

                                      sub-paths for high availability mode



  Options for default mode:

     -D <property=value>             Allows specifying multiple generic

                                     configuration options. The available

                                     options can be found at

                                     https://ci.apache.org/projects/flink/flink-

                                     docs-stable/ops/config.html

     -m,--jobmanager <arg>           Address of the JobManager to which to

                                     connect. Use this flag to connect to a

                                     different JobManager than the one specified

                                     in the configuration. Attention: This

                                     option is respected only if the

                                     high-availability configuration is NONE.

     -z,--zookeeperNamespace <arg>   Namespace to create the Zookeeper sub-paths

                                     for high availability mode







Action "savepoint" triggers savepoints for a running job or disposes existing ones.



  Syntax: savepoint [OPTIONS] <Job ID> [<target directory>]

  "savepoint" action options:

     -d,--dispose <arg>       Path of savepoint to dispose.

     -j,--jarfile <jarfile>   Flink program JAR file.

  Options for Generic CLI mode:

     -D <property=value>   Allows specifying multiple generic configuration

                           options. The available options can be found at

                           https://ci.apache.org/projects/flink/flink-docs-stabl

                           e/ops/config.html

     -e,--executor <arg>   DEPRECATED: Please use the -t option instead which is

                           also available with the "Application Mode".

                           The name of the executor to be used for executing the

                           given job, which is equivalent to the

                           "execution.target" config option. The currently

                           available executors are: "remote", "local",

                           "kubernetes-session", "yarn-per-job", "yarn-session".

     -t,--target <arg>     The deployment target for the given application,

                           which is equivalent to the "execution.target" config

                           option. For the "run" action the currently available

                           targets are: "remote", "local", "kubernetes-session",

                           "yarn-per-job", "yarn-session". For the

                           "run-application" action the currently available

                           targets are: "kubernetes-application",

                           "yarn-application".



  Options for yarn-cluster mode:

     -m,--jobmanager <arg>            Set to yarn-cluster to use YARN execution

                                      mode.

     -yid,--yarnapplicationId <arg>   Attach to running YARN session

     -z,--zookeeperNamespace <arg>    Namespace to create the Zookeeper

                                      sub-paths for high availability mode



  Options for default mode:

     -D <property=value>             Allows specifying multiple generic

                                     configuration options. The available

                                     options can be found at

                                     https://ci.apache.org/projects/flink/flink-

                                     docs-stable/ops/config.html

     -m,--jobmanager <arg>           Address of the JobManager to which to

                                     connect. Use this flag to connect to a

                                     different JobManager than the one specified

                                     in the configuration. Attention: This

                                     option is respected only if the

                                     high-availability configuration is NONE.

     -z,--zookeeperNamespace <arg>   Namespace to create the Zookeeper sub-paths

                                     for high availability mode

 

上一篇:查看服务器剩余内存的正确姿势


下一篇:Swfit 常用特性(Attribute)关键字