You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 15 Current »

MapReduce Streaming Job — POST mapreduce/streaming

Description

Create and queue a Hadoop streaming MapReduce job.

Version: Hive 0.13.0 and later

As of Hive 0.13.0, GET version/hadoop displays the Hadoop version used for the MapReduce job.

URL

http://www.myserver.com/templeton/v1/mapreduce/streaming

Parameters

Name

Description

Required?

Default

input

Location of the input data in Hadoop.

Required

None

output

Location in which to store the output data. If not specified, WebHCat will store the output in a location that can be discovered using the queue resource.

Optional

See description

mapper

Location of the mapper program in Hadoop.

Required

None

reducer

Location of the reducer program in Hadoop.

Required

None

file

Add an HDFS file to the distributed cache.

Optional

None

define

Set a Hadoop configuration variable using the syntax define=NAME=VALUE

Optional

None

cmdenv

Set an environment variable using the syntax cmdenv=NAME=VALUE

Optional

None

arg

Set a program argument.

Optional

None

statusdir

A directory where WebHCat will write the status of the Map Reduce job. If provided, it is the caller's responsibility to remove this directory when done.

Optional

None

enablelog

If statusdir is set and enablelog is "true", collect Hadoop job configuration and logs into a directory named $statusdir/logs after the job finishes. Both completed and failed attempts are logged. The layout of subdirectories in $statusdir/logs is: 

logs/$job_id (directory for $job_id) 
logs/$job_id/job.xml.html 
logs/$job_id/$attempt_id (directory for $attempt_id) 
logs/$job_id/$attempt_id/stderr 
logs/$job_id/$attempt_id/stdout 
logs/$job_id/$attempt_id/syslog 

This parameter was introduced in Hive 0.12.0. (See HIVE-4531.)

Optional in Hive 0.12.0+None

callback

Define a URL to be called upon job completion. You may embed a specific job ID into this URL using $jobId. This tag will be replaced in the callback URL with this job's job ID.

Optional

None

The standard parameters are also supported.

Results

Name

Description

id

A string containing the job ID similar to "job_201110132141_0001".

info

A JSON object containing the information returned when the job was queued. See the Hadoop documentation (Class TaskController) for more information.

Example

Code and Data Setup

% cat mydata/file01 mydata/file02
Hello World Bye World
Hello Hadoop Goodbye Hadoop

% hadoop fs -put mydata/ .

% hadoop fs -ls mydata
Found 2 items
-rw-r--r--   1 ctdean supergroup         23 2011-11-11 13:29 /user/ctdean/mydata/file01
-rw-r--r--   1 ctdean supergroup         28 2011-11-11 13:29 /user/ctdean/mydata/file02

Curl Command

% curl -s -d input=mydata \
       -d output=mycounts \
       -d mapper=/bin/cat \
       -d reducer="/usr/bin/wc -w" \
       'http://localhost:50111/templeton/v1/mapreduce/streaming?user.name=ekoifman'

Version information

Prior to Hive 0.13.0, user.name was specified in POST requests as a form parameter: curl -d user.name=<user>.

In Hive 0.13.0 onward, user.name should be specified in the query string (as shown above): 'http://.../templeton/v1/mapreduce/streaming?user.name=<name>'. Specifying user.name as a form parameter is deprecated.

JSON Output

{
 "id": "job_201111111311_0008",
 "info": {
          "stdout": "packageJobJar: [] [/Users/ctdean/var/hadoop/hadoop-0.20.205.0/share/hadoop/contrib/streaming/hadoop-streaming-0.20.205.0.jar...
                    templeton-job-id:job_201111111311_0008
                    ",
          "stderr": "11/11/11 13:26:43 WARN mapred.JobClient: Use GenericOptionsParser for parsing the arguments
                    11/11/11 13:26:43 INFO mapred.FileInputFormat: Total input paths to process : 2
                    ",
          "exitcode": 0
         }
}

Example Results

% hadoop fs -ls mycounts
Found 3 items
-rw-r--r--   1 ctdean supergroup          0 2011-11-11 13:27 /user/ctdean/mycounts/_SUCCESS
drwxr-xr-x   - ctdean supergroup          0 2011-11-11 13:26 /user/ctdean/mycounts/_logs
-rw-r--r--   1 ctdean supergroup         10 2011-11-11 13:27 /user/ctdean/mycounts/part-00000

% hadoop fs -cat mycounts/part-00000
      8
  • No labels