[
https://issues.apache.org/jira/browse/MAPREDUCE-5965?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14554197#comment-14554197
]
Wilfred Spiegelenburg commented on MAPREDUCE-5965:
--------------------------------------------------
[~amalakar] thank you for the assignment. The comment should be added back,
I'll do that with an updated patch. The move to keep it in the same method was
to make the change as simple as possible.
[~rchiang]
The streaming configuration does not really have a *-default.xml file. There is
documentation (markdown) that shows some of the settings and options: adding it
to the FAQ would probably be the correct place. There is a help that is printed
in the main StreamJob code that shows most of the options. I will update the
two files and explain the setting that is available. I can upload a new patch
with that added before I do lets get the other points finalised.
A white list or black list is possible but what would we exclude or include? In
the job configuration there could be any value which could be too long, a user
could set something he wants. It will be really difficult to filter that
consistently and be sure that we have a fix with limited impact.
Making the lenLimit configurable is possible. However I do not see what we
would win with making the length configurable. The data is not used anywhere
and lowering or increasing the size at which we cut it off will not give us
anything extra. If you really want to make it configurable the easiest way
would be to roll the two settings in one. We could make the
stream.truncate.long.jobconf.values an integer: -1 do not truncate otherwise
truncate at the length given.
> Hadoop streaming throws error if list of input files is high. Error is:
> "error=7, Argument list too long at if number of input file is high"
> --------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: MAPREDUCE-5965
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-5965
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Reporter: Arup Malakar
> Assignee: Wilfred Spiegelenburg
> Attachments: MAPREDUCE-5965.1.patch, MAPREDUCE-5965.2.patch,
> MAPREDUCE-5965.patch
>
>
> Hadoop streaming exposes all the key values in job conf as environment
> variables when it forks a process for streaming code to run. Unfortunately
> the variable mapreduce_input_fileinputformat_inputdir contains the list of
> input files, and Linux has a limit on size of environment variables +
> arguments.
> Based on how long the list of files and their full path is this could be
> pretty huge. And given all of these variables are not even used it stops user
> from running hadoop job with large number of files, even though it could be
> run.
> Linux throws E2BIG if the size is greater than certain size which is error
> code 7. And java translates that to "error=7, Argument list too long". More:
> http://man7.org/linux/man-pages/man2/execve.2.html I suggest skipping
> variables if it is greater than certain length. That way if user code
> requires the environment variable it would fail. It should also introduce a
> config variable to skip long variables, and set it to false by default. That
> way user has to specifically set it to true to invoke this feature.
> Here is the exception:
> {code}
> Error: java.lang.RuntimeException: Error in configuring object at
> org.apache.hadoop.util.ReflectionUtils.setJobConf(ReflectionUtils.java:109)
> at org.apache.hadoop.util.ReflectionUtils.setConf(ReflectionUtils.java:75) at
> org.apache.hadoop.util.ReflectionUtils.newInstance(ReflectionUtils.java:133)
> at org.apache.hadoop.mapred.MapTask.runOldMapper(MapTask.java:426) at
> org.apache.hadoop.mapred.MapTask.run(MapTask.java:342) at
> org.apache.hadoop.mapred.YarnChild$2.run(YarnChild.java:168) at
> java.security.AccessController.doPrivileged(Native Method) at
> javax.security.auth.Subject.doAs(Subject.java:415) at
> org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1548)
> at org.apache.hadoop.mapred.YarnChild.main(YarnChild.java:163) Caused by:
> java.lang.reflect.InvocationTargetException 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.ReflectionUtils.setJobConf(ReflectionUtils.java:106)
> ... 9 more Caused by: java.lang.RuntimeException: Error in configuring object
> at
> org.apache.hadoop.util.ReflectionUtils.setJobConf(ReflectionUtils.java:109)
> at org.apache.hadoop.util.ReflectionUtils.setConf(ReflectionUtils.java:75) at
> org.apache.hadoop.util.ReflectionUtils.newInstance(ReflectionUtils.java:133)
> at org.apache.hadoop.mapred.MapRunner.configure(MapRunner.java:38) ... 14
> more Caused by: java.lang.reflect.InvocationTargetException 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.ReflectionUtils.setJobConf(ReflectionUtils.java:106)
> ... 17 more Caused by: java.lang.RuntimeException: configuration exception at
> org.apache.hadoop.streaming.PipeMapRed.configure(PipeMapRed.java:222) at
> org.apache.hadoop.streaming.PipeMapper.configure(PipeMapper.java:66) ... 22
> more Caused by: java.io.IOException: Cannot run program
> "/data/hadoop/hadoop-yarn/cache/yarn/nm-local-dir/usercache/oo-analytics/appcache/application_1403599726264_13177/container_1403599726264_13177_01_000006/./rbenv_runner.sh":
> error=7, Argument list too long at
> java.lang.ProcessBuilder.start(ProcessBuilder.java:1041) at
> org.apache.hadoop.streaming.PipeMapRed.configure(PipeMapRed.java:209) ... 23
> more Caused by: java.io.IOException: error=7, Argument list too long at
> java.lang.UNIXProcess.forkAndExec(Native Method) at
> java.lang.UNIXProcess.<init>(UNIXProcess.java:135) at
> java.lang.ProcessImpl.start(ProcessImpl.java:130) at
> java.lang.ProcessBuilder.start(ProcessBuilder.java:1022) ... 24 more
> Container killed by the ApplicationMaster. Container killed on request. Exit
> code is 143 Container exited with a non-zero exit code 143
> {code}
> Hive does a similar trick: HIVE-2372 I have a patch for this, will soon
> submit a patch.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
