0

While Running the terasort application by modifying the parameters I'm getting the following Error.

15/05/24 21:41:42 ERROR terasort.TeraSort: Input path does not exist: maprfs:/user/user01/–DXmx1024m

Command I'm running for executing the jog

$ hadoop jar /opt/mapr/hadoop/hadoop-0.20.2/hadoop-0.20.2-dev-\
examples.jar terasort –DXmx1024m –Dmapred.reduce.tasks=2 \
-Dio.sort.mb=1 /user/user01/6/TERA_IN /user/user01/6/TERA_OUT_2

If I remove these Parameters –DXmx1024m –Dmapred.reduce.tasks=2 -Dio.sort.mb=1 then everything is working fine.

I think i'm facing the below bug which is posted on jira but its for windows https://issues.apache.org/jira/browse/HADOOP-8536

JHobern
  • 866
  • 1
  • 13
  • 20
smali
  • 4,687
  • 7
  • 38
  • 60

2 Answers2

2

It is just a copy/paste issue from the MapR LAB GUIDE.

Just change '–' to '-' from the VM input parameters.

See the execution with copy/paste:

-bash-4.1$ hadoop jar /opt/mapr/hadoop/hadoop-0.20.2/hadoop-0.20.2-dev-\
> examples.jar terasort –DXmx1024m –Dmapred.reduce.tasks=2 \
> -Dio.sort.mb=1 /user/user01/6/TERA_IN /user/user01/6/TERA_OUT_2
15/10/26 15:24:57 INFO terasort.TeraSort: starting
15/10/26 15:24:59 ERROR terasort.TeraSort: Input path does not exist: maprfs:/user/user01/–DXmx1024m
-bash-4.1$

See the execution with the command line fixed:

-bash-4.1$ hadoop jar /opt/mapr/hadoop/hadoop-0.20.2/hadoop-0.20.2-dev-examples.jar terasort -DXmx1024m -Dmapred.reduce.tasks=2 -Dio.sort.mb=1 /user/user01/6/TERA_IN /user/user01/6/TERA_OUT_2
15/10/26 15:25:26 INFO terasort.TeraSort: starting
15/10/26 15:25:27 INFO input.FileInputFormat: Total input paths to process : 2
....
Rodolfo De Andrade Vasconcelos
  • 36
  • 4
0

Do you run teragen before you run terasort?

And what filesystem do you used? Why "maprfs". If it isn't on purpose, check your config files

15/05/24 21:41:42 ERROR terasort.TeraSort: Input path does not exist: **maprfs**:/user/user01/–DXmx1024m
gwgyk
  • 460
  • 4
  • 13
  • I'm Using Mapr Distribution thats why its using maprfs, and I have executed teragen before executing terasort. and as i mention in my question If i remove the parameters then everything is executing fine. – smali May 25 '15 at 12:24
  • "io.sort.mb" is used to set the map output buffer, default is 100M. But you set it 1M, it's to small. Try to remove this parameter. – gwgyk May 26 '15 at 00:57