Grokbase Groups Pig user May 2009
FAQ
You are running in map reduce mode, but you are not attaching to your
hadoop cluster. It's running it locally. That's what the "Connecting
to hadoop file system at file:///" means. If you were connecting to a
cluster it would saying "hdfs://yournamenode" instead of "file:///"
Is the directory containing your hadoop-site.xml in your classpath
when executing the pig command? See http://hadoop.apache.org/pig/docs/r0.2.0/tutorial.html
, the section "Running the Pig Scripts in Hadoop Mode".

Alan.
On Jun 9, 2009, at 11:03 PM, George Pang wrote:

I think it's not in mapreduce mode. Because I also found the error,
again:
INFO executionengine.HExecutionEngine: Connecting to hadoop file
system at:
file:///
09/06/09 21:18:00 INFO jvm.JvmMetrics: Initializing JVM Metrics with
processName=JobTracker, sessionId=

George

2009/6/9 George Pang <p0941p@gmail.com>
Now I can run id.hadoop(from the official tutorial
http://hadoop.apache.org/pig/docs/r0.2.0/quickstart.html) as an
embedded
Java program, and I can get the result from HDFS. But one line of
the
console message before the "Success! " reads:
WARN mapReduceLayer.MapReduceLauncher: Jobs not found in the
JobClient.
Please try to use Local, Hadoop Distributed or Hadoop MiniCluster
modes
instead of Hadoop LocalExecution

What does it mean or does it matter? Am my program running in map-
reduce
mode at all? Thanks for any idea!

George


2009/6/3 George Pang <p0941p@gmail.com>

Hi Ankur,
Everything runs in the command line, the error only happens when I
use
Eclipse. My Eclipse version is 4.3.2. When I run the embedded
java
program, it gave me the error
"INFO executionengine.HExecutionEngine: Connecting to hadoop file
system
at: file:/// INFO jvm.JvmMetrics: Initializing JVM Metrics with
processName=JobTracker, sessionId="

The environment variables are set, is it something to do with
where the
data files are put? Thank you.

George

2009/6/3 Ankur Goel <gankur@yahoo-inc.com>

Make sure you have the following parameters set:-
PIGDIR=your/pig/dir

# you will need to set this, else pig assumes the version to be 17
# and may not be able to find/connect your namenode/jobtracker
PIG_HADOOP_VERSION=18

HADOOPDIR=your/hadoop/dir/conf

PIG_CLASSPATH=$PIGDIR/pig.jar:$HADOOPDIR

Also make sure you have yahoo specific lines at the bottom
commented out
in pig.properties
under PIGDIR/conf.

-Ankur

----- Original Message -----
From: "George Pang" <p0941p@gmail.com>
To: pig-user@hadoop.apache.org
Sent: Wednesday, June 3, 2009 6:05:22 AM GMT +05:30 Chennai,
Kolkata,
Mumbai, New Delhi
Subject: Re: Error on running pig-embedded Java code

Any one trying to answer this one?
Thanks

George

2009/5/30 George Pang <p0941p@gmail.com>
Dear users,

I compiled and ran the pig-embedded Java code from the "pig quick start"
example on Eclipse. I got the following error:

INFO executionengine.HExecutionEngine: Connecting to hadoop file
system at:
file:///
INFO jvm.JvmMetrics: Initializing JVM Metrics with
processName=JobTracker,
sessionId=

Obviously it can't find the HDFS or Hadoop. But I have set the
PIG_CLASSPATH as
/usr/share/cloudera/apps/pig-0.1.1/pig-0.1.1-core.jar:/usr/share/
cloudera/apps/hadoop-0.18.3-patched/conf
and other environments under Run Configurations / Environment
Is there anything I forgot to do? Any idea is much appreciated!

Pig: 0.1.1
Hadoop: 0.18.3
Eclipse: 3.4.2

George

Search Discussions

Discussion Posts

Previous

Follow ups

Related Discussions

Discussion Navigation
viewthread | post
posts ‹ prev | 10 of 13 | next ›
Discussion Overview
groupuser @
categoriespig, hadoop
postedMay 31, '09 at 2:44a
activeJun 11, '09 at 12:56a
posts13
users4
websitepig.apache.org

People

Translate

site design / logo © 2021 Grokbase