Problems running GridMix2

[ps...@gmail.com]

Hi,

1 - I'm trying to run GridMix2 (rungridmix_2) in a cluster, but it happens  
nothing. A Job isn't created. It simple appears the message:

[code]
GridMix results:
Total num of Jobs: 0
ExecutionTime: 0
[/code]

Is there a way to know what is happenning with gridmix?


Here's the gridmix-env-2 config file:
[code]
#!/usr/bin/env bash

## Environment configuration
# Hadoop installation
export HADOOP_VERSION=hadoop-0.20.1
HADOOP_INSTALL_HOME=/opt
export HADOOP_HOME=${HADOOP_INSTALL_HOME}/hadoop
#export HADOOP_CONF_DIR=${HADOOP_HOME}/src/test/hadoop-site.xml (I also  
tried with this option)
#export HADOOP_CONF_DIR=${HADOOP_HOME}/conf/mapred-site.xml (I also tried  
with this option)
#export USE_REAL_DATASET=TRUE

export APP_JAR=${HADOOP_HOME}/${HADOOP_VERSION}-test.jar
export EXAMPLE_JAR=${HADOOP_HOME}/hadoop-0.20.2-dev-examples.jar
export  
STREAMING_JAR=${HADOOP_HOME}/contrib/streaming/${HADOOP_VERSION}-streaming.jar

[code]

2 - Running this example in my local machine, sometimes it runs, sometimes  
it don't. When the GridMix2 runs, I never saw it ending, but maybe because  
it's too many jobs to execute on my poor little PC. :( Am I right with this  
deduction?



Thanks,
Pedro

Re: Problems running GridMix2

[Pedro Costa <ps...@gmail.com>]

The command that I launch to run GridMix2 is :
$>./rungridmix_2
Here's the content of the file:
[code]
#!/usr/bin/env bash

## Environment configuration

GRID_DIR=$(dirname "$0")
GRID_DIR=$(cd "$GRID_DIR"; pwd)
source $GRID_DIR/gridmix-env-2

Date=$(date +%F-%H-%M-%S-%N)
echo $Date >  $1_start.out

export HADOOP_CLASSPATH=${APP_JAR}:${EXAMPLE_JAR}:${STREAMING_JAR}
export LIBJARS=${APP_JAR},${EXAMPLE_JAR},${STREAMING_JAR}
GRIDMIX2_JAR=${HADOOP_HOME}/src/benchmarks/gridmix2/build/gridmix.jar
set -x
${HADOOP_HOME}/bin/hadoop jar $GRIDMIX2_JAR
org.apache.hadoop.mapred.GridMixRunner -libjars ${LIBJARS}
set +x

Date=$(date +%F-%H-%M-%S-%N)
echo $Date >  $1_end.out
[/code]

Before running GridMix2, I executed successfully the generateGridmix2data.sh.







On Tue, Nov 30, 2010 at 4:05 PM,  <ps...@gmail.com> wrote:
> Hi,
>
> 1 - I'm trying to run GridMix2 (rungridmix_2) in a cluster, but it happens
> nothing. A Job isn't created. It simple appears the message:
>
> [code]
> GridMix results:
> Total num of Jobs: 0
> ExecutionTime: 0
> [/code]
>
> Is there a way to know what is happenning with gridmix?
>
>
> Here's the gridmix-env-2 config file:
> [code]
> #!/usr/bin/env bash
>
> ## Environment configuration
> # Hadoop installation
> export HADOOP_VERSION=hadoop-0.20.1
> HADOOP_INSTALL_HOME=/opt
> export HADOOP_HOME=${HADOOP_INSTALL_HOME}/hadoop
> #export HADOOP_CONF_DIR=${HADOOP_HOME}/src/test/hadoop-site.xml (I also
> tried with this option)
> #export HADOOP_CONF_DIR=${HADOOP_HOME}/conf/mapred-site.xml (I also tried
> with this option)
> #export USE_REAL_DATASET=TRUE
>
> export APP_JAR=${HADOOP_HOME}/${HADOOP_VERSION}-test.jar
> export EXAMPLE_JAR=${HADOOP_HOME}/hadoop-0.20.2-dev-examples.jar
> export
> STREAMING_JAR=${HADOOP_HOME}/contrib/streaming/${HADOOP_VERSION}-streaming.jar
>
> [code]
>
> 2 - Running this example in my local machine, sometimes it runs, sometimes
> it don't. When the GridMix2 runs, I never saw it ending, but maybe because
> it's too many jobs to execute on my poor little PC. :( Am I right with this
> deduction?
>
>
>
> Thanks,
> Pedro



-- 
Pedro