Created
February 26, 2017 14:20
-
-
Save mfernest/c9875b395552a19cf10dd3cbdf3a5eaa to your computer and use it in GitHub Desktop.
Script to test wall time of teragen/terasort exercise
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#!/bin/sh | |
## Test for optimal mix of mappers, reducers, and container allocations | |
## Modified from ripped-off script, attributed to BVT and excised Cloudera docs | |
## Paths below correspond to active parcel installation | |
MR=/opt/cloudera/parcels/CDH/lib/hadoop-0.20-mapreduce | |
CDH=/opt/cloudera/parcels/CDH/bin | |
## 51,200,000 records @ 100 bytes/record = 5.12 GB | |
RECORDS=51200000 | |
# Mark start of the loop | |
echo Testing loop started on `date` | |
# Mapper tasks | |
# Set this value to some multiple of worker nodes | |
for maps in 8 | |
do | |
# Reducer tasks | |
# Set this value to some factor of worker nodes | |
for reds in 1 | |
do | |
# Container memory | |
for mem in 512 1024 | |
do | |
# Set JVM heap | |
mheap=`echo "(${mem}*0.8)/1" | bc` | |
rheap=`echo "(${mem}*0.8)/1" | bc` | |
time ${CDH}/hadoop jar ${MR}/hadoop-examples.jar teragen \ | |
-Dmapreduce.job.maps=${maps} \ | |
-Dmapreduce.map.memory.mb=${mem} \ | |
-Dmapreduce.map.java.opts.max.mem=${mheap} \ | |
${RECORDS} /results/tg-10GB-${maps}-${mem} \ | |
1>tera_${maps}_${mem}.out \ | |
2>tera_${maps}_${mem}.err | |
time ${CDH}/hadoop jar ${MR}/hadoop-examples.jar terasort \ | |
-Dmapreduce.job.maps=${maps} \ | |
-Dmapreduce.job.reduces=${reds} \ | |
-Dmapreduce.map.memory.mb=${mem} \ | |
-Dmapreduce.map.java.opts.max.mem=${mheap} \ | |
-Dmapreduce.reduce.memory.mb=${mem} \ | |
-Dmapreduce.reduce.java.opts.max.mem=${rheap} \ | |
/results/tg-10GB-${maps}-${mem} \ | |
/results/ts-10GB-${maps}-${reds}-${mem} \ | |
1>>tera_${maps}_${reds}_${mem}.out \ | |
2>>tera_${maps}_${reds}_${mem}.err | |
$HADOOP/hadoop fs -rm -r -skipTrash /results/tg-10GB-${maps}-${mem} | |
$HADOOP/hadoop fs -rm -r -skipTrash /results/ts-10GB-${maps}-${red}-${mem} | |
done | |
done | |
done | |
echo Testing loop ended on `date` |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment