Changes between Initial Version and Version 1 of III120211/Lab10


Ignore:
Timestamp:
Feb 11, 2012, 11:34:05 PM (13 years ago)
Author:
jazz
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • III120211/Lab10

    v1 v1  
     1◢ <[wiki:III120211/Lab9 實作九]> | <[wiki:III120211 回課程大綱]> ▲ | <[wiki:III120211/Lab11 實作十一]> ◣
     2
     3= 實作十 Lab10 =
     4
     5{{{
     6#!html
     7<div style="text-align: center;"><big style="font-weight: bold;"><big>HDFS、MapReduce 與 Hadoop Streaming 觀念驗證<br/>Running Hadoop Streaming with HDFS and MapReduce</big></big></div>
     8}}}
     9
     10== 準備輸入資料集 Input Dataset ==
     11
     12 * 首先準備輸入,包含兩部份:(1) 供 Velvet 計算的 *.fa 檔案,這裡為了方便示範起見,採用 test_long.fa 當範本,並複製 99 份不同檔名,當輸入 (2) 供 Mapper 運算的輸入檔案(內含 HDFS 的檔名路徑)
     13{{{
     14~$ cp /usr/share/doc/velvet-example/examples/data/test_long.fa.gz .
     15~$ gunzip test_long.fa.gz
     16~$ for ((i=1;i<100;i++)); do hadoop fs -put test_long.fa sample-$i.fa; done
     17~$ for ((i=1;i<20;i++)); do echo /user/$(whoami)/sample-$i.fa; done > sample-01.txt
     18~$ for ((i=20;i<40;i++)); do echo /user/$(whoami)/sample-$i.fa; done > sample-02.txt
     19~$ for ((i=40;i<60;i++)); do echo /user/$(whoami)/sample-$i.fa; done > sample-03.txt
     20~$ for ((i=60;i<80;i++)); do echo /user/$(whoami)/sample-$i.fa; done > sample-04.txt
     21~$ for ((i=80;i<100;i++)); do echo /user/$(whoami)/sample-$i.fa; done > sample-05.txt
     22~$ hadoop fs -mkdir lab10_input
     23~$ hadoop fs -put sample-0* lab10_input
     24}}}
     25 * 檢查輸入檔案
     26{{{
     27~$ hadoop fs -ls
     28~$ hadoop fs -ls lab10_input
     29}}}
     30
     31== 觀察 Hadoop Streaming 執行身份與工作目錄 ==
     32
     33 * 撰寫 testmapper.sh
     34{{{
     35#!sh
     36#!/bin/bash
     37
     38id="h998"
     39mkdir -p /tmp/$id
     40host=`hostname`
     41pwd=`pwd`
     42uid=`whoami`
     43
     44while read line; do
     45  input=$line
     46  filename=`basename $input`
     47  echo "$uid@$host:$pwd> hadoop fs -get $input /tmp/$id/$filename"
     48  echo "$uid@$host:$pwd> velveth output-$filename 17 -fasta -short /tmp/$id/$filename"
     49  echo "$uid@$host:$pwd> hadoop fs -put output-$filename ."
     50done
     51rm -rf /tmp/$id
     52}}}
     53 * 接著,讓我們在本地端先驗證一下 testmapper.sh 的運作
     54{{{
     55~$ head -n 2 sample-01.txt > sample-00.txt
     56~$ cat > testmapper.sh << EOF
     57#!/bin/bash
     58
     59id="`whoami`"
     60mkdir -p /tmp/\$id
     61host=\`hostname\`
     62pwd=\`pwd\`
     63uid=\`whoami\`
     64
     65while read line; do
     66  input=\$line
     67  filename=\`basename \$input\`
     68  echo "\$uid@\$host:\$pwd> hadoop fs -get \$input /tmp/\$id/\$filename"
     69  echo "\$uid@\$host:\$pwd> velveth output-\$filename 17 -fasta -short /tmp/\$id/\$filename"
     70  echo "\$uid@\$host:\$pwd> hadoop fs -put output-\$filename ."
     71done
     72rm -rf /tmp/\$id
     73EOF
     74~$ chmod a+x testmapper.sh
     75~$ cat sample-00.txt | ./testmapper.sh
     76h998@hadoop:/home/h998> hadoop fs -get /user/h998/sample-1.fa /tmp/h998/sample-1.fa
     77h998@hadoop:/home/h998> velveth output-sample-1.fa 17 -fasta -short /tmp/h998/sample-1.fa
     78h998@hadoop:/home/h998> hadoop fs -put output-sample-1.fa .
     79h998@hadoop:/home/h998> hadoop fs -get /user/h998/sample-2.fa /tmp/h998/sample-2.fa
     80}}}
     81 * 讓我們用 Hadoop Streaming 的方式來執行 testmapper.sh
     82{{{
     83~$ hadoop jar hadoop-streaming.jar -input lab10_input -output lab10_out1 -mapper testmapper.sh -file testmapper.sh
     84}}}
     85 * 觀察 lab10_out1 的結果,看與本機執行有何不同呢?
     86{{{
     87~$ hadoop fs -cat /user/$(whoami)/lab10_out1/part-00000 | head
     88hadoop@hadoop104:/var/lib/hadoop/cache/hadoop/mapred/local/taskTracker/jobcache/job_201106041247_1820/attempt_201106041247_1820_m_000002_0/work> hadoop fs -get /user/h998/sample-60.fa /tmp/h998/sample-60.fa 
     89hadoop@hadoop104:/var/lib/hadoop/cache/hadoop/mapred/local/taskTracker/jobcache/job_201106041247_1820/attempt_201106041247_1820_m_000002_0/work> hadoop fs -get /user/h998/sample-61.fa /tmp/h998/sample-61.fa 
     90hadoop@hadoop104:/var/lib/hadoop/cache/hadoop/mapred/local/taskTracker/jobcache/job_201106041247_1820/attempt_201106041247_1820_m_000002_0/work> hadoop fs -get /user/h998/sample-62.fa /tmp/h998/sample-62.fa 
     91}}}
     92
     93== 實作透過 Hadoop Streaming 執行 99 組 velvet 運算 ==
     94
     95 * 撰寫 mapper.sh
     96{{{
     97#!sh
     98#!/bin/bash
     99
     100id="h998"
     101mkdir -p /tmp/$id
     102host=`hostname`
     103pwd=`pwd`
     104uid=`whoami`
     105
     106while read line; do
     107  input=$line
     108  filename=`basename $input`
     109  echo "$uid@$host> hadoop fs -get $input /tmp/$id/$filename"
     110  hadoop fs -get $input /tmp/$id/$filename
     111  echo "$uid@$host> velveth output-$filename 17 -fasta -short /tmp/$id/$filename"
     112  velveth output-$filename 17 -fasta -short /tmp/$id/$filename
     113  echo "$uid@$host> hadoop fs -put output-$filename /user/$id/."
     114  hadoop fs -put output-$filename /user/$id/.
     115  hadoop fs -chown $id /user/$id/output-$filename
     116done
     117rm -rf /tmp/$id
     118}}}
     119 * 於本機測試 mapper.sh
     120{{{
     121~$ cat > mapper.sh << EOF
     122#!/bin/bash
     123
     124id="`whoami`"
     125mkdir -p /tmp/\$id
     126host=\`hostname\`
     127pwd=\`pwd\`
     128uid=\`whoami\`
     129
     130while read line; do
     131  input=\$line
     132  filename=\`basename \$input\`
     133  echo "\$uid@\$host> hadoop fs -get \$input /tmp/\$id/\$filename"
     134  hadoop fs -get \$input /tmp/\$id/\$filename
     135  echo "\$uid@\$host> velveth output-\$filename 17 -fasta -short /tmp/\$id/\$filename"
     136  velveth output-\$filename 17 -fasta -short /tmp/\$id/\$filename
     137  echo "\$uid@\$host> hadoop fs -put output-\$filename /user/\$id/."
     138  hadoop fs -put output-\$filename /user/\$id/.
     139  hadoop fs -chown \$id /user/\$id/output-\$filename
     140done
     141rm -rf /tmp/\$id
     142EOF
     143~$ chmod a+x mapper.sh
     144~$ cat sample-00.txt | ./mapper.sh
     145~$ hadoop fs -rmr output-*
     146~$ rm -rf output-sample-*
     147}}}
     148 * 接著用 hadoop streaming 來執行
     149{{{
     150~$ hadoop jar hadoop-streaming.jar -input lab10_input -output lab10_out2 -mapper mapper.sh -file mapper.sh
     151}}}
     152