hadoop 1.1.2 installation on windows

Question

I am trying to install hadoop 1.1.2 on window machine with cygwin.
From on-line videos and tutorials, I have set up all most everything.
Now problem is when i try to create folder with commands

cd /usr/local/hadoop-1.1.2/bin         --> this works proper and set proper path then
./hadoop dfs -mkdir input      --> when this execute i get error

in error i get issue related to JAVA_HOME path is not set proper also show me text like /Java/jre7/bin/bin/java which looks wrong path.

but i have set JAVA_HOME path properly its here,

I have set same path with /bin in path variable.
I don't know where i have made mistake.

EDIT
full Error

./hadoop: line 320: C:/Java/jre7/bin/bin/java: No such file or directory
./hadoop: line 390: C:/Java/jre7/bin/bin/java: No such file or directory
./hadoop: line 390: exec: C:/Java/jre7/bin/bin/java: cannot execute: No such file or directory

Answer 1

problems with its solution

1. JAVA PATH ISSUE

first is JAVA_PATH issue

Note: Have to use JDK not JRE.
For Hadoop try to use folder name without space

In Environment variable

JAVA_HOME = C:\Java\jdk1.7.0_25

In path variable add below code with othres with ;sepration

%JAVA_HOME%\bin

In hadoop-env.sh file (you can find this file from C:\cygwin\usr\local\hadoop-1.1.2\conf if you are using windows machine).

Note that, remove # from starting of line and USE \\ twice in file

export JAVA_HOME=C:\\Java\\jdk1.7.0_25

if everything OK with JAVA_PATH you can check from CYGWIN consol
try below code to get javaPath which hadoop will use

echo $JAVA_HOME

here you will get java version path
also can set JAVAPATH from runtime,try below code on CYGWIN treminal

export JAVA_HOME=C:/JAVA/Jdk1.7.0_25

2. USER ISSUE

first of all Note that when you start with hadoop installation use same USER for MASTER and SALVE.

If you have different user then have to generate one extra file with name config(without extension)

if your MASTER's machine name is jubin-pc and username is jubinp and SLAVES machine name is trainees11 and username is trainees(have to do vice-verse for both)

config file(for MASTER) Location C:\cygwin\home\jubinp\.ssh\

Host trainees11
User trainees

config file(for SLAVE) Location C:\cygwin\home\trainees\.ssh\

Host jubin-pc
User jubinp

Answer 2

Solution for hadoop-2.6.0 and early:

1. Be sure that path to JDK does not contain backspaces.
   (my variant C:\Java\jdk1.8.0_25)
2. Add JAVA_HOME to path
   • My Computer -> Properties -> Advanced -> Environment Variables -> Create
     JAVA_HOME
     C:\Java\jdk1.8.0_25
   • Add ;%JAVA_HOME%\bin to system Path
3. Open hadoop-env.sh
   (It located in C:\hadoop-2.6.0\etc\hadoop for my hadoop-2.6.0)
   and add line export JAVA_HOME=C:/Java/Jdk1.8.0_25

Answer 3

Quit Cygwin.

Your path to the bin folder of JAVA contains is in another folder named bin ? I don't think so.

Install a JDK (not a JRE) properly in a Path with no blanks. For example : C:\jdk1.7.0_21

In Windows : Add an environnement variable JAVA_HOME to C:\jdk1.7.0_21 Then, add the JAVA_HOME/bin to your PATH.

Edit hadoop/conf/hadoop-env.sh : Uncomment the JAVA_HOME Export. For my example :

export JAVA_HOME=/cygdrive/c/jdk1.7.0_21/