Hadoop: Not able to find the start-dfs.sh start-mapred.sh scripts

Question

I installed it using ppa.

When I reached step 12 : Starting Hadoop Cluster.

I got stuck.

When I do cd ~/tasdik/hadoop/bin/ It says no such file or directory.

Any suggestions?

Which ppa? because ppa:hadoop-ubuntu/stable does not provide 14.10 packages (only 14.04) — Sylvain Pineau, Feb 27 '15 at 15:24

Sylvain Pineau · Accepted Answer · 2015-02-27T16:15:16.750

3

The tutorial step 2 asked to create a new user: hduser.

So if you decided to adopt this username for the Hadoop user, the path of start-all.sh should be:

cd /home/hduser/hadoop/bin/  
start-all.sh

Or if installed system-wide:

/usr/lib/hadoop/bin/

edited Feb 27 '15 at 16:15

answered Feb 27 '15 at 15:09

Sylvain Pineau

If that works, please suggest an edit to the tuturial to fix the paths – Sylvain Pineau Feb 27 '15 at 15:10
I did not create a new user for it. But it's not inside that, or I am looking inside the wrong place maybe – Tasdik Rahman Feb 27 '15 at 15:17
@prodicus you could also try sudo updatedb && locate start-all.sh – Sylvain Pineau Feb 27 '15 at 15:38
Found it. Was inside /usr/lib/hadoop/bin/ – Tasdik Rahman Feb 27 '15 at 15:43

Avinash Raj · Answer 2 · 2015-06-21T04:25:25.440

1

I have found the start-all.sh file on /hadoop-x.x.x/sbin directory.

path/to/hadoop-x.x.x/sbin/start-all.sh

where hadoop-x.x.x is a directory which was created on extracting the downloaded the hadoop.tar.gz file.

If you initiate the file, it would show

$ bash /home/foo/hadoop-2.7.0/sbin/start-all.sh
This script is Deprecated. Instead use start-dfs.sh and start-yarn.sh

So, it's better to invoke, start-dfs.sh located on the same directory itself.

bash /home/foo/hadoop-2.7.0/sbin/start-dfs.sh

edited Jun 21 '15 at 04:25

answered Jun 21 '15 at 04:15

Avinash Raj

2 Answers2