diff --git a/README.md b/README.md
index 442f641243f46ef81e4a9de24e703ba220f09154..bae8815ac1ecccf1b8ab1f313dc16847336cac5a 100644
--- a/README.md
+++ b/README.md
@@ -82,11 +82,15 @@ Then run the script [ansible-launch.sh](ansible-launch.sh) :
 ## Deploy cluster HDFS with multiple VMs
 1. Set your nodes' IP address in [VagrantFile](vagrant/cluster/Vagrantfile)
 2. Declare those IP for ansible provision in [vars](playbook/roles/hosts-file/vars/main.yml)
-3. in cli : start your multiple VM from this [directory : vagrant/cluster](vagrant/cluster) :
+3. Configure your own computer to access to your nodes using their hostname
+	```shell
+	vim /etc/hosts
+	```	
+4. in cli : start your multiple VM from this [directory : vagrant/cluster](vagrant/cluster) :
 	```shell
 	vagrant up
 	```
-4. Format HDFS :
+5. Format HDFS :
  	* ssh on namenode
  	* in cli : as user hadoop : change directory & format HDFS
  	```shell
@@ -94,16 +98,16 @@ Then run the script [ansible-launch.sh](ansible-launch.sh) :
  	cd /usr/local/hadoop/bin/
  	hdfs namenode -format
  	```
- 5. Start HDFS deamon on your cluser
+6. Start HDFS deamon on your cluser
   	* ssh on namenode
  	* in cli : as root : start service hadoop
  	 ```shell
  	sudo systemctl start hadoop
  	```
  	* **WORK In Progress** : systemd will tell you something wrong happens but cluster is working anyway.
- 6. Verify your cluster is up: 
+7. Verify your cluster is up: 
  	* on your own device, use a webbrowser
  	* go on [IP-of-your-namenode]:9870
  	if default : http://10.0.0.10:9870
 ## Deploy cluster HDFS on servers
-work in progress
\ No newline at end of file
+work in progress