Question

Failed to restart Hadoop namenode using cloudera quickstart

Answer 1

You can use Cloudera Manager to manage all the services in your Cloudera quickstart VM. Cloudera Manager is the admin tool that provides a nice UI for managing the services. You can check the health status, start, stop, or restart any service with just a few clicks.

By default, Cloudera Manager is disabled when you start the Cloudera quickstart VM. To enable it you need to issue the following command on the terminal:

sudo /home/cloudera/cloudera-manager --force --express

It will take a few minutes so as to start the Cloudera manager for you.

Now, to open Cloudera manager you need to visit the URL as shown in the above screenshot.

To start a specific service click the dropdown icon corresponding the service and then click on start in the dropdown menu as shown in the image below: 

I hope it helps.

Answer 2

Fixed my own problem

Check status of all services: for service in /etc/init.d/hadoop-hdfs-*; do $service status; done; , output should be:

Hadoop datanode is running [  OK  ]  
Hadoop namenode is dead and pid file exists [FAILED]  
Hadoop secondarynamenode is running  [  OK  ]

To get the namenode running do the following:

• stop all services: for service in /etc/init.d/hadoop-hdfs-*; do $service stop; done;
• clear cache from cache directory: sudo rm -rf /var/lib/hadoop-hdfs/cache/*
• reformat name node: sudo -u hdfs hdfs namenode -format
• start all services: for service in /etc/init.d/hadoop-hdfs-*; do $service start; done;
• check status: for service in /etc/init.d/hadoop-hdfs-*; do $service status; done;