Common setup for all HPC nodes

Home > Ubuntu > Ubuntu HPC setup with slurm and linux containers > Common setup for all HPC nodes

Install Ubuntu 22.04

Install Ubuntu 22.04 Server on all nodes with most default values.

Create a non-root user (e.g., admin) during setup.

Login Using the Admin User

Login to each node using the admin user.

Configure History Retention

Enable storing date and time along with each command in history, as explained in the guide:

Storing date / time along with commands in history

Install Essential Packages

 sudo su - apt update apt -y install openssh-server vim htop stress munge 

Set Root Password

 passwd 

Enable Root SSH Access

Edit the SSH configuration file:

 vim /etc/ssh/sshd_config 

Locate and change the following line:

 PermitRootLogin yes 

Restart SSH Service

 systemctl restart sshd 

Stop and disable OS firewall services

systemctl stop ufw
systemctl disable ufw

Add IP Address and Hostname Mapping

On all nodes, including containers and VM/bare-metal systems, add the IP address and hostname mapping entries into the `/etc/hosts` file.

Also, make sure to comment out the default hostname entry like this:

#127.0.1.1    <hostname>

Example

#127.0.1.1    infra
192.168.2.5   infra.local       infra
192.168.2.3   node2.local       node2
192.168.2.4   node1.local       node1
192.168.2.6   slurm-login.local      slurm-login
192.168.2.7   slurm-db.local         slurm-dbsrv
192.168.2.8   slurm-master.local     slurm-master
192.168.2.9   slurm-ldap.local       slurm-ldapsrv

Install Environment Modules

 apt install -y environment-modules 

Re-login for Modules to Work

After installation, logout and login again into each node to ensure module command works.

Configure Module Path

Edit the module path configuration file:

 vim /etc/environment-modules/modulespath 

Add the following line:

 /export/modules 

Home > Ubuntu > Ubuntu HPC setup with slurm and linux containers > Common setup for all HPC nodes