I am Computer Science Junior from India. I recently took up a project for automating the creation of a virtual PBS cluster. Now its easy to make an install script for one node, how can I extend this process to automatically form a cluster across all the nodes in my lab.
Any guidance will be appreciated.
Thanks for your interests in PBS Pro OSS , please find the below use cases
Use case 1:
- Store the PBS RPM in a common location accessible by all the nodes
- use pdsh to install PBS execution components at once
- use pdsh to copy the pbs.conf and mom_priv/config to all the nodes
- use pdsh to restart all the pbs_mom services
Use case 2:
Create a scripts that copies (scp / ssh) the rpm to the compute nodes and deploys and configures it
Use case 3:
- Use provisioning tools warewulf , rocks and others
Thanks. I’ll go through the above steps