Contributed Nodes at HPC

Information about contributed nodes at HPC IIT Mandi

The contributor/school who purchase a node will get high preferential access to their nodes. In case these nodes are idle this can be used by any user (this will be configured in the job scheduler). The nodes will be identical to that of other nodes of the cluster in features like OS and the jobs are submitted by the scheduler. The advantage of pooling the resources is that you might be able to buy a node in much cheaper prices than available under rate contracts or individual purchases. These nodes are maintained by single administration that is under single contract between vendor and IIT Mandi with 3 year on site warranty. If any problems occurs in any of nodes this will be reported to the vendor directly by staff in charge of the HPC at IIT Mandi. And there is also advantage of saving space since nodes occupy a small space and accessed in a central location. In addition nodes purchased from the projects, school funds etc. will be accessible to the computing purpose of all users when they are idle.

Node Contribution Policy at HPC IIT Mandi

• Request through HPC email: To contribute a node to the HPC cluster at IIT Mandi, the owner should submit a request through email to it_helpdesk@iitmandi.ac.in. The request should include information about the hardware specifications of the node, compatibility with HPC usage.
• Academic/Research and optimal time utilization: The contributed node should be used for research/academic/teaching purposes and should be made available for use by other users when not in use by the owner.
• Maintenance by HPC team: The HPC team at IIT Mandi should be responsible for the maintenance of the contributed node. The team should ensure that the node is functioning properly and address any issues that may arise. The maintenance should be carried out till the end of the node’s life, based on the repair costing and other factors.
• Integration of node after owner leaves the institute: If the owner of the contributed node leaves the institute, the node will be owned by HPC. The HPC team should ensure that the node is used effectively and efficiently, and prioritize research projects that require high computing power.
• Regular monitoring and review: The HPC team should regularly monitor the contributed node to ensure its proper functioning and to address any issues that may arise. The team should also periodically review the contributed node policy to ensure its effectiveness and make any necessary changes.
• Updating and Operating system: It is important to have the same OS in all machines in the cluster. So the same OS is required in the node. This ensures compatibility and reduces potential issues that could arise from running different operating systems. After the same operating system is installed on all the machines, it needs to be integrated with the cluster. This means configuring the cluster to communicate effectively and efficiently between all machines. Regular updates to the operating system are crucial to maintain its security and performance. Periodic updates ensure that any vulnerabilities or bugs are fixed, and the cluster continues to function optimally. Overall, keeping the operating system up-to-date is essential for ensuring the smooth and secure operation of the cluster.
• Software’s and Databases in the nodes: All the software and databases should be available on every node in the cluster. This means that all nodes in the cluster will be equivalent in terms of software and database access.
• No return policy: Once the node is contributed to the cluster, it will not be returned.