The modular nature of Pod systems means that parts can be easily replaced. There are also some parts of the IPU-Machine that can be easily replaced in the case of a fault. This guide describes how to perform these maintenance tasks.
These maintenance tasks can generally be completed while the Pod is still powered on, but you should not have any machine learning jobs running on the Pod.
This guide describes:
This guide is for properly trained service personnel and technicians who are required to replace parts in a Pod system or on an IPU-Machine.
In this guide, the term IPU-Machine refers to the blades installed in your Pod, so Bow-2000 in Bow Pod systems or IPU-M2000 in IPU-POD systems.