Advent of Code 2022 – Day 2

In this video, we look into Ceph erasure coding. This is a feature where you can use more of the available space in your cluster without losing too much redundancy. It’s more complex and gives you the option to have more or less safety.

We look at configuring and running training in the cloud.

We talk about creating directories with the right permissions, downloading source files and verifying the validity of the downloaded files.

We look into high availability using Heartbeat and DRBD. Heartbeat is a software to monitor and failover when a machine is unresponsive and DRBD can help you replicate data between machines making your setup highly available. This solution can help you doing maintenance work without downtime.

We use JCuda to run some workloads on the GPU and see the difference in speed.

I show off an open-source project I started writing a RabbitMQ executor in Java. This little tool is something you can deploy on a machine in order to execute workloads from a RabbitMQ queue. The executor will run the commands defined by a file and return the result to another queue.