how to avoid deadlocks when reading from a process
Sometimes we need to run a child process and read its output: it could be ffmpeg to convert a video to another format or apply a watermark, or run pgdump to backup a database. When the process we are running is noisy, it often can lead to unexpected deadlocks, which is the point of this post.
helping librarians to preserve all human knowledge
While reading “Service Model”, a book by Adrian Tchaikovsky, I was first fascinated by the librarians, and then later awestruck by the elegance and silliness of their idea. This post will explore a potential implementation of the algorithm, a little bit of exploration of Async in Ruby, and some spoilers for that part of the book. You have been warned.
how not to shoot yourself in the foot
We often use temporary files when building software: to store a report before sending it to a customer or uploading it to AWS S3; to download a CSV to parse into a database; or to store a picture of your beloved cat before sending it to the generative AI to put a silly hat on the unsuspecting feline. Today, I will talk about common mistakes with temporary files.
multi-arch Docker images without immutable tags
Most of the base images in the Docker registry do not offer immutable tags. This means that if you’re building your Docker image based on, for example, an official Ruby image
ruby:3.3.0-slim, it might change overnight, and the machines that have already downloaded this tag will never receive an updated image.
a story of a change, that broke Sidekiq's reliability promises
Now and then you receive a report about something not working as expected. This time it was scarier than usual: a job, killed by Docker after consuming too much memory with an OOM error, was disappearing from the queue without a trace. In production, we deploy
sidekiq-profor its reliability guarantees, specificallysuper_fetchstrategy for pulling work from the queue, and normally should re-queue the job after a restart.