rm -rf: The Typo That Deleted GitLab's Production Database (ep. 489)
Автор: DatabaseHistory
Загружено: 2026-01-04
Просмотров: 14
In 2017, GitLab suffered one of the most infamous engineering disasters in modern DevOps history. A tired engineer, working late into the night, accidentally deleted 300 gigabytes of live production data with a single command. But the real horror came next: GitLab discovered that their automated backups—believed to be running for months—had silently failed.
Instead of hiding the incident, GitLab did something unheard of. They opened a global livestream and let the entire world watch their engineers scramble to recover the platform in real time. Every command, every mistake, every moment of panic was broadcast publicly. It was radical transparency at its rawest.
In the end, GitLab was saved by an accidental snapshot—an unexpected lifeline that restored the site, though several hours of user data were lost forever. The incident became a defining moment in DevOps culture, proving the value of blame‑free postmortems and the absolute necessity of testing your backups before you need them.
This is the story of the night GitLab almost died… and the lessons every engineer should take from it.
I’m Mr. Ed, and this is DatabaseHistory.
Доступные форматы для скачивания:
Скачать видео mp4
-
Информация по загрузке:
