Rebuilding partition tables in production

An exciting story about how we managed to save customer data at SourceLair, when our disks failed because they ran out of available inodes.
Wed, 28 Feb 2024 | DevOops Berlin
Thumbnail for video 9FM7issQa_4

In this talk, Antonis Kalipetis shares an interesting DevOps story from SourceLair — our first company and an online IDE - that turned into a DevOops moment. It all started when JavaScript's node_modules caused storage issues, leading to broken disks and backup failures. The recovery process to save customer data involved partition table fixes, block-level disk copies and hard lessons on backup strategies.

Spoiler: It was all because of excessive node modules by our users!

Build your next project with LOGIC

You will care about the "what", we will take care of the "how".