TIL: Snaps Are Secretly Eating Your Disk Space (Here’s How to Fix It) 🐧

On 8 May, a Linux developer published something that felt less like a discovery and more like a reckoning. Ubuntu's Snap packaging system, they explained, had been doing something entirely by design that most users had never consciously agreed to: keeping old versions of packages around after updates, quietly, persistently, in ways that add up to real gigabytes over real time. The mechanism is almost logical once you see it. When a Snap updates, the previous version isn't removed — it's retained as a rollback point. Then the version before that. The default retention policy keeps two old revisions per package, which means that on a system running a dozen or so Snaps, you can accumulate a surprising volume of disk you never asked to spend. The fix, once you know the shape of the problem, is a single command: `snap set system snapshots.automatic.retention=2` adjusted alongside `snap list --all` to identify stale revisions, followed by `snap remove --revision` for each one. Unglamorous. Effective. The kind of thing that shouldn't require a Dev.to post to surface.

If confirmed at scale, here is what this means. For developers running Ubuntu or Ubuntu-derived distributions — which describes a substantial portion of the Linux desktop and server population — this is a quiet tax on storage that compounds with every update cycle. On a system with limited disk, perhaps a VPS or an older laptop, the accumulation isn't a minor inconvenience; it's the kind of thing that causes cryptic failures when a build runs out of space at the worst possible moment. The broader implication cuts toward Canonical's packaging philosophy: Snap was designed to solve dependency hell and improve security isolation, and it does both reasonably well, but the tradeoff in disk overhead has never been prominently communicated. Users who trusted the system to manage itself were not wrong to do so — they simply weren't told what "managing itself" involved. That gap between design intent and user expectation is where most of the frustration lives.

Watch for Canonical's own documentation to be updated or community pressure to shift the default retention policy from two revisions to one — either would signal that this quiet cost has finally become loud enough to act on.

NewsHive monitors these sources continuously. All signal titles above link to the original reporting.

Intelligence by NewsHive. Need help navigating what this means for your business? Contact GeekyBee →