Pieces OS — full database snapshots every ~30 min, no retention cap — quietly filled my disk to zero free space

[nexus-infinity]

Hey team — first off, when Pieces is humming, it's honestly one of
the more useful pieces of my daily toolkit, and I want that on
record before I get into the gripe, because the good outweighs the
friction when it's working right.

I've been running it for years now, through a few rough patches
with sign-in and stability, and we always found our way back to it
being worth keeping installed. This one's worth flagging though
because it's a quiet failure — it doesn't throw an error, it just
keeps eating.

Here's the structural piece of it: ~/Library/com.pieces.os/production/Backups/Snapshots/
is taking a full snapshot of the local database roughly every 30
minutes, and nothing downstream of that is pruning the older ones.
I caught two from the same day, 30 minutes apart, both ~5.1GB:

snapshot_20260622_113006_291 — 5.1GB
snapshot_20260622_120009_880 — 5.1GB

For reference, the live dataset in production/Pieces/ is about
5.2GB — so every cycle is duplicating the whole thing, with no cap,
no rotation, nothing retiring the superseded copy. End of that
road, for me, was a full ENOSPC — zero bytes free on the boot
volume, which took out my dev environment until I tracked it down.

Versions: Pieces OS 12.3.10, CLI 1.20.1, macOS/Apple Silicon,
installed via Homebrew.

What I'd ask for — not as a complaint, more a "where's the lever"
question — is whether there's already a retention setting I've
missed, or if this needs a max-snapshot-count / max-total-size cap
added on your end. Either way works for me, I just don't want to be
the guy who finds out the hard way again in three months.

Workaround for now: deleted the older snapshot, kept the newest as
a safety net.

Appreciate you building something I keep coming back to — just
want it to keep being safe to leave running in the background.

[mark-at-pieces]

Hey @nexus-infinity — first off, thank you for being a long-time Pieces user and for the kind words. Really appreciate the thoughtful write-up here too.

So a couple of things to clarify on the backup behavior:

There is a retention cap — we keep a maximum of 2 snapshot backups at any given time. Older snapshots beyond that limit are automatically pruned. So what you're seeing (two snapshots co-existing) is actually the system working as intended — it's just that with a ~5.2GB live database, those 2 snapshots add up to ~10GB of backup space, which is understandably a lot. If you are seeing more than 2, then there is likely a bug that will require a little bit of digging.

When does it trigger? — A backup is created each time PiecesOS is closed and then reopened. It's not a 30-minute timer per se; if you happened to restart PiecesOS twice within 30 minutes, that would produce two snapshots (and the oldest beyond 2 would be cleaned up). This is a safeguard we have in place so that if anything happens to your data — corruption, crash, etc. — we can recover from the most recent known-good state.

Is there a user-facing lever? — Not currently. There isn't a setting to adjust the retention count or disable backups. We hear you that for users with larger databases, even 2 full copies can eat significant disk space, and that's valid feedback we'll take back to the team.

Your workaround is solid — keeping the newest and deleting the older one is exactly what the system does automatically. You're safe to leave it as-is going forward; it should never exceed 2 snapshots.

If you'd like to dig deeper into your setup or chat through any of this, feel free to grab a time on our support calendar — would love to hear more about your workflow and how Pieces fits into it.

Thanks again for flagging this!