Stop automatic snapshots from filling the disk
The generic half checks out and the Omarchy half is wrong in a way that makes the fix harmful. Confirmed against the cited Arch wiki Snapper page: the default timeline really is 10 hourly, 10 daily, 10 monthly and 10 yearly, the package really enables neither timer, the retention keys and TIMELINE_CREATE="no" are correct, and man snapper confirms delete number1-number2 range syntax. On Omarchy 4 none of the premise survives. Confirmed on this machine: /etc/snapper/configs/root is owned by no package and is Omarchy's template from /usr/share/omarchy/default/snapper/root, carrying TIMELINE_CREATE="no", NUMBER_CLEANUP="yes" and NUMBER_LIMIT="5". systemctl is-enabled reports snapper-timeline.timer disabled and snapper-cleanup.timer enabled, running hourly, so the cause's claim that the cleanup timer is not enabled automatically is false here. pacman -Q snap-pac reports the package absent, so there is no pre/post pair either. One number snapshot is taken per update by /usr/share/omarchy/bin/omarchy-snapshot, called from omarchy-update. The worst defect is the fix instructing systemctl enable --now snapper-timeline.timer. Upstream's install/config/snapper.sh on quattro explicitly runs systemctl disable --now snapper-timeline.timer, so that step restarts the exact mechanism that caused the reported problem and is reverted on the next repair. The retention-key edit is also futile there, because the same script does install -m 0644 "$template" "$SNAPPER_CONFIG_PATH" and /usr/share/omarchy/migrations/1781984677.sh re-runs it whenever the config is missing or the units drift. The real Omarchy cause and fix were absent: leaked timeline-marked snapshots that number cleanup cannot reap, drained by /usr/share/omarchy/migrations/1784809452.sh in batches of 20 to survive a D-Bus timeout, whose contract is pinned by test/shell.d/snapper-timeline-leak-test.sh in the quattro tree. That migration's own comment puts the cost at hundreds of snapshots pinning over 100 GB. The danger also missed the bootloader entirely: limine-list on this machine lists each snapshot as a boot entry, and /usr/lib/systemd/system/snapper-cleanup.service.d/limine-snapper-override.conf runs limine-snapper-sync --no-force-save afterwards, so deletions do resync and leave no dead entry while limine-snapper-sync.service is active, which it is here. Both cited URLs resolve and support the generic branch, so nothing was removed. Severity high and frequency common are left unchanged: on Omarchy the migration now makes this self-healing, but the record also applies to four distros that still ship the generic default, and upstream shipped a migration precisely because real users hit it. NOT exercised: no sudo, so snapper list, any delete, any cleanup run and the migration itself were all read rather than run, and I could not enumerate this machine's actual snapshots beyond the two boot entries limine-list shows.
Deleting snapshots is permanent and you lose the ability to roll back to those points. Never delete the snapshot you are currently booted into. On Omarchy 4 every snapshot is also a Limine boot entry, listed by limine-list, so deleting the number snapshots removes the pre-update recovery entries that omarchy-snapshot restore offers. limine-snapper-sync.service resyncs the entry list after snapper-cleanup.service, so a deletion leaves no dead boot entry behind while that service is enabled and active. Delete in batches of about 20, because one large delete can fail on a D-Bus timeout partway through and leave the rest untouched. Do not enable snapper-timeline.timer on Omarchy 4 to fix this: it restarts the mechanism that created the problem, and Omarchy disables it again on the next repair. On plain Arch, running a cron daemon alongside the systemd timers produces duplicate snapshots, so enable one mechanism, not both.
Symptom
My root filesystem keeps filling up over weeks even though I have not added files. sudo btrfs filesystem usage / shows most of the disk in use and snapper -c root list shows dozens or hundreds of snapshots going back months.
Cause
On plain Arch, snapper's default timeline keeps 10 hourly, 10 daily, 10 monthly and 10 yearly snapshots per config, and the package enables neither snapper-timeline.timer nor snapper-cleanup.timer, so nothing reaps what accumulates. Each snapshot pins the blocks of every file that has since changed. If snap-pac is installed, every package update adds a pre/post pair on top.
On Omarchy 4 the defaults differ and the mechanism is narrower. Omarchy installs /etc/snapper/configs/root from /usr/share/omarchy/default/snapper/root with TIMELINE_CREATE="no", NUMBER_CLEANUP="yes" and NUMBER_LIMIT="5", disables snapper-timeline.timer, and enables snapper-cleanup.timer and limine-snapper-sync.service. snap-pac is not installed, and one number snapshot is taken per omarchy update by omarchy-snapshot create. A current install is therefore capped at five root snapshots. The accumulation happens on machines installed under Omarchy's earlier defaults, which did take hourly timeline snapshots. The newer config stopped creating them, but the existing ones carry cleanup algorithm timeline, and number cleanup does not touch those because the two algorithms are separate. They pile up untouched, pinning extents that nothing on the machine will ever release.
Fix
See what you have, and which cleanup algorithm each snapshot carries, because that is what decides whether anything will ever reap it:
sudo snapper -c root list
sudo snapper -c root --csvout list --columns number,cleanup
sudo btrfs filesystem usage /
**Omarchy 4.** Do not enable snapper-timeline.timer, and do not edit the retention keys. Omarchy sets TIMELINE_CREATE="no" and NUMBER_LIMIT="5" deliberately, install/config/snapper.sh reinstalls that file from its template, and a migration re-runs that script whenever the config or the units drift. Confirm the config and the units first:
cat /etc/snapper/configs/root
systemctl is-enabled snapper-cleanup.timer limine-snapper-sync.service
systemctl is-enabled snapper-timeline.timer # should report disabled
If the list is long, those are leaked timeline snapshots from an earlier default. omarchy update drains them through a migration that deletes only the timeline ones:
omarchy update
omarchy update refuses to start with less than 10 GiB free on /, so if the disk is already that full, drain by hand in batches of about 20. One large delete can die on a D-Bus timeout partway:
sudo snapper -c root --csvout list --columns number,cleanup |
awk -F, '$2 == "timeline" { print $1 }' | head -20 |
xargs sudo snapper -c root delete --sync
Repeat until no timeline rows remain. Leave the number snapshots alone: those are the pre-update recovery points, and snapper-cleanup.timer already holds them at five.
**Plain Arch, EndeavourOS, CachyOS, Manjaro.** Delete a range you do not need:
sudo snapper -c root delete --sync 20-140
Then tighten retention in /etc/snapper/configs/root:
TIMELINE_MIN_AGE="1800"
TIMELINE_LIMIT_HOURLY="5"
TIMELINE_LIMIT_DAILY="7"
TIMELINE_LIMIT_WEEKLY="0"
TIMELINE_LIMIT_MONTHLY="0"
TIMELINE_LIMIT_YEARLY="0"
The package enables neither timer, so enable the ones that create and reap snapshots:
sudo systemctl enable --now snapper-timeline.timer snapper-cleanup.timer
systemctl list-timers 'snapper*'
If you do not want timeline snapshots at all, set TIMELINE_CREATE="no" in the same file and enable only snapper-cleanup.timer.
Verify
On Omarchy 4, sudo snapper -c root --csvout list --columns number,cleanup shows no timeline rows and at most five number ones. On plain Arch it shows only the number your policy allows once snapper-cleanup.timer has run. In both cases sudo btrfs filesystem usage / shows Used fallen:
sudo snapper -c root --csvout list --columns number,cleanup
sudo btrfs filesystem usage /
Btrfs frees the extents of a deleted snapshot in the background, so the space appears a little after the delete returns unless you passed --sync.