SystemRequires admin/root

Pacman Package Cache

Pacman stores downloaded package archives in /var/cache/pacman/pkg so reinstalls, downgrades, and repeated installs do not need to fetch the same files again. That cache consists mainly of package tar archives such as .pkg.tar.zst files and partially downloaded package files left behind by interrupted transactions; over time it can grow with superseded versions that pacman no longer needs for normal upgrades. Kudu removes unneeded cached package archives and incomplete downloads from this root-owned cache without touching installed packages, system settings, accounts, or personal files.

Why clean Pacman Package Cache?

  • Old package archives accumulate after every system upgrade, so /var can quietly consume multiple gigabytes and you notice low disk space warnings or failed updates due to insufficient space
  • Interrupted or canceled downloads leave partial package files in the cache, which can confuse later installs and show up as repeated re-downloads or transaction errors for the same package
  • Keeping many superseded versions of large kernels, browsers, and desktop packages bloats backup size and makes filesystem snapshots noticeably larger and slower to create
  • A full pacman cache can block new package transactions when there is not enough room to download and unpack updates, leading to 'no space left on device' during routine system maintenance
  • On small root partitions, cached package tarballs compete with logs and package database writes, so the system may feel fine until an upgrade suddenly fails midway because /var filled up
  • Stale cached copies are not used forever because pacman verifies package integrity, but retaining obsolete archives still wastes disk and makes it harder to see which versions are actually worth keeping for rollback
What gets cleaned

Cache paths Kudu targets

Linux

/var/cache/pacman/pkg
Frequently asked

Common questions about Pacman Package Cache

Free & open source

Download Kudu and reclaim your disk space.

Available on Windows, macOS, and Linux. No account required, no feature gates, no telemetry without consent. All cleaning targets are open source and community-auditable.