OpenZFS and Linux
summary
OpenZFS on Linux now provides native encryption (since 0.8), replaces mdadm/luks/lvm combos, and ZoL is the main upstream contributor after BTRFS lost momentum. Creates pools from vdevs (mirrors, raidz1/2/3), pools are portable between hosts, and snapshots support snapshots via zfs send/receive for incremental offsite backup. Key tunables: ARC uses up to half RAM, L2ARC on SSDs, SLOG intent log for faster sync writes, ~232 kernel knobs and ~75 dataset attributes, compression (lz4 default) helps space; dedup is RAM‑heavy; default recordsize 128K (MySQL tuned to 16K).
AI-written summary · about
chapters (17)
- 0:00OpenZFS and Linux intro
- 1:24Licensing challenges
- 2:52ZFS game-changer for storage
- 5:21ZFS design for admins
- 7:37Copy-on-write architecture
- 9:47ZFS vdevs and pools
- 13:22Datasets, volumes, swap
- 14:55ARC, L2ARC, SLOG explained
- 17:35ZFS commands: zpool zfs
- 19:05Stable disk vdevs
- 20:33RAID-Z levels
- 21:40Snapshots and their importance
- 22:43ZFS send/receive backups
- 24:15ZFS compression dedup
- 26:29Tuning ZFS performance
- 28:41Tuning ZFS for DBs VMs
- 29:31Controversies and final recommendations
related talks
16mOpenZFS- 16mOpenZFS
44mReinventing the File System: OpenZFS
35mVindicating ZFS with PostgreSQL: Unleashing the Power of Scalability
33mEncryption for filesystems with advanced features new fscrypt functionality
23mBtrfs RAID on zoned devices
23mImproving data placement for Zoned Linux File systems
52mHow ZNS SSD's Are Changing the Landscape