GATE E
ARCHIVED
intake record retained · source can be wiped
0
items here now
—
aggregate size
What this gate means
An item is at ARCHIVED once it's fully retired: source decommissioned, intake record retained for audit. The data lives in its canonical home (RAID + R2 mirror); the source — whether that's a cloud archive about to expire, a 14-year-old hard drive on the desk, or a duplicated backup generation — can now be wiped or returned to circulation.
Sub-routines · operational playbook
Source decommission · drive wipe
diskutil secureErase 1 (zero-pass, sufficient for spinners)
For mechanical hard drives, a single-pass zero from
diskutil secureErase or dd if=/dev/zero is sufficient. Multi-pass DoD wipes are theatre on modern drives — recoverable data after one zero pass requires lab-grade scanning electron microscopy. SSDs need a different approach (blkdiscard or the drive's secure-erase ATA command).diskutil secureErase 1 /dev/diskN # 1 = zero-out, one pass
when: Once item is verified safe in canonical home + R2 mirror
Cloud account · optional source deletion
delete original Takeout / archive on the source platform
Once a Google Takeout / iCloud archive has been pulled, checksummed, sorted, and mirrored, the source archive on Google/Apple can be deleted — they expire anyway. Keep the account itself; we just delete the prepared export package.
# manual via takeout.google.com / privacy.apple.com
when: Optional — Google/Apple expire archives automatically anyway
Intake record · audit trail
set state: archived in migrations.yml, re-bake
Don't delete the migrations.yml entry. Flip state to
archived and leave the record so you have a history of what landed when, where it lives now, and what source it came from. The dashboard items-by-gate ARCHIVED section becomes the audit log.$EDITOR migrations.yml # state: archived
./bake.py && git commit -am 'archive: '
when: After source decommission
Backup posture · 3-2-1
RAID + R2 derivatives + offsite cold copy
Archive doesn't mean 'one copy on the RAID and done'. The 3-2-1 rule: 3 copies (working + backup + offsite), 2 different media (RAID + R2 / cold drive), 1 offsite (R2 IS offsite; ideally also a rotating external in a different physical location). Verify periodically — silent corruption on a backup you never read is the same as no backup.
# quarterly: scrub RAID parity (zfs scrub / mdadm check)
# semi-annual: rotate cold backup drive offsite
when: Ongoing posture, not a one-shot
Items at this gate
No items currently at this gate.
Gate-exit checklist
Verify before moving items into (no next gate — terminal state):
- Bytes at canonical home AND mirrored to R2 AND verifiable
- Source wiped (if physical) or marked for natural expiry (if cloud)
- migrations.yml record retained with state: archived
- Quarterly RAID scrub on the calendar
Gate E · ARCHIVED · baked 2026-05-29 from
migrations.yml + GATE_DETAIL