- sam:doctor: health report (secret key, address, HTTPS, SMTP and SSO
reachable, Google Drive, migrations, disk, jobs, backup/check/off-site/
procedures/audit recency); exits 1 on a failure
- records:check: backup now + history and signature check; exits 1 on a problem
- backups:list / backups:restore[file]: checks the copy (nightly or Drive
.gz), refuses while the server runs, moves the current DB aside, asks
(CONFIRM=yes to skip)
- users:sign_out_all, sam:records_package[tail], email:test[address],
offsite:backup, offsite:restore_test
- Admin guide: maintenance commands table, restore steps use backups:restore
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Admin → Off-site Backup: connect a Google account with your own OAuth
client (drive.file scope; refresh token and secret encrypted)
- Nightly after the verified database backup: upload the gzipped snapshot
(newest N kept, Settings) and every uploaded file not yet copied
- Restore test downloads the newest copy, opens and checks it, and checks
sampled files against their checksums; failures emailed to admins
- About, procedures document, policy completeness and audits use it
- offsite:restore_files rake task; restore steps in the admin guide
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>