Repository navigation
fix(backup): deploy dumps no longer push out the nightly backups - #181
Merged
Merged
Conversation
prune() kept the newest 14 backups of any reason. Pre-deploy, rollback, manual, reset-content and pre-restore safety dumps used the same 14 slots, so a day with several deploys pushed the nightly backups out and the daily recovery window shrank without anyone deciding it. Now two pools: nightly backups (reason "scheduled", or no reason at all for old manifests) keep the admin setting; every other reason keeps the newest BACKUP_KEEP_OTHER (new setting, default 5, 1 to 100). Prune never leaves zero verified nightly backups: if the newest ones all failed verify, the newest verified one is kept beyond the limit. The Backups page states both numbers. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Settings > Backup still called the number "حداکثر نسخهها" (maximum copies), but since the two retention pools it limits only the nightly backups. On PostgreSQL installs the label now says nightly automatic backups, and the help text says the other backups (manual, before an update, and so on) are counted apart with their own number. SQLite installs, which have one pool, keep the old wording. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Brings in main up to a38acb6 (#180, off-site SFTP copies and their admin panel). Conflicts, all additive, both sides kept: - .env.example and app/config.py: #180's OFFSITE_SFTP_* / OFFSITE_GPG_* / OFFSITE_REMOTE_KEEP settings first (they continue the off-site block), then BACKUP_KEEP_OTHER. - static/admin/js/infra_backups.js: load() now calls both renderSchedule(data.schedule, data.keep_other) and renderOffsite(data.offsite). The router merged on its own: the list response carries both keep_other and offsite. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Root cause
app/services/pg_backup.pyprune(keep)kept the newestkeep(default 14) backups of ANY reason. Pre-deploy dumps,rollback dumps, manual ones, reset-content dumps and the panel restore's safety dump all counted in the same 14. A day
with several deploys pushed out nightly backups, so the daily recovery window shrank without anyone deciding it.
Part of Daneshbonyan re-evaluation track 7 (evaluator reason 4: database maturity).
Design doc
docs/features/db-maturity/RESEARCH.md section 5.3 (retention bullet) and ADR-028 in docs/engineering/DECISIONS.md.
Decision made by the lead agent and stated in the task contract (the owner may change it in review): two pools.
What changed:
prune(keep, keep_other=None): scheduled backups (reasonscheduled, empty, missing, or an unreadable manifest)keep
configured_keep(); every other reason keepsBACKUP_KEEP_OTHER(new setting, default 5, clamped 1..100,documented in
.env.example).newest verified one is kept even beyond the keep number.
its number limits the nightly backups and that other backups are counted apart.
Tests
(
test_backup_retention_oracle.py, not committed): 13 failed / 3 passed on the base; 17/17 on this branch (its onepage case was corrected to render the PostgreSQL engine answer, plus a SQLite control).
tests/test_backup_retention_pools.py(36 + 2 cases): a deploy dump never evicts a scheduled one; the 6th"other" dump evicts the oldest "other" one; manifests without a reason count as scheduled; the safety rule; both page
texts. Red first (33 failed), then green; 3 deliberate mutations each caught.
Security
No security impact: no new route or door; the list response gains one number for the existing admin-only page.
AI assistance
Written by Claude agents in a Herdr team: implementer
drill-t1-impl(Opus 5.5) wrote the code, tests and docs;reviewer
drill-t1-rev(Opus 5.5) wrote the oracle first and reviewed (accepted, 1 low label fix applied); leadforeman7(Opus 5.5) re-ran the oracle on the base and the patch. Log entry indocs/engineering/AI_ASSISTANCE_LOG.md.Human review