Stop passing prune options into the legacy scheduled backup engine from the
Django backup command. Record the completed snapshot first, then apply retention
through the SQL-backed retention service so pruning sees the same SnapshotRecord
state as the admin and retention command.
Also record prune failures on BackupRun.result instead of leaving the run in an
ambiguous state.
Add a Django-native snapshot discovery service and management command
that scans backup directories, reads snapshot metadata, and idempotently
upserts SnapshotRecord rows. Expose it through the pobsync command
wrapper, update admin/docs, and cover discovery behavior with tests.
Retire the old YAML and cron oriented pobsync CLI commands and expose a
SQL-first Django-backed command surface instead. Add schedule and
retention management commands, move shared defaults/parsing out of legacy
commands, remove obsolete command modules, and update documentation and
tests for the new workflow.
Build runtime pobsync configuration exclusively from structured SQL
fields, leaving legacy JSON only for import and audit context. Add
SQL-first management commands for global and host configuration and
cover them with tests.
Add explicit Django model fields for global and host backup settings,
including SSH, rsync, source, excludes, and retention configuration.
Populate them from legacy JSON during migration, make the config
repository prefer structured fields, and update import/admin/tests around
the SQL-first configuration model.
Allow retention planning and pruning to use the same ConfigSource
abstraction as scheduled backups. This removes the remaining SQL-to-YAML
export dependency from Django backup runs with pruning, keeping YAML only
as a legacy CLI compatibility path.
Introduce a ConfigSource interface so scheduled backups no longer need
to load host configuration directly from runtime YAML. Add a Django-backed
config source for SQL-driven backup runs, keep file-based config as the
CLI default, and make scheduled prune execution actually apply retention
after successful runs.
Treat SQL-backed Django models as the source of truth for pobsync
configuration, exporting runtime YAML only as a compatibility layer for
the existing engine. Add a database-driven scheduler command, Docker
scheduler services, schedule run-state fields, and tests for scheduler,
config export, and retention behavior.
Add a Django admin-backed management layer for pobsync configs, runs,
snapshots, and schedules. Keep the existing CLI engine as the execution
source of truth, add import/run management commands, and provide SQLite
default plus optional MariaDB Docker Compose support.
