Files
nextcloud-ng_backup/lib/Service/PruneService.php
T
dillard 6cbc5d79d4
CI / unit (8.2) (push) Has been cancelled
CI / unit (8.3) (push) Has been cancelled
CI / unit (8.4) (push) Has been cancelled
CI / psalm (push) Has been cancelled
CI / integration (mariadb, 32) (push) Has been cancelled
CI / integration (mariadb, 33) (push) Has been cancelled
CI / integration (mariadb, 34) (push) Has been cancelled
CI / integration (mariadb, 35) (push) Has been cancelled
CI / integration (pgsql, 32) (push) Has been cancelled
CI / integration (pgsql, 33) (push) Has been cancelled
CI / integration (pgsql, 34) (push) Has been cancelled
CI / integration (pgsql, 35) (push) Has been cancelled
CI / integration (sqlite, 32) (push) Has been cancelled
CI / integration (sqlite, 33) (push) Has been cancelled
CI / integration (sqlite, 34) (push) Has been cancelled
CI / integration (sqlite, 35) (push) Has been cancelled
CI / integration-targets (push) Has been cancelled
CI / integration-user (push) Has been cancelled
Phase 2: per-user backup and restore via user_migration
Wires the existing RepositoryExportDestination/RepositoryImportSource (see
lib/UserMigration/) into user_migration's own UserMigrationService, resolved on demand
through a new UserMigrationServiceLocator since the app is an optional dependency
(PLAN.md 3.8) -- without it, only this feature is unavailable, the full backup still
works.

New: UserBackupService::backupUser() (occ backup:backup:user) streams a user_migration
export straight into the repository, under the same shared location lock as a regular
backup. UserRestoreService::restoreUser() (occ backup:restore:user) never imports over a
live account: --mode=replace safety-exports the current state first, then deletes and
recreates the account from the backup under the same uid; --mode=as-backup restores into
a new "<uid>-bak" account instead, since Nextcloud cannot rename a uid.

The catalog now also tracks user export manifests (Repository::recordUserExport/
listUserExports/forgetUserExport/isRecordedUserExport), rollback-protected like a
snapshot: listUserExports() verifies each one is still present on the location, and
restore only ever reads a manifest the verified catalog actually recorded, never an
arbitrary caller-supplied path. prune()'s existing raw users/ scan for blob retention is
unchanged. A new short exclusive "catalog" lock (PruneService::catalogLockKey())
serializes the catalog write itself, so two concurrent per-user backups (or one
alongside a full backup) can no longer race to write the same next generation and
silently drop one of them -- the broader per-target shared lock only ever protected
against a concurrent cleanup, not against other shared holders.

Two real bugs found by testing this against a live Nextcloud instance, not by unit
tests (RepositoryExportDestination/RepositoryImportSource implement OCP interfaces that
plain PHPUnit can't load, see tests/Unit/Repository/UserExportTest.php's docblock; the
new tests/integration/user-roundtrip.sh covers the real round trip instead):
- RepositoryImportSource::copyToFolder() double-closed the stream it handed to
  putContent()/newFile() (both already close it themselves), which throws a TypeError on
  PHP 8 and broke every restored file with content.
- A replace-mode restore failed on NotPermittedException: IRootFolder::getUserFolder()
  caches a Folder per uid for the process lifetime, so after delete+recreate of the same
  uid the import still got the deleted account's (now storage-less) folder. Fixed by
  dispatching FilesystemTornDownEvent, which Root already listens for to clear exactly
  that cache -- no internal OC\ classes needed.

Local review (CodeRabbit + Codex) against this diff found three more real issues, all
fixed: the catalog write race above, a failed as-backup import leaving a broken stub
account behind (now deleted on failure), and an export stream leak on a read/backend
failure in RepositoryExportDestination::copyFolder() (now closed in a finally).

Deliberately left out: wiring per-user export into the automatic backup schedule
(on-demand via occ only, for now) and retention/pruning of old user exports -- both open
design questions, not yet decided. The lease covering a single export/import still has a
fixed TTL (3600s) with no mid-call refresh, since UserMigrationService exposes no
progress; a resumable, step-based per-user backup (like the main BackupService run)
would remove that cap but is a bigger change than this pass.
2026-10-05 17:36:46 +02:00

177 lines
6.8 KiB
PHP

<?php
declare(strict_types=1);
// SPDX-FileCopyrightText: 2026 Dillard Blom
// SPDX-License-Identifier: AGPL-3.0-or-later
namespace OCA\NgBackup\Service;
use OCA\NgBackup\AppInfo\Application;
use OCA\NgBackup\Db\RunMapper;
use OCA\NgBackup\Db\SnapshotMapper;
use OCA\NgBackup\Db\Target;
use OCP\IAppConfig;
use OCP\Lock\ILockingProvider;
use OCP\Lock\LockedException;
/**
* Applies the retention policy to a location: forgets snapshots the policy does not keep and
* prunes data no remaining snapshot uses. Never runs on append-only locations, never while a
* backup to the same location is running (the location lock is shared with BackupService).
*/
final class PruneService {
public function __construct(
private TargetService $targets,
private RunMapper $runs,
private SnapshotMapper $snapshots,
private IAppConfig $appConfig,
private ILockingProvider $locking,
private AlertService $alerts,
private KeyService $keys,
private LeaseService $leases,
) {
}
public function untrash(Target $target, string $snapshotId): void {
$lease = $this->lockExclusive($target, 300);
try {
$this->doUntrash($target, $snapshotId);
} finally {
$this->leases->release($lease);
}
}
/**
* Run $fn under the location's exclusive lock (to create a run atomically with respect to a
* cleanup, which takes the same lock and then refuses while a run exists).
*/
public function underStartLock(Target $target, callable $fn): mixed {
try {
return $this->leases->with(self::lockKey($target), LeaseService::EXCLUSIVE, 60, fn () => $fn());
} catch (LockedException) {
throw new \RuntimeException('A cleanup of ' . $target->getName() . ' is in progress; try again later');
}
}
/**
* Exclusive lock on the location, then make sure no run of any kind (backup, restore, export)
* is in progress: their data may not be referenced by a snapshot yet, or is being read.
*/
private function lockExclusive(Target $target, int $ttl): string {
try {
$lease = $this->leases->acquire(self::lockKey($target), LeaseService::EXCLUSIVE, $ttl);
} catch (LockedException) {
throw new \RuntimeException('A backup, restore or cleanup of ' . $target->getName() . ' is in progress; try again later');
}
if ($this->runs->findRunning($target->getId()) !== []) {
$this->leases->release($lease);
throw new \RuntimeException('A backup or restore of ' . $target->getName() . ' is running; try again later');
}
return $lease;
}
private function doUntrash(Target $target, string $snapshotId): void {
$repo = $this->targets->repository($target);
if (!array_key_exists($snapshotId, $repo->trash())) {
throw new \InvalidArgumentException('That snapshot is not in the trash (or was already deleted)');
}
$repo->untrash($snapshotId);
// Put it back in the local snapshot list as well.
$meta = $repo->snapshotMeta($snapshotId);
$s = new \OCA\NgBackup\Db\Snapshot();
$s->setTargetId($target->getId());
$s->setSnapshotId($snapshotId);
$s->setKind((string)($meta['kind'] ?? 'full'));
$s->setLabel(($meta['label'] ?? '') !== '' ? $meta['label'] : null);
$s->setCreatedAt((int)strtotime((string)($meta['time'] ?? 'now')));
$s->setFiles((int)($meta['stats']['files'] ?? 0));
$s->setBytes((int)($meta['stats']['bytes'] ?? 0));
try {
$this->snapshots->insert($s);
} catch (\Throwable) {
}
$this->alerts->securityEvent('snapshot_untrashed', ['target' => $target->getName(), 'snapshot' => $snapshotId]);
}
public static function lockKey(Target $target): string {
return 'ng_backup/target/' . $target->getId();
}
/**
* A separate, short-lived exclusive lock serializing writes to the catalog itself. The
* per-target lock above only keeps readers/writers out while a cleanup runs; it does not
* stop two SHARED holders (e.g. two concurrent per-user backups, or one alongside a full
* backup) from both reading the same catalog generation and racing to write the next one,
* silently dropping whichever's change loses. The full-backup path avoids this today only
* because a Run row makes concurrent runs for the same target mutually exclusive; per-user
* backup/restore has no such row, so it takes this lock around the catalog write itself.
*/
public static function catalogLockKey(Target $target): string {
return self::lockKey($target) . '/catalog';
}
public function policy(): RetentionPolicy {
return RetentionPolicy::fromArray(json_decode($this->appConfig->getValueString(Application::APP_ID, 'retention', '{}'), true) ?: []);
}
public function setPolicy(RetentionPolicy $policy): void {
if ($policy->last < 1) {
throw new \InvalidArgumentException('Keep at least the last snapshot');
}
$old = $this->policy()->toArray();
$this->appConfig->setValueString(Application::APP_ID, 'retention', json_encode($policy->toArray(), JSON_THROW_ON_ERROR));
if ($old !== $policy->toArray()) {
$this->alerts->securityEvent('retention_changed', ['policy' => json_encode($policy->toArray())]);
}
}
/**
* Snapshots the policy drops go to the trash on the location; trash entries older than the
* deletion delay are purged, and only then is their data freed.
*
* @return array{kept: array<string, list<string>>, forgotten: list<string>, purged: list<string>, prune: array}
*/
public function apply(Target $target, bool $dryRun = false, bool $scheduled = false): array {
if ($target->getAppendOnly()) {
throw new \RuntimeException('Location ' . $target->getName() . ' is append-only: NG Backup does not delete anything there');
}
$lease = $this->lockExclusive($target, 1800);
try {
$repo = $this->targets->repository($target);
$times = [];
foreach ($repo->listSnapshots() as $id) {
$meta = $repo->snapshotMeta($id);
$times[$id] = (int)strtotime((string)($meta['time'] ?? '@0'));
}
$keep = $this->policy()->keep($times);
$forget = array_values(array_diff(array_keys($times), array_keys($keep)));
$actor = $scheduled ? 'retention policy' : $this->alerts->actor();
$purged = [];
if (!$dryRun) {
foreach ($forget as $id) {
$repo->forget($id, $actor);
try {
$this->snapshots->delete($this->snapshots->findOne($target->getId(), $id));
} catch (\Throwable) {
}
}
}
// Also removes leftovers of interrupted runs. A dry run counts the snapshots it would forget as gone.
$delay = $this->keys->deleteDelayDays();
if (!$dryRun && $forget !== []) {
// Scheduled cleanup follows the policy: audit log only, no notification every week.
$this->alerts->securityEvent('snapshots_forgotten', ['target' => $target->getName(), 'count' => (string)count($forget),
'until' => date('Y-m-d', time() + $delay * 86400)], !$scheduled);
}
if (!$dryRun) {
$purged = $repo->purgeTrash($delay * 86400);
}
$prune = $repo->prune($dryRun);
return ['kept' => $keep, 'forgotten' => $forget, 'purged' => $purged, 'prune' => $prune];
} finally {
$this->leases->release($lease);
}
}
}