Merge pull request #20 from dillardblom/rc6

Release 1.0.0-rc.6

CHANGELOG.md

@@ -7,6 +7,35 @@ the app store accepts a `-beta` suffix, but Nextcloud servers then only offer th
 beta/daily update channel, which would hide the app from most admins. The stability stage is
 instead carried by the app description, the git tag, the release, and this file.
 
+## [1.0.0-rc.6] - 2026-10-24
+
+Sixth release candidate for 1.0.0; the app version is 0.9.12. Fixes from two disaster recovery
+simulations and a user report. If no problems are reported for about a week, 1.0.0 will be this
+same code.
+
+### Fixed
+- The daily backup time was taken as UTC (Nextcloud runs PHP in UTC): 03:00 ran at 05:00 in
+  Berlin (#19). The time is now saved with the administrator's time zone and shown with it. A time
+  saved before is converted once to the zone of the server or an administrator, so it shows as it
+  effectively ran (02:30 UTC becomes 04:30 Europe/Berlin in summer) and nothing moves.
+- A full restore over SMB stopped after 17 hours with "Decryption failed": a read over a long
+  line handed out wrong bytes (the backup itself was fine, the same files restored identically
+  afterwards). A blob that does not decrypt as first read, of a file or of the database, is now
+  read once more on a request of its own, and only a second failure stops the restore.
+- A full restore onto MySQL/MariaDB stopped with "Incorrect string value" when it created a table
+  (of an app that is not installed here) whose rows hold emoji or other 4-byte characters: the
+  table got 3-byte utf8. Tables the restore creates now get this installation's character set and
+  collation (utf8mb4). Found in a disaster recovery simulation onto MariaDB.
+
+### Changed
+- The admin page uses more of the screen (sections up to 1400 pixels instead of 900), so the
+  list of backup runs fits without scrolling sideways.
+- README: a roadmap, the tested disaster recoveries, measurements for OVH S3, the Storage Box over
+  SFTP and SMB, how to build the SMB extension when PECL is down, and what works without shell
+  access. The guide lists shell access among what you need.
+- The full-restore integration test checks a 4-byte character in a table the restore creates, and
+  retries app installs from the app store.
+
 ## [1.0.0-rc.5] - 2026-10-22
 
 Fifth release candidate for 1.0.0; the app version is 0.9.11. A full restore of 1 TB over SFTP

PLAN.md

@@ -395,11 +395,53 @@ in README "Known limitations":
 - **1.0.1:**
   1. (Naar rc4: herstelsnelheid over SFTP.) S3 leest al 16 tegelijk bij index en tree, maar een
      restore leest de packs nog één voor één: ook vooruitlezen via de S3-client.
-  2. Waarschuwing in de beheerpagina en `backup:status` als achtergrondjobs niet op systeem-cron
-     staan, en "waiting for background jobs" als een lopende backup lang geen stap zette (issue
-     #13). Cron zelf aanzetten kan de app niet: zonder echte cron-regel stopt dan alles.
+  2. Achtergrondjobs zonder systeem-cron (issue #13; besluit 2026-10-19). De PR die dit
+     oplost zegt "Fixes #13" (bevestigd: #13 was de cron; besluit 2026-10-21):
+     a. **Waarschuwen** in de beheerpagina en `backup:status`: modus uit `core backgroundjobs_mode`
+        (`ajax`/`webcron`/`cron`) en de laatste run uit `core lastcron`. Niet op `cron`, of de
+        laatste run te lang geleden: melding met de cron-regel erbij ("Backups only move while
+        someone has Nextcloud open"), en "waiting for background jobs" als een lopende backup lang
+        geen stap zette.
+     b. **De pagina laat de backup lopen** zolang ze openstaat: in `ajax`/`webcron` voert de
+        statusaanroep die de pagina toch al herhaalt ook een korte, begrensde stap uit (zoals de
+        korte stappen die AJAX-cron nu al krijgt). "Back up now" werkt dan ook zonder cron zolang
+        het tabblad openstaat; de melding zegt dat geplande backups echte cron nodig hebben.
+        De melding zegt (zonder werkende systeem-cron): backups hebben systeem-cron nodig, ook de
+        gewone nachtelijke; zo stel je cron in (link naar de Nextcloud-documentatie over
+        achtergrondjobs voor de versie van de server, nu
+        https://docs.nextcloud.com/server/34/admin_manual/configuration_server/background_jobs_configuration.html,
+        versienummer uit de server, niet vast); zolang er geen cron is, loopt een backup die je
+        met de knop start alleen door terwijl de NG Backup-pagina openstaat.
+     c. **Voortgang in de databasefase** (issue #13): nu staat er alleen "running (db)", wat er
+        hetzelfde uitziet als vastzitten. Toon "running (db: 47/183 tables)" zoals de bestandsfase
+        een percentage toont, en "waiting for background jobs (last ran 12 min ago)" als een run
+        een tijd geen stap zette, ook bij een cron die is blijven staan.
+     Niet doen: de cron-modus vanuit de app omzetten (zonder echte cron-regel stopt alles, en
+     `cron.php` zet de modus zelf op `cron`), en geen `exec php occ backup:run &` vanuit de knop
+     (vaak uitgeschakeld, niet op hosted, buiten Nextcloud om lastig te volgen en te stoppen).
   3. Volledige DR-test met een backup die al de tabelstructuur bevat (gemaakt met rc2 of later),
      zonder `--skip-missing-tables`: dat pad is op 1 TB nog niet getest.
+  4. **Label bij "Back up now"** (besluit 2026-10-19): een optioneel tekstveld naast de knop; de
+     tekst gaat als label van de snapshot mee (nu kan dat alleen met `occ backup:run -l`; de knop
+     stuurt een leeg label, geplande runs krijgen "scheduled"). Raakt alleen de UI en een
+     bestaande parameter.
+  5. **`--salvage`** voor `backup:restore` en `backup:restore:full` (besluit 2026-10-22): terughalen
+     wat kan, een bestand dat niet te lezen of te ontsleutelen is overslaan en aan het eind
+     opsommen, in plaats van stoppen.
+  6. **Tijdelijke bestanden van een afgebroken restore opruimen** (gevonden in de SMB-DR van
+     rc6, 2026-10-24): de rc5-poging die stopte liet `VID_0009.mp4.ngb-restore-<id>` staan; de
+     geslaagde restore daarna haalde het niet weg, en `files:scan` maakte het zichtbaar in Files.
+     Een restore ruimt `*.ngb-restore-*` in de mappen die hij terugzet op.
+  7. **Eén bestand terugzetten met `backup:restore`**: een pad naar een bestand zet 0 bestanden
+     terug (er komt een `/` achter); nu werkt alleen een map.
+- **DR-tests over databasegrenzen** (besluit 2026-10-20; productie nc01 draait PostgreSQL): na
+  de MariaDB→PostgreSQL-fout (tellers, rc5) elke overstap in de integratietest (`to-pgsql`,
+  `to-mysql`, `to-sqlite`, ook in CI). Volle DR-tests met echte data: vóór 1.0.0 de S3-simulatie
+  (MariaDB-bron, 138 GB) nogmaals naar een MariaDB-doel en de SMB-simulatie naar een SQLite-doel;
+  **na 1.0.0** de 1 TB van nc01 (PostgreSQL) naar een MariaDB-doel.
+- **Polish:** `target:add` toonde bij het eerste indexbestand al "about 642 min left" (S3,
+  4.568 bestanden, daarna binnen een minuut klaar): een schatting pas na ~5% of na meer dan
+  een paar bestanden.
 - **1.1.0:**
   1. `backup:key:import` zonder kit: de sleutel-slots van de locatie (`config`) lezen, met de kit
      als reserve als `config` ontbreekt of niet opent. Vergt een andere volgorde van
@@ -412,6 +454,22 @@ in README "Known limitations":
   3. Snelheid: lezen en uploaden laten overlappen, meerdere bestanden tegelijk lezen, eventueel
      een eigen worker (`occ background-job:worker`). Gemeten op nc01: 55% wachten op de schijf,
      33% op de SMB-upload, ~7% CPU. Vergt extra PHP-processen; botst mogelijk met uitgangspunt 1.
+  4. **Schema per locatie** (besluit 2026-10-19): in de instellingen van een locatie "uit",
+     "dagelijks om HH:MM" of "elke N uur (vanaf HH:MM)". Nu is er één `schedule_time` voor alle
+     locaties, één keer per dag. Bij de update wordt de huidige tijd het schema van elke
+     bestaande locatie (voor bestaande gebruikers verandert er niets).
+  5. **Locaties na elkaar laten starten** in plaats van tegelijk: twee runs die dezelfde schijven
+     lezen zijn samen niet sneller (gezien op nc01); wachtende runs zichtbaar als "queued".
+  7. **SMB sneller** (gemeten 2026-10-21): lokaal netwerk zonder vertraging (Samba-container op de
+     laptop, 1,8 GB) back-up 24,9 MB/s, restore 27,6 MiB/s; vanaf Canada naar de Storage Box (130 ms)
+     2,6 MiB/s, omdat SMB per blok op het antwoord wacht. Grotere schrijfblokken, en vooruitlezen
+     bij een restore zoals SFTP (IReadsAhead) of meerdere packs tegelijk. Een NAS met 1 Gbit haalt
+     ~110 MB/s: daar is winst.
+     Ook de index: `target:add` over SMB vanaf Canada haalt ~1 indexbestand per seconde (4.927
+     bestanden ~85 min), want SMB krijgt geen `getMany()` met verzoeken tegelijk zoals S3 en SFTP.
+  6. **Notitie achteraf bij een restorepunt**, te wijzigen in de lijst. Het label in de snapshot
+     is versleuteld en ligt vast; de notitie apart op de locatie bewaren (bijv. in de catalogus),
+     zodat een nieuwe server na een DR hem ook ziet.
 - **Bevindingen DR-test op PVE02 (2026-10-17, snapshot van nc01, 1 TB, AIO → gewone NC 34 met
   PostgreSQL 18)** — voor **1.0.0-rc.2**:
   1. Tabellen van apps die bij de backup **niet aanstonden** (restanten: Deck, ONLYOFFICE,

README.md

@@ -146,10 +146,31 @@ What you should do on the location (strongly recommended):
 - **Files and folders** of a user: in the web interface or `occ backup:restore`, into a new folder
   ("Restored <date>"), merged into the original place, or replacing it.
 - **Raw restore** of any path (for example `config/`) to a local directory: `occ backup:restore ... --to-directory=DIR`.
-- **Whole server** (disaster recovery onto a fresh installation): `occ backup:restore:full`, see the how-to.
+- **Whole server** (disaster recovery onto a fresh installation): `occ backup:restore:full`, see
+  [docs/DISASTER-RECOVERY.md](docs/DISASTER-RECOVERY.md). It runs in maintenance mode, so that
+  nobody and no background job writes into the instance while it is being replaced.
 - **One user** (account, settings, files, calendars, contacts via user_migration): `occ backup:user:backup` and
   `occ backup:user:restore`, replacing the account or restoring it as `<uid>-bak`. See the how-to.
 
+**Without shell access** (a hosted Nextcloud where you cannot run `occ`): files and folders can be
+restored in the web interface. Restoring a whole user or the whole server needs `occ`, so on such
+hosting it is something your provider (or whoever has shell access) runs. A restore mode that
+works without maintenance mode, and so from the web interface, is on the roadmap.
+
+## Tested disaster recoveries
+
+Full restores onto a fresh server, following only [the guide](docs/DISASTER-RECOVERY.md):
+
+| Backup | Location | Database | Result |
+|---|---|---|---|
+| Production, about 1 TB, 150,000 files (Nextcloud All-in-One) | Hetzner Storage Box (SFTP), Germany → Canada | PostgreSQL → PostgreSQL | 150,622 files restored in about 25 h (13-14 MiB/s), sample identical |
+| 138 GB simulation: 20 users, calendars, contacts, mail, shares, team folders with an ACL, versions, trash | OVH S3, Warsaw → Canada | MariaDB → PostgreSQL | all 15,772 files and every database count identical |
+| The same | OVH S3 | MariaDB → MariaDB | all 15,772 files and every database count identical |
+| The same | Hetzner Storage Box (SMB), Germany → Canada | MariaDB → SQLite | all 15,772 files and every database count identical (27.6 h, about 1.5 MB/s) |
+| 51 GB test instance | OVH S3 | PostgreSQL → PostgreSQL | 23,519/23,519 checksums identical |
+
+The integration tests restore across every pair of PostgreSQL, MariaDB and SQLite.
+
 ## Storage backend experiences
 
 Notes from actually running a backend, beyond what's in files_external's own documentation.
@@ -206,12 +227,64 @@ Contributions welcome via PR — add what you ran into with your own provider.
   can be read on more connections: `occ config:app:set ng_backup restore_sftp_connections
   --value=4` (at most 8; the server must allow that many plus one at once for the user).
 
+- **OVH Object Storage (S3, Warsaw, 2026-10):** works with the `amazons3` backend (path style,
+  region `waw`). A full backup of 1 TB from Germany ran at about 7 MiB/s, from Canada about
+  6 MiB/s with BBR; a check of a 1 TB backup took 3.5 minutes (index files are fetched 16 at a
+  time); a restore of 138 GB from Warsaw to Canada took about 6 hours. Turn on versioning and
+  Object Lock on the bucket (see above); OVH charges no egress.
+
+- **Hetzner Storage Box over SFTP, full restore:** about 13-14 MiB/s from Germany to Canada
+  (122 ms) with the read-ahead of 1.0.0-rc.4; without it about 2.8 MiB/s. A Storage Box gets
+  slower with several connections reading at once (4-6 MiB/s for 2-4 connections, it reads from
+  disks), so one extra connection is the default. A Storage Box drops an SFTP connection that
+  stayed quiet for a while without telling the client; NG Backup checks and replaces it.
+
+- **SMB speed:** on a local network (Samba, no delay) a backup ran at about 25 MB/s and a
+  restore at about 28 MiB/s. Over a long line SMB waits for an answer to each request: from
+  Canada to a Storage Box in Germany (130 ms) about 2.6 MiB/s for a backup and 1.6 MiB/s for a
+  restore, and connecting a location with an existing backup reads its index at about one file a
+  second. SMB is meant for a local network; use SFTP or S3 for a location far away. When
+  `pecl install smbclient` fails (pecl.php.net was unreachable for hours on 2026-10-21), the
+  extension builds from its source: `libsmbclient-php` on GitHub, `phpize && ./configure && make
+  && make install`, then `docker-php-ext-enable smbclient`.
+
 - **NFS is not a files_external backend at all.** Mount the NFS export on the server yourself
   (e.g. via `/etc/fstab` or a systemd `.mount` unit) and point ng_backup's `local` backend at the
   mount point; ng_backup never mounts or unmounts anything on its own, so an NFS mount dropping
   mid-run behaves like any other local-path failure (the run resumes). See PLAN.md 3.5 for why NFS
   can't be driven from PHP directly.
 
+## Roadmap
+
+What is planned, without dates. Ideas and votes are welcome in the
+[issues](https://github.com/dillardblom/ng_backup/issues); if something you need is listed here,
+a 👍 on its issue helps decide what comes first.
+
+**1.0.1**
+- A warning on the admin page and in `occ backup:status` when background jobs do not run on a
+  system cron, with how to set one up. Without cron, a backup started with "Back up now" moves on
+  while the NG Backup page is open ([#13](https://github.com/dillardblom/ng_backup/issues/13)).
+- An optional label for a backup started with "Back up now".
+- Faster full restores from S3 and SMB, reading ahead as restores over SFTP do since 1.0.0-rc.4.
+
+**1.1.0**
+- Restore without the recovery kit: the backup key from the location itself, opened with one of
+  your passphrases.
+- A schedule per location, also "every N hours".
+- Locations backed up one after the other instead of at the same time.
+- A note on a restore point that you can change afterwards.
+- An option to pause heavy background jobs (such as Memories, Recognize or the Preview
+  Generator) while a backup runs.
+- Faster backups: reading and uploading at the same time.
+
+**Later**
+- A full restore onto the existing installation, without first installing a fresh Nextcloud, in a
+  restore mode of its own instead of maintenance mode, so that it can also run from the web
+  interface (for hosted Nextcloud without shell access).
+- Users restoring their own files from the Files app.
+- Primary object storage and External Storage mounts (see Known limitations).
+- An option not to keep the backup key on the location (the recovery kit is then the only copy).
+
 ## Known limitations
 
 On the backlog; none of these block the beta or 1.0.0, but know them before you rely on it:
@@ -236,8 +309,11 @@ On the backlog; none of these block the beta or 1.0.0, but know them before you
 - Location passwords passed with `-o password=...` are visible in the process list and shell
   history while the command runs.
 - Temporary files of a restore into a user's files end up in that user's trash.
-- Restoring the database does not order tables by foreign keys; apps with cascading foreign keys
-  could lose rows (Nextcloud core does not use them).
+- **Restoring a whole user or the whole server needs `occ`** (shell access); see "Without shell
+  access" above.
+- **`--install-apps` depends on the app store.** Nextcloud downloads the whole app list (about
+  38 MB) and gives up after 120 seconds; on a slow day apps are not installed. The restore then
+  lists them: install them afterwards with `occ app:install`, their data is restored already.
 - A pack whose index upload failed stays on the location until removed by hand.
 
 ## Commands

appinfo/info.xml

@@ -42,7 +42,7 @@ welcome: use the "Share your experience" link on the admin page.
 
 Developed and sponsored by [Ensembia](https://www.ensembia.com). We can offer (paid) assistance if
 you need help with restoring a backup after a crash.]]></description>
-    <version>0.9.11</version>
+    <version>0.9.12</version>
     <licence>agpl</licence>
     <author mail="dillard.blom@ensembia.com">Dillard Blom</author>
     <namespace>NgBackup</namespace>
@@ -53,9 +53,9 @@ you need help with restoring a backup after a crash.]]></description>
     <website>https://github.com/dillardblom/ng_backup</website>
     <bugs>https://github.com/dillardblom/ng_backup/issues</bugs>
     <repository type="git">https://github.com/dillardblom/ng_backup</repository>
-    <screenshot>https://raw.githubusercontent.com/dillardblom/ng_backup/v1.0.0-rc.5/screenshots/1-backups.png</screenshot>
-    <screenshot>https://raw.githubusercontent.com/dillardblom/ng_backup/v1.0.0-rc.5/screenshots/2-restore.png</screenshot>
-    <screenshot>https://raw.githubusercontent.com/dillardblom/ng_backup/v1.0.0-rc.5/screenshots/3-settings.png</screenshot>
+    <screenshot>https://raw.githubusercontent.com/dillardblom/ng_backup/v1.0.0-rc.6/screenshots/1-backups.png</screenshot>
+    <screenshot>https://raw.githubusercontent.com/dillardblom/ng_backup/v1.0.0-rc.6/screenshots/2-restore.png</screenshot>
+    <screenshot>https://raw.githubusercontent.com/dillardblom/ng_backup/v1.0.0-rc.6/screenshots/3-settings.png</screenshot>
     <dependencies>
         <php min-version="8.1"/>
         <lib>sodium</lib>

docs/DISASTER-RECOVERY.md

@@ -10,6 +10,8 @@ Nextcloud back from an NG Backup location. Do them in this order.
 - [ ] A new server with the **same Nextcloud major version** as the old one (for example 34),
       a database (PostgreSQL, MySQL/MariaDB or SQLite; it may differ from the old one) and room for
       the whole backup plus a margin.
+- [ ] **Shell access** to the new server, to run `occ` (on a hosted Nextcloud without it, your
+      provider runs these steps).
 - [ ] A list of the apps that are **not in the app store** (your own builds, apps copied in by
       hand). Keep their packages somewhere safe beforehand.
 

lib/BackgroundJob/BackupJob.php

@@ -32,6 +32,7 @@ final class BackupJob extends TimedJob {
 		private RunMapper $runs,
 		private IAppConfig $appConfig,
 		private LoggerInterface $logger,
+		private \OCA\NgBackup\Service\Schedule $schedule,
 	) {
 		parent::__construct($time);
 		// Well below the usual 5-minute cron interval: the last run is stamped a moment after cron
@@ -101,14 +102,10 @@ final class BackupJob extends TimedJob {
 		$this->appConfig->setValueString(Application::APP_ID, 'last_prune_week', $due);
 	}
 
-	/** Daily schedule: appconfig ng_backup/schedule_time "HH:MM" (server time); empty = no schedule. */
+	/** Daily schedule: "HH:MM" in the schedule's time zone (see Schedule); empty = no schedule. */
 	private function startScheduled(): void {
-		$at = $this->appConfig->getValueString(Application::APP_ID, 'schedule_time', '');
-		if (!preg_match('/^([01]\d|2[0-3]):[0-5]\d$/', $at)) {
-			return;
-		}
-		$todayAt = (int)strtotime(date('Y-m-d') . ' ' . $at);
-		if (time() < $todayAt) {
+		$todayAt = $this->schedule->todayAt();
+		if ($todayAt === null || time() < $todayAt) {
 			return;
 		}
 		foreach ($this->targets->list() as $t) {

lib/Controller/AdminApiController.php

@@ -47,6 +47,7 @@ final class AdminApiController extends Controller {
 		private \OCA\NgBackup\Service\AlertService $alerts,
 		private \OCA\NgBackup\Service\KeySlotService $slots,
 		private \OCA\NgBackup\Service\PruneService $prune,
+		private \OCA\NgBackup\Service\Schedule $schedule,
 	) {
 		parent::__construct(Application::APP_ID, $request);
 	}
@@ -72,7 +73,8 @@ final class AdminApiController extends Controller {
 			'runs' => $runs,
 			'olderRuns' => $allRuns ? 0 : $this->runs->countAll() - count($shown),
 			'recentDays' => \OCA\NgBackup\Command\Status::RECENT_DAYS,
-			'schedule' => $this->appConfig->getValueString(Application::APP_ID, 'schedule_time', ''),
+			'schedule' => $this->schedule->get()['time'],
+			'scheduleTimezone' => $this->schedule->get()['timezone'],
 		]);
 	}
 
@@ -240,12 +242,14 @@ final class AdminApiController extends Controller {
 		return new JSONResponse($this->runJson($run));
 	}
 
-	public function setSchedule(string $time): JSONResponse {
-		if ($time !== '' && !preg_match('/^([01]\d|2[0-3]):[0-5]\d$/', $time)) {
-			return new JSONResponse(['error' => 'Use HH:MM or empty'], Http::STATUS_BAD_REQUEST);
+	/** @param string $timezone the administrator's zone (the browser's); empty: keep the current one */
+	public function setSchedule(string $time, string $timezone = ''): JSONResponse {
+		try {
+			$this->schedule->set($time, $timezone !== '' ? $timezone : $this->schedule->get()['timezone']);
+		} catch (\InvalidArgumentException $e) {
+			return new JSONResponse(['error' => $e->getMessage()], Http::STATUS_BAD_REQUEST);
 		}
-		$this->appConfig->setValueString(Application::APP_ID, 'schedule_time', $time);
-		return new JSONResponse(['schedule' => $time]);
+		return new JSONResponse(['schedule' => $time, 'scheduleTimezone' => $this->schedule->get()['timezone']]);
 	}
 
 	public function snapshots(int $targetId): JSONResponse {

lib/Db/DbDumper.php

@@ -159,7 +159,7 @@ final class DbDumper {
 		// everywhere); tables it does not know are left out, as asked.
 		foreach ($missing as $table) {
 			if (isset($manifest['tables'][$table]['schema'])) {
-				$this->db->createSchemaManager()->createTable(self::tableFromStructure($table, $manifest['tables'][$table]['schema']));
+				$this->db->createSchemaManager()->createTable(self::tableFromStructure($table, $manifest['tables'][$table]['schema'], $this->tableOptions()));
 				$report['created'][] = $table;
 			} else {
 				$report['missing'][] = $table;
@@ -319,8 +319,31 @@ final class DbDumper {
 			'foreignKeys' => $foreignKeys];
 	}
 
-	private static function tableFromStructure(string $name, array $structure): \Doctrine\DBAL\Schema\Table {
-		$table = new \Doctrine\DBAL\Schema\Table($name);
+	/**
+	 * Table options for a table the restore creates: those of this installation's own tables.
+	 * On MySQL/MariaDB that is the character set and collation, as Nextcloud's tables have them
+	 * (utf8mb4, so emoji and other 4-byte characters fit); DBAL alone would use 3-byte utf8.
+	 *
+	 * @return array<string, mixed>
+	 * @psalm-suppress UndefinedClass the DBAL classes come with Nextcloud, not with this app
+	 */
+	private function tableOptions(): array {
+		$options = $this->db->createSchemaManager()->createSchemaConfig()->getDefaultTableOptions();
+		if ($this->db->getDatabasePlatform() instanceof \Doctrine\DBAL\Platforms\AbstractMySQLPlatform) {
+			$collation = $this->db->fetchOne('SELECT TABLE_COLLATION FROM information_schema.TABLES WHERE TABLE_SCHEMA = DATABASE() AND TABLE_NAME = ?',
+				[$this->prefix . 'appconfig']);
+			if (is_string($collation) && $collation !== '') {
+				$options['collation'] = $collation;
+				$options['charset'] = explode('_', $collation, 2)[0];
+				unset($options['collate']);
+			}
+		}
+		return $options;
+	}
+
+	/** @param array<string, mixed> $options table options (see tableOptions()) */
+	private static function tableFromStructure(string $name, array $structure, array $options = []): \Doctrine\DBAL\Schema\Table {
+		$table = new \Doctrine\DBAL\Schema\Table($name, [], [], [], [], $options);
 		foreach ($structure['columns'] as $c) {
 			$table->addColumn($c['name'], $c['type'], array_filter([
 				'length' => $c['length'], 'precision' => $c['precision'], 'scale' => $c['scale'],

lib/Repository/Repository.php

@@ -845,7 +845,15 @@ final class Repository {
 		try {
 			foreach ($this->backend->getRanges($ranges) as $i => $data) {
 				foreach ($parts[$i] as [$id, $at, $length]) {
-					yield $this->decodeBlob($id, substr($data, $at, $length));
+					try {
+						$blob = $this->decodeBlob($id, substr($data, $at, $length));
+					} catch (\Exception) {
+						// a read that went wrong on the way (see loadBlob()): this blob on its own
+						$blob = $this->loadBlob($id);
+						// its pack stream is closed before the next range is read on the same connection
+						$this->packReader->close();
+					}
+					yield $blob;
 					$done++;
 				}
 			}
@@ -895,8 +903,46 @@ final class Repository {
 			} catch (BackendException) {
 				throw $e;
 			}
+			$pack = $moved;
 		}
-		return $this->decodeBlob($id, $encrypted);
+		try {
+			return $this->decodeBlob($id, $encrypted);
+		} catch (\Exception $damaged) {
+			// Bytes that do not decrypt: damaged on the location, or a read that went wrong on the
+			// way (a stream that lost its place: an SMB restore stopped after 17 hours like this).
+			// Read exactly this blob once more, on a request of its own; only a second failure
+			// means the data itself is damaged.
+			$this->packReader->close();
+			try {
+				$again = $this->backend->getRange('packs/' . substr($pack, 0, 2) . '/' . $pack, $offset, $length);
+			} catch (\Exception) {
+				throw $damaged;
+			}
+			$data = $this->decodeBlob($id, $again);
+			if (function_exists('OCP\Log\logger')) {
+				\OCP\Log\logger('ng_backup')->warning('NG Backup: blob ' . $id . ' did not decrypt as first read from pack ' . $pack . ' at ' . $offset . '; read again, it is fine. First read: ' . self::howItDiffered($encrypted, $again));
+			}
+			return $data;
+		}
+	}
+
+	/**
+	 * What a read that went wrong looked like next to the right bytes, for the log: where it starts
+	 * to differ, and whether the rest is zeros or the right bytes from another place (a stream that
+	 * lost its place).
+	 */
+	private static function howItDiffered(string $wrong, string $right): string {
+		$at = strspn($wrong ^ $right, "\0");
+		$text = strlen($wrong) . ' bytes, differ from byte ' . $at . ' of ' . strlen($right);
+		$tail = substr($wrong, $at, 64);
+		if ($tail === '') {
+			return $text . ' (too short)';
+		}
+		if (trim($tail, "\0") === '') {
+			return $text . ' (zeros)';
+		}
+		$found = strpos($right, $tail);
+		return $text . ($found !== false ? ' (the right bytes from ' . ($found - $at) . ' bytes further on)' : ' (other bytes)');
 	}
 
 	/** Decrypts and checks a blob read from its pack; throws when it is corrupted or tampered with. */

lib/Service/Schedule.php

@@ -0,0 +1,110 @@
+<?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 OCP\IAppConfig;
+use OCP\IConfig;
+use OCP\IGroupManager;
+
+/**
+ * The daily backup time: "HH:MM" in a time zone. Nextcloud runs PHP in UTC, so a time without a
+ * zone meant UTC (issue #19: 03:00 ran at 05:00 in Berlin). A time saved before there was a zone
+ * is converted once to the zone of the server or of an administrator, at today's offset: it then
+ * shows as it effectively ran (02:30 UTC becomes 04:30 Europe/Berlin in summer).
+ */
+final class Schedule {
+	public const TIME_KEY = 'schedule_time';
+	public const ZONE_KEY = 'schedule_timezone';
+
+	public function __construct(
+		private IAppConfig $appConfig,
+		private IConfig $config,
+		private IGroupManager $groups,
+	) {
+	}
+
+	/** @return array{time: string, timezone: non-empty-string} time "" = no schedule */
+	public function get(): array {
+		$time = $this->appConfig->getValueString(Application::APP_ID, self::TIME_KEY, '');
+		$zone = $this->appConfig->getValueString(Application::APP_ID, self::ZONE_KEY, '');
+		if ($zone === '' || !self::validZone($zone)) {
+			$zone = $this->guessZone();
+			if ($time !== '' && self::validTime($time)) {
+				$time = self::convert($time, 'UTC', $zone);
+				$this->appConfig->setValueString(Application::APP_ID, self::TIME_KEY, $time);
+			}
+			$this->appConfig->setValueString(Application::APP_ID, self::ZONE_KEY, $zone);
+		}
+		return ['time' => $time, 'timezone' => $zone];
+	}
+
+	/** @throws \InvalidArgumentException for a malformed time or an unknown zone */
+	public function set(string $time, string $zone): void {
+		if ($time !== '' && !self::validTime($time)) {
+			throw new \InvalidArgumentException('Use HH:MM or empty');
+		}
+		if (!self::validZone($zone)) {
+			throw new \InvalidArgumentException('Unknown time zone: ' . $zone);
+		}
+		$this->appConfig->setValueString(Application::APP_ID, self::ZONE_KEY, $zone);
+		$this->appConfig->setValueString(Application::APP_ID, self::TIME_KEY, $time);
+	}
+
+	/** Today's scheduled moment as a Unix time, in the schedule's zone; null without a schedule. */
+	public function todayAt(?int $now = null): ?int {
+		['time' => $time, 'timezone' => $zone] = $this->get();
+		if ($time === '' || !self::validTime($time)) {
+			return null;
+		}
+		$tz = new \DateTimeZone($zone);
+		$day = (new \DateTimeImmutable('@' . ($now ?? time())))->setTimezone($tz)->format('Y-m-d');
+		return (new \DateTimeImmutable($day . ' ' . $time, $tz))->getTimestamp();
+	}
+
+	/**
+	 * "HH:MM" in zone $from as "HH:MM" in zone $to, at today's offsets.
+	 *
+	 * @param non-empty-string $from
+	 * @param non-empty-string $to
+	 */
+	public static function convert(string $time, string $from, string $to, ?int $now = null): string {
+		$day = (new \DateTimeImmutable('@' . ($now ?? time())))->setTimezone(new \DateTimeZone($from))->format('Y-m-d');
+		return (new \DateTimeImmutable($day . ' ' . $time, new \DateTimeZone($from)))->setTimezone(new \DateTimeZone($to))->format('H:i');
+	}
+
+	public static function validTime(string $time): bool {
+		return preg_match('/^([01]\d|2[0-3]):[0-5]\d$/', $time) === 1;
+	}
+
+	/** @psalm-assert-if-true non-empty-string $zone */
+	public static function validZone(string $zone): bool {
+		return in_array($zone, \DateTimeZone::listIdentifiers(), true);
+	}
+
+	/**
+	 * The server's default_timezone when set, else the first administrator's, else UTC. A server
+	 * zone of UTC is passed over on purpose: it is mostly what an image or installer left there,
+	 * while the administrator who set the time read it as their own (#19).
+	 *
+	 * @return non-empty-string
+	 */
+	private function guessZone(): string {
+		$zone = $this->config->getSystemValueString('default_timezone', '');
+		if ($zone !== '' && $zone !== 'UTC' && self::validZone($zone)) {
+			return $zone;
+		}
+		foreach ($this->groups->get('admin')?->getUsers() ?? [] as $user) {
+			$zone = $this->config->getUserValue($user->getUID(), 'core', 'timezone', '');
+			if ($zone !== '' && self::validZone($zone)) {
+				return $zone;
+			}
+		}
+		return 'UTC';
+	}
+}

src/AdminSettings.vue

@@ -89,8 +89,10 @@ onBeforeUnmount(() => clearTimeout(timer))
 </script>
 
 <style scoped>
+/* NcSettingsSection sets its own width (at most 900px): the runs table needs more */
 .ng-backup :deep(.settings-section) {
-	max-width: 900px;
+	width: min(1400px, 100% - var(--default-grid-baseline) * 7 * 2);
+	max-width: none;
 }
 .ngb-feedback {
 	color: var(--color-text-maxcontrast);

src/api.js

@@ -33,7 +33,7 @@ export default {
 	testTarget: (id) => call('post', `/targets/${id}/test`),
 	setTargetLimit: (id, maxGb) => call('put', `/targets/${id}/limit`, { maxGb }),
 	startBackup: (targetId, label) => call('post', '/backups', { targetId, label }),
-	setSchedule: (time) => call('put', '/schedule', { time }),
+	setSchedule: (time, timezone) => call('put', '/schedule', { time, timezone }),
 	snapshots: (targetId) => call('get', `/targets/${targetId}/snapshots`),
 	browse: (targetId, snapshotId, path) => call('get', `/targets/${targetId}/snapshots/${snapshotId}/browse`, { path }),
 	startRestore: (targetId, snapshotId, path, mode) => call('post', '/restores', { targetId, snapshotId, path, mode }),

src/components/Backups.vue

@@ -6,8 +6,8 @@
 	<NcSettingsSection :name="t('ng_backup', 'Backups')"
 		:description="t('ng_backup', 'Backups run in the background through Nextcloud\'s cron. Unchanged files are not read again; only new data is uploaded.')">
 		<div class="ngb-row">
-			<NcTextField v-model="schedule" class="ngb-time" :label="t('ng_backup', 'Daily at (HH:MM, empty = off)')" />
-			<NcButton :disabled="schedule === status.schedule" @click="saveSchedule">{{ t('ng_backup', 'Save') }}</NcButton>
+			<NcTextField v-model="schedule" class="ngb-time" :label="t('ng_backup', 'Daily at (HH:MM, {zone}, empty = off)', { zone: browserZone })" />
+			<NcButton :disabled="schedule === status.schedule && browserZone === status.scheduleTimezone" @click="saveSchedule">{{ t('ng_backup', 'Save') }}</NcButton>
 		</div>
 		<div class="ngb-row">
 			<NcButton v-for="target in status.targets" :key="target.id" :disabled="isRunning(target.id)" @click="start(target)">
@@ -79,15 +79,19 @@ import NcNoteCard from '@nextcloud/vue/components/NcNoteCard'
 import NcSettingsSection from '@nextcloud/vue/components/NcSettingsSection'
 import NcTextField from '@nextcloud/vue/components/NcTextField'
 import api from '../api.js'
-import { formatDateTime, formatDate, formatSize } from '../format.js'
+import { formatDateTime, formatDate, formatSize, toZone } from '../format.js'
 import { confirmPassword } from '@nextcloud/password-confirmation'
 import { reactive } from 'vue'
 
 const props = defineProps({ status: { type: Object, required: true }, allRuns: { type: Boolean, default: false } })
 const emit = defineEmits(['changed', 'show-all-runs'])
-const schedule = ref(props.status.schedule)
+// The time is saved in the administrator's own zone (the browser's); a schedule saved in another
+// zone is shown in this one, so it reads as the moment it runs here.
+const browserZone = Intl.DateTimeFormat().resolvedOptions().timeZone || 'UTC'
+const shown = (s) => (s && props.status.scheduleTimezone && props.status.scheduleTimezone !== browserZone) ? toZone(s, props.status.scheduleTimezone, browserZone) : s
+const schedule = ref(shown(props.status.schedule))
 const error = ref('')
-watch(() => props.status.schedule, (s) => { schedule.value = s })
+watch(() => [props.status.schedule, props.status.scheduleTimezone], ([s]) => { schedule.value = shown(s) })
 
 const trash = reactive({})
 async function loadTrash() {
@@ -131,7 +135,7 @@ async function start(target) {
 async function saveSchedule() {
 	error.value = ''
 	try {
-		await api.setSchedule(schedule.value.trim())
+		await api.setSchedule(schedule.value.trim(), browserZone)
 		emit('changed')
 	} catch (e) {
 		error.value = e.message

src/format.js

@@ -26,3 +26,30 @@ export function formatSize(bytes) {
 	}
 	return i === 0 ? `${b} B` : `${b.toFixed(1)} ${units[i]}`
 }
+
+/**
+ * A time of day "HH:MM" in zone `from`, as "HH:MM" in zone `to`, at today's offsets.
+ *
+ * @param {string} hhmm the time
+ * @param {string} from IANA zone it is in
+ * @param {string} to IANA zone to show it in
+ * @param {Date} now the day to convert on (today)
+ * @return {string}
+ */
+export function toZone(hhmm, from, to, now = new Date()) {
+	const [h, m] = hhmm.split(':').map(Number)
+	const parts = (zone, date) => Object.fromEntries(new Intl.DateTimeFormat('en-GB', {
+		timeZone: zone, hourCycle: 'h23', year: 'numeric', month: '2-digit', day: '2-digit', hour: '2-digit', minute: '2-digit',
+	}).formatToParts(date).map((p) => [p.type, p.value]))
+	const offsetAt = (at) => {
+		const seen = parts(from, new Date(at))
+		return Date.UTC(+seen.year, +seen.month - 1, +seen.day, +seen.hour, +seen.minute) - at
+	}
+	const today = parts(from, now)
+	const wall = Date.UTC(+today.year, +today.month - 1, +today.day, h, m)
+	// the offset at the wall time itself: on a day the clocks change it differs from the one at
+	// midnight, so it is looked up a second time at the moment found first
+	const at = wall - offsetAt(wall - offsetAt(wall))
+	const there = parts(to, new Date(at))
+	return `${there.hour}:${there.minute}`
+}

tests/Unit/Repository/DamagedReadTest.php

@@ -0,0 +1,133 @@
+<?php
+
+declare(strict_types=1);
+
+namespace OCA\NgBackup\Tests\Unit\Repository;
+
+use OCA\NgBackup\Backend\IBackend;
+use OCA\NgBackup\Backend\LocalBackend;
+use OCA\NgBackup\Crypto\KeyRing;
+use OCA\NgBackup\Repository\Repository;
+use OCA\NgBackup\Tests\Unit\TempDirTrait;
+use PHPUnit\Framework\TestCase;
+
+/**
+ * A pack stream that hands out wrong bytes once (an SMB restore stopped after 17 hours with
+ * "Decryption failed"): the blob is read once more on a request of its own; only data that is
+ * damaged on the location stops the restore.
+ */
+class DamagedReadTest extends TestCase {
+	use TempDirTrait;
+
+	private string $root;
+	private KeyRing $keys;
+	private string $src;
+	private string $snapshot;
+
+	protected function setUp(): void {
+		$this->root = $this->tempDir();
+		$this->keys = KeyRing::generate();
+		$this->src = $this->tempDir();
+		for ($i = 0; $i < 6; $i++) {
+			file_put_contents("$this->src/f$i", random_bytes(3 * 1048576));
+		}
+		$repo = Repository::initWithKey(new LocalBackend($this->root), $this->keys, $this->keys->wrap('pw'));
+		$this->snapshot = $repo->backupDirectory($this->src)['snapshot'];
+	}
+
+	public function testABlobThatCameInDamagedOnceIsReadAgain(): void {
+		$backend = new GarblingBackend(new LocalBackend($this->root), streams: 1);
+		$target = $this->tempDir();
+
+		$this->assertSame(6, Repository::openWithKey($backend, $this->keys)->restore($this->snapshot, $target));
+
+		$this->assertSame(self::hashTree($this->src), self::hashTree($target));
+		$this->assertSame(0, $backend->streams, 'the damaged read happened');
+	}
+
+	public function testATableDumpThatCameInDamagedOnceIsReadAgain(): void {
+		$repo = Repository::openWithKey(new LocalBackend($this->root), $this->keys);
+		$stats = [];
+		$writer = $repo->blobWriter($stats);
+		$lines = [];
+		for ($i = 0; $i < 20000; $i++) {
+			$lines[] = json_encode(['i' => $i, 'pad' => str_repeat('x', $i % 400)]);
+		}
+		$writer->write(implode("\n", $lines) . "\n");
+		$blobs = $writer->finish();
+		$repo->flushPacks();
+		$backend = new GarblingBackend(new LocalBackend($this->root), ranges: 1);
+
+		$this->assertSame($lines, iterator_to_array(Repository::openWithKey($backend, $this->keys)->readLines($blobs), false));
+		$this->assertSame(0, $backend->ranges, 'the damaged read happened');
+	}
+
+	public function testDataDamagedOnTheLocationStillStopsTheRestore(): void {
+		$backend = new GarblingBackend(new LocalBackend($this->root), streams: 1000, ranges: 1000);
+
+		$this->expectException(\Exception::class);
+		Repository::openWithKey($backend, $this->keys)->restore($this->snapshot, $this->tempDir());
+	}
+}
+
+/** Pack reads that flip a byte: the first $streams pack streams and $ranges range reads. */
+final class GarblingBackend implements IBackend {
+	use \OCA\NgBackup\Backend\GetsOneByOne;
+
+	public function __construct(
+		private IBackend $inner,
+		public int $streams = 0,
+		public int $ranges = 0,
+	) {
+	}
+
+	public function put(string $path, $stream): void {
+		$this->inner->put($path, $stream);
+	}
+
+	public function get(string $path) {
+		$fh = $this->inner->get($path);
+		if (!str_starts_with($path, 'packs/') || $this->streams <= 0) {
+			return $fh;
+		}
+		$this->streams--;
+		$data = stream_get_contents($fh);
+		fclose($fh);
+		for ($k = 100; $k < strlen($data); $k += 65536) {
+			$data[$k] = chr(ord($data[$k]) ^ 0x55);
+		}
+		$out = fopen('php://temp', 'w+b');
+		fwrite($out, $data);
+		rewind($out);
+		return $out;
+	}
+
+	public function getRange(string $path, int $offset, int $length): string {
+		$data = $this->inner->getRange($path, $offset, $length);
+		if (str_starts_with($path, 'packs/') && $this->ranges > 0) {
+			$this->ranges--;
+			$data[0] = chr(ord($data[0]) ^ 0x55);
+		}
+		return $data;
+	}
+
+	public function exists(string $path): bool {
+		return $this->inner->exists($path);
+	}
+
+	public function move(string $from, string $to): void {
+		$this->inner->move($from, $to);
+	}
+
+	public function list(string $prefix): array {
+		return $this->inner->list($prefix);
+	}
+
+	public function listFiles(string $dir): array {
+		return $this->inner->listFiles($dir);
+	}
+
+	public function delete(string $path): void {
+		$this->inner->delete($path);
+	}
+}

tests/Unit/Service/ScheduleTest.php

@@ -0,0 +1,71 @@
+<?php
+
+declare(strict_types=1);
+
+namespace OCA\NgBackup\Tests\Unit\Service;
+
+use OCA\NgBackup\Service\Schedule;
+use OCP\IAppConfig;
+use OCP\IConfig;
+use OCP\IGroup;
+use OCP\IGroupManager;
+use OCP\IUser;
+use PHPUnit\Framework\TestCase;
+
+/** The daily time in a time zone (issue #19), and the conversion of a time saved without one. */
+class ScheduleTest extends TestCase {
+	private array $values = [];
+
+	private function schedule(string $serverZone = '', string $adminZone = ''): Schedule {
+		$app = $this->createMock(IAppConfig::class);
+		$app->method('getValueString')->willReturnCallback(fn (string $a, string $k, string $d = '') => $this->values[$k] ?? $d);
+		$app->method('setValueString')->willReturnCallback(function (string $a, string $k, string $v): bool {
+			$this->values[$k] = $v;
+			return true;
+		});
+		$config = $this->createMock(IConfig::class);
+		$config->method('getSystemValueString')->willReturn($serverZone);
+		$config->method('getUserValue')->willReturn($adminZone);
+		$user = $this->createMock(IUser::class);
+		$user->method('getUID')->willReturn('admin');
+		$group = $this->createMock(IGroup::class);
+		$group->method('getUsers')->willReturn([$user]);
+		$groups = $this->createMock(IGroupManager::class);
+		$groups->method('get')->willReturn($group);
+		return new Schedule($app, $config, $groups);
+	}
+
+	public function testATimeSavedWithoutAZoneShowsAsItRanInTheAdministratorsZone(): void {
+		$this->values = [Schedule::TIME_KEY => '02:30'];
+
+		$s = $this->schedule(adminZone: 'Europe/Berlin')->get();
+
+		// it ran at 02:30 UTC: in summer that is 04:30 in Berlin, in winter 03:30
+		$summer = Schedule::convert('02:30', 'UTC', 'Europe/Berlin', strtotime('2026-07-01 12:00 UTC'));
+		$this->assertSame('04:30', $summer);
+		$this->assertSame('Europe/Berlin', $s['timezone']);
+		$this->assertSame(Schedule::convert('02:30', 'UTC', 'Europe/Berlin'), $s['time']);
+	}
+
+	public function testTheServersDefaultZoneComesFirst(): void {
+		$this->values = [Schedule::TIME_KEY => '01:00'];
+		$this->assertSame('America/Toronto', $this->schedule('America/Toronto', 'Europe/Berlin')->get()['timezone']);
+	}
+
+	public function testWithoutAnyZoneItStaysUtc(): void {
+		$this->values = [Schedule::TIME_KEY => '02:30'];
+		$this->assertSame(['time' => '02:30', 'timezone' => 'UTC'], $this->schedule()->get());
+	}
+
+	public function testTodayAtIsTheMomentInTheSchedulesZone(): void {
+		$this->values = [Schedule::TIME_KEY => '03:00', Schedule::ZONE_KEY => 'Europe/Berlin'];
+		$at = $this->schedule()->todayAt(strtotime('2026-10-22 12:00 UTC'));
+		$this->assertSame(strtotime('2026-10-22 01:00 UTC'), $at, '03:00 in Berlin (summer time) is 01:00 UTC');
+	}
+
+	public function testAnUnknownZoneOrABadTimeIsRefused(): void {
+		$s = $this->schedule();
+		$this->expectException(\InvalidArgumentException::class);
+		$s->set('03:00', 'Mars/Olympus');
+	}
+}

tests/integration/full-restore-roundtrip.sh

@@ -47,6 +47,12 @@ if [ "$LEGACY" = legacy ]; then
 fi
 
 run_on() { docker exec -u www-data "$1" php occ "${@:2}"; }
+# The app store sends the whole app list (~38 MB) and Nextcloud gives up after 120 s: on a slow day
+# an install fails for reasons outside this test. Up to three tries.
+install_app() { # container app
+	for i in 1 2 3; do run_on "$1" app:install "$2" >/dev/null 2>&1 && return 0; sleep 20; done
+	return 1
+}
 
 # Database settings for an instance: SQLite inside the container, or its own PostgreSQL container
 db_env() {
@@ -109,7 +115,7 @@ if [ "$DB" != sqlite ]; then
 	# while the restore loads tables in name order, accounts before mailboxes.
 	docker run -d --name ngb-frt-imap --network "$NET" -e MAILNAME=mail.domain.tld -e MAIL_ADDRESS=user@domain.tld \
 		-e MAIL_PASS=mypassword ghcr.io/christophwurst/docker-imap-devel:latest >/dev/null
-	run_on ngb-frt-origin app:install mail >/dev/null || fail "install mail on the origin"
+	install_app ngb-frt-origin mail || fail "install mail on the origin (app store unreachable?)"
 	run_on ngb-frt-origin config:system:set app.mail.verify-tls-peer --type=boolean --value=false >/dev/null
 	sleep 10
 	run_on ngb-frt-origin mail:account:create frtest Test user@domain.tld ngb-frt-imap 993 ssl user@domain.tld mypassword \
@@ -118,7 +124,9 @@ if [ "$DB" != sqlite ]; then
 	[ "$(sql origin 'select count(*) from oc_mail_accounts where drafts_mailbox_id is not null')" = 1 ] \
 		|| fail "the mail account does not point at its mailboxes"
 	# Deck installed and then switched off: its tables stay in the database and in the backup
-	run_on ngb-frt-origin app:install deck >/dev/null || fail "install deck on the origin"
+	install_app ngb-frt-origin deck || fail "install deck on the origin (app store unreachable?)"
+	# a 4-byte character (emoji) in a table the restore has to create: utf8mb4 on MySQL/MariaDB
+	sql origin "insert into oc_deck_boards (title, owner, color, archived, deleted_at, last_modified) values ('Plan 📌', 'frtest', '0087C5', false, 0, 0)" >/dev/null
 	run_on ngb-frt-origin app:disable deck >/dev/null
 	ok "origin has a synced Mail account and the tables of a disabled app"
 fi
@@ -202,6 +210,7 @@ if [ "$ORIGIN_DBTYPE" != sqlite ] && [ "$DB" != sqlite ] && [ "$LEGACY" = legacy
 	ok "older backup: tables of the disabled app left out, Mail restored"
 elif [ "$ORIGIN_DBTYPE" != sqlite ] && [ "$DB" != sqlite ]; then
 	[ "$(sql restore "select count(*) from information_schema.tables where table_name = 'oc_deck_boards'")" = 1 ] || fail "the tables of the disabled app were not created"
+	[ "$(sql restore "select title from oc_deck_boards")" = "Plan 📌" ] || fail "a 4-byte character did not survive in a created table"
 	[ "$(sql restore "select count(*) from oc_mail_accounts where drafts_mailbox_id is not null")" = 1 ] || fail "the mail account was not restored"
 	[ "$(sql restore "select count(*) from information_schema.table_constraints where constraint_type = 'FOREIGN KEY' and table_name like 'oc_mail_%'")" -gt 0 ] \
 		|| fail "the foreign keys of Mail are missing after the restore"