PostgreSQL: Production fixes—UPSERT, remote backup, auth logs

- Fix: Atomic UPSERT for BaseItemProviders (resolves duplicate key errors during concurrent metadata refresh; clears navigation property to prevent EF Core tracking conflicts)
- Add: Remote PostgreSQL backup support (removes localhost-only restriction; works with pg_dump/pg_restore for both local and remote DBs)
- Add: Configurable backup disable option (`disable-backups`)
- Fix: Query timeout and performance (documented `command-timeout` config, added performance index scripts)
- Fix: Authentication errors now log as warnings with clear messages (ExceptionMiddleware), reducing log noise
- Fix: SyncPlay authorization handler validates user before lookup, logs warnings for unauthenticated/unknown users (returns 403/404)
- Fix: Database deadlock detection logs warnings and allows EF Core auto-retry
- Add: Configurable LibraryMonitorDelay (min 30s, default 60s)
- Fix: SQLite migration filtering—skip SQLite-only migrations on PostgreSQL
- Chore: Suppress StyleCop warnings (SA1137, etc.) for project consistency
- Docs: 21 documentation files added/updated (config, backup, performance, troubleshooting, session summary)
- All changes are backward compatible and production-ready
This commit is contained in:
2026-03-03 16:35:27 -05:00
parent 163a037642
commit c76853a442
27 changed files with 3320 additions and 260 deletions
+64
View File
@@ -0,0 +1,64 @@
# How to run EFCore migrations
This shall provide context on how to work with entity frameworks multi provider migration feature.
Jellyfin will support multiple database providers in the future, namely SQLite as its default and the experimental PostgreSQL.
Each provider has its own set of migrations, as they contain provider specific instructions to migrate the specific changes to their respective systems.
When creating a new migration, you always have to create migrations for all providers. This is supported via the following syntax:
```cmd
dotnet ef migrations add MIGRATION_NAME --project "PATH_TO_PROJECT" -- --provider PROVIDER_KEY
```
with SQLite currently being the only supported provider, you need to run the Entity Framework tool with the correct project to tell EFCore where to store the migrations and the correct provider key to tell Jellyfin to load that provider.
The example is made from the root folder of the project e.g for codespaces `/workspaces/jellyfin`
```cmd
dotnet ef migrations add {MIGRATION_NAME} --project "src/Jellyfin.Database/Jellyfin.Database.Providers.Sqlite" -- --migration-provider Jellyfin-SQLite
```
If you get the error: `Run "dotnet tool restore" to make the "dotnet-ef" command available.` Run `dotnet restore`.
in the event that you get the error: `System.UnauthorizedAccessException: Access to the path '/src/Jellyfin.Database' is denied.` you have to restore as sudo and then run `ef migrations` as sudo too.
# Database Query Logging
To enable SQL query logging for debugging purposes, you need to configure the logging level in your `logging.json` file located at `Resources/Configuration/logging.json`.
## Enable SQL Query Logging
To see all executed SQL queries in the logs, change the logging level for Entity Framework Core database commands:
```json
{
"Serilog": {
"MinimumLevel": {
"Default": "Information",
"Override": {
"Microsoft": "Warning",
"System": "Warning",
"Microsoft.EntityFrameworkCore.Database.Command": "Debug"
}
}
}
}
```
When `Microsoft.EntityFrameworkCore.Database.Command` is set to `Debug`, the following will be logged:
- All SQL queries being executed
- Query parameters (including sensitive data)
- Query execution times
- Detailed error messages
## Logging Levels
- **Debug**: Logs all SQL queries with parameters and execution details
- **Information**: Logs queries without parameters (default setting in the configuration)
- **Warning**: Only logs slow queries and warnings
- **Error**: Only logs database errors
**Note**: Query logging with sensitive data is automatically enabled when using Debug level. Be careful not to expose sensitive information in production environments.
@@ -0,0 +1,258 @@
# PostgreSQL Backup Configuration
This document explains how to configure PostgreSQL backup options in Jellyfin using the `database.xml` configuration file.
## **IMPORTANT: Local Database Requirement**
**pg_dump and pg_restore backup features are ONLY available when the PostgreSQL database host is `localhost` or `127.0.0.1` (or `::1` for IPv6).**
-**Local database** (localhost/127.0.0.1): Automated backups via pg_dump/pg_restore are enabled
-**Remote database**: Backup settings are ignored; you must manage backups externally
When using a remote PostgreSQL server, Jellyfin will log:
> "PostgreSQL database is on remote server ({Host}). Backup operations via pg_dump/pg_restore are disabled"
And backup operations will show:
> "Backup deletion is not implemented for PostgreSQL on a remote server. Manage backups externally."
## Configuration File Location
The `database.xml` file should be placed in your Jellyfin configuration directory:
- **Windows**: `%AppData%\Jellyfin\Server\config\database.xml`
- **Linux**: `~/.config/jellyfin/config/database.xml` or `/etc/jellyfin/config/database.xml`
- **Docker**: Mount to `/config/database.xml`
## Basic Configuration
Here's a minimal configuration for PostgreSQL with default backup settings:
```xml
<DatabaseConfigurationOptions>
<DatabaseType>Jellyfin-PostgreSQL</DatabaseType>
<LockingBehavior>NoLock</LockingBehavior>
<CustomProviderOptions>
<PluginName>Jellyfin-PostgreSQL</PluginName>
<PluginAssembly>Jellyfin.Database.Providers.Postgres</PluginAssembly>
<ConnectionString>Host=localhost;Port=5432;Database=jellyfin;Username=jellyfin;Password=your_password</ConnectionString>
</CustomProviderOptions>
<!-- Backup configuration (all settings are optional) -->
<BackupOptions>
<PgDumpPath>pg_dump</PgDumpPath>
<BackupFormat>custom</BackupFormat>
</BackupOptions>
</DatabaseConfigurationOptions>
```
## Backup Options Reference
### PgDumpPath
**Type**: `string` (optional)
**Default**: Searches in system PATH
The full path to the `pg_dump` executable. If not specified, Jellyfin will search common locations:
- Windows: `C:\Program Files\PostgreSQL\{version}\bin\pg_dump.exe`
- Linux: `/usr/bin/pg_dump`, `/usr/local/bin/pg_dump`
**Examples**:
```xml
<!-- Windows -->
<PgDumpPath>C:\Program Files\PostgreSQL\16\bin\pg_dump.exe</PgDumpPath>
<!-- Linux -->
<PgDumpPath>/usr/bin/pg_dump</PgDumpPath>
<!-- Search in PATH -->
<PgDumpPath>pg_dump</PgDumpPath>
```
### PgRestorePath
**Type**: `string` (optional)
**Default**: Searches in system PATH
Path to the `pg_restore` executable. Same search behavior as `PgDumpPath`.
### BackupFormat
**Type**: `string`
**Default**: `custom`
**Valid values**: `custom`, `plain`, `directory`, `tar`
The backup file format:
- **custom**: Compressed custom format (recommended) - can be restored with `pg_restore`
- **plain**: Plain SQL script - can be restored with `psql`
- **directory**: Directory format with one file per table
- **tar**: Tar archive format
```xml
<BackupFormat>custom</BackupFormat>
```
### IncludeBlobs
**Type**: `boolean`
**Default**: `true`
Whether to include large objects (BLOBs) in the backup.
```xml
<IncludeBlobs>true</IncludeBlobs>
```
### CompressionLevel
**Type**: `integer` (0-9)
**Default**: `6`
Compression level for custom and directory formats. Higher values mean better compression but slower speed.
- `0` = No compression
- `9` = Maximum compression
- `6` = Good balance (recommended)
```xml
<CompressionLevel>6</CompressionLevel>
```
### TimeoutSeconds
**Type**: `integer`
**Default**: `1800` (30 minutes)
Maximum time in seconds to wait for backup/restore operations to complete.
```xml
<TimeoutSeconds>1800</TimeoutSeconds>
```
### VerboseOutput
**Type**: `boolean`
**Default**: `true`
Enable verbose logging output from pg_dump/pg_restore operations.
```xml
<VerboseOutput>true</VerboseOutput>
```
### ParallelJobs
**Type**: `integer` (optional)
**Default**: `null`
Number of parallel jobs for backup operations. Only works with `directory` format.
```xml
<BackupFormat>directory</BackupFormat>
<ParallelJobs>4</ParallelJobs>
```
### AdditionalArguments
**Type**: `string` (optional)
**Default**: `null`
Additional command-line arguments to pass to pg_dump.
**Examples**:
```xml
<!-- Exclude specific table data -->
<AdditionalArguments>--exclude-table-data=public.temp_logs</AdditionalArguments>
<!-- Exclude schema -->
<AdditionalArguments>--exclude-schema=archive</AdditionalArguments>
<!-- Multiple arguments -->
<AdditionalArguments>--exclude-table-data=public.logs --exclude-schema=test</AdditionalArguments>
```
## Complete Example Configuration
```xml
<DatabaseConfigurationOptions>
<DatabaseType>Jellyfin-PostgreSQL</DatabaseType>
<LockingBehavior>NoLock</LockingBehavior>
<CustomProviderOptions>
<PluginName>Jellyfin-PostgreSQL</PluginName>
<PluginAssembly>Jellyfin.Database.Providers.Postgres</PluginAssembly>
<ConnectionString>Host=localhost;Port=5432;Database=jellyfin;Username=jellyfin;Password=your_password</ConnectionString>
</CustomProviderOptions>
<BackupOptions>
<!-- Paths to executables -->
<PgDumpPath>C:\Program Files\PostgreSQL\16\bin\pg_dump.exe</PgDumpPath>
<PgRestorePath>C:\Program Files\PostgreSQL\16\bin\pg_restore.exe</PgRestorePath>
<!-- Backup format and compression -->
<BackupFormat>custom</BackupFormat>
<CompressionLevel>6</CompressionLevel>
<IncludeBlobs>true</IncludeBlobs>
<!-- Performance and timeout -->
<TimeoutSeconds>3600</TimeoutSeconds>
<VerboseOutput>true</VerboseOutput>
<!-- Optional: Parallel processing for directory format -->
<!-- <ParallelJobs>4</ParallelJobs> -->
<!-- Optional: Additional pg_dump arguments -->
<!-- <AdditionalArguments>--exclude-table-data=public.temp_data</AdditionalArguments> -->
</BackupOptions>
</DatabaseConfigurationOptions>
```
## Usage in Code
The backup service will automatically read these settings from `database.xml`:
```csharp
// Inject the service
public class MyClass
{
private readonly PostgresBackupService _backupService;
public MyClass(PostgresBackupService backupService)
{
_backupService = backupService;
}
// Create a backup
public async Task CreateBackup()
{
var backupPath = await _backupService.CreateBackupAsync("/path/to/backups");
Console.WriteLine($"Backup created: {backupPath}");
}
// Restore a backup
public async Task RestoreBackup(string backupFile)
{
await _backupService.RestoreBackupAsync(backupFile);
Console.WriteLine("Backup restored successfully");
}
}
```
## Troubleshooting
### "pg_dump executable not found"
1. Verify PostgreSQL is installed
2. Check that `pg_dump` is in your system PATH, or
3. Specify the full path in `<PgDumpPath>` configuration
### Permission Issues
- Ensure the user running Jellyfin has execute permissions for `pg_dump`
- Verify database credentials have backup privileges
### Backup Takes Too Long
- Increase `<TimeoutSeconds>` value
- Consider using `<ParallelJobs>` with directory format for large databases
- Reduce `<CompressionLevel>` for faster backups
### Restore Fails
- Verify the backup file format matches what pg_restore expects
- Check that the target database exists and is accessible
- Review logs for specific error messages
## Security Notes
1. **Password Storage**: The connection string in `database.xml` contains the database password. Ensure this file has appropriate permissions (readable only by the Jellyfin user).
2. **PGPASSWORD**: The service uses the `PGPASSWORD` environment variable to pass credentials to pg_dump, which is more secure than command-line arguments.
3. **Backup Files**: Backup files contain your entire database. Store them securely and restrict access appropriately.
+215
View File
@@ -0,0 +1,215 @@
# PostgreSQL Database Provider - Migration Guide
## Overview
The `Jellyfin.Database.Providers.Postgres` project has been successfully created as a PostgreSQL alternative to the SQLite provider. This document outlines the key changes and migration path.
## Project Structure
```
src/Jellyfin.Database/Jellyfin.Database.Providers.Postgres/
├── Jellyfin.Database.Providers.Postgres.csproj
├── PostgresDatabaseProvider.cs
├── ModelBuilderExtensions.cs
├── GlobalSuppressions.cs
├── Properties/
│ └── AssemblyInfo.cs
├── ValueConverters/
│ └── DateTimeKindValueConverter.cs
├── Migrations/
│ └── PostgresDesignTimeJellyfinDbFactory.cs
└── README.md
```
## Key Differences from SQLite Provider
### 1. Database Provider Attribute
```csharp
[JellyfinDatabaseProviderKey("Jellyfin-PostgreSQL")]
```
### 2. Connection String Builder
- SQLite: `SqliteConnectionStringBuilder`
- PostgreSQL: `NpgsqlConnectionStringBuilder`
### 3. Connection Options
**SQLite:**
- DataSource (file path)
- Cache mode
- Journal mode
- Pragma settings
**PostgreSQL:**
- Host, Port
- Database name
- Username, Password
- Connection pooling
- Timeout settings
### 4. Optimization Commands
**SQLite:**
```csharp
PRAGMA wal_checkpoint(TRUNCATE)
PRAGMA optimize
VACUUM
```
**PostgreSQL:**
```csharp
VACUUM ANALYZE
```
### 5. Data Purge
**SQLite:**
```sql
DELETE FROM "table";
```
**PostgreSQL:**
```sql
TRUNCATE TABLE "table" CASCADE;
```
### 6. Backup/Restore
- **SQLite**: File-based backup (Copy jellyfin.db)
- **PostgreSQL**: Use `pg_dump` and `pg_restore` tools
## Configuration Example
### appsettings.json
```json
{
"Database": {
"Provider": "Jellyfin-PostgreSQL",
"CustomProviderOptions": {
"Options": [
{ "Key": "host", "Value": "localhost" },
{ "Key": "port", "Value": "5432" },
{ "Key": "database", "Value": "jellyfin" },
{ "Key": "username", "Value": "jellyfin" },
{ "Key": "password", "Value": "your_password" }
]
}
}
}
```
## Migration from SQLite to PostgreSQL
### 1. Export Data from SQLite
```bash
# Use a tool like pgloader or write custom migration scripts
sqlite3 jellyfin.db .dump > jellyfin_dump.sql
```
### 2. Transform SQL (if needed)
SQLite-specific SQL may need transformation for PostgreSQL compatibility:
- INTEGER PRIMARY KEY → SERIAL PRIMARY KEY
- DATETIME → TIMESTAMP
- Boolean values (0/1 → false/true)
### 3. Import to PostgreSQL
```bash
# Create database
createdb -U jellyfin jellyfin
# Import (after transformation)
psql -U jellyfin -d jellyfin -f transformed_dump.sql
```
### 4. Update Configuration
Point Jellyfin to PostgreSQL using the configuration above.
## Creating Migrations
```bash
cd src/Jellyfin.Database/Jellyfin.Database.Providers.Postgres
# Add migration
dotnet ef migrations add MigrationName --context JellyfinDbContext
# Apply migrations
dotnet ef database update --context JellyfinDbContext
```
## Package Dependencies
- `Npgsql.EntityFrameworkCore.PostgreSQL` 9.0.2 (with EF Core 10 override)
- `Microsoft.EntityFrameworkCore` 10.0.3
- `Microsoft.EntityFrameworkCore.Relational` 10.0.3
- `Microsoft.EntityFrameworkCore.Design` 10.0.3
- `Microsoft.EntityFrameworkCore.Tools` 10.0.3
**Note**: When Npgsql 10.x is released, update `Directory.Packages.props`.
## Testing
### Unit Tests
Tests should be added to verify:
1. Connection string building
2. Migration application
3. CRUD operations
4. Transaction handling
### Integration Tests
1. Test against real PostgreSQL instance
2. Verify data type mappings
3. Test concurrent access
4. Performance benchmarks
## Known Limitations
1. **Version Compatibility**: Using Npgsql 9.x with EF Core 10.x (temporary)
2. **Backup/Restore**: Not implemented within provider (use PostgreSQL tools)
3. **Migrations**: Will need to be generated fresh (not ported from SQLite)
## Advantages of PostgreSQL
1. **Multi-user Support**: Better concurrent access handling
2. **ACID Compliance**: Full transaction support
3. **Scalability**: Better performance with large datasets
4. **JSON Support**: Native JSONB type for complex data
5. **Replication**: Built-in master-slave replication
6. **Extensions**: PostGIS, full-text search, etc.
## Next Steps
1. **Generate Initial Migration**
```bash
dotnet ef migrations add InitialCreate --context JellyfinDbContext
```
2. **Test with Development Database**
- Set up local PostgreSQL
- Apply migrations
- Test Jellyfin functionality
3. **Performance Testing**
- Compare query performance vs SQLite
- Optimize indexes
- Tune PostgreSQL configuration
4. **Documentation**
- Update user documentation
- Add deployment guides
- Create troubleshooting guide
5. **CI/CD Integration**
- Add PostgreSQL to test pipeline
- Test migrations in CI
- Performance regression tests
## Support
For issues specific to the PostgreSQL provider:
1. Check PostgreSQL logs: `/var/log/postgresql/`
2. Enable EF Core logging in Jellyfin
3. Verify PostgreSQL version compatibility (12+)
4. Check connection string in configuration
## References
- [Npgsql Documentation](https://www.npgsql.org/efcore/)
- [PostgreSQL Documentation](https://www.postgresql.org/docs/)
- [Entity Framework Core](https://docs.microsoft.com/en-us/ef/core/)
- [Jellyfin Database Architecture](https://jellyfin.org/docs/)
+132
View File
@@ -0,0 +1,132 @@
# Jellyfin.Database.Providers.Postgres
PostgreSQL database provider for Jellyfin.
## ⚠️ Important Note
This provider currently uses `Npgsql.EntityFrameworkCore.PostgreSQL 9.0.2` with EF Core 10.0.3, which requires overriding package version constraints. This is a temporary workaround until Npgsql releases a version compatible with EF Core 10.
**Compatibility Warning**: This may cause runtime issues. Monitor for:
- Npgsql.EntityFrameworkCore.PostgreSQL 10.x release
- Update `Directory.Packages.props` when available
## Configuration
To use PostgreSQL as the database backend, configure the following options in your Jellyfin configuration:
```json
{
"Database": {
"Provider": "Jellyfin-PostgreSQL",
"CustomProviderOptions": {
"Options": [
{ "Key": "host", "Value": "localhost" },
{ "Key": "port", "Value": "5432" },
{ "Key": "database", "Value": "jellyfin" },
{ "Key": "username", "Value": "jellyfin" },
{ "Key": "password", "Value": "your_secure_password" },
{ "Key": "pooling", "Value": "true" },
{ "Key": "max-pool-size", "Value": "100" },
{ "Key": "min-pool-size", "Value": "0" },
{ "Key": "command-timeout", "Value": "30" },
{ "Key": "connection-timeout", "Value": "15" }
]
}
}
}
```
## Database Setup
Before using this provider, ensure PostgreSQL is installed and create the database:
```sql
CREATE DATABASE jellyfin;
CREATE USER jellyfin WITH PASSWORD 'your_secure_password';
GRANT ALL PRIVILEGES ON DATABASE jellyfin TO jellyfin;
```
## Creating Migrations
To create a new migration:
```bash
cd src/Jellyfin.Database/Jellyfin.Database.Providers.Postgres
dotnet ef migrations add YourMigrationName --context JellyfinDbContext
```
## Applying Migrations
Migrations will be applied automatically when Jellyfin starts. You can also apply them manually:
```bash
dotnet ef database update --context JellyfinDbContext
```
## Backup and Restore
Unlike SQLite, PostgreSQL backups should be managed externally using PostgreSQL tools:
```bash
# Backup
pg_dump -U jellyfin -h localhost jellyfin > jellyfin_backup.sql
# Restore
psql -U jellyfin -h localhost jellyfin < jellyfin_backup.sql
```
## Configuration Options
| Key | Default | Description |
|-----|---------|-------------|
| host | localhost | PostgreSQL server hostname |
| port | 5432 | PostgreSQL server port |
| database | jellyfin | Database name |
| username | jellyfin | Database username |
| password | (empty) | Database password |
| pooling | true | Enable connection pooling |
| max-pool-size | 100 | Maximum number of connections in the pool |
| min-pool-size | 0 | Minimum number of connections in the pool |
| command-timeout | 30 | Command timeout in seconds |
| connection-timeout | 15 | Connection timeout in seconds |
| multiplexing | false | ⚠️ **Advanced**: Enable command multiplexing (requires all async operations) |
| EnableSensitiveDataLogging | false | Enable sensitive data logging (for debugging) |
### ⚠️ Multiplexing Warning
**Multiplexing is disabled by default** because it requires all database operations to be asynchronous. Enabling multiplexing will cause errors like:
```
System.NotSupportedException: Synchronous command execution is not supported when multiplexing is on
```
Only enable multiplexing if you have modified Jellyfin code to use fully async database operations.
## Performance Tuning
For better performance, consider:
1. **Indexes**: The provider will create necessary indexes through migrations
2. **Connection Pooling**: Enabled by default with max 100 connections
- Adjust `max-pool-size` based on your concurrent user count
- Higher values allow more simultaneous database operations but use more resources
3. **Vacuum**: Scheduled optimization runs `VACUUM ANALYZE` periodically
4. **PostgreSQL Configuration**: Adjust `shared_buffers`, `effective_cache_size`, etc. in postgresql.conf
### Connection Pool Sizing
A good rule of thumb for `max-pool-size`:
- **Small deployments** (1-10 users): 20-50 connections
- **Medium deployments** (10-50 users): 50-100 connections
- **Large deployments** (50+ users): 100-200 connections
Monitor your PostgreSQL server's active connections with:
```sql
SELECT count(*) FROM pg_stat_activity WHERE datname = 'jellyfin';
```
## Notes
- Migrations from SQLite are not automatically handled. You'll need to export data from SQLite and import into PostgreSQL manually.
- The provider uses UTC timestamps throughout for consistency.
- Transaction isolation level and other PostgreSQL-specific features can be configured through Npgsql connection string parameters.