SiaFoundation / renterd

A renter for Sia
https://sia.tech/software/renterd
MIT License
68 stars 20 forks source link
blockchain decentralized-storage golang hacktoberfest storage

Sia

GoDoc

renterd is an advanced Sia renter engineered by the Sia Foundation. Designed to cater to both casual users seeking straightforward data storage and developers requiring a robust API for building apps on Sia

Overview

renterd is the successor to siad, offering feature parity while extending its capabilities with new features like an enhanced web UI and an autopilot. That said, renterd does not support backwards compatibility with siad metadata. Consequently, files uploaded via siad cannot currently be migrated to renterd. Our immediate focus is on refining renterd to enhance its stability, scalability, and performance, ensuring it serves as a robust foundation for new Sia applications. Useful links:

Database

renterd requires a database to store its operational data. We support both SQLite and MySQL, with SQLite set as the default due to its ease of setup. SQLite is ideal for testing and development purposes, whereas MySQL is recommended for production environments.

Configuration

renterd can be configured in various ways, through the use of a yaml file, CLI flags or environment variables. Settings that are configured multiple times will be evaluated in this order. In the CLI, use the help command to see an overview of all settings configurable through the CLI.

Name Description Default Value CLI Flag Environment Variable YAML Path
HTTP.Address Address for serving the API :9980 --http - http.address
HTTP.Password Password for the HTTP server - - RENTERD_API_PASSWORD http.password
Directory Directory for storing node state . --dir - directory
Seed Seed for the node - - RENTERD_SEED seed
AutoOpenWebUI Automatically open the web UI on startup true --openui - autoOpenWebUI
Network Network to run on (mainnet/zen/anagami) mainnet --network RENTERD_NETWORK network
ShutdownTimeout Timeout for node shutdown 5m --node.shutdownTimeout - shutdownTimeout
Log.Level Global logger level (debug|info|warn|error). Defaults to 'info' info --log.level RENTERD_LOG_LEVEL log.level
Log.File.Enabled Enables logging to disk. Defaults to 'true' true --log.file.enabled RENTERD_LOG_FILE_ENABLED log.file.enabled
Log.File.Format Format of log file (json|human). Defaults to 'json' json --log.file.format RENTERD_LOG_FILE_FORMAT log.file.format
Log.File.Path Path of log file. Defaults to 'renterd.log' within the renterd directory renterd.log --log.file.path RENTERD_LOG_FILE_PATH log.file.path
Log.StdOut.Enabled Enables logging to stdout. Defaults to 'true' true --log.stdout.enabled RENTERD_LOG_STDOUT_ENABLED log.stdout.enabled
Log.StdOut.Format Format of log output (json|human). Defaults to 'human' human --log.stdout.format RENTERD_LOG_STDOUT_FORMAT log.stdout.format
Log.StdOut.EnableANSI Enables ANSI color codes in log output. Defaults to 'true' on non-Windows systems true (false on Windows) --log.stdout.enableANSI RENTERD_LOG_STDOUT_ENABLE_ANSI log.stdout.enableANSI
Log.Database.Enabled Enable logging database queries. Defaults to 'true' true --log.database.enabled RENTERD_LOG_DATABASE_ENABLED log.database.enabled
Log.Database.Level Logger level for database queries (info|warn|error). Defaults to 'warn' warn --log.database.level RENTERD_LOG_DATABASE_LEVEL, RENTERD_LOG_LEVEL log.database.level
Log.Database.IgnoreRecordNotFoundError Enable ignoring 'not found' errors resulting from database queries. Defaults to 'true' true --log.database.ignoreRecordNotFoundError RENTERD_LOG_DATABASE_IGNORE_RECORD_NOT_FOUND_ERROR log.database.ignoreRecordNotFoundError
Log.Database.SlowThreshold Threshold for slow queries in logger. Defaults to 100ms 100ms --log.database.slowThreshold RENTERD_LOG_DATABASE_SLOW_THRESHOLD log.database.slowThreshold
Database.MySQL.URI Database URI for the bus - --db.uri RENTERD_DB_URI database.mysql.uri
Database.MySQL.User Database username for the bus renterd --db.user RENTERD_DB_USER database.mysql.user
Database.MySQL.Password Database password for the bus - - RENTERD_DB_PASSWORD database.mysql.password
Database.MySQL.Database Database name for the bus renterd --db.name RENTERD_DB_NAME database.mysql.database
Database.MySQL.MetricsDatabase Database for metrics renterd_metrics --db.metricsName RENTERD_DB_METRICS_NAME database.mysql.metricsDatabase
Database.SQLite.Database SQLite database name - - - database.sqlite.database
Database.SQLite.MetricsDatabase SQLite metrics database name - - - database.sqlite.metricsDatabase
Bus.AllowPrivateIPs Allows hosts with private IPs - --bus.allowPrivateIPs - bus.allowPrivateIPs
Bus.AnnouncementMaxAgeHours Max age for announcements 8760h (1 year) --bus.announcementMaxAgeHours - bus.announcementMaxAgeHours
Bus.Bootstrap Bootstraps gateway and consensus modules true --bus.bootstrap - bus.bootstrap
Bus.GatewayAddr Address for Sia peer connections :9981 --bus.gatewayAddr RENTERD_BUS_GATEWAY_ADDR bus.gatewayAddr
Bus.RemoteAddr Remote address for the bus - - RENTERD_BUS_REMOTE_ADDR bus.remoteAddr
Bus.RemotePassword Remote password for the bus - - RENTERD_BUS_API_PASSWORD bus.remotePassword
Bus.UsedUTXOExpiry Expiry for used UTXOs in transactions 24h --bus.usedUTXOExpiry - bus.usedUtxoExpiry
Bus.SlabBufferCompletionThreshold Threshold for slab buffer upload 4096 --bus.slabBufferCompletionThreshold RENTERD_BUS_SLAB_BUFFER_COMPLETION_THRESHOLD bus.slabBufferCompletionThreshold
Worker.BusFlushInterval Interval for flushing data to bus 5s --worker.busFlushInterval - worker.busFlushInterval
Worker.DownloadMaxOverdrive Max overdrive workers for downloads 5 --worker.downloadMaxOverdrive - worker.downloadMaxOverdrive
Worker.DownloadMaxMemory Max memory for downloads 1GiB --worker.downloadMaxMemory RENTERD_WORKER_DOWNLOAD_MAX_MEMORY worker.downloadMaxMemory
Worker.ID Unique ID for worker worker --worker.id RENTERD_WORKER_ID worker.id
Worker.DownloadOverdriveTimeout Timeout for overdriving slab downloads 3s --worker.downloadOverdriveTimeout - worker.downloadOverdriveTimeout
Worker.UploadMaxMemory Max amount of RAM the worker allocates for slabs when uploading 1GiB --worker.uploadMaxMemory RENTERD_WORKER_UPLOAD_MAX_MEMORY worker.uploadMaxMemory
Worker.UploadMaxOverdrive Max overdrive workers for uploads 5 --worker.uploadMaxOverdrive - worker.uploadMaxOverdrive
Worker.UploadOverdriveTimeout Timeout for overdriving slab uploads 3s --worker.uploadOverdriveTimeout - worker.uploadOverdriveTimeout
Worker.Enabled Enables/disables worker true --worker.enabled RENTERD_WORKER_ENABLED worker.enabled
Worker.AllowUnauthenticatedDownloads Allows unauthenticated downloads - --worker.unauthenticatedDownloads RENTERD_WORKER_UNAUTHENTICATED_DOWNLOADS worker.allowUnauthenticatedDownloads
Worker.ExternalAddress Address of the worker on the network, only necessary when the bus is remote - - RENTERD_WORKER_EXTERNAL_ADDR worker.externalAddress
Worker.RemoteAddrs List of remote worker addresses (semicolon delimited) - - RENTERD_WORKER_REMOTE_ADDRS worker.remotes
Worker.RemotePassword API password for the remote workers - - RENTERD_WORKER_API_PASSWORD worker.remotes
Autopilot.Enabled Enables/disables autopilot true --autopilot.enabled RENTERD_AUTOPILOT_ENABLED autopilot.enabled
Autopilot.AccountsRefillInterval Interval for refilling workers' account balances 24h --autopilot.accountRefillInterval - autopilot.accountsRefillInterval
Autopilot.Heartbeat Interval for autopilot loop execution 30m --autopilot.heartbeat - autopilot.heartbeat
Autopilot.MigrationHealthCutoff Threshold for migrating slabs based on health 0.75 --autopilot.migrationHealthCutoff - autopilot.migrationHealthCutoff
Autopilot.RevisionBroadcastInterval Interval for broadcasting contract revisions 168h (7 days) --autopilot.revisionBroadcastInterval RENTERD_AUTOPILOT_REVISION_BROADCAST_INTERVAL autopilot.revisionBroadcastInterval
Autopilot.ScannerBatchSize Batch size for host scanning 1000 --autopilot.scannerBatchSize - autopilot.scannerBatchSize
Autopilot.ScannerInterval Interval for scanning hosts 24h --autopilot.scannerInterval - autopilot.scannerInterval
Autopilot.ScannerNumThreads Number of threads for scanning hosts 100 - - autopilot.scannerNumThreads
Autopilot.MigratorParallelSlabsPerWorker Parallel slab migrations per worker 1 --autopilot.migratorParallelSlabsPerWorker RENTERD_MIGRATOR_PARALLEL_SLABS_PER_WORKER autopilot.migratorParallelSlabsPerWorker
S3.Address Address for serving S3 API :9982 --s3.address RENTERD_S3_ADDRESS s3.address
S3.DisableAuth Disables authentication for S3 API false --s3.disableAuth RENTERD_S3_DISABLE_AUTH s3.disableAuth
S3.Enabled Enables/disables S3 API true --s3.enabled RENTERD_S3_ENABLED s3.enabled
S3.HostBucketBases Enables bucket rewriting in the router for the provided bases - --s3.hostBucketBases RENTERD_S3_HOST_BUCKET_BASES s3.hostBucketBases
S3.HostBucketEnabled Enables bucket rewriting in the router - --s3.hostBucketEnabled RENTERD_S3_HOST_BUCKET_ENABLED s3.hostBucketEnabled
Explorer.Disable Disables explorer service false --explorer.disable RENTERD_EXPLORER_DISABLE explorer.disable
Explorer.URL URL of service to retrieve data about the Sia network https://api.siascan.com --explorer.url RENTERD_EXPLORER_URL explorer.url

Single-Node Setup

A single-node setup involves running all components (bus, worker, and autopilot) on the same machine. This is ideal for testing, development, or small-scale deployments. This setup is the default when running renterd without any flags.

Cluster Setup

In a cluster setup, the bus, worker, and autopilot run on separate nodes. This setup is ideal for large-scale deployments where you want to horizontally scale your renter. The worker nodes can be spread across multiple machines, and the autopilot can be run on a separate machine.

Bus Node Configuration

The bus is the only node that exposes the UI. To run the bus separately, the autopilot and worker have to be disabled using the --autopilot.enabled and --worker.enabled flags. The only other requirement to run a bus is the (walet) seed.

Worker Node Configuration

To configure the worker as a standalone node, the autopilot has to be disabled using the --autopilot.enabled flag, and the bus has to be disabled. There's no flag to explicitly disable the bus, it's implied by configuring a remote address for the bus using the --bus.remoteAddr and --bus.remotePassword flags. When the bus is remote, the worker has to be configured with an external address of the form http://<worker-ip>:<port>, on localhost however this can be the same as the worker's HTTP address. The worker needs to know its location on the network because it relies on some webhooks it needs to register with the bus, which in turn needs to know how to reach the worker when certain events occur. Therefor it is important to start the worker after the bus is reachable.

Autopilot Node Configuration

To run the autopilot separately, the worker has to be disabled using the --worker.enabled flag. Similar to the worker, the autopilot has to be configured with a remote bus for the node not to start a bus itself. Alongside with knowing where the bus is located, the autopilot also has to be aware of the workers. These remote workers can be configured through yaml under the option worker.remotes, or through environment variables (RENTERD_WORKER_REMOTE_ADDRS and RENTERD_WORKER_API_PASSWORD).

Example docker-compose with minimal configuration

version: '3.9'

services:
  bus:
    image: ghcr.io/siafoundation/renterd:master
    container_name: renterd_bus
    environment:
      - RENTERD_SEED=<enter seed here>
      - RENTERD_API_PASSWORD=bus-pass
    ports:
      - "9980:9980"
      - "9981:9981"

  worker-1:
    image: ghcr.io/siafoundation/renterd:master
    container_name: renterd_worker-1
    environment:
      - RENTERD_AUTOPILOT_ENABLED=false
      - RENTERD_SEED=<enter seed here>
      - RENTERD_API_PASSWORD=worker-pass
      - RENTERD_BUS_API_PASSWORD=bus-pass
      - RENTERD_BUS_REMOTE_ADDR=http://bus:9980/api/bus
      - RENTERD_WORKER_EXTERNAL_ADDR=http://worker-1:9980/api/worker
    ports:
      - "9982:9980"
      - "8082:8080"
    depends_on:
      - bus

  worker-2:
    image: ghcr.io/siafoundation/renterd:master
    container_name: renterd_worker-2
    environment:
      - RENTERD_SEED=<enter seed here>
      - RENTERD_API_PASSWORD=worker-pass
      - RENTERD_BUS_API_PASSWORD=bus-pass
      - RENTERD_BUS_REMOTE_ADDR=http://bus:9980/api/bus
      - RENTERD_WORKER_EXTERNAL_ADDR=http://worker-2:9980/api/worker
    ports:
      - "9983:9980"
      - "8083:8080"
    depends_on:
      - bus

  autopilot:
    image: ghcr.io/siafoundation/renterd:master
    container_name: renterd_autopilot
    environment:
      - RENTERD_API_PASSWORD=autopilot-pass
      - RENTERD_BUS_API_PASSWORD=bus-pass
      - RENTERD_BUS_REMOTE_ADDR=http://bus:9980/api/bus
      - RENTERD_WORKER_API_PASSWORD=<worker-password>
      - RENTERD_WORKER_REMOTE_ADDRS=http://worker-1:9980/api/worker;http://worker-2:9980/api/worker
    ports:
      - "9984:9980"
    depends_on:
      - bus
      - worker-1
      - worker-2

Tweaking Performance

Depending on hardware specs, you can change the configuration to better utilize it and gain more performance out of renterd. This section highlights some of the more obvious tweaks one can apply.

Increase/Decrease memory

By default, renterd uses reasonable limits for RAM consumed by uploads and downloads. Especially when downloading or uploading single large files, more RAM can make a difference since it allows for processing the download/upload in parallel. To change the max RAM renterd is going to use update the Worker.DownloadMaxMemory and Worker.UploadMaxMemory settings.

Overdrive

Both uploads and downloads have a setting we call "overdrive". Since renterd operates in a trustless environment, we can't rely on all of our hosts being reliable and of high quality. So when uploading n shards of some data to the network (or downloading from it), the process is bottlenecked by the slowest host. That is where the overdrive comes in.

Worker.UploadMaxOverdrive and Worker.DownloadMaxOverdrive can be used to configure how many additional hosts to the number we need to upload/download we use to reduce the chance of getting hung up on a slow one. The default is 3 which means up to 3 hosts can get stuck with the upload/download remaining mostly unaffected. Worker.UploadOverdriveTimeout and Worker.DownloadOverdriveTimeout specify the time that needs to pass before we launch the overdrive uploads/downloads.

Two conditions need to be met before the overdrive launches:

  1. When uploading/downloading to/from n hosts (without overdrive), n - overdriveHosts pieces need to finish.
  2. Once condition 1. is met, the configured overdrive timeout needs to pass

What this means is that there is a tradeoff between using/paying for more bandwidth and the ability to compensate for slow/stuck hosts. If you handpick hosts you trust to be reliable, you can set the max overdrive to 0 for more max efficiency while you can also increase the overdrive to 30 hosts after 100ms for faster uploads at the cost of uploading more data than necessary and overpaying. Regardless, we recommend that you perform your own benchmarking to see what works best for your set of hosts, budget and use-case.

Backups

This section provides a step-by-step guide covering the procedures for creating and restoring backups for renterd. Regularly backing up your renter's metadata and verifying its restorability are essential practices to ensure your data's integrity and availability in the event of a disaster scenario. Ensure backups are stored on a different machine than the one running renterd to avoid data loss in case of hardware failure.


IMPORTANT NOTE

It is important to note that having a backup is not enough to ensure the data can be recovered from the network. The renter has to be online enough of the time to ensure data gets migrated away from hosts that went offline.


Creating a backup

Step 1: shut down renter

It's strongly recommended to shut down the renter before creating a backup to ensure data consistency. This precaution addresses the "upload packing" feature (enabled by default), which relies on on-disk buffers. To capture a consistent state between the database and on-disk data, shut down renterd first. Even if the feature is not enabled, it is best to shut down the renter before taking a backup to be on the safe side and ensure consistency.

Step 2: backing up the database

renterd uses two SQL databases: a main database for contracts, host metadata and object metadata essential for data retrieval, and a metrics database for UI features. Both are critical but prioritize the main database for file recovery.

Depending on how the renter is configured, the databases are either SQLite (default) or MySQL databases. By default, the SQLite databases are called db.sqlite and metrics.sqlite and are located in a folder called db, right in the renter's root directory.

SQLite

Use the .backup command to create a backup of the SQLite databases.

sqlite3 db.sqlite ".backup 'db.bkp'"
sqlite3 metrics.sqlite ".backup 'metrics.bkp'"

There should only be two files in the db folder, if you encounter write-ahead log files or index files (usually named -wal or -shm) it indicates the renter was not shut down gracefully. In that case it's best to restart the renter and shut it down again.

MySQL

Use the mysqldump command to create a backup of the MySQL databases. It's a utility provided by MySQL to backup or transfer a MySQL database and it's usually installed alongside the MySQL client tools. Replace placeholders with actual user and password.

mysqldump -u [RENTERD_DB_USER] --password=[RENTERD_DB_PASSWORD] renterd > renterd_bkp.sql
mysqldump -u [RENTERD_DB_USER] --password=[RENTERD_DB_PASSWORD] renterd_metrics > renterd_metrics_bkp.sql

Step 3: backing up partial slabs

If "upload packing" is enabled, back up the partial_slabs folder located in the renter's root directory. These files contain data that has not been uploaded to the network yet, losing these files means an immediate loss of your data.

tar -cvf partial_slabs.tar partial_slabs/

Restoring from a backup

If the goal is to install renterd from a backup on a new machine, the easiest way is to do a fresh renterd install and then overwrite the empty database with the backup. Use the same RENTERD_SEED as the original installation.

Step 1: shutdown renter

Same as before we advise to shut down the renter before reinstating a backup to ensure consistency. It's a good idea to backup the database right before trying to restore a backup to be safe and have a way out in case overwriting the database renders it corrupt somehow.

Step 2: restore the database backup

SQLite

For SQLite we can reinstate the database by replacing both .sqlite files with our backups. Make sure to rename them to their original filename db.sqlite and metrics.sqlite. These filenames are configurable, so make sure you match the configured values.

MySQL

Depending on when the backup was taken its schema might be out of date. To maximize the chance the schema migrations go smoothly, it's advisable to recreate the databases before importing the backup. Take a backup before doing this.

# log in to MySQL shell
mysql -u [RENTERD_DB_USER] -p

# recreate renterd database
DROP DATABASE IF EXISTS renterd;
CREATE DATABASE renterd;

# recreate renterd_metrics database
DROP DATABASE IF EXISTS renterd_metrics;
CREATE DATABASE renterd_metrics;

The backups can then be imported using the following commands:

cat renterd_bkp.sql | mysql -u [RENTERD_DB_USER] --password=[RENTERD_DB_PASSWORD] renterd
cat renterd_metrics_bkp.sql | mysql -u [RENTERD_DB_USER] --password=[RENTERD_DB_PASSWORD] renterd_metrics

Step 3: restore the partial slabs

If applicable, remove the contents of the partial_slabs directory and replace it with your backup.

rm -rf partial_slabs
tar -xvf partial_slabs.tar

Step 4: start the renter

After starting the renter it is possible it has to run through migrations to its database schema. Depending on when the backup was taken, this might take some time. If we restored the backup on a fresh renterd install, it will take some time for consensus to sync.

Docker Support

renterd includes a Dockerfile which can be used for building and running renterd within a docker container. For the time being it is provided as-is without any compatibility guarantees as it will change over time and be extended with more configuration options. The image can also be pulled from ghcr.io/siafoundation/renterd.

Docker Compose

version: "3.9"
services:
  renterd:
    image: ghcr.io/siafoundation/renterd:master
    environment:
      - RENTERD_SEED=put your seed here
      - RENTERD_API_PASSWORD=test
    ports:
      - 9980:9980
      - 9981:9981
      - 8080:8080
    volumes:
      - ./data:/data
    restart: unless-stopped
    stop_grace_period: 5m

Build Image

From within the root of the repo run the following command to build an image of renterd tagged renterd.

docker build -t renterd:master -f ./docker/Dockerfile .

Run Container

Run renterd in the background as a container named renterd that exposes its API to the host system and the gateway to the world.

Mainnet

docker run -d --name renterd -e RENTERD_API_PASSWORD="<PASSWORD>" -e RENTERD_SEED="<SEED>" -p 127.0.0.1:9980:9980/tcp -p :9981:9981/tcp ghcr.io/siafoundation/renterd:master

Testnet

To run renterd on testnet use the RENTERD_NETWORK environment variable.

docker run -d --name renterd -e RENTERD_API_PASSWORD="<PASSWORD>" -e RENTERD_NETWORK="<network>" -e RENTERD_SEED="<SEED>" -p 127.0.0.1:9980:9980/tcp -p :9981:9981/tcp ghcr.io/siafoundation/renterd:master

Currently available values for <network> are:

Architecture

renterd distinguishes itself from siad through a unique architecture comprised of three main components: the autopilot, the bus, and one or more workers.

Instead of adopting another Electron app bundle, renterd incorporates an embedded web UI. This approach caters to developers and power users who prefer a streamlined experience. For those who do not require the UI, renterd can be compiled without it, which reduces software bloat and simplifies the build process.

Autopilot

The autopilot in renterd automates high-level functions like host scanning, ranking, and the management of contracts and data. This ensures efficient host selection for data storage. While it offers convenience for most users, those desiring full control can disable the autopilot via a CLI flag. Remarkably, the autopilot is designed over the public renterd HTTP API, allowing for straightforward customization and language porting.

Bus

Serving as the central nervous system of renterd, the bus handles data persistence and connections with Sia's peer-to-peer network. While renterd defaults to an SQLite database, users have the option to configure it for MySQL use. SQLite is the default choice due to its seamless initial experience, but switching to MySQL is recommended for enhanced performance.

Worker(s)

Workers are the direct interface for users, managing tasks such as file uploads/downloads and contract handling. They depend on the bus for consistent data persistence.

Usage

The Web UI streamlines the initial setup and configuration for newcomers. However, if manual configuration is necessary, the subsequent sections outline a step-by-step guide to achieving a functional renterd instance.

Wallet

Make sure the wallet is funded, it's a good rule of thumb to have at least twice the allowance in the wallet. Fetch the wallet's address and transfer some money. Verify the wallet's balance is not zero using the following endpoint:

The autopilot will automatically redistribute the wallet funds over a certain number of outputs that make sense with regards to the autopilot's configuration. Contract formation and renewals work best when the autopilot has a good amount of outputs at its disposal. It's definitely a good idea to verify whether this is the case because if not it means that it's likely the autopilot is misconfigured, in which case the logs should be of help.

Consensus

In order for the contracts to get formed, your node has to be synced with the blockchain. If you are not bootstrapping your node this can take a while. Verify your node's consensus state using the following endpoint:

Config

The configuration can be updated through the UI or by using the following endpoints:

The autopilot will not perform any tasks until it is configured. An example configuration can be found below. Especially the contracts section is important, make sure the amount is set to the amount of hosts with which you want to form a contract. The allowance is the amount of money the autopilot can spend per period, make sure it is not set to zero or contracts won't get formed.

{
    "hosts": {
        "allowRedundantIPs": false,
        "maxDowntimeHours": 1440,
        "maxConsecutiveScanFailures": 20,
        "scoreOverrides": {}
    },
    "contracts": {
        "set": "autopilot",
        "amount": 50,
        "allowance": "10000000000000000000000000000",
        "period": 6048,
        "renewWindow": 2016,
        "download": 1099511627776, // 1TiB
        "upload": 1099511627776, // 1TiB
        "storage": 1099511627776 // 1TiB
    }
}

Blocklist

Unfortunately the Sia blockchain is subject to hosts that announced themselves with faulty parameters and/or bad intentions, something which is unavoidable of course in a decentralized environment. We added a blocklist to give users the power to block hosts by IP or domain.

The Sia Foundation does not ship renterd with a default blocklist, the following entries exclude a decent amount of bad/old/malicious hosts: