Bacalhau supports running programs that are compiled to WebAssembly (WASM). With the Bacalhau client, you can upload WASM programs, retrieve data from public storage, read and write data, receive program arguments, and access environment variables.

Prerequisites and Limitations

1. Supported WebAssembly System Interface (WASI) Bacalhau can run compiled WASM programs that expect the WebAssembly System Interface (WASI) Snapshot 1. Through this interface, WebAssembly programs can access data, environment variables, and program arguments.

2. Networking Restrictions All ingress/egress networking is disabled – you won't be able to pull data/code/weights etc. from an external source. WASM jobs can say what data they need using URLs or CIDs (Content IDentifier) and can then access the data by reading from the filesystem.

3. Single-Threading There is no multi-threading as WASI does not expose any interface for it.

Onboarding Your Workload

Step 1: Replace network operations with filesystem reads and writes

If your program typically involves reading from and writing to network endpoints, follow these steps to adapt it for Bacalhau:

1. Replace Network Operations: Instead of making HTTP requests to external servers (e.g., example.com), modify your program to read data from the local filesystem.

2. Input Data Handling: Specify the input data location in Bacalhau using the --input flag when running the job. For instance, if your program used to fetch data from example.com, read from the /inputs folder locally, and provide the URL as input when executing the Bacalhau job. For example, --input http://example.com.

3. Output Handling: Adjust your program to output results to standard output (stdout) or standard error (stderr) pipes. Alternatively, you can write results to the filesystem, typically into an output mount. In the case of WASM jobs, a default folder at /outputs is available, ensuring that data written there will persist after the job concludes.

By making these adjustments, you can effectively transition your program to operate within the Bacalhau environment, utilizing filesystem operations instead of traditional network interactions.

Step 2: Configure your compiler to output WASI-compliant WebAssembly

You will need to compile your program to WebAssembly that expects WASI. Check the instructions for your compiler to see how to do this.

For example, Rust users can specify the wasm32-wasi target to rustup and cargo to get programs compiled for WASI WebAssembly. See the Rust example for more information on this.

Step 3: Upload the input data

Data is identified by its content identifier (CID) and can be accessed by anyone who knows the CID. You can use either of these methods to upload your data:

• Copy data from a URL to public storage

• Pin Data to public storage

• Copy Data from S3 Bucket to public storage.

Step 4: Run your program

You can run a WebAssembly program on Bacalhau using the bacalhau wasm run command.

bacalhau wasm run

Run Locally Compiled Program:

If your program is locally compiled, specify it as an argument. For instance, running the following command will upload and execute the main.wasm program:

bacalhau wasm run main.wasm

Alternative Program Specification:

You can use a Content IDentifier (CID) for a specific WebAssembly program.

bacalhau wasm run Qmajb9T3jBdMSp7xh2JruNrqg3hniCnM6EUVsBocARPJRQ

Input Data Specification:

Make sure to specify any input data using --input flag.

bacalhau wasm run --input http://example.com

This ensures the necessary data is available for the program's execution.

Program arguments

You can give the WASM program arguments by specifying them after the program path or CID. If the WASM program is already compiled and located in the current directory, you can run it by adding arguments after the file name:

bacalhau wasm run echo.wasm hello world

For a specific WebAssembly program, run:

bacalhau wasm run Qmajb9T3jBdMSp7xh2JruNrqg3hniCnM6EUVsBocARPJRQ hello world

bacalhau wasm run prog.wasm /inputs/data.txt

Environment variables

You can also specify environment variables using the -e flag.

$ bacalhau wasm run prog.wasm -e HELLO=world

Examples

See the Rust example for a workload that leverages WebAssembly support.

Support

If you have questions or need support or guidance, please reach out to the Bacalhau team via Slack (#general channel).

What is Bacalhau?

Bacalhau is a platform for fast, cost efficient, and secure computation by running jobs where the data is generated and stored. With Bacalhau, you can streamline your existing workflows without the need of extensive rewriting by running arbitrary Docker containers and WebAssembly (wasm) images as tasks. This architecture is also referred to as Compute Over Data (or CoD). Bacalhau was coined from the Portuguese word for salted Cod fish.

Bacalhau seeks to transform data processing for large-scale datasets to improve cost and efficiency, and to open up data processing to larger audiences. Our goals is to create an open, collaborative compute ecosystem that enables unparalleled collaboration. We (Expanso.io) offer a demo network so you can try out jobs without even installing. Give it a shot!

Why Bacalhau?

⚡️ Bacalhau simplifies the process of managing compute jobs by providing a unified platform for managing jobs across different regions, clouds, and edge devices.

🤝 Bacalhau provides reliable and network-partition resistant orchestration, ensuring that your jobs will complete even if there are network disruptions.

🚨 Bacalhau provides a complete and permanent audit log of exactly what happened, so you can be confident that your jobs are being executed securely.

🔐 You can run private workloads to reduce the chance of leaking private information or inadvertently sharing your data outside of your organization.

💸 Bacalhau reduces ingress/egress costs since jobs are processed closer to the source.

🤓 You can mount your data anywhere on your machine, and Bacalhau will be able to run against that data.

💥 You can integrate with services running on nodes to run a jobs, such as on DuckDB.

📚 Bacalhau operates at scale over parallel jobs. You can batch process petabytes (quadrillion bytes) of data.

How it works

Bacalhau concists of a peer-to-peer network of nodes that enables decentralized communication between computers. The network consists of two types of nodes:

Requester Node: responsible for handling user requests, discovering and ranking compute nodes, forwarding jobs to compute nodes, and monitoring the job lifecycle.

Compute Node: responsible for executing jobs and producing results. Different compute nodes can be used for different types of jobs, depending on their capabilities and resources.

The goal of the Bacalhau project is to make it easy to perform distributed computation next to where the data resides. In order to do this, first you need to ingest some data.

Data ingestion

Data is identified by its content identifier (CID) and can be accessed by anyone who knows the CID. Here are some options that can help you mount your data:

• Copy data from a URL to public storage

• Pin Data to public storage

• Copy Data from S3 Bucket to public storage

Security in Bacalhau

All workloads run under restricted Docker or WASM permissions on the node. Additionally, you can use existing (locked down) binaries that are pre-installed through Pluggable Executors.

Best practices in 12-factor apps is to use environment variables to store sensitive data such as access tokens, API keys, or passwords. These variables can be accessed by Bacalhau at runtime and are not visible to anyone who has access to the code or the server.

Alternatively, you can pre-provision credentials to the nodes and access those on a node by node basis.

Finally, endpoints (such as vaults) can also be used to provide secure access to Bacalhau. This way, the client can authenticate with Bacalhau using the token without exposing their credentials.

Workloads Bacalhau is best suited for

Bacalhau can be used for a variety of data processing workloads, including machine learning, data analytics, and scientific computing. It is well-suited for workloads that require processing large amounts of data in a distributed and parallelized manner.

Use Cases

Once you have more than 10 devices generating or storing around 100GB of data, you're likely to face challenges with processing that data efficiently. Traditional computing approaches may struggle to handle such large volumes, and that's where distributed computing solutions like Bacalhau can be extremely useful. Bacalhau can be used in various industries, including security, web serving, financial services, IoT, Edge, Fog, and multi-cloud. Bacalhau shines when it comes to data-intensive applications like data engineering, model training, model inference, molecular dynamics, etc.

Here are some example tutorials on how you can process your data with Bacalhau:

• Stable Diffusion AI

• Generate Realistic Images using StyleGAN3 and Bacalhau

• Object Detection with YOLOv5 on Bacalhau

• Running Genomics on Bacalhau

• Training Pytorch Model with Bacalhau

Community

Bacalhau has a very friendly community and we are always happy to help you get started:

• GitHub Discussions – ask anything about the project, give feedback or answer questions that will help other users.

• Join the Slack Community and go to #bacalhau channel – it is the easiest way engage with other members in the community and get help.

• Contributing – learn how to contribute to the Bacalhau project.

Next Steps

👉 Continue with Bacalhau Getting Started guide to learn how to install and run a job with the Bacalhau client.

👉 Or jump directly to try out the different Examples that showcases Bacalhau abilities.

Different jobs may require different amounts of resources to execute. Some jobs may have specific hardware requirements, such as GPU. This page describes how to specify hardware requirements for your job.

Docker Executor

The following table describes how to specify hardware requirements for the Docker executor.

How it Works

When you specify hardware requirements, the job will be offered out to the network to see if there are any nodes that can satisfy the requirements. If there are, the job will be scheduled on the node and the executor will be started.

GPU Setup

Bacalhau supports GPU workloads. Learn how to run a job using GPU workloads with the Bacalhau client.

Prerequisites

• The Bacalhau network must have an executor node with a GPU exposed

• Your container must include the CUDA runtime (cudart) and must be compatible with the CUDA version running on the node

Usage

Use following command to see available resources amount:

bacalhau node list --show=capacity

To submit a request for a job that requires more than the standard set of resources, add the --cpu and --memory flags. For example, for a job that requires 2 CPU cores and 4Gb of RAM, use --cpu=2 --memory=4Gb, e.g.:

bacalhau docker run ubuntu echo Hello World --cpu=2 --memory=4Gb

To submit a GPU job request, use the --gpu flag under the docker run command to select the number of GPUs your job requires. For example:

bacalhau docker run --gpu=1 nvidia/cuda:11.0.3-base-ubuntu20.04 nvidia-smi

Limitations

The following limitations currently exist within Bacalhau.

1. Maximum CPU and memory limits depend on the participants in the network

2. For GPU:

   1. NVIDIA, Intel or AMD GPUs only

   2. Only the Docker Executor supports GPUs

When running a node, you can choose which jobs you want to run by using configuration options, environment variables or flags to specify a job selection policy.

Job selection probes

If you want more control over making the decision to take on jobs, you can use the --job-selection-probe-exec and --job-selection-probe-http flags.

These are external programs that are passed the following data structure so that they can make a decision about whether or not to take on a job:

The exec probe is a script to run that will be given the job data on stdin, and must exit with status code 0 if the job should be run.

The http probe is a URL to POST the job data to. The job will be rejected if the HTTP request returns a non-positive status code (e.g. >= 400).

For example, the following response will reject the job:

If the HTTP response is not a JSON blob, the content is ignored and any non-error status code will accept the job.

Docker Workloads

Bacalhau executes jobs by running them within containers. Bacalhau employs a syntax closely resembling Docker, allowing you to utilize the same containers. The key distinction lies in how input and output data are transmitted to the container via IPFS, enabling scalability on a global level.

This section describes how to migrate a workload based on a Docker container into a format that will work with the Bacalhau client.

Requirements

Here are few things to note before getting started:

1. Container Registry: Ensure that the container is published to a public container registry that is accessible from the Bacalhau network.

2. Architecture Compatibility: Bacalhau supports only images that match the host node's architecture. Typically, most nodes run on linux/amd64, so containers in arm64 format are not able to run.

3. Input Flags: The --input ipfs://... flag supports only directories and does not support CID subpaths. The --input https://... flag supports only single files and does not support URL directories. The --input s3://... flag supports S3 keys and prefixes. For example, s3://bucket/logs-2023-04* includes all logs for April 2023.

Runtime Restrictions

To help provide a safe, secure network for all users, we add the following runtime restrictions:

1. Limited Ingress/Egress Networking:

All ingress/egress networking is limited as described in the documentation. You won't be able to pull data/code/weights/ etc. from an external source.

1. Data Passing with Docker Volumes:

A job includes the concept of input and output volumes, and the Docker executor implements support for these. This means you can specify your CIDs, URLs, and/or S3 objects as input paths and also write results to an output volume. This can be seen in the following example:

The above example demonstrates an input volume flag -i s3://mybucket/logs-2023-04*, which mounts all S3 objects in bucket mybucket with logs-2023-04 prefix within the docker container at location /input (root).

Output volumes are mounted to the Docker container at the location specified. In the example above, any content written to /output_folder will be made available within the apples folder in the job results CID.

Once the job has run on the executor, the contents of stdout and stderr will be added to any named output volumes the job has used (in this case apples), and all those entities will be packaged into the results folder which is then published to a remote location by the publisher.

Onboarding Your Workload

Step 1 - Read Data From Your Directory

If you need to pass data into your container you will do this through a Docker volume. You'll need to modify your code to read from a local directory.

We make the assumption that you are reading from a directory called /inputs, which is set as the default.

Step 2 - Write Data to the Your Directory

If you need to return data from your container you will do this through a Docker volume. You'll need to modify your code to write to a local directory.

We make the assumption that you are writing to a directory called /outputs, which is set as the default.

Step 3 - Build and Push Your Image To a Registry

For example:

Step 4 - Test Your Container

To test your docker image locally, you'll need to execute the following command, changing the environment variables as necessary:

Let's see what each command will be used for:

Exports the current working directory of the host system to the LOCAL_INPUT_DIR variable. This variable will be used for binding a volume and transferring data into the container.

Exports the current working directory of the host system to the LOCAL_OUTPUT_DIR variable. Similarly, this variable will be used for binding a volume and transferring data from the container.

Creates an array of commands CMD that will be executed inside the container. In this case, it is a simple command executing 'ls' in the /inputs directory and writing text to the /outputs/stdout file.

Launches a Docker container using the specified variables and commands. It binds volumes to facilitate data exchange between the host and the container.

For example:

The result of the commands' execution is shown below:

Step 5 - Upload the Input Data

Data is identified by its content identifier (CID) and can be accessed by anyone who knows the CID. You can use either of these methods to upload your data:

You can choose to

Step 6 - Run the Workload on Bacalhau

To launch your workload in a Docker container, using the specified image and working with input data specified via IPFS CID, run the following command.

To check the status of your job, run the following command.

To get more information on your job, you can run the following command.

To download your job, run.

To put this all together into one would look like the following.

This outputs the following.

Alternatively, you can run your workload with a publicly accessible http(s) URL, which will download the data temporarily into your public storage:

Troubleshooting

If you run into this compute error while running your docker image

This can often be resolved by re-tagging your docker image

Support

Bacalhau employs the and libraries for configuration management. Users can configure their Bacalhau node through a combination of command-line flags, environment variables, and the dedicated configuration file.

The Bacalhau Repo

Bacalhau manages its configuration, metadata, and internal state within a specialized repository named .bacalhau. Serving as the heart of the Bacalhau node, this repository holds the data and settings that determine node behavior. It's located on the filesystem, and by default, Bacalhau initializes this repository at $HOME/.bacalhau, where $HOME is the home directory of the user running the bacalhau process.

To customize this location, users can:

1. Set the BACALHAU_DIR environment variable to specify their desired path.

2. Utilize the --repo command line flag to specify their desired path.

Upon executing a Bacalhau command for the first time, the system will initialize the .bacalhau repository. If such a repository already exists, Bacalhau will seamlessly access its contents.

Structure of a Newly Initialized .bacalhau Repository

Below is the structure of a freshly initialized .bacalhau repository:

This repository comprises four directories and seven files:

Files

1. user_id.pem:

   • This file houses the Bacalhau node user's cryptographic private key, used for signing requests sent to a Requester Node.

   • Format: PEM.

2. repo.version:

   • Indicates the version of the Bacalhau node's repository.

   • Format: JSON, e.g., {"Version":1}.

3. libp2p_private_key:

   • Format: Base64 encoded RSA private key.

4. config.yaml:

   • Contains configuration settings for the Bacalhau node.

   • Format: YAML.

5. update.json:

   • A file containing the date/time when the last version check was made.

   • Format: JSON, e.g., {"LastCheck":"2024-01-24T11:06:14.631816Z"}

6. tokens.json:

   • A file containing the tokens obtained through authenticating with bacalhau clusters.

Directories

1. QmdGUjsMHEgtAfdtw7U62yPEcAZFtA33tKMsczLToegZtv-compute:

   • Note: The segment QmdGUjsMHEgtAfdtw7U62yPEcAZFtA33tKMsczLToegZtv is a unique NodeID for each Bacalhau node, derived from the libp2p_private_key.

2. QmdGUjsMHEgtAfdtw7U62yPEcAZFtA33tKMsczLToegZtv-requester:

   • Note: NodeID derivation is similar to the Compute directory.

3. executor_storages:

   • Storage for data handled by Bacalhau storage drivers.

4. plugins:

   • Houses binaries that allow the Compute node to execute specific tasks.

   • Note: This feature is currently experimental and isn't active during standard node operations.

Configuring a Bacalhau Node

Within a .bacalhau repository, a config.yaml file may be present. This file serves as the configuration source for the bacalhau node and adheres to the YAML format.

Although the config.yaml file is optional, its presence allows Bacalhau to load custom configurations; otherwise, Bacalhau is configured with built-in default values, environment variables and command line flags.

Modifications to the config.yaml file will not be dynamically loaded by the Bacalhau node. A restart of the node is required for any changes to take effect. Bacalhau determines its configuration based on the following precedence order, with each item superseding the subsequent:

1. Command-line Flag

2. Environment Variable

3. Config File

4. Defaults

Relationship Between config.yaml and Bacalhau Environment Variables

Bacalhau establishes a direct relationship between the value-bearing keys within the config.yaml file and corresponding environment variables. For these keys that have no further sub-keys, the environment variable name is constructed by capitalizing each segment of the key, and then joining them with underscores, prefixed with BACALHAU_.

For example, a YAML key with the path Node.IPFS.Connect translates to the environment variable BACALHAU_NODE_IPFS_CONNECT and is represented in a file like:

There is no corresponding environment variable for either Node or Node.IPFS. Config values may also have other environment variables that set them for simplicity or to maintain backwards compatibility.

Environments

• Bacalhau leverages the BACALHAU_ENVIRONMENT environment variable to determine the specific environment configuration when initializing a repository. Notably, if a .bacalhau repository has already been initialized, the BACALHAU_ENVIRONMENT setting will be ignored.

  By default, if the BACALHAU_ENVIRONMENT variable is not explicitly set by the user, Bacalhau will adopt the production environment settings.

  Below is a breakdown of the configurations associated with each environment:

  1. Production (public network)

  • Environment Variable: BACALHAU_ENVIRONMENT=production

  • Configurations:

    • Node.ClientAPI.Host: "bootstrap.production.bacalhau.org"

    • Node.Client.API.Host: 1234

    • ...other configurations specific to this environment...

  2. Staging (staging network)

  • Environment Variable: BACALHAU_ENVIRONMENT=staging

  • Configurations:

    • Node.ClientAPI.Host: "bootstrap.staging.bacalhau.org"

    • Node.Client.API.Host: 1234

    • ...other configurations specific to this environment...

  3. Development (development network)

  • Environment Variable: BACALHAU_ENVIRONMENT=development

  • Configurations:

    • Node.ClientAPI.Host: "bootstrap.development.bacalhau.org"

    • Node.Client.API.Host: 1234

    • ...other configurations specific to this environment...

  4. Local (private or local networks)

  • Environment Variable: BACALHAU_ENVIRONMENT=local

  • Configurations:

    • Node.ClientAPI.Host: "0.0.0.0"

    • Node.Client.API.Host: 1234

    • ...other configurations specific to this environment...

  ---

Usage Examples

How to initialize a Bacalhau Server for a local private network

How to initialize a Bacalhau Server with a custom repo path

Or

How to start a Bacalhau Server with DEBUG logs

Authentication and authorization

Bacalhau includes a flexible auth system that supports multiple methods of auth that are appropriate for different deployment environments.

By default

With no specific authentication configuration supplied, Bacalhau runs in "anonymous mode" – which allows unidentified users limited control over the system. "Anonymous mode" is only appropriate for testing or evaluation setups.

In anonymous mode, Bacalhau will allow:

• Users identified by a self-generated private key to submit any job and cancel their own jobs.

• Users not identified by any key to access other read-only endpoints, such as to read job lists, describe jobs, and query node or agent information.

Restricting anonymous access

Bacalhau auth is controlled by policies. Configuring the auth system is done by supplying a different policy file.

Restricting API access to only users that have authenticated requires specifying a new authorization policy. You can download a policy that restricts anonymous access and install it by using:

Once the node is restarted, accessing the node APIs will require the user to be authenticated, but by default will still allow users with a self-generated key to authenticate themselves.

Restricting the list of keys that can authenticate to only a known set requires specifying a new authentication policy. You can download a policy that restricts key-based access and install it by using:

Then, modify the allowed_clients variable in challange_ns_no_anon.rego to include acceptable client IDs, found by running bacalhau id.

Once the node is restarted, only keys in the allowed list will be able to access any API.

Username and password access

Users can authenticate using a username and password instead of specifying a private key for access. Again, this requires installation of an appropriate policy on the server.

Passwords are not stored in plaintext and are salted. The downloaded policy expects password hashes and salts generated by scrypt. To generate a salted password, the helper script in pkg/authn/ask/gen_password can be used:

This will ask for a password and generate a salt and hash to authenticate with it. Add the encoded username, salt and hash into the ask_ns_password.rego.

Writing custom policies

In principle, Bacalhau can implement any auth scheme that can be described in a structured way by a policy file.

Custom authentication policies

Bacalhau will pass information pertinent to the current request into every authentication policy query as a field on the input variable. The exact information depends on the type of authentication used.

challenge authentication

challenge authentication uses identifies the user by the presence of a private key. The user is asked to sign an input phrase to prove they have the key they are identifying with.

Policies used for challenge authentication do not need to actually implement the challenge verification logic as this is handled by the core code. Instead, they will only be invoked if this verification passes.

Policies for this type will need to implement these rules:

• bacalhau.authn.token: if the user should be authenticated, an access token they should use in subsequent requests. If the user should not be authenticated, should be undefined.

They should expect as fields on the input variable:

• clientId: an ID derived from the user's private key that identifies them uniquely

• nodeId: the ID of the requester node that this user is authenticating with

• signingKey: the private key (as a JWK) that should be used to sign any access tokens to be returned

The simplest possible policy might therefore be this policy that returns the same opaque token for all users:

ask authentication

ask authentication uses credentials supplied manually by the user as identification. For example, an ask policy could require a username and password as input and check these against a known list. ask policies do all the verification of the supplied credentials.

Policies for this type will need to implement these rules:

• bacalhau.authn.token: if the user should be authenticated, an access token they should use in subsequent requests. If the user should not be authenticated, should be undefined.

• bacalhau.authn.schema: a static JSON schema that should be used to collect information about the user. The type of declared fields may be used to pick the input method, and if a field is marked as writeOnly then it will be collected in a secure way (e.g. not shown on screen). The schema rule does not receive any input data.

They should expect as fields on the input variable:

• ask: a map of field names from the JSON schema to strings supplied by the user. The policy should validate these credentials.

• nodeId: the ID of the requester node that this user is authenticating with

• signingKey: the private key (as a JWK) that should be used to sign any access tokens to be returned

The simplest possible policy might therefore be one that asks for no data and returns the same opaque token for every user:

Custom authorization policies

Authorization policies do not vary depending on the type of authentication used – Bacalhau uses one authz policy for all API requests.

Authz policies are invoked for every API request. Authz policies should check the validity of any supplied access tokens and issue an authz decision for the requested API endpoint. It is not required that authz policies enforce that an access token is present – they may choose to grant access to unauthorized users.

Policies will need to implement these rules:

• bacalhau.authz.token_valid: true if the access token in the request is "valid" (but does not necessarily grant access for this request), or false if it is invalid for every request (e.g. because it has expired) and should be discarded.

• bacalhau.authz.allow: true if the user should be permitted to carry out the input request, false otherwise.

They should expect as fields on the input variable for both rules:

• http: details of the user's HTTP request:

  • host: the hostname used in the HTTP request

  • method: the HTTP method (e.g. GET, POST)

  • path: the path requested, as an array of path components without slashes

  • query: a map of URL query parameters to their values

  • headers: a map of HTTP header names to arrays representing their values

  • body: a blob of any content submitted as the body

• constraints: details about the receiving node that should be used to validate any supplied tokens:

  • cert: keys that the input token should have been signed with

  • iss: the name of a node that this node will recognize as the issuer of any signed tokens

  • aud: the name of this node that is receiving the request

Notably, the constraints data is appropriate to be passed directly to the Rego io.jwt.decode_verify method which will validate the access token as a JWT against the given constraints.

The simplest possible authz policy might be this one that allows all users to access all endpoints:

These are the flags that control the capacity of the Bacalhau node, and the limits for jobs that might be run.

  --limit-job-cpu string                 Job CPU core limit for single job (e.g. 500m, 2, 8).
  --limit-job-gpu string                 Job GPU limit for single job (e.g. 1, 2, or 8).
  --limit-job-memory string              Job Memory limit for single job  (e.g. 500Mb, 2Gb, 8Gb).
  --limit-total-cpu string               Total CPU core limit to run all jobs (e.g. 500m, 2, 8).
  --limit-total-gpu string               Total GPU limit to run all jobs (e.g. 1, 2, or 8).
  --limit-total-memory string            Total Memory limit to run all jobs  (e.g. 500Mb, 2Gb, 8Gb).

The --limit-total-* flags control the total system resources you want to give to the network. If left blank, the system will attempt to detect these values automatically.

The --limit-job-* flags control the maximum amount of resources a single job can consume for it to be selected for execution.

Resource limits are not supported for Docker jobs running on Windows. Resource limits will be applied at the job bid stage based on reported job requirements but will be silently unenforced. Jobs will be able to access as many resources as requested at runtime.

Bacalhau supports GPUs out of the box and defaults to allowing execution on all GPUs installed on the node.

Prerequisites

Bacalhau makes the assumption that you have installed all the necessary drivers and tools on your node host and have appropriately configured them for use by Docker.

In general for GPUs from any vendor, the Bacalhau client requires:

1. Docker

2. Permission to access Docker

Nvidia

1. NVIDIA GPU Drivers

2. NVIDIA Container Toolkit (nvidia-docker2)

3. Verify installation by Running a Sample Workload

4. nvidia-smi installed and functional

AMD

1. AMD GPU drivers

2. rocm-smi tool installed and functional

See the Running ROCm Docker containers for guidance on how to run Docker workloads on AMD GPU.

Intel

1. Intel GPU drivers

2. xpu-smi tool installed and functional

See the Running on GPU under docker for guidance on how to run Docker workloads on Intel GPU.

GPU Node Configuration

Access to GPUs can be controlled using resource limits. To limit the number of GPUs that can be used per job, set a job resource limit. To limit access to GPUs from all jobs, set a total resource limit.

By default, the requester node APIs used by the Bacalhau CLI are accessible over HTTP, but it is possible to configure it to use Transport Level Security (TLS) so that they are accessible over HTTPS instead. There are several ways to obtain the necessary certificates and keys, and Bacalhau supports obtaining them via ACME and Certificate Authorities or even self-signing them.

Getting a certificate from Let's Encrypt with ACME

Automatic Certificate Management Environment (ACME) is a protocol that allows for automating the deployment of Public Key Infrastructure, and is the protocol used to obtain a free certificate from the Let's Encrypt Certificate Authority.

Using the --autocert [hostname] parameter to the CLI (in the serve and devstack commands), a certificate is obtained automatically from Lets Encrypt. The provided hostname should be a comma-separated list of hostnames, but they should all be publicly resolvable as Lets Encrypt will attempt to connect to the server to verify ownership (using the ACME HTTP-01 challenge). On the very first request this can take a short time whilst the first certificate is issued, but afterwards they are then cached in the bacalhau repository.

Alternatively, you may set these options via the environment variable, BACALHAU_AUTO_TLS. If you are using a configuration file, you can set the values inNode.ServerAPI.TLS.AutoCert instead.

sudo setcap CAP_NET_BIND_SERVICE+ep $(which bacalhau)

A cache of ACME data is held in the config repository, by default ~/.bacalhau/autocert-cache, and this will be used to manage renewals to avoid rate limits.

Getting a certificate from a Certificate Authority

Obtaining a TLS certificate from a Certificate Authority (CA) without using the Automated Certificate Management Environment (ACME) protocol involves a manual process that typically requires the following steps:

1. Choose a Certificate Authority: First, you need to select a trusted Certificate Authority that issues TLS certificates. Popular CAs include DigiCert, GlobalSign, Comodo (now Sectigo), and others. You may also consider whether you want a free or paid certificate, as CAs offer different pricing models.

2. Generate a Certificate Signing Request (CSR): A CSR is a text file containing information about your organization and the domain for which you need the certificate. You can generate a CSR using various tools or directly on your web server. Typically, this involves providing details such as your organization's name, common name (your domain name), location, and other relevant information.

3. Submit the CSR: Access your chosen CA's website and locate their certificate issuance or order page. You'll typically find an option to "Submit CSR" or a similar option. Paste the contents of your CSR into the provided text box.

4. Verify Domain Ownership: The CA will usually require you to verify that you own the domain for which you're requesting the certificate. They may send an email to one of the standard domain-related email addresses (e.g., admin@yourdomain.com, webmaster@yourdomain.com). Follow the instructions in the email to confirm domain ownership.

5. Complete Additional Verification: Depending on the CA's policies and the type of certificate you're requesting (e.g., Extended Validation or EV certificates), you may need to provide additional documentation to verify your organization's identity. This can include legal documents or phone calls from the CA to confirm your request.

6. Payment and Processing: If you're obtaining a paid certificate, you'll need to make the payment at this stage. Once the CA has received your payment and completed the verification process, they will issue the TLS certificate.

Once you have obtained your certificates, you will need to put two files in a location that bacalhau can read them. You need the server certificate, often called something like server.cert or server.cert.pem, and the server key which is often called something like server.key or server.key.pem.

Once you have these two files available, you must start bacalhau serve which two new flags. These are tlscert and tlskey flags, whose arguments should point to the relevant file. An example of how it is used is:

bacalhau serve --node-type=requester --tlscert=server.cert --tlskey=server.key

Alternatively, you may set these options via the environment variables, BACALHAU_TLS_CERT and BACALHAU_TLS_KEY. If you are using a configuration file, you can set the values inNode.ServerAPI.TLS.ServerCertificate and Node.ServerAPI.TLS.ServerKey instead.

Self-signed certificates

If you wish, it is possible to use Bacalhau with a self-signed certificate which does not rely on an external Certificate Authority. This is an involved process and so is not described in detail here although there is a helpful script in the Bacalhau github repository which should provide a good starting point.

Once you have generated the necessary files, the steps are much like above, you must start bacalhau serve which two new flags. These are tlscert and tlskey flags, whose arguments should point to the relevant file. An example of how it is used is:

bacalhau serve --node-type=requester --tlscert=server.cert --tlskey=server.key

Alternatively, you may set these options via the environment variables, BACALHAU_TLS_CERT and BACALHAU_TLS_KEY. If you are using a configuration file, you can set the values inNode.ServerAPI.TLS.ServerCertificate and Node.ServerAPI.TLS.ServerKey instead.

If you use self-signed certificates, it is unlikely that any clients will be able to verify the certificate when connecting to the Bacalhau APIs. There are three options available to work around this problem:

1. Provide a CA certificate file of trusted certificate authorities, which many software libraries support in addition to system authorities.

2. Install the CA certificate file in the system keychain of each machine that needs access to the Bacalhau APIs.

3. Instruct the software library you are using not to verify HTTPS requests.

Bacalhau supports the three main 'pillars' of observability - logging, metrics, and tracing. Bacalhau uses the OpenTelemetry Go SDK for metrics and tracing, which can be configured using the standard environment variables. Exporting metrics and traces can be as simple as setting the OTEL_EXPORTER_OTLP_PROTOCOL and OTEL_EXPORTER_OTLP_ENDPOINT environment variables. Custom code is used for logging as the OpenTelemetry Go SDK currently doesn't support logging.

Logging

Logging in Bacalhau outputs in human-friendly format to stderr at INFO level by default, but this can be changed by two environment variables:

• LOG_LEVEL - Can be one of trace, debug, error, warn or fatal to output more or fewer logging messages as required

• LOG_TYPE - Can be one of the following values:

  • default - output logs to stderr in a human-friendly format

  • json - log messages outputted to stdout in JSON format

  • combined - log JSON formatted messages to stdout and human-friendly format to stderr

Log statements should include the relevant trace, span and job ID so it can be tracked back to the work being performed.

Metrics

Bacalhau produces a number of different metrics including those around the libp2p resource manager (rcmgr), performance of the requester HTTP API and the number of jobs accepted/completed/received.

Tracing

Traces are produced for all major pieces of work when processing a job, although the naming of some spans is still being worked on. You can find relevant traces covering working on a job by searching for the jobid attribute.

Viewing

The metrics and traces can easily be forwarded to a variety of different services as we use OpenTelemetry, such as Honeycomb or Datadog.

To view the data locally, or simply to not use a SaaS offering, you can start up Jaeger and Prometheus placing these three files into a directory then running docker compose start while running Bacalhau with the OTEL_EXPORTER_OTLP_PROTOCOL=grpc and OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:4317 environment variables.

version: "2"
services:

  jaeger-all-in-one:
    image: "jaegertracing/all-in-one:1.42"
    restart: "always"
    ports:
      - "16686:16686" # Jaeger UI
      - "14250:14250" # Jaeger gRPC endpoint

  otel-collector:
    image: "otel/opentelemetry-collector:0.70.0"
    restart: "always"
    command: ["--config=/etc/otel-collector-config.yaml"]
    volumes:
      - "./otel-collector-config.yaml:/etc/otel-collector-config.yaml"
    ports:
      - "8888:8888"   # Prometheus metrics exposed by the collector
      - "8889:8889"   # Prometheus exporter metrics
      - "13133:13133" # health_check extension
      - "4317:4317"   # OTLP gRPC receiver
    depends_on:
      - "jaeger-all-in-one"
      - "prometheus"

  prometheus:
    container_name: "prometheus"
    image: "prom/prometheus:v2.42.0"
    restart: "always"
    volumes:
      - "./prometheus.yaml:/etc/prometheus/prometheus.yml"
    ports:
      - "9090:9090" # Prometheus UI

receivers:
  otlp:
    protocols:
      grpc:

exporters:
  prometheus:
    endpoint: "0.0.0.0:8889"

  jaeger:
    endpoint: jaeger-all-in-one:14250
    tls:
      insecure: true

processors:
  batch:

extensions:
  health_check:

service:
  extensions: [health_check]
  pipelines:
    traces:
      receivers: [otlp]
      processors: [batch]
      exporters: [jaeger]
    metrics:
      receivers: [otlp]
      processors: [batch]
      exporters: [prometheus]

scrape_configs:
  - job_name: 'otel-collector'
    scrape_interval: 10s
    static_configs:
      - targets: ['otel-collector:8889']
      - targets: ['otel-collector:8888']

Install the Bacalhau CLI

In this tutorial, you'll learn how to install and run a job with the Bacalhau client using the Bacalhau CLI or Docker.

Step 1 - Install the Bacalhau Client

The Bacalhau client is a command-line interface (CLI) that allows you to submit jobs to the Bacalhau. The client is available for Linux, macOS, and Windows. You can also run the Bacalhau client in a Docker container.

Step 1.1 - Install the Bacalhau CLI

curl -sL https://get.bacalhau.org/install.sh | bash

docker image rm -f ghcr.io/bacalhau-project/bacalhau:latest # Remove old image if it exists
docker pull ghcr.io/bacalhau-project/bacalhau:latest

Step 1.2 - Verify the Installation

To verify installation and check the version of the client and server, use the version command. To run a Bacalhau client command with Docker, prefix it with docker run ghcr.io/bacalhau-project/bacalhau:latest.

bacalhau version

docker run -it ghcr.io/bacalhau-project/bacalhau:latest version

If you're wondering which server is being used, the Bacalhau Project has a demo network that's shared with the community. This network allows you to familiarize with Bacalhau's capabilities and launch jobs from your computer without maintaining a compute cluster on your own.

Step 2 - Submit a Hello World job

To submit a job in Bacalhau, we will use the bacalhau docker run command. The command runs a job using the Docker executor on the node. Let's take a quick look at its syntax:

bacalhau docker run [FLAGS] IMAGE[:TAG] [COMMAND]

bacalhau docker run ubuntu echo Hello World

Job successfully submitted. Job ID: f8e7789d-8e76-4e6c-8e71-436e2d76c72e
Checking job status... (Enter Ctrl+C to exit at any time, your job will continue running):

    Communicating with the network  ................  done ✅  0.2s
       Creating job for submission  ................  done ✅  0.7s
                   Job in progress  ................  done ✅  2.1s

To download the results, execute:
    bacalhau get f8e7789d-8e76-4e6c-8e71-436e2d76c72e

To get more details about the run, execute:
    bacalhau describe f8e7789d-8e76-4e6c-8e71-436e2d76c72e

Job successfully submitted. Job ID: 9d20bbad-c3fc-48f8-907b-1da61c927fbd
Checking job status...

docker run -t ghcr.io/bacalhau-project/bacalhau:latest \
                docker run \
                --id-only \
                --wait \
                ubuntu:latest -- \
                sh -c 'uname -a && echo "Hello from Docker Bacalhau!"'

14:02:25.992 | INF pkg/repo/fs.go:81 > Initializing repo at '/root/.bacalhau' for environment 'production'
19b105c9-4cb5-43bd-a12f-d715d738addd

Step 3 - Checking the State of your Jobs

After having deployed the job, we now can use the CLI for the interaction with the network. The jobs were sent to the public demo network, where it was processed and we can call the following functions. The job_id will differ for every submission.

Step 3.1 - Job status:

You can check the status of the job using bacalhau list command adding the --id-filter flag and specifying your job id.

bacalhau list --id-filter 9d20bbad

Let's take a look at the results of the command execution in the terminal:

CREATED   ID          JOB                                       STATE      PUBLISHED
15:24:31  0ed7617d    Type:"docker",Params:"map[Entrypoint:<ni  Completed
                       l> EnvironmentVariables:[] Image:ubuntu:
                       latest Parameters:[sh -c uname -a && ech
                       o "Hello from Docker Bacalhau!"] Working
                       Directory:]"

When it says Completed, that means the job is done, and we can get the results.

Step 3.2 - Job information:

You can find out more information about your job by using bacalhau describe.

bacalhau describe 9d20bbad

Let's take a look at the results of the command execution in the terminal:

Job:
    APIVersion: V1beta2
    Metadata:
        ClientID: 0ff57b2521334a92e9ddab4b2f8202c887b1eaa35d2aa945ab0e247d3bc0aa88
        CreatedAt: "2023-12-21T15:24:31.750306239Z"
        ID: 0ed7617d-d5ff-40f7-8411-89830b3f3058
        Requester:
        RequesterNodeID: QmbxGSsM6saCTyKkiWSxhJCt6Fgj7M9cns1vzYtfDbB5Ws
        RequesterPublicKey: CAASpgIwggEiMA0GCSqGSIb3DQEBAQUAA4IBDwAwggEKAoIBAQDEHTUAD1JzO0130W9vsaDGhU0PVgpcNjG3fYlE0sJ1BiBWENFuP4jx3Q9alcjNGhdRFdju0Mb/fidTOtJcPhxTdb+H6JxFP6HsADGes9jU4ylBU2SL2vfdb0KXzKdXjNHGGf4BuCGTcH07Oqxp209diK/cT7takL2fLjcgs1tM+6PzlfGzFqCPxvh9Sa0ek34mdmHjcp1XH8yjF1OKOuHvD+pYphqvOBL/2LEN+EBC4fz/QUnhUajCmKYO83MJcNUXSGxb4AN6K3DpVV+cJph7fj9ADdP7i996o2S4Gkz8W4Wpt/jICaPpkUjmyU3Jgcw7MHkZaYEzWxnnO2J936+pAgMBAAE=
    Spec:
    ...

This outputs all information about the job, including stdout, stderr, where the job was scheduled, and so on.

Step 3.3 - Job download:

You can download your job results directly by using bacalhau get.

bacalhau get 9d20bbad

This results in

Fetching results of job '0ed7617d'...
Results for job '0ed7617d' have been written to...
/Users/test/job-0ed7617d

In the command below, we created a directory called myfolder and download our job output to be stored in that directory.

Fetching results of job '0ed7617d'...
Results for job '0ed7617d' have been written to...
/myfolder

After the download has finished you should see the following contents in the results directory.

job-0ed7617d
├── exitCode
├── outputs
├── stderr
└── stdout

Step 4 - Viewing your Job Output

cat job-9d20bbad/stdout

That should print out the string Hello World.

Hello world

Step 5 - Where to go next?

Here are few resources that provide a deeper dive into running jobs with Bacalhau:

How Bacalhau works, Setting up Bacalhau, Examples & Use Cases

Support

If you have questions or need support or guidance, please reach out to the Bacalhau team via Slack (#general channel).

Both compute nodes, and requester nodes, maintain state. How that state is maintained is configurable, although the defaults are likely adequate for most use-cases. This page describes how to configure the persistence of compute and requester nodes should the defaults not be suitable.

Compute node persistence

The computes nodes maintain information about the work that has been allocated to them, including:

• The current state of the execution, and

• The original job that resulted in this allocation

This information is used by the compute and requester nodes to ensure allocated jobs are completed successfully. By default, compute nodes store their state in a bolt-db database and this is located in the bacalhau repository along with configuration data. For a compute node whose ID is "abc", the database can be found in ~/.bacalhau/abc-compute/executions.db.

In some cases, it may be preferable to maintain the state in memory, with the caveat that should the node restart, all state will be lost. This can be configured using the environment variables in the table below.

Requester node persistence

When running a requester node, it maintains state about the jobs it has been requested to orchestrate and schedule, the evaluation of those jobs, and the executions that have been allocated. By default, this state is stored in a bolt db database that, with a node ID of "xyz" can be found in ~/.bacalhau/xyz-requester/jobs.db.

In this tutorial we will go over the components and the architecture of Bacalhau. You will learn how it is built, what components are used, how you could interact and how you could use Bacalhau.

Chapter 1 - Architecture

Bacalhau is a peer-to-peer network of nodes that enables decentralized communication between computers. The network consists of two types of nodes, which can communicate with each other.

The requester and compute nodes together form a p2p network and use gossiping to discover each other, share information about node capabilities, available resources and health status. Bacalhau is a peer-to-peer network of nodes that enables decentralized communication between computers.

To interact with the Bacalhau network, users can use the Bacalhau CLI (command-line interface) to send requests to a requester node in the network. These requests are sent using the JSON format over HTTP, a widely-used protocol for transmitting data over the internet. Bacalhau's architecture involves two main sections which are the core components and interfaces.

Core Components

The core components are responsible for handling requests and connecting different nodes. The network includes two different components:

Interfaces

The interfaces handle the distribution, execution, storage and publishing of jobs. In the following all the different components are described and their respective protocols are shown.

Chapter 2 - Job cycle

Job preparation

You can create jobs in the Bacalhau network using various job types introduced in version 1.2. Each job may need specific variables, resource requirements and data details that are described in the Job Specification.

Job Submission

You should use the Bacalhau client to send a task to the network. The client transmits the job information to the Bacalhau network via established protocols and interfaces. Jobs submitted via the Bacalhau CLI are forwarded to a Bacalhau network node at http://bootstrap.production.bacalhau.org/ via port 1234 by default. This Bacalhau node will act as the requester node for the duration of the job lifecycle.

Bacalhau provides an interface to interact with the server via a REST API. Bacalhau uses 127.0.0.1 as the localhost and 1234 as the port by default.

bacalhau create [flags]

Endpoint: `PUT /api/v1/orchestrator/jobs` 

You can use the bacalhau docker run command to start a job in a Docker container. Below, you can see an excerpt of the commands:

bacalhau docker run [flags] IMAGE[:TAG|@DIGEST] [COMMAND] [ARG...]

Flags:
    --concurrency int                  How many nodes should run the job (default 1)
    --cpu string                       Job CPU cores (e.g. 500m, 2, 8).
    --disk string                      Job Disk requirement (e.g. 500Gb, 2Tb, 8Tb).
    --do-not-track                     When true the job will not be tracked(?) TODO BETTER DEFINITION
    --domain stringArray               Domain(s) that the job needs to access (for HTTP networking)
    --download                         Should we download the results once the job is complete?
    --download-timeout-secs duration   Timeout duration for IPFS downloads. (default 5m0s)
    --dry-run                          Do not submit the job, but instead print out what will be submitted
    --entrypoint strings               Override the default ENTRYPOINT of the image
-e, --env strings                      The environment variables to supply to the job (e.g. --env FOO=bar --env BAR=baz)
-f, --follow                           When specified will follow the output from the job as it runs
-g, --gettimeout int                   Timeout for getting the results of a job in --wait (default 10)
    --gpu string                       Job GPU requirement (e.g. 1, 2, 8).
-h, --help                             help for run
    --id-only                          Print out only the Job ID on successful submission.
-i, --input storage                    Mount URIs as inputs to the job. Can be specified multiple times. Format: src=URI,dst=PATH[,opt=key=value]
                                    Examples:
                                    # Mount IPFS CID to /inputs directory
                                    -i ipfs://QmeZRGhe4PmjctYVSVHuEiA9oSXnqmYa4kQubSHgWbjv72
                                    # Mount S3 object to a specific path
                                    -i s3://bucket/key,dst=/my/input/path
                                    # Mount S3 object with specific endpoint and region
                                    -i src=s3://bucket/key,dst=/my/input/path,opt=endpoint=https://s3.example.com,opt=region=us-east-1
    --ipfs-connect string              The ipfs host multiaddress to connect to, otherwise an in-process IPFS node will be created if not set.
    --ipfs-serve-path string           path local Ipfs node will persist data to
    --ipfs-swarm-addrs strings         IPFS multiaddress to connect the in-process IPFS node to - cannot be used with --ipfs-connect. (default [/ip4/35.245.161.250/tcp/4001/p2p/12D3KooWAQpZzf3qiNxpwizXeArGjft98ZBoMNgVNNpoWtKAvtYH,/ip4/35.245.161.250/udp/4001/quic/p2p/12D3KooWAQpZzf3qiNxpwizXeArGjft98ZBoMNgVNNpoWtKAvtYH,/ip4/34.86.254.26/tcp/4001/p2p/12D3KooWLfFBjDo8dFe1Q4kSm8inKjPeHzmLBkQ1QAjTHocAUazK,/ip4/34.86.254.26/udp/4001/quic/p2p/12D3KooWLfFBjDo8dFe1Q4kSm8inKjPeHzmLBkQ1QAjTHocAUazK,/ip4/35.245.215.155/tcp/4001/p2p/12D3KooWH3rxmhLUrpzg81KAwUuXXuqeGt4qyWRniunb5ipjemFF,/ip4/35.245.215.155/udp/4001/quic/p2p/12D3KooWH3rxmhLUrpzg81KAwUuXXuqeGt4qyWRniunb5ipjemFF,/ip4/34.145.201.224/tcp/4001/p2p/12D3KooWBCBZnXnNbjxqqxu2oygPdLGseEbfMbFhrkDTRjUNnZYf,/ip4/34.145.201.224/udp/4001/quic/p2p/12D3KooWBCBZnXnNbjxqqxu2oygPdLGseEbfMbFhrkDTRjUNnZYf,/ip4/35.245.41.51/tcp/4001/p2p/12D3KooWJM8j97yoDTb7B9xV1WpBXakT4Zof3aMgFuSQQH56rCXa,/ip4/35.245.41.51/udp/4001/quic/p2p/12D3KooWJM8j97yoDTb7B9xV1WpBXakT4Zof3aMgFuSQQH56rCXa])
    --ipfs-swarm-key string            Optional IPFS swarm key required to connect to a private IPFS swarm
-l, --labels strings                   List of labels for the job. Enter multiple in the format '-l a -l 2'. All characters not matching /a-zA-Z0-9_:|-/ and all emojis will be stripped.
    --memory string                    Job Memory requirement (e.g. 500Mb, 2Gb, 8Gb).
    --network network-type             Networking capability required by the job. None, HTTP, or Full (default None)
    --node-details                     Print out details of all nodes (overridden by --id-only).
-o, --output strings                   name:path of the output data volumes. 'outputs:/outputs' is always added unless '/outputs' is mapped to a different name. (default [outputs:/outputs])
    --output-dir string                Directory to write the output to.
    --private-internal-ipfs            Whether the in-process IPFS node should auto-discover other nodes, including the public IPFS network - cannot be used with --ipfs-connect. Use "--private-internal-ipfs=false" to disable. To persist a local Ipfs node, set BACALHAU_SERVE_IPFS_PATH to a valid path. (default true)
-p, --publisher publisher              Where to publish the result of the job (default ipfs)
    --raw                              Download raw result CIDs instead of merging multiple CIDs into a single result
-s, --selector string                  Selector (label query) to filter nodes on which this job can be executed, supports '=', '==', and '!='.(e.g. -s key1=value1,key2=value2). Matching objects must satisfy all of the specified label constraints.
    --target all|any                   Whether to target the minimum number of matching nodes ("any") (default) or all matching nodes ("all") (default any)
    --timeout int                      Job execution timeout in seconds (e.g. 300 for 5 minutes)
    --wait                             Wait for the job to finish. Use --wait=false to return as soon as the job is submitted. (default true)
    --wait-timeout-secs int            When using --wait, how many seconds to wait for the job to complete before giving up. (default 600)
-w, --workdir string                   Working directory inside the container. Overrides the working directory shipped with the image (e.g. via WORKDIR in Dockerfile).

You can also use the bacalhau wasm run command to run a job compiled into the (WASM) format. Below, you can find an excerpt of the commands in the Bacalhau CLI:

bacalhau wasm run {cid-of-wasm | } [--entry-point ] [wasm-args ...] [flags]

Flags:
    --concurrency int                  How many nodes should run the job (default 1)
    --cpu string                       Job CPU cores (e.g. 500m, 2, 8).
    --disk string                      Job Disk requirement (e.g. 500Gb, 2Tb, 8Tb).
    --do-not-track                     When true the job will not be tracked(?) TODO BETTER DEFINITION
    --domain stringArray               Domain(s) that the job needs to access (for HTTP networking)
    --download                         Should we download the results once the job is complete?
    --download-timeout-secs duration   Timeout duration for IPFS downloads. (default 5m0s)
    --dry-run                          Do not submit the job, but instead print out what will be submitted
    --entry-point string               The name of the WASM function in the entry module to call. This should be a zero-parameter zero-result function that
                                will execute the job. (default "_start")
-e, --env strings                      The environment variables to supply to the job (e.g. --env FOO=bar --env BAR=baz)
-f, --follow                           When specified will follow the output from the job as it runs
-g, --gettimeout int                   Timeout for getting the results of a job in --wait (default 10)
    --gpu string                       Job GPU requirement (e.g. 1, 2, 8).
-h, --help                             help for run
    --id-only                          Print out only the Job ID on successful submission.
-U, --import-module-urls url           URL of the WASM modules to import from a URL source. URL accept any valid URL supported by the 'wget' command, and supports both HTTP and HTTPS.
-I, --import-module-volumes cid:path   CID:path of the WASM modules to import from IPFS, if you need to set the path of the mounted data.
-i, --input storage                    Mount URIs as inputs to the job. Can be specified multiple times. Format: src=URI,dst=PATH[,opt=key=value]
                                        Examples:
                                        # Mount IPFS CID to /inputs directory
                                        -i ipfs://QmeZRGhe4PmjctYVSVHuEiA9oSXnqmYa4kQubSHgWbjv72
                                        # Mount S3 object to a specific path
                                        -i s3://bucket/key,dst=/my/input/path
                                        # Mount S3 object with specific endpoint and region
                                        -i src=s3://bucket/key,dst=/my/input/path,opt=endpoint=https://s3.example.com,opt=region=us-east-1
    --ipfs-connect string              The ipfs host multiaddress to connect to, otherwise an in-process IPFS node will be created if not set.
    --ipfs-serve-path string           path local Ipfs node will persist data to
    --ipfs-swarm-addrs strings         IPFS multiaddress to connect the in-process IPFS node to - cannot be used with --ipfs-connect. (default [/ip4/35.245.161.250/tcp/4001/p2p/12D3KooWAQpZzf3qiNxpwizXeArGjft98ZBoMNgVNNpoWtKAvtYH,/ip4/35.245.161.250/udp/4001/quic/p2p/12D3KooWAQpZzf3qiNxpwizXeArGjft98ZBoMNgVNNpoWtKAvtYH,/ip4/34.86.254.26/tcp/4001/p2p/12D3KooWLfFBjDo8dFe1Q4kSm8inKjPeHzmLBkQ1QAjTHocAUazK,/ip4/34.86.254.26/udp/4001/quic/p2p/12D3KooWLfFBjDo8dFe1Q4kSm8inKjPeHzmLBkQ1QAjTHocAUazK,/ip4/35.245.215.155/tcp/4001/p2p/12D3KooWH3rxmhLUrpzg81KAwUuXXuqeGt4qyWRniunb5ipjemFF,/ip4/35.245.215.155/udp/4001/quic/p2p/12D3KooWH3rxmhLUrpzg81KAwUuXXuqeGt4qyWRniunb5ipjemFF,/ip4/34.145.201.224/tcp/4001/p2p/12D3KooWBCBZnXnNbjxqqxu2oygPdLGseEbfMbFhrkDTRjUNnZYf,/ip4/34.145.201.224/udp/4001/quic/p2p/12D3KooWBCBZnXnNbjxqqxu2oygPdLGseEbfMbFhrkDTRjUNnZYf,/ip4/35.245.41.51/tcp/4001/p2p/12D3KooWJM8j97yoDTb7B9xV1WpBXakT4Zof3aMgFuSQQH56rCXa,/ip4/35.245.41.51/udp/4001/quic/p2p/12D3KooWJM8j97yoDTb7B9xV1WpBXakT4Zof3aMgFuSQQH56rCXa])
    --ipfs-swarm-key string            Optional IPFS swarm key required to connect to a private IPFS swarm
-l, --labels strings                   List of labels for the job. Enter multiple in the format '-l a -l 2'. All characters not matching /a-zA-Z0-9_:|-/ and all emojis will be stripped.
    --memory string                    Job Memory requirement (e.g. 500Mb, 2Gb, 8Gb).
    --network network-type             Networking capability required by the job. None, HTTP, or Full (default None)
    --node-details                     Print out details of all nodes (overridden by --id-only).
-o, --output strings                   name:path of the output data volumes. 'outputs:/outputs' is always added unless '/outputs' is mapped to a different name. (default [outputs:/outputs])
    --output-dir string                Directory to write the output to.
    --private-internal-ipfs            Whether the in-process IPFS node should auto-discover other nodes, including the public IPFS network - cannot be used with --ipfs-connect. Use "--private-internal-ipfs=false" to disable. To persist a local Ipfs node, set BACALHAU_SERVE_IPFS_PATH to a valid path. (default true)
-p, --publisher publisher              Where to publish the result of the job (default ipfs)
    --raw                              Download raw result CIDs instead of merging multiple CIDs into a single result
-s, --selector string                  Selector (label query) to filter nodes on which this job can be executed, supports '=', '==', and '!='.(e.g. -s key1=value1,key2=value2). Matching objects must satisfy all of the specified label constraints.
    --target all|any                   Whether to target the minimum number of matching nodes ("any") (default) or all matching nodes ("all") (default any)
    --timeout int                      Job execution timeout in seconds (e.g. 300 for 5 minutes)
    --wait                             Wait for the job to finish. Use --wait=false to return as soon as the job is submitted. (default true)
    --wait-timeout-secs int            When using --wait, how many seconds to wait for the job to complete before giving up. (default 600)

Job Acceptance

When a job is submitted to a requester node, it selects compute nodes that are capable and suitable to execute the job, and communicate with them directly. The compute node has a collection of named executors, storage sources, and publishers, and it will choose the most appropriate ones based on the job specifications.

Job execution

The selected compute node receives the job and starts its execution inside a container. The container can use different executors to work with the data and perform the necessary actions. A job can use the docker executor, WASM executor or a library storage volumes. Use Docker Engine Specification to view the parameters to configure the Docker Engine. If you want tasks to be executed in a WebAssembly environment, pay attention to WebAssembly Engine Specification.

Results publishing

When the Compute node completes the job, it publishes the results to S3's remote storage, IPFS.

Bacalhau's seamless integration with IPFS ensures that users have a decentralized option for publishing their task results, enhancing accessibility and resilience while reducing dependence on a single point of failure. View IPFS Publisher Specification to get the detailed information.

Bacalhau's S3 Publisher provides users with a secure and efficient method to publish task results to any S3-compatible storage service. This publisher supports not just AWS S3, but other S3-compatible services offered by cloud providers like Google Cloud Storage and Azure Blob Storage, as well as open-source options like MinIO. View S3 Publisher Specification to get the detailed information.

Chapter 3 - Returning Information

The Bacalhau client receives updates on the task execution status and results. A user can access the results and manage tasks through the command line interface.

Get Job Results

To Get the results of a job you can run the following command.

bacalhau get [id] [flags]

One can choose from a wide range of flags, from which a few are shown below.

Usage:
  bacalhau get [id] [flags]

Flags:
      --download-timeout-secs duration   Timeout duration for IPFS downloads. (default 5m0s)
  -h, --help                             help for get
      --ipfs-connect string              The ipfs host multiaddress to connect to, otherwise an in-process IPFS node will be created if not set.
      --ipfs-serve-path string           path local Ipfs node will persist data to
      --ipfs-swarm-addrs strings         IPFS multiaddress to connect the in-process IPFS node to - cannot be used with --ipfs-connect. (default [/ip4/35.245.161.250/tcp/4001/p2p/12D3KooWAQpZzf3qiNxpwizXeArGjft98ZBoMNgVNNpoWtKAvtYH,/ip4/35.245.161.250/udp/4001/quic/p2p/12D3KooWAQpZzf3qiNxpwizXeArGjft98ZBoMNgVNNpoWtKAvtYH,/ip4/34.86.254.26/tcp/4001/p2p/12D3KooWLfFBjDo8dFe1Q4kSm8inKjPeHzmLBkQ1QAjTHocAUazK,/ip4/34.86.254.26/udp/4001/quic/p2p/12D3KooWLfFBjDo8dFe1Q4kSm8inKjPeHzmLBkQ1QAjTHocAUazK,/ip4/35.245.215.155/tcp/4001/p2p/12D3KooWH3rxmhLUrpzg81KAwUuXXuqeGt4qyWRniunb5ipjemFF,/ip4/35.245.215.155/udp/4001/quic/p2p/12D3KooWH3rxmhLUrpzg81KAwUuXXuqeGt4qyWRniunb5ipjemFF,/ip4/34.145.201.224/tcp/4001/p2p/12D3KooWBCBZnXnNbjxqqxu2oygPdLGseEbfMbFhrkDTRjUNnZYf,/ip4/34.145.201.224/udp/4001/quic/p2p/12D3KooWBCBZnXnNbjxqqxu2oygPdLGseEbfMbFhrkDTRjUNnZYf,/ip4/35.245.41.51/tcp/4001/p2p/12D3KooWJM8j97yoDTb7B9xV1WpBXakT4Zof3aMgFuSQQH56rCXa,/ip4/35.245.41.51/udp/4001/quic/p2p/12D3KooWJM8j97yoDTb7B9xV1WpBXakT4Zof3aMgFuSQQH56rCXa])
      --ipfs-swarm-key string            Optional IPFS swarm key required to connect to a private IPFS swarm
      --output-dir string                Directory to write the output to.
      --private-internal-ipfs            Whether the in-process IPFS node should auto-discover other nodes, including the public IPFS network - cannot be used with --ipfs-connect. Use "--private-internal-ipfs=false" to disable. To persist a local Ipfs node, set BACALHAU_SERVE_IPFS_PATH to a valid path. (default true)
      --raw                              Download raw result CIDs instead of merging multiple CIDs into a single result

Describe a Job

To describe a specific job, inserting the ID to the CLI or API gives back an overview of the job.

bacalhau describe [id] [flags]

Endpoint: `GET /api/v1/orchestrator/jobs/:jobID`

List of Jobs

If you run more then one job or you want to find a specific job ID

bacalhau list [flags]

Endpoint: `GET /api/v1/orchestrator/jobs`

Job Executions

To list executions follow the following commands.

bacalhau job executions [id] [flags]

Endpoint: `GET /api/v1/orchestrator/jobs/:jobID/executions`

Chapter 4 - Monitoring and Management

The Bacalhau client provides the user with tools to monitor and manage the execution of jobs. You can get information about status, progress and decide on next steps. View the Bacalhau Agent APIs if you want to know the node's health, capabilities, and deployed Bacalhau version. To get information about the status and characteristics of the nodes in the cluster use Nodes API Documentation.

Stop a Job

bacalhau cancel [id] [flags]

Endpoint: `DELETE /api/v1/orchestrator/jobs/:jobID`

Job History

bacalhau job history [id] [flags]

Endpoint: `GET /api/v1/orchestrator/jobs/:jobID/history`

Job Logs

bacalhau logs [flags] [id]

You can use this command to retrieve the log output (stdout, and stderr) from a job. If the job is still running it is possible to follow the logs after the previously generated logs are retrieved.

Bacalhau has two ways to make use of external storage providers: Sources and Publishers. Sources storage resources consumed as inputs to jobs. And Publishers storage resources created with the results of jobs.

Sources

S3

Bacalhau allows you to use S3 or any S3-compatible storage service as an input source. Users can specify files or entire prefixes stored in S3 buckets to be fetched and mounted directly into the job execution environment. This capability ensures that your jobs have immediate access to the necessary data. See the for more details.

To use the S3 source, you will have to to specify the mandatory name of the S3 bucket and the optional parameters Key, Filter, Region, Endpoint, VersionID and ChechsumSHA256.

Below is an example of how to define an S3 input source in YAML format:

IPFS

You can then configure your Bacalhau node to use this IPFS server by passing the --ipfs-connect argument to the serve command:

Below is an example of how to define an IPFS input source in YAML format:

To use a local data source, you will have to to:

1. Enable the use of local data when configuring the node itself by using the --allow-listed-local-paths flag for bacalhau serve, specifying the file path and access mode. For example

1. In the job description specify parameters SourcePath - the absolute path on the compute node where your data is located and ReadWrite - the access mode.

Below is an example of how to define a Local input source in YAML format:

To use a URL data source, you will have to to specify only URL parameter, as in the part of the declarative job description below:

Here’s an example of the part of the declarative job description that outlines the process of using the S3 Publisher with Bacalhau:

To speed up the download or to retrieve results from a private IPFS node, pass the swarm multiaddress to bacalhau get to download results.

Pass the swarm key to bacalhau get if the IPFS swarm is a private swarm.

And part of the declarative job description with an IPFS publisher will look like this:

Here is an example of part of the declarative job description with a local publisher:

Bacalhau can limit the total time a job spends executing. A job that spends too long executing will be cancelled and no results will be published.

By default, a Bacalhau node does not enforce any limit on job execution time. Both node operators and job submitters can supply a maximum execution time limit. If a job submitter asks for a longer execution time than permitted by a node operator, their job will be rejected.

Applying job timeouts allows node operators to more fairly distribute the work submitted to their nodes. It also protects users from transient errors that results in their jobs waiting indefinitely.

Configuring execution time limits for a job

Job submitters can pass the --timeout flag to any Bacalhau job submission CLI to set a maximum job execution time. The supplied value should be a whole number of seconds with no unit.

The timeout can also be added to an existing job spec by adding the Timeout property to the Spec.

Configuring execution time limits for a node

Node operators can pass the --max-job-execution-timeout flag to bacalhau serve to configure the maximum job time limit. The supplied value should be a numeric value followed by a time unit (one of s for seconds, m for minutes or h for hours).

Node operators can also use configuration properties to configure execution limits.

Compute nodes will use the properties:

Requester nodes will use the properties:

Before you join the main Bacalhau network, you can test locally.

To test, you can use the bacalhau devstack command, which offers a way to get a 3 node cluster running locally.

In another window, export the following environment variables so that the Bacalhau client binary connects to our local development cluster:

You can now interact with Bacalhau - all jobs are running by the local devstack cluster.

Docker Workloads

Bacalhau executes jobs by running them within containers. Bacalhau employs a syntax closely resembling Docker, allowing you to utilize the same containers. The key distinction lies in how input and output data are transmitted to the container via IPFS, enabling scalability on a global level.

This section describes how to migrate a workload based on a Docker container into a format that will work with the Bacalhau client.

Requirements

Here are few things to note before getting started:

1. Container Registry: Ensure that the container is published to a public container registry that is accessible from the Bacalhau network.

2. Architecture Compatibility: Bacalhau supports only images that match the host node's architecture. Typically, most nodes run on linux/amd64, so containers in arm64 format are not able to run.

3. Input Flags: The --input ipfs://... flag supports only directories and does not support CID subpaths. The --input https://... flag supports only single files and does not support URL directories. The --input s3://... flag supports S3 keys and prefixes. For example, s3://bucket/logs-2023-04* includes all logs for April 2023.

Runtime Restrictions

To help provide a safe, secure network for all users, we add the following runtime restrictions:

1. Limited Ingress/Egress Networking:

All ingress/egress networking is limited as described in the documentation. You won't be able to pull data/code/weights/ etc. from an external source.

1. Data Passing with Docker Volumes:

A job includes the concept of input and output volumes, and the Docker executor implements support for these. This means you can specify your CIDs, URLs, and/or S3 objects as input paths and also write results to an output volume. This can be seen in the following example:

The above example demonstrates an input volume flag -i s3://mybucket/logs-2023-04*, which mounts all S3 objects in bucket mybucket with logs-2023-04 prefix within the docker container at location /input (root).

Output volumes are mounted to the Docker container at the location specified. In the example above, any content written to /output_folder will be made available within the apples folder in the job results CID.

Once the job has run on the executor, the contents of stdout and stderr will be added to any named output volumes the job has used (in this case apples), and all those entities will be packaged into the results folder which is then published to a remote location by the publisher.

Onboarding Your Workload

Step 1 - Read Data From Your Directory

If you need to pass data into your container you will do this through a Docker volume. You'll need to modify your code to read from a local directory.

We make the assumption that you are reading from a directory called /inputs, which is set as the default.

Step 2 - Write Data to the Your Directory

If you need to return data from your container you will do this through a Docker volume. You'll need to modify your code to write to a local directory.

We make the assumption that you are writing to a directory called /outputs, which is set as the default.

Step 3 - Build and Push Your Image To a Registry

For example:

Step 4 - Test Your Container

To test your docker image locally, you'll need to execute the following command, changing the environment variables as necessary:

Let's see what each command will be used for:

For example:

The result of the commands' execution is shown below:

Step 5 - Upload the Input Data

Data is identified by its content identifier (CID) and can be accessed by anyone who knows the CID. You can use either of these methods to upload your data:

Step 6 - Run the Workload on Bacalhau

To launch your workload in a Docker container, using the specified image and working with input data specified via IPFS CID, run the following command:

To check the status of your job, run the following command:

To get more information on your job,run:

To download your job, run:

For example, running:

outputs:

Alternatively, you can run your workload with a publicly accessible http(s) URL, which will download the data temporarily into your public storage:

Troubleshooting

If you run into this compute error while running your docker image

This can often be resolved by re-tagging your docker image

Support

Running a Windows-based node is not officially supported, so your mileage may vary. Some features (like ) are not present in Windows-based nodes.

Bacalhau currently makes the assumption that all containers are Linux-based. Users of the Docker executor will need to manually ensure that their Docker engine is running and to support Linux containers, e.g. using the WSL-based backend.

This tutorial serves as an introduction to Bacalhau. In this example, you'll be executing a simple "Hello, World!" Python script hosted on a website on Bacalhau.

Prerequisites​

To get started, you need to install the Bacalhau client, see more information here

1. Running Python Locally​

We'll be using a very simple Python script that displays the traditional first greeting. Create a file called hello-world.py:

# hello-world.py
print("Hello, world!")

Running the script to print out the output:

python3 hello-world.py

After the script has run successfully locally we can now run it on Bacalhau.

2. Running a Bacalhau Job​

To submit a workload to Bacalhau you can use the bacalhau docker run command. This command allows passing input data into the container using content identifier (CID) volumes, we will be using the --input URL:path argument for simplicity. This results in Bacalhau mounting a data volume inside the container. By default, Bacalhau mounts the input volume at the path /inputs inside the container.

export JOB_ID=$(bacalhau docker run \
    --id-only \
    --input https://raw.githubusercontent.com/bacalhau-project/examples/151eebe895151edd83468e3d8b546612bf96cd05/workload-onboarding/trivial-python/hello-world.py \
    python:3.10-slim \
    -- python3 /inputs/hello-world.py)

Structure of the command​

1. bacalhau docker run: call to Bacalhau

2. --id-only: specifies that only the job identifier (job_id) will be returned after executing the container, not the entire output

3. --input https://raw.githubusercontent.com/bacalhau-project/examples/151eebe895151edd83468e3d8b546612bf96cd05/workload-onboarding/trivial-python/hello-world.py \: indicates where to get the input data for the container. In this case, the input data is downloaded from the specified URL, which represents the Python script "hello-world.py".

4. python:3.10-slim: the Docker image that will be used to run the container. In this case, it uses the Python 3.10 image with a minimal set of components (slim).

5. --: This double dash is used to separate the Bacalhau command options from the command that will be executed inside the Docker container.

6. python3 /inputs/hello-world.py: running the hello-world.py Python script stored in /inputs.

When a job is submitted, Bacalhau prints out the related job_id. We store that in an environment variable so that we can reuse it later on.

Declarative job description​

The same job can be presented in the declarative format. In this case, the description will look like this:

name: Running Trivial Python
type: batch
count: 1
tasks:
  - name: My main task
    Engine:
      type: docker
      params:
        Image: python:3.10-slim
        Entrypoint:
          - /bin/bash
        Parameters:
          - -c
          - python3 /inputs/hello-world.py
    InputSources:
      - Target: /inputs
        Source:
          Type: urlDownload
          Params:
            URL: https://raw.githubusercontent.com/bacalhau-project/examples/151eebe895151edd83468e3d8b546612bf96cd05/workload-onboarding/trivial-python/hello-world.py
            Path: /inputs/hello-world.py

The job description should be saved in .yaml format, e.g. helloworld.yaml, and then run with the command:

bacalhau job run helloworld.yaml

3. Checking the State of your Jobs​

Job status: You can check the status of the job using bacalhau list.

bacalhau list --id-filter ${JOB_ID} --no-style

When it says Published or Completed, that means the job is done, and we can get the results.

Job information: You can find out more information about your job by using bacalhau describe.

bacalhau describe ${JOB_ID}

Job download: You can download your job results directly by using bacalhau get. Alternatively, you can choose to create a directory to store your results. In the command below, we created a directory (results) and downloaded our job output to be stored in that directory.

rm -rf results && mkdir results
bacalhau get ${JOB_ID} --output-dir results

4. Viewing your Job Output​

To view the file, run the following command:

cat results/stdout

Support​

If you have questions or need support or guidance, please reach out to the Bacalhau team via Slack (#general channel).

Overview

The Bacalhau WebUI offers an intuitive interface for interacting with the Bacalhau network. This guide provides comprehensive instructions for setting up, deploying, and utilizing the WebUI.

For contributing to the WebUI's development, please refer to the Bacalhau WebUI GitHub Repository.

Spinning Up the WebUI Locally

Prerequisites

• Ensure you have a Bacalhau v1.1.7 or later installed.

Running the WebUI

To launch the WebUI locally, execute the following command:

bacalhau serve --node-type=requester,compute --web-ui

This command initializes a requester and compute node, configured to listen on HOST=0.0.0.0 and PORT=1234.

Accessing the Local WebUI

Once started, the WebUI is accessible at http://127.0.0.1/. This local instance allows you to interact with your local Bacalhau network setup.

Accessing the WebUI from the Browser

For observational purposes, a development version of the WebUI is available at bootstrap.development.bacalhau.org. This instance displays jobs from the development server.

N.b. The development version of the WebUI is for observation only and may not reflect the latest changes or features available in the local setup.

You can use official Docker containers for each language, like R or Python. In this example, we will use the official R container and run it on Bacalhau.

In this tutorial example, we will run a "hello world" R script on Bacalhau.

Prerequisites​

To get started, you need to install the Bacalhau client, see more information here

1. Running an R Script Locally​

To install R follow these instructions A Installing R and RStudio | Hands-On Programming with R. After R and RStudio are installed, create and run a script called hello.R:

# hello.R
print("hello world")

Run the script:

Rscript hello.R

Next, upload the script to your public storage (in our case, IPFS). We've already uploaded the script to IPFS and the CID is: QmVHSWhAL7fNkRiHfoEJGeMYjaYZUsKHvix7L54SptR8ie. You can look at this by browsing to one of the HTTP IPFS proxies like ipfs.io or w3s.link.

2. Running a Job on Bacalhau​

Now it's time to run the script on Bacalhau:

export JOB_ID=$(bacalhau docker run \
    --wait \
    --id-only \
    -i ipfs://QmQRVx3gXVLaRXywgwo8GCTQ63fHqWV88FiwEqCidmUGhk:/hello.R \
    r-base \
    -- Rscript hello.R)

Structure of the command​

1. bacalhau docker run: call to Bacalhau

2. i ipfs://QmQRVx3gXVLaRXywgwo8GCTQ63fHqWV88FiwEqCidmUGhk:/hello.R: Mounting the uploaded dataset at /inputs in the execution. It takes two arguments, the first is the IPFS CID (QmQRVx3gXVLaRXywgwo8GCTQ63fHqWV88FiwEqCidmUGhk) and the second is file path within IPFS (/hello.R)

3. r-base: docker official image we are using

4. Rscript hello.R: execute the R script