BOINC¶

Category: Computing Type: Workload Template Tags: computing · science · distributed-computing

Overview¶

BOINC (Berkeley Open Infrastructure for Network Computing) is an open-source volunteer computing platform that uses your computer's CPU/GPU for scientific research — curing diseases, studying climate change, discovering pulsars, and more. This workload template deploys the LinuxServer.io BOINC desktop container, providing a web GUI to manage BOINC client projects and work units.

Each workload provisions a BOINC client with a persistent volume at /config to retain configuration across restarts. On startup, the BOINC client automatically attaches to the configured project using the provided project URL and weak account key — no manual setup needed. The web GUI is accessible via browser at the workload's endpoint URL for monitoring and project management. Optional GPU acceleration is available for compatible BOINC projects.

Architecture support: amd64 and arm64 — use tag latest (multi-arch) or pin with arm64v8-latest for ARM clusters.

How It Works¶

Workload Template — Installs the BOINC schema into Genesis. Once installed, the BOINC type appears in Genesis on the Workloads page, where it can be authored into a workload template. Users can then launch and provision BOINC manager instances on demand within a project through Hubble. After launch, access the BOINC Manager GUI through the Hubble endpoint list to attach to projects, monitor progress, and configure settings.

Prerequisites¶

A StorageClass must be available for the BOINC config PVC (mounted at /config)
A BOINC weak account key for the target project (obtain this from the project's website)
For GPU-accelerated workloads: NVIDIA GPU Operator plugin must be installed and nodes must have the nvidia runtime class configured

Installation¶

Open Terra and navigate to the Plugin Marketplace
Search for "BOINC"
Click Install
Click Confirm to deploy (no install-time fields required)

Once installed, the BOINC schema is available in Genesis. From the Workloads page, author the template — users can then launch and provision BOINC manager instances through Hubble.

Configuration¶

Install-Time Fields¶

No install-time configuration is required for this plugin.

Workload Launch Fields¶

These fields are configured when authoring the workload template in Genesis and used each time a user provisions a BOINC instance through Hubble:

Field	Details
`registry`	string · Required · Default: `lscr.io` Container registry for the BOINC image
`repo`	string · Required · Default: `linuxserver/boinc` BOINC image repository
`tag`	string · Required · Default: `latest` Image tag — use `latest` for current, `arm64v8-latest` for ARM64
`project_url`	string · Required BOINC project URL (e.g. `https://universeathome.pl/universe/`)
`account_key`	string · Required Weak account key for authenticating with the project
`gpu`	boolean · Required · Default: `false` Attach a GPU; sets `runtimeClassName: nvidia`
`storage_class`	k8sStorageClass · Required Storage class for the BOINC config PVC mounted at `/config`
`storage_size`	string · Optional · Default: `10Gi` Size of the BOINC config persistent volume
`arch`	select · Required · Default: `amd64` Target CPU architecture — `amd64`, `arm64`, or `either` (no constraint)

Resource Configuration¶

CPU and memory limits are automatically applied from the workload's Kubernetes resource requests/limits — no manual tuning needed.

Resource	Source	BOINC Setting
CPU	Pod `limits.cpu` via cgroup `cpu.max`	`<ncpus>` in `cc_config.xml` — limits concurrent tasks to the pod's CPU quota
Memory	Pod `limits.memory` via cgroup `memory.max`	`ram_max_used_busy_pct: 80` in `global_prefs_override.xml` — caps BOINC memory usage at 80% of the pod's limit

Both files are written during container initialization, before the BOINC client starts. Configuration is static — changing the workload's resource limits after launch requires a pod recycle (restart) to take effect.

Account Setup¶

Each BOINC project requires an account. To create one and connect it to the workload:

Create an account on the project's website, e.g.:
Asteroids@home — click "Join" and fill in the registration form
Universe@home — click "Join"
Einstein@home — click "Participate"
Get your weak account key — after creating the account, log in and visit your account page. Look for a link like "Account keys" or "Weak account key". It will be a long string like:
```
123456_abcdef1234567890abcdef1234567890ab
```
Configure the workload — when authoring the workload template in Genesis, set the project_url and account_key fields:
project_url: the project's BOINC URL (e.g. https://asteroidsathome.net/boinc/)
account_key: the weak account key from step 2

The BOINC client will automatically attach to the project on startup using these credentials. See the BOINC wiki for more details on weak account keys.

Custom Environment Variables¶

Genesis has a built-in env field for adding arbitrary environment variables at workload launch. These are commonly useful for the BOINC workload:

Variable	Description
`PASSWORD`	Password for the BOINC web GUI — enables HTTP basic auth
`TZ`	Timezone for log timestamps (e.g. `America/New_York`)

Notes¶

GPU-accelerated BOINC requires the NVIDIA GPU Operator plugin and cluster nodes with NVIDIA GPUs
The config PVC is mounted at /config — BOINC state and project configuration persist across pod restarts
Access the BOINC web GUI through the workload endpoint in Hubble after launch
The BOINC Manager GUI is used to attach to projects, manage work units, and monitor progress
This workload uses the LinuxServer.io BOINC image — see the upstream documentation for advanced configuration
Both amd64 and arm64 architectures are supported

plugins/boinc/terra.yaml
resource_id: boinc
name: BOINC
icon: https://raw.githubusercontent.com/juno-fx/Terra-Official-Plugins/refs/heads/main/plugins/boinc/assets/logo.png
description: Distributed computing client for BOINC projects
category: Computing
tags:
  - cluster-level
  - workload
  - computing
  - science
fields: []