Server Installation

The Pankha Fan Control server is deployed with Docker Compose: one app container (dashboard + control logic) and one PostgreSQL container (history). This page is the full setup reference - if you just want fans spinning, the Quick Start covers the fast path.

Prerequisites

Docker Engine with the Docker Compose plugin.
A machine that’s always on - the server is the brain; when it’s down, agents fall back to their failsafe speed.
Resources are modest: a few hundred MB of RAM; disk grows with your data-retention setting.

1. Download the Release Files

Every release ships the production compose.yml and an example.env. Use them as-is:

mkdir pankha && cd pankha
wget https://github.com/Anexgohan/pankha/releases/latest/download/compose.yml
wget https://github.com/Anexgohan/pankha/releases/latest/download/example.env -O .env

Don’t hand-edit compose.yml - all normal customization happens in .env. The one exception is the image tag (see Release Channels).

2. Configure `.env`

Open .env in an editor. The top section is marked REQUIRED CONFIGURATION:

Variable	Required?	What it does
`PANKHA_HUB_IP`	Yes	This server’s LAN IP or hostname (e.g. `192.168.1.100`). Agent install commands are built from it - the shipped placeholder will not work.
`POSTGRES_USER` / `POSTGRES_PASSWORD`	Yes	Your database credentials - pick your own.
`PANKHA_PORT`	No (default `3143`)	The port the dashboard and agents connect to.
`TIMEZONE`	No (default UTC)	Uncomment and set (e.g. `Asia/Kolkata`) so dashboard times and logs match your clock.
`LOG_LEVEL`	No (default `info`)	Server log verbosity - also changeable later from the dashboard’s Settings.
`POSTGRES_DB` / `POSTGRES_HOST` / `POSTGRES_PORT`	No	Leave at defaults for the bundled database container.
`PANKHA_STAGING_DIR`	No	Where staged agent binaries live inside the container - leave as-is.
`POSTGRES_MAX_WAL_SIZE` and friends	No	PostgreSQL disk-usage tuning - leave as-is unless you know why.

External PostgreSQL (advanced): the backend builds its database connection from the POSTGRES_* variables, so pointing POSTGRES_HOST/POSTGRES_PORT at an existing PostgreSQL instance works - load backend/src/database/schema.sql into it yourself, and don’t start the bundled pankha-postgres service.

3. Start It

docker compose pull && docker compose up -d

The dashboard is at http://<server-ip>:3143 (or your PANKHA_PORT).

Deployment Architecture

graph TD
    User([You]) -->|docker compose up -d| Docker[Docker Compose]

    subgraph "Server machine"
        Docker --> App["pankha_app (:3143)<br/>dashboard + control logic"]
        Docker --> DB[("pankha_postgres<br/>history")]
        App <--> DB
    end

    User -.->|browser| App

Verification

docker compose ps          # both containers Up (healthy)
curl http://localhost:3143/health

Both services have healthchecks - the app waits for the database to be healthy before starting.

Release Channels

The image tag in compose.yml picks your channel - this is the one line in the compose file you may want to edit:

Tag	Meaning
`anexgohan/pankha:latest`	Stable releases (default)
`anexgohan/pankha:beta`	Absolute latest
`anexgohan/pankha:testing`	Pre-releases

Where Data Lives

Everything persistent sits next to your compose.yml, so backing up the folder backs up the system:

pankha/
├── compose.yml
├── .env
└── docker-data/
    ├── backend/database/postgres_data/   # PostgreSQL data
    └── staging/                          # Staged agent binaries (Deployment Center)

The PostgreSQL container’s ports are deliberately not exposed to the network - only the app talks to it.

Updating

docker compose down && docker compose pull && docker compose up -d

Your .env, database, and staged binaries are untouched by updates.

Next Steps

Quick Start: deploy your first agent and put a fan under control.
Server Configuration: environment variables and runtime settings in depth.
Deployment Center: roll agents out to your machines.