Installation
Follow these simple steps to install a new instance of Supervisely to your server.
Pre-requirements
Before you start the installation of Supervisely to your host, please make sure your server meets the hardware & software requirements below.
Hardware
| Component | Minimum | Recommended | Videos |
|---|---|---|---|
CPU | 4 | 8 | 8+ |
RAM | 8Gb | 32Gb | 32Gb+ |
Disk | 100Gb SSD | 200Gb SSD | 200Gb SSD + Cloud storage |
Choose the disk size according to your needs. There is also an option to attach any S3 compatible storage to distribute the data in your cloud.
Make sure you are using SSD disk for OS, docker data and DATA_PATH/db. Otherwise, the performance will be significantly lower
If you don't have enough free space on a system drive, but you have another one mounted, you can change the place where the data is stored by changing DATA_PATH in .env file.
Please check this article for more details.
Never set DATA_PATH pointing to a network share (NFS/SMB/ESB/etc), because it affects the performance significantly. Set DATA_PATH pointing to an actual directory on your local disk and then mount storage subfolder of DATA_PATH directory as a network share - this way "hot" subfolders like db or proxy_cache will be on a local disk and image / videos storage on a network share.
Network
By default Supervisely exposes 2 ports on its host. Both ports are configurable.
PROXY_PORT is used for the HTTP server. Default to 80
if you're using Supervisely's built-in HTTPS then port 443 will be exposed as well
NET_SERVER_PORT is used for WireGuard communication between the server and the Apps. Default to 51822 UDP
Make sure that the agent can reach the WireGuard server via {SERVER_ADDRESS}:{NET_SERVER_PORT}. Otherwise the GUI Apps won't work
You can find the configuration file .env in the Supervisely workdir:
After changing the values you have to redeploy Supervisely services:
GPU instance deployment
If you plan to use Smart Tool (AI powered semantic segmentation) or Neural Networks (human-in-the-loop) you will need a server with GPU. The requirements are as follows:
GPU Memory: 8 GB or more (GeForce GTX 1080, Tesla K80, Tesla P100)
You can install Supervisely and run GPU computations on different machines using agents (see How to connect agents).
AWS & cloud providers
We recommend to use the following EC2-instances for deployment in AWS:
1 × m5.2xlarge — the platform itself
1 × p2.xlarge — (for Neural Network Module only)
Software
To run Supervisely you will need Linux OS with kernel 3.10 or newer. The following configuration is recommended:
OS: Ubuntu 18.04 or later
NVIDIA graphics driver: 460.27.03 or later (for Neural Network module only)
nvidia-docker2 (for Neural Network module only)
All pre-requirements can be automatically installed via supervisely-cli utility for Ubuntu 18.04 and later. In case of a different Linux distribution, you can install Supervisely and pre-requirements manually, but we won't be able to provide installation support.
Installation
When all requirements above have been installed, you can deploy Supervisely. Please follow those steps one by one. Read each step carefully and complete it fully before going to the next step.
Step 1. Get your unique key from us
Before the installation we will send you a license key and a supervisely-cli installation command. Please run it on the machine where you want to install Supervisely.
Step 2. Install pre-requirements
Run sudo supervisely install-all in your terminal. We will detect the necessary dependencies and install them. The following software may be installed:
NVIDIA Driver (for Training Module)
NVIDIA Container Toolkit (for Training Module)
If you don't have NVIDIA drivers, you will be asked if you need to install it. Choose y if you have GPU on your server. Your computer will be rebooted during the installation of the NVIDIA driver.
In case install-all doesn't support your OS, you can install the dependencies manually. This script should work on most of the Linux distributions:
Step 3. Run the installation command
Run sudo supervisely init "<your license key>". If this is your first time running this command, you will see three prompts, asking you to choose where to store the CLI configuration file (first prompt), Supervisely configuration files (second prompt) and backups (the last prompt).
You can select one of the suggested options or enter your custom location. If you are not sure which option to choose, you can select the default (/opt/supervisely).
You can always check where your configuration and backups locations are stored using supervisely where command.
Step 4. Wait for the installation to complete
Now, the CLI will automatically do the following steps, required to set your Supervisely instance:
Download docker-compose configuration files
Login to our docker registry
Pull and run docker containers
Initilize the database and download some sample data
Install the license key
Deploy the default Supervisely Agent
Step 5. Start using Supervisely!
Now you can refresh the current page. You will see the login box:
Enter the default credentials: login and password admin. Now you are in. Welcome!
We strongly advice you to complete the post-installation steps after the initial setup.
Troubleshooting
I open "http://localhost" but see an error
Database is still initializing. Wait a minute and try again.
I try to login but see "You have no permissions for this action. Please, contact administrator."
One of our internal services haven't got the license key. Refresh the page and try again.
My "admin / Main Node" agent is stuck in "Waiting" status
It seems the agent has tried to wait for Supervisely to start. Restart agent manually: sudo supervisely deploy-agent.
When i deploy an agent i see a "runtime" error
Nvidia container runtime has not been installed. Try to run command docker run --runtime=nvidia --rm nvidia/cuda nvidia-smi to see if it works.
If you don't want to run train & inference tasks, please remove checkbox "Use nvidia runtime" under advances settings in the new node modal window and run the deploy command again.
Last updated
