Installation
It is recommended to visit the hardware reqirements page to learn more about deployment options and hardware requirements.
Tip
It is recommended to have a high speed network connection, as multiple GBs of data will be downloaded during the installation.
Prerequisites
The system can be installed on any OS that supports docker, however the recommended OS is Ubuntu 22.04. The steps described below assume an Ubuntu 22.04 installation.
Info
At the moment, U-Query software components are only available from a private dockerhub repository. Contact Ultinous to get access.
Info
The installation will require you to have sudo rights.
1. System update
Make sure your system is up-to-date.
sudo apt update
sudo apt upgrade -y
2. Install docker
Both indexer nodes and query nodes require docker. If your system does not have it installed, follow these steps before continuing.
3. Install NVIDIA driver and NVIDIA docker runtime (indexer node only)
If your system does not have NVIDIA docker runtime installed, follow these steps before continuing.
Installation steps
1. Download deployment scripts
Download and extract the configuration scripts to the u-query folder.
wget https://u-query.ultinous.com/docs/r2/u_query_install.tar -P /tmp
sudo tar -xvf /tmp/u_query_install.tar -C /
sudo chgrp -R $(id -g) /opt/u-query
2. Set database password
Create a unique password for the database, and set it in docker-compose's .env file. Examples below;
- Using
pwgen
:
sed -i "s/\(DB_POSTGRES_PASSWORD=\)/\1$(pwgen -s 27 1)/g" /opt/u-query/.env
- Alternatively, using 'standard' tools:
sed -i "s/\(DB_POSTGRES_PASSWORD=\)/\1$(< /dev/urandom tr -dc 'A-Za-z0-9' | head -c 27)/g" /opt/u-query/.env
3. Login to docker
Login to dockerhub to be able to download software components.
Info
Contact Ultinous to get access.
docker login
4. Model optimization (indexer node only)
If the installed node is an indexer node (or standalone installation), model optimization need to be performed. This process optimizes deep learning models for the current GPU.
Execute the following commands to start model optimization.
cd /opt/u-query
docker compose -f build_trt.yaml pull
docker compose -f build_trt.yaml up -d
Tip
Model optimization can take up to 20 minutes. Use docker logs -f build_trt
to see how it progresses.
5. Download software components
cd /opt/u-query
docker compose pull
docker logout
6. Start U-Query services
cd /opt/u-query
docker compose up -d