First manual deployment | Tdg
Administrator’s guide Deployment First manual deployment

First manual deployment

This guide explains how to quickly deploy Tarantool Data Grid (TDG) manually for the first time. As a result, you will get a single node TDG cluster deployed locally.


For TDG deployment, you need Linux OS (CentOS 7/RHEL 7 are preferable). Otherwise, you’ll need to set up a virtual machine with Linux OS first.

Getting a TGZ file for deployment

To deploy Tarantool Data Grid, you need an RPM (.rpm), TGZ (tar.gz), or Docker image (docker-image.tar.gz) file. For the first deployment, a TGZ file will do just fine. It is easier to deploy and does not require root access.

Download a TGZ file of the latest version at the customer zone of Make sure your browser did not unarchive the downloaded file: the file extension should be tar.gz.

If you do not have access to the customer zone, you can get one by applying this form or writing to


  1. Unpack tar.gz file:

    $ tar xzf tdg-<VERSION>.tar.gz # change <VERSION> for the TDG version that you've downloaded
  2. Run a single node TDG cluster inside the unpacked tar.gz file:

    $ ./tarantool ./init.lua --bootstrap true

    If you already have Tarantool installed, make sure that now, while deploying TDG, you use the Tarantool version that is packed in the just downloaded tar.gz archive.

  3. Go to to check the deployed TDG:

    Сконфигурированный экземпляр

    By running tarantool ./init.lua --bootstrap true, you’ve deployed a configured instance with assigned roles. If you want to try and assign roles by yourself, run:

    $ tarantool ./init.lua

    As a result, you’ll get an unconfigured TDG instance:

    Экземпляр без заданной конфигурации

    In case you want to start over and deploy TDG from scratch, don’t forget to delete the configuration, xlog, and snap files that TDG created during the first deployment:

    $ rm -rf ./dev/output
Found what you were looking for?