Install Inmanta

This page explains how to install the Inmanta orchestrator software and setup an orchestration server. Regardless what platform you installed it on, Inmanta requires at least the latest Python 3.6 or 3.7 and git.

Install the software

For CentOS use yum:

sudo tee /etc/yum.repos.d/inmanta_oss_stable.repo <<EOF
[inmanta-oss-stable]
name=Inmanta OSS stable
baseurl=https://pkg.inmanta.com/inmanta-oss-stable/el7/
gpgcheck=1
gpgkey=https://pkg.inmanta.com/inmanta-oss-stable/inmanta-oss-stable-public-key
repo_gpgcheck=1
enabled=1
enabled_metadata=1
EOF

sudo yum install -y epel-release
sudo yum install -y python3-inmanta python3-inmanta-server python3-inmanta-agent

The first package (python3-inmanta) contains all the code and the commands. The server and the agent packages install config files and systemd unit files. The dashboard is installed with the server package.

First make sure Python >= 3.6 and git are installed. Inmanta requires many dependencies so it is recommended to create a virtual env. Next install inmanta with pip install in the newly created virtual env.

# Install python3 >= 3.6 and git
sudo python3 -m venv /opt/inmanta
sudo /opt/inmanta/bin/pip install inmanta
sudo /opt/inmanta/bin/inmanta --help

The misc folder in the source distribution contains systemd service files for both the server and the agent. Also install inmanta.cfg from the misc folder in /etc/inmanta/inmanta.cfg

If you want to use the dashboard you need to install it as well. Get the source from our github page Next, build and install the dashboard. For this you need to have yarn and grunt:

tar xvfz inmanta-dashboard-20xx.x.x.tar.gz
cd inmanta-dashboard-20xx.x.x
yarn install
grunt dist

This creates a dist.tgz file in the current directory. Unpack this tarball in /opt/inmanta/dashboard and point the server in /etc/inmanta/inmanta.cfg to this location: set dashboard.path to /opt/inmanta/dashboard

On Windows only the compile and export commands are supported. This is useful in the Push to server deployment mode of inmanta. First make sure you have Python >= 3.6 and git. Inmanta requires many dependencies so it is recommended to create a virtual env. Next install inmanta with pip install in the newly created virtual env.

# Install python3 >= 3.6 and git
python3 -m venv C:\inmanta\env
C:\inmanta\env\Script\pip install inmanta
C:\inmanta\env\Script\inmanta --help

Get the source either from our release page on github or clone/download a branch directly.

git clone https://github.com/inmanta/inmanta-core.git
cd inmanta
pip install -c requirements.txt .

Warning

When you use Inmanta modules that depend on python libraries with native code, python headers and a working compiler are required as well.

Configure server

This guide goes through the steps to set up an Inmanta service orchestrator server. This guide assumes a RHEL 7 or CentOS 7 server is used. The rpm packages install the server configuration file in /etc/inmanta/inmanta.cfg.

Optional step 1: Setup SSL and authentication

Follow the instructions in Setting up authentication to configure both SSL and authentication. While not mandatory, it is highly recommended you do so.

Step 2: Install PostgreSQL 10

PostgreSQL 10 can be installed by following the installation guide for your platform.

Step 3: Setup a PostgreSQL database for the Inmanta server

Initialize the PostgreSQL server:

sudo /usr/pgsql-10/bin/postgresql-10-setup initdb

Start the PostgreSQL database and make sure it is started at boot.

sudo systemctl enable --now postgresql-10

Create a inmanta user and an inmanta database by executing the following command. This command will request you to choose a password for the inmanta database.

sudo -u postgres -i sh -c "createuser --pwprompt inmanta; createdb -O inmanta inmanta"

Change the authentication method for local connections to md5 by changing the following lines in the /var/lib/pgsql/10/data/pg_hba.conf file

# IPv4 local connections:
host    all             all             127.0.0.1/32            ident
# IPv6 local connections:
host    all             all             ::1/128                 ident

to

# IPv4 local connections:
host    all             all             127.0.0.1/32            md5
# IPv6 local connections:
host    all             all             ::1/128                 md5

Restart the PostgreSQL server to apply the changes made in the pg_hba.conf file:

sudo systemctl restart postgresql-10

Step 4: Set the database connection details

Add a /etc/inmanta/inmanta.d/database.cfg file as such that it contains the correct database connection details. That file should look as follows:

[database]
host=<ip-address-database-server>
name=inmanta
username=inmanta
password=<password>

Replace <password> in the above-mentioned snippet with the password of the inmanta database. By default Inmanta tries to connect to the local server and uses the database inmanta. See the database section in the configfile for other options.

Step 5: Set the server address

When virtual machines are started by this server that install the inmanta agent, the correct server.server-address needs to be configured. This address is used to create the correct boot script for the virtual machine.

Set this value to the hostname or IP address that other systems use to connect to the server in the configuration file stored at /etc/inmanta/inmanta.d/server.cfg.

[server]
server-address=<server-ip-address-or-hostname>

Note

If you deploy configuration models that modify resolver configuration it is recommended to use the IP address instead of the hostname.

Step 6: Configure ssh of the inmanta user

The inmanta user that runs the server needs a working ssh client. This client is required to checkout git repositories over ssh and if the remote agent is used.

  1. Provide the inmanta user with one or more private keys:

  1. Generate a new key with ssh-keygen as the inmanta user: sudo -u inmanta ssh-keygen -N ""

  2. Install an exiting key in /var/lib/inmanta/.ssh/id_rsa

  3. Make sure the permissions and ownership are set correctly.

ls -l /var/lib/inmanta/.ssh/id_rsa

-rw-------. 1 inmanta inmanta 1679 Mar 21 13:55 /var/lib/inmanta/.ssh/id_rsa
  1. Configure ssh to accept all host keys or white list the hosts that are allowed or use signed host keys (depends on your security requirements). This guide configures ssh client for the inmanta user to accept all host keys. Create /var/lib/inmanta/.ssh/config and create the following content:

Host *
    StrictHostKeyChecking no
    UserKnownHostsFile=/dev/null

Ensure the file belongs to the inmanta user:

sudo chown inmanta:inmanta /var/lib/inmanta/.ssh/config
  1. Add the public key to any git repositories and save if to include in configuration models that require remote agents.

  2. Test if you can login into a machine that has the public key and make sure ssh does not show you any prompts to store the host key.

Step 7: Configure the server bind address

By default the server only listens on localhost, port 8888. This can be changed by altering the server.bind-address and server.bind-port options in the /etc/inmanta/inmanta.d/server.cfg file.

[server]
bind-address=<server-bind-address>
bind-port=<server-bind-port>

Step 8: Start the Inmanta server

Start the Inmanta server and make sure it is started at boot.

sudo systemctl enable --now inmanta-server

The server dashboard is now available on the port and host configured in step 7.

Optional Step 9: Setup influxdb for collection of performance metrics

Follow the instructions in Performance Metering to send performance metrics to influxdb. This is only recommended for production deployments.

Optional Step 10: Configure logging

Logging can be configured by following the instructions in Logging.