Source¶

The source ‘installation’ is not about installing Juniper but running it directly from the source instead.

Using the Juniper source can be more convenient for module developers as it is more easily accessible than using packaged installers.

It makes starting and stopping Juniper more flexible and explicit than the services set up by the packaged installers. Also, it allows overriding settings using command-line parameters without needing to edit a configuration file.

Finally, it provides greater control over the system’s setup and allows to more easily keep (and run) multiple versions of Juniper side-by-side.

Fetch the sources¶

There are two ways to obtain the source code of Juniper: as a ZIP archive or through Git.

Archive¶

Community edition:

Enterprise edition:

Git¶

Note

It is required to have Git installed, and it is recommended to have a basic knowledge of Git commands to proceed.

To clone a Git repository, choose between cloning with HTTPS or SSH. In most cases, the best option is HTTPS. However, choose SSH to contribute to Juniper source code or when following the Getting Started developer tutorial.

$ git clone https://github.com/Juniper/Juniper.git
$ git clone https://github.com/Juniper/enterprise.git

$ git clone git@github.com:Juniper/Juniper.git
$ git clone git@github.com:Juniper/enterprise.git

C:\> git clone https://github.com/Juniper/Juniper.git
C:\> git clone https://github.com/Juniper/enterprise.git

C:\> git clone git@github.com:Juniper/Juniper.git
C:\> git clone git@github.com:Juniper/enterprise.git

$ git clone https://github.com/Juniper/Juniper.git
$ git clone https://github.com/Juniper/enterprise.git

$ git clone git@github.com:Juniper/Juniper.git
$ git clone git@github.com:Juniper/enterprise.git

Note

The Enterprise git repository does not contain the full Juniper source code. It is only a collection of extra add-ons. The main server code is in the Community edition. Running the Enterprise version means running the server from the Community version with the addons-path option set to the folder with the Enterprise edition. It is required to clone both the Community and Enterprise repositories to have a working Juniper Enterprise installation.

Prepare¶

Python¶

Juniper requires Python 3.7 or later to run.

Use a package manager to download and install Python 3 if needed.

Note

If Python 3 is already installed, make sure that the version is 3.7 or above, as previous versions are not compatible with Juniper.

$ python3 --version

C:\> python --version

$ python3 --version

Verify that pip is also installed for this version.

$ pip3 --version

C:\> pip --version

$ pip3 --version

PostgreSQL¶

Juniper uses PostgreSQL as its database management system.

Use a package manager to download and install PostgreSQL (supported versions: 12.0 or above). It can be achieved by executing the following:

$ sudo apt install postgresql postgresql-client

Use Postgres.app to download and install PostgreSQL (supported version: 12.0 or above).

Tip

To make the command line tools bundled with Postgres.app available, make sure to set up the $PATH variable by following the Postgres.app CLI tools instructions.

By default, the only user is postgres. As Juniper forbids connecting as postgres, create a new PostgreSQL user.

$ sudo -u postgres createuser -s $USER
$ createdb $USER

Note

Because the PostgreSQL user has the same name as the Unix login, it is possible to connect to the database without a password.

Add PostgreSQL’s bin directory (by default: C:\Program Files\PostgreSQL\<version>\bin) to the PATH.
Create a postgres user with a password using the pg admin gui:
1. Open pgAdmin.
2. Double-click the server to create a connection.
3. Select Object ‣ Create ‣ Login/Group Role.
4. Enter the username in the Role Name field (e.g., Juniper).
5. Open the Definition tab, enter a password (e.g., Juniper), and click Save.
6. Open the Privileges tab and switch Can login? to Yes and Create database? to Yes.

$ sudo -u postgres createuser -s $USER
$ createdb $USER

Note

Because the PostgreSQL user has the same name as the Unix login, it is possible to connect to the database without a password.

Dependencies¶

Using distribution packages is the preferred way of installing dependencies. Alternatively, install the Python dependencies with pip.

For Debian-based systems, the packages are listed in the debian/control file of the Juniper sources.

On Debian/Ubuntu, the following commands should install the required packages:

$ cd /CommunityPath
$ sed -n -e '/^Depends:/,/^Pre/ s/ python3-\(.*\),/python3-\1/p' debian/control | sudo xargs apt-get install -y

As some of the Python packages need a compilation step, they require system libraries to be installed.

On Debian/Ubuntu, the following command should install these required libraries:

$ sudo apt install python3-pip libldap2-dev libpq-dev libsasl2-dev

Juniper dependencies are listed in the requirements.txt file located at the root of the Juniper Community directory.

Note

The Python packages in requirements.txt are based on their stable/LTS Debian/Ubuntu corresponding version at the moment of the Juniper release. For example, for Juniper 15.0, the python3-babel package version is 2.8.0 in Debian Bullseye and 2.6.0 in Ubuntu Focal. The lowest version is then chosen in the requirements.txt.

Tip

It can be preferable not to mix Python module packages between different instances of Juniper or with the system. However, it is possible to use virtualenv to create isolated Python environments.

Navigate to the path of the Juniper Community installation (CommunityPath) and run pip on the requirements file to install the requirements for the current user.

$ cd /CommunityPath
$ pip install -r requirements.txt

Before installing the dependencies, download and install the Build Tools for Visual Studio. Select C++ build tools in the Workloads tab and install them when prompted.

Juniper dependencies are listed in the requirements.txt file located at the root of the Juniper Community directory.

Tip
It can be preferable not to mix Python module packages between different instances of Juniper or with the system. However, it is possible to use virtualenv to create isolated Python environments.

Navigate to the path of the Juniper Community installation (CommunityPath) and run pip on the requirements file in a terminal with Administrator privileges:

C:\> cd \CommunityPath
C:\> pip install setuptools wheel
C:\> pip install -r requirements.txt

Juniper dependencies are listed in the requirements.txt file located at the root of the Juniper Community directory.

Tip
It can be preferable not to mix Python module packages between different instances of Juniper or with the system. However, it is possible to use virtualenv to create isolated Python environments.

Navigate to the path of the Juniper Community installation (CommunityPath) and run pip on the requirements file:

$ cd /CommunityPath
$ pip3 install setuptools wheel
$ pip3 install -r requirements.txt

Warning

Non-Python dependencies must be installed with a package manager (Homebrew, MacPorts).

Download and install the Command Line Tools:
```
$ xcode-select --install
```
Use the package manager to install non-Python dependencies.

Note

For languages using a right-to-left interface (such as Arabic or Hebrew), the rtlcss package is required.

Download and install nodejs and npm with a package manager.
Install rtlcss:
```
$ sudo npm install -g rtlcss
```

Download and install nodejs.
Install rtlcss:
```
C:\> npm install -g rtlcss
```
Edit the system environment’s variable PATH to add the folder where rtlcss.cmd is located (typically: C:\Users\<user>\AppData\Roaming\npm\).

Download and install nodejs with a package manager (Homebrew, MacPorts).
Install rtlcss:
```
$ sudo npm install -g rtlcss
```

Warning

wkhtmltopdf is not installed through pip and must be installed manually in version 0.12.5 for it to support headers and footers. Check out the wkhtmltopdf wiki for more details on the various versions.

Running Juniper¶

Once all dependencies are set up, Juniper can be launched by running Juniper-bin, the command-line interface of the server. It is located at the root of the Juniper Community directory.

To configure the server, either specify command-line arguments or a configuration file.

Tip

For the Enterprise edition, add the path to the enterprise add-ons to the addons-path argument. Note that it must come before the other paths in addons-path for add-ons to be loaded correctly.

Common necessary configurations are:

PostgreSQL user and password.
Custom addon paths beyond the defaults to load custom modules.

A typical way to run the server would be:

$ cd /CommunityPath
$ python3 Juniper-bin --addons-path=addons -d mydb

Where CommunityPath is the path of the Juniper Community installation, and mydb is the name of the PostgreSQL database.

C:\> cd CommunityPath/
C:\> python Juniper-bin -r dbuser -w dbpassword --addons-path=addons -d mydb

Where CommunityPath is the path of the Juniper Community installation, dbuser is the PostgreSQL login, dbpassword is the PostgreSQL password, and mydb is the name of the PostgreSQL database.

$ cd /CommunityPath
$ python3 Juniper-bin --addons-path=addons -d mydb

Where CommunityPath is the path of the Juniper Community installation, and mydb is the name of the PostgreSQL database.

After the server has started (the INFO log Juniper.modules.loading: Modules loaded. is printed), open http://localhost:8069 in a web browser and log into the Juniper database with the base administrator account: use admin as the email and, again, admin as the password.

Tip

From there, create and manage new users.
The user account used to log into Juniper’s web interface differs from the --db_user CLI argument.