This project defines some minimal code needed to start a server that stores most of the information required by the FROG v.4 architecture:
- NF template: file that describes the characteristics of the network function, such as the its capapbility (e.g., firewall, NAT), required resources (e.g., amount of CPU, amount of memory), required execution environment (e.g., KVM hypervisor, Docker, etc), number and type of virtual interfaces, and more. Examples of templates, which have to follow a proper schema, are available in sample-templates;
- NF image: raw image of the NF (e.g, VM disk, archive file). The NF image can be installed either directly in the Datastore, or in a different backend (e.g., OpenStack Glance);
- NF-FG: file that describes a network functions forwaring graph, written according to a proper schema.
- YANG model: schema used to validate the configuration forwarded to the VNF by means of the configuration service
- YIN model: it is the JSON representation of a YANG model, it is used by the GUI in order to provide the users a simple interface for the management of their services
Moreover it provides a configuration service exploited by FROGv4 in order to retrieve/push configuration from/into VNFs.
sudo apt-get install mysql-server libmysqlclient-dev libffi-dev python-virtualenv python-pip libpython2.7-dev
During the installation process, you have to chose the root password for the SQL database on the machine where you are going to install/create the Datastore.
To create and initialize the SQL database:
mysql -u root -p
mysql> CREATE DATABASE frog4_datastore;
mysql> GRANT ALL PRIVILEGES ON frog4_datastore.* TO 'datastore'@'localhost' \
IDENTIFIED BY 'datastorePWD';
mysql> GRANT ALL PRIVILEGES ON frog4_datastore.* TO 'datastore'@'%' \
IDENTIFIED BY 'datastorePWD';
mysql> exit
where datastore and datastorePWD can be replaced respectively by the username and the password that the Datastore will use to access to the SQL database. In case you change them, you have also to edit these two lines in the python code.
Now you have to clone this repository and all the submodules using the following command:
$ git clone https://github.com/netgroup-polito/frog4-datastore
$ cd frog4-datastore
$ git submodule init && git submodule update
In order to properly configure the Datastore, edit its configuration file. Particularly, in this file you have to set the TCP port to be used to interact with the Datastore through its REST API.
Creating a private environment in which we run Django 1.8; unfortunately API changed in 1.9, hence it is better to use this specific version only in our project and start the server:
$ cd [frog4-datastore]
$ virtualenv .env
$ source .env/bin/activate
$ pip install django==1.8.2 djangorestframework==3.6.3 MySQL-python django-rest-swagger==0.3.5 django-chunked-upload==1.1.3 django-cors-headers==2.2.0 wrapt bcrypt pyang xmltodict jsonschema
$ python manage.py makemigrations datastore
$ python manage.py migrate
$ python manage.py runserver [--d datastore_config.ini]
$ cd [frog4-datastore]
$ source .env/bin/activate
$ python manage.py runserver [--d datastore_config.ini]
A description of the API exposed by the datastore is available at the URL: ip_address:port/docs (e.g., 127.0.0.1:8081/docs).
It is also included a management command in order to clear uncompleted NF image uploads which has expired, both from the datastore's DB and from disk. You can set how long (in hours) an upload is valid after its creation in the configuration file of the datastore (i.e. deafault-config.ini) by means of the variable upload_expiration_hrs.
You can launch the command manually from your virtual environment like this:
$ python manage.py delete_expired_uploads
You can also set a cron-job:
crontab -e
Add this line and save:
*/60 * * * * cd [frog4-datastore] && .env/bin/python manage.py delete_expired_uploads > /dev/null 2>&1
Change [frog4-datastore] with the full path where the datastore is located (i.e. /home/user/frog4-datastore).
You can verify if your cron-job is installed with the command
crontab -l