a very simple django app that tracks logins and logouts from public computers in our library. it collects data recording a machine name, its status, and a timestamp. its API is exposed by tastypie.

these instructions are optimized for running on a newly installed Ubuntu 12.04 LTS host.

$ sudo apt-get install python-dev python-virtualenv libapache2-mod-wsgi

same for whichever db you want to use:

$ sudo apt-get install postgresql-9.1 postgresql-server-dev-9.1
[or]
$ sudo apt-get install mysql-server libmysqld-dev
[etc]

Install Git

$ sudo apt-get install git-core

create the database you'll use (pgsql example included, do equiv for mysql):

% sudo su - postgres
(postgres)% psql
postgres=# create user MYUSERNAME with createdb password 'MYPASS';
CREATE ROLE
postgres=# \q
(postgres)% exit
% createdb -U MYUSERNAME MYDBNAME -W
Password: <'MYPASS'>

set up a directory to host the deployed code, and a deployment name inside there (setting perms as needed):

$ mkdir -p /loggins/tag_foo
$ cd /loggins/tag_foo

get the code for this application, loggins, from github, then check out the tag to deploy:

$ git clone https://github.com/gwu-libraries/loggins.git
$ cd loggins
$ git checkout tag_foo

create and activate a virtualenv:

$ virtualenv --no-site-packages ENV
$ source ENV/bin/activate

install dependencies (note: based on db server choice, use psycopg2 or MySQL-python as appropriate):

$ pip install -r requirements.txt

copy over the local settings template to a real local settings file, then add in correct ADMINS, DATABASE and SNMP_COMMUNITY_STRING info. Add host info in ALLOWED_HOSTS. Note: If you don't have a SNMP community string, just enter 'public'

$ cd loggins/loggins
$ cp local_settings.py.template local_settings.py
[edit edit edit]

copy over the wsgi template to a real wsgi file, uncomment the virtualenv details and specify the path to your virtualenv:

$ cp wsgi.py.template wsgi.py
[edit edit edit]

pull in the static files (for admin, etc.):

$ cd ..
$ ./manage.py collectstatic

set up the db with django. WARNING: Be sure you are still using your virtualenv. DO NOT create a superuser when prompted!

$ ./manage.py syncdb

Migrate the database to the latest updates

$ ./manage.py migrate

Create the database super user

$ ./manage.py createsuperuser

at this point you should be able to run the app in debugging mode:

$ ./manage.py runserver

if your high ports are closed and you need to look at it locally at the commandline, use links:

$ links http://127.0.0.1:8000/

if it's working, you're good to go!

to install into apache, copy and install your apache config into apache's sites, then edit it to use proper paths and hostname info, optionally enabling SSL (cert/key handling follows standard approach see http://httpd.apache.org/docs/2.2/ssl/ for details):

$ sudo cp loggins/apache.conf.template /etc/apache2/sites-available/loggins_beta_001
[EDIT EDIT EDIT]

[optionally disable the default site, then] enable the loggins site, then reload apache:

$ sudo a2dissite default
$ sudo a2ensite loggins
$ sudo /etc/init.d/apache2 reload

you should be up and running. to test it out, use curl:

curl script to record a login event
$ curl -i -H "Content-Type: application/json" -H "Accept: application/json" -H "Authorization: ApiKey <YOURUSERNAME>:<YOURAPIKEY>" -X PUT -d "{\"ip_address\":\"<IPADDRESS>\",\"state\":\"i\",\"observation_time\":\"$(date +%Y-%m-%dT%H:%M:%S)\"}" http://<YOUR.SERVER.NAME>:<PORT>/api/v1/location/<HOSTNAME>/

curl script to record a logout event
$ curl -i -H "Content-Type: application/json" -H "Accept: application/json" -H "Authorization: ApiKey <YOURUSERNAME>:<YOURAPIKEY>" -X PUT -d "{\"ip_address\":\"<IPADDRESS>\",\"state\":\"a\",\"observation_time\":\"$(date +%Y-%m-%dT%H:%M:%S)\"}" http://<YOUR.SERVER.NAME>:<PORT>/api/v1/location/<HOSTNAME>/

Additional Location attributes that can be updated using the API are, "building", "floor", "hostname", "os", and "station_name" to create user/accounts/api keys, go to the django admin site at /admin, and under Users create new users as needed. new api keys are generated when you create new accounts. the generated api key for each user is listed at the bottom of their account page in the admin ui.

to update existing api keys, go to the django admin page for the user account in question and update the api key value at the bottom.

to generate some test data, use the login_monkey management command. this will generate new records, one per random host, along with a set of initial testing hosts, at an interval you can specify. by default, it creates 20 "monkey" hosts, spread throughout the first floors of three of our libraries ('g', 'e', and 'v'). for example, this will add a new record every five seconds:

$ ./manage.py login_monkey --interval 5 --verbose

To set up live status updates from the computers on the network, the scripts in the /scripts folder must be deployed as follows:

• map_network_hosts.sh should be run as a cron job, with a recommended interval of 1 day.
• capture_snmp_state.sh should be run as a cron job, at the interval desired for state updates (e.g. 30 seconds)
• update_mac_statuses.sh should be run as a cron job, at the interval desired for state updates (e.g. 30 seconds)
• mac_loggins-login.sh should be deployed to each Mac workstation as a shell script to run upon logging in
• mac_loggins-logout.sh should be deployed to each Mac workstation as a shell script to run upon logging out

Note that the scripts must be edited to configure values for LOGGINS_PATH_TO_ENV, LOGGINS_PROJECT_DIR, NETWORK_PREFIXES, API_KEY, API_USERNAME, API_URL