NS News & Views

By Clive Norman
List all 36 articles

Tracking devices in UniFi WiFi

| Tags: JavaScript UniFi SysAdmin DevOps

Several years ago, we migrated from a Cisco Wi-Fi infrastructure to a UniFi setup.

UniFi (from Ubiquiti Networks) offer a pretty remarkable WiFi solution, especially when it comes down to cost.  I shan’t go into the in’s and out’s of UniFi, as apart from anything else, I can’t claim to be an expert in the technology.  In many ways, it just works!

What’s interesting, is the database supporting the UniFi controller is MongoDB (a NoSQL database, that can be queried using JavaScript).

Again, whilst I wouldn’t claim to be a MongoDB expert, I do like the idea of being able to use JavaScript to query data.


Lost iPad

Recently, we had a situation where an iPad had gone missing.  Although it is possible to establish an approximate location of a given device, using UniFi’s controller system, clearly this is only effective whilst the device is connected; the next best option is a singular ‘last seen’ log entry.

So in this particular instance, there wasn’t any real ‘tracking’ data available to work with.  However, UniFi was logging connection information, so I concluded that there must be a way of extracting this information from the backend.

The first thing I did was download and install a GUI based MongoDB management tool called RoboMongo (ok I could have consoled my way to success, but the GUI just felt a little easier).

Installing and configuring RoboMogo is really very easy.  In our instance, it was just install and connect to ‘localhost’ using port 27117.

Tracking Script

Once RoboMongo was installed, it was then a matter of scripting out an appropriate query.

Using the beg, borrow and manipulate principle – I ended up with an effective script.

The script takes two parameters – a device ID (which is a wildcard by default) and the total number of hours that you wish to search back.

It produces a list of results, detailing the device (or devices), along with every AP the device(s) has connected to, complete with the time at this location; all going back for as long you declared in hours (or have enough data to search).

// How many hours to summarize
var hours = 1400;
// User to summarize
var user = "Ipad";
var users = db.user.find({
    "hostname": new RegExp(user)
});

for (var j = 0; j < users.length(); j++) {
    var user = users[j];
    var sessions = db.session.find({
        "mac": user.mac,
        disassoc_time: {
            $gt: new Date().getTime() / 1000 - hours * 3600
        }

    }).sort({
        disassoc_time: -1
    });


    for (var k = 0; k < sessions.length(); k++) {
        var session = sessions[k];
        var hostname = user.hostname + " ";
        var aps = db.device.find({
            "mac": session.ap_mac
        });
        for (var a = 0; a < aps.length(); a++) {
            var ap = aps[a];
        }
        print("AP: " + ap.name + " | DateTime: " + new Date(session.disassoc_time * 1000) + " | " + "Username: " + hostname.substr(0, 29));
    }
}

Result

In our particular instance, the results were astonishing; we could trace a complete history of the device’s journey (even though it was no longer turned on).

I am also very happy to report that the device in question, was found safe and well.