Mdadm is a Linux utility to manage and monitor software RAID. The mdadm monitoring plugin keeps track of failed disks in a RAID setup.

Mdadm metrics

- Faulty
- Active
- Resync
- Read only

Mdadm dependancies

This plugin needs the mdstat python plugin, you can install this by running pip install mdstat.
To fetch the status we also have to give this plugin sudo access, edit /etc/sudoers and add the following at the end of the file.

nixstats ALL=(ALL) NOPASSWD: /usr/local/bin/mdjson

This is considering mdjson is located at /usr/local/bin/mdjson to verify it's location run whereis mdjson

Testing the mdadm plugin

Run sudo -u nixstats nixstatsagent test mdstat to check if it's returning any data.

root@nixstats:~# sudo -u nixstats nixstatsagent test mdstat
mdstat:
{
    "md0": {
        "active": 1,
        "faulty": 0,
        "read_only": 0,
        "resync": 0
    },
    "md1": {
        "active": 1,
        "faulty": 0,
        "read_only": 0,
        "resync": 0
    },
    "md2": {
        "active": 1,
        "faulty": 0,
        "read_only": 0,
        "resync": 0
    }
}

Enable the plugin

Open /etc/nixstats.ini and append the following lines at the end of the file.

[mdstat]
enabled = yes

Restart the plugin by running service nixstatsagent restart

Creating charts

Click on the Metrics link on the top menu, now select "mdstat" as metric type and active/faulty/read_only/resync as metric, choose the servers you would like to graph and save it to your dashboard.

You can create an alert for in case the faulty metric is higher than zero, indicating that a server has a bad drive.

Updated on: 15 / 03 / 2018