Monitor disk health with HDSentinel and collectd

CC-BY-4.0 by Peter Molnar () at

A little shell script to connect HDSentinel to collectd.

HDSentinel1 is a brilliant little freeware capable of showing S.M.A.R.T. data from the disks in various ways. Since these values are important in monitoring and I’m using collecd as monitoring system, I needed a way to integrate the two. I ended up using the Exec plugin of collectd.

The steps:

chmod 0755 HDSentinel
mv HDSentinel /usr/bin/hdsentinel

Add the monitor scripts:

vim /etc/collect/scripts/
#!/usr/bin/env bash

HOSTNAME=`uname -n`

disks=`ls /dev/sd[a-z]`

while sleep "${INTERVAL}"; do
    hdsentinel=`/usr/bin/sudo /usr/bin/hdsentinel`
    for disk in ${disks}; do
        data=`echo "${hdsentinel}" | grep -A11 ${disk}`
        name=`echo "${data}" | grep "HDD Device" | awk '{print $4}'`

        temp=`echo "${data}" | grep "Temperature" | awk '{print $3}'`
        echo "PUTVAL "${HOSTNAME}/hdsentinel-${disk}/temperature" interval=$INTERVAL N:${temp}"

        tempmax=`echo "${data}" | grep "Highest Temp" | awk '{print $3}'`
        echo "PUTVAL "${HOSTNAME}/hdsentinel-${disk}/temperature-max" interval=$INTERVAL N:${tempmax}"

        health=`echo "${data}" | grep "Health" | awk '{print $3}'`
        echo "PUTVAL "${HOSTNAME}/hdsentinel-${disk}/percent-health" interval=$INTERVAL N:${health}"

        performance=`echo "${data}" | grep "Performance" | awk '{print $3}'`
        echo "PUTVAL "${HOSTNAME}/hdsentinel-${disk}/percent-performance" interval=$INTERVAL N:${performance}"


Note: “percent” and “temperature” are types of collectd. Thx deadite662!

Add the collectd plugin:

vim /etc/collectd/collectd.conf
LoadPlugin exec
<plugin exec>
  Exec "nobody" "/etc/collectd/scripts/"

Add “nobody” to sudoers, but only for hdsentinel ( install sudo if you need to, hdsentinel needs root access to run, but collectd will not exec with root user ):

vim /etc/sudoers
# add to the end
nobody    ALL=(ALL) NOPASSWD: /usr/bin/hdsentinel