Observium HDD temperature

Leave a Reply

Comment as a guest.

  1. Hi,

    What about hddtemp? It’s installable on Linux and has a supported Unix agent script 🙂

    PS: Can’t fill out the Name and Email fields on Firefox, had to switch to Chrome… :-/

    1. Have been wanting to check out more of the Observium unix agents scripts, have only used the nginx, mysql & powerdns check_mk plugins. Converting my MailScanner / Spamassassin Observium stats into a check_mk plugin is on the todo, but need to find some time 🙂

  2. Hi there,
    I looked at the wiki page with hddtemp but there’s hardly any info and I’m not sure how that should work.

    I already have smartmontools setup and getting the temperatures in Observium was a piece of cake. But how do I add the rest of the SMART counters? Temperatures are the least of our worries when it comes down to HDD reliability, prefailure counters are at least as important. Can you help me set these up in Observium?


    1. Hi Pablo,

      We have a working solution for the other values, however, we implemented it via the Observium application ‘interface’. This writes directly to the RRD graphing files, and does not store the values in the Observium database. This was the cleanest solution, however, you will not be able to use the values with the Observium Alerting system.

      Let me know if you are still interested, then I’ll clean up the code, and put it online.

      1. Yes, definitely interested if its not too much trouble.

        I do wonder how you deal with reinstalls/crashes etc, as your values are not stored in the database you will loose historic data and trend analysis.

        How hard is it to use Glen Pitt-Pladdy’s way for the other values too? That stores everything in its DB doesn’t it? All SMART counters are dumped and the perl script can pull any one of them. Isn’t it just a matter of making that custom include php file use it?

        I tried staring at that file intensely and even attempted to frown at it too, but it refused to make sense to me. 😳


  3. Glen Pitt-Pladdy perl script just parses the smartctl output, and provides an easy way to get the values.

    The complete flow (using this post),

    1) cronjob generates intermediary files (/var/cache/snmp/smart-*) with the smartctl output

    2) when an snmp client connects, snmpd executes Glen Pitt-Pladdy’s perl script with the smart ‘ID’ as a parameter. The script will then parse the intermediary files and gets the specific value for that smart ID.

    3) because Observium already has a ‘temperature’ sensor, it is reused for this solution. And because of this, it is stored in the Observium db.

    4) for the other values (error rate, seek errors, etc) there are no ‘easy’ solutions in Observium. It would require quite some additional code to make it happen.

    Thus, the easy way to implement is not using Observium’s sensors, and use the applications framework.

    This stores all the data inside of the RRD files.

    If you backup the Observium RRD files, and the database dump, it normally should keep the historic data even after reinstalls.

    I’ll provide a new post within some days with the solution.

    1. Alright, I get that 😀 .

      Thanks for explaining. As long as I can keep historic data I don’t mind if its in the DB or not so it will be fine.

      Looking forward to your solution, really appreciate you’re taking the time to do this!

      Thanks again!

    2. Don’t mean to push my luck here, but while you’re at it, do you think its possible to use device ID’s (/dev/disk/by-id) instead of sda, sdb, if you’re not already doing that? I suppose if you’re monitoring more than a few disks youself this would be helpful for you as well.

      Several reasons for this really, for instance sometimes after a reboot the names sda, sdb etc can get mingled. If they change, graphs will display the wrong info. Or when you need to replace a disk it could continue in an old graph. Using device ID’s would prevent this and makes it easier to identify which disk is doing what etc.

      Thanks again!

Read Next

Sliding Sidebar