Monitor several OMV servers

    • Monitor several OMV servers

      I have a bunch of OMV-servers running on ODROID HC2s. Two at the moment. Perhaps four soon.

      I would like to have the ability to monitor them a little more easily than having to login to each web interface.

      When the HC2s are under heavy utilization, sustained copying of GBs or even TBs, I worry about CPU and HDD temperatures.

      Before I start looking for software/methods myself, I thought I should ask if someone here can recommend some simple monitoring software?

      I have one HC2 dedicated as "control", with an SSD, that can run the monitoring software along with other software.

      Ideally the monitoring software should have a web interface so I can open a page on a browser on the phone, or an app, and at a glance see basic things like:

      General health.
      HDD temperature.
      Share utilization.
      CPU temperature.
      CPU load.

      ... for all servers, perhaps also for the last few hours.

      Can someone recommend something suitable? Small. Easy to config. Open source. Free.

      Or is this a case for DIY hacking? Perhaps with Node-Red? Does OMV have an API or interface for things like this?

      Should OMV 5 have an API for this?
      OMV 4, 5 x ODROID HC2, 2 x 12TB, 2 x 8 TB, 1 x 500 GB SSD, GbE, WiFi mesh
    • Yes, I have seen it. But I feel it is a slight overkill... And I believe it is considered to be difficult to configure. And I want to run other apps on the control server as well.

      Also it is not free and open source, even if you can run it for free on 7 nodes. Hm...
      OMV 4, 5 x ODROID HC2, 2 x 12TB, 2 x 8 TB, 1 x 500 GB SSD, GbE, WiFi mesh
    • Adoby wrote:

      When the HC2s are under heavy utilization, sustained copying of GBs or even TBs, I worry about CPU and HDD temperatures
      My take on this: Test worst case conditions first, then never care again. The most interesting thing with HC2 is IMO whether contact area between the Exynos SoC and the heatsink is sufficient. One of my HC2 showed high temperatures. Shutting down, gently massaging the PCB (and the thermal compound below) and I had lower temps (less than 10°C IIRC)

      Adoby wrote:

      Or is this a case for DIY hacking?

      Since you're only interested in 5 different numbers per host you could have a look at RPi-Monitor too. Draws graphs and keeps data in RRDs. You can easily tweak templates and getting remote data sources into the system is possible via various options.

      But if I would fear temperatures exceeding sane tresholds I would prefer a real monitoring solution that generates alerts in such cases. OpenNMS is most probably overkill too but I love the auto discovery feature (new hosts in your network are detected automagically and added. If you prepare stuff a bit then new devices are added with appropriate 'templates' -- we use it at customers to monitor even client macOS machines -- they get a tweaked /etc/snmpd.conf when enrolled and then simply magically appear in the OpenNMS system with all relevant data so support staff often resolves issues prior to the user taking notice)
    • Adoby wrote:

      Also it is not free and open source, even if you can run it for free on 7 nodes. Hm...
      Uh, nagios core is free and open source - github.com/NagiosEnterprises/nagioscore
      omv 4.1.15 arrakis | 64 bit | 4.15 proxmox kernel | omvextrasorg 4.1.13
      omv-extras.org plugins source code and issue tracker - github

      Please read this before posting a question and this and this for docker questions.
      Please don't PM for support... Too many PMs!
    • @Adoby: Just realized you're an NFS user and that /proc and /sys also can be exported via NFS.

      A pretty lightweight alternative might then also be RPi-Monitor which has a background task that collects data in RRDs and a frontend task displaying the stuff in the browser (moving all the actual rendering into the browser so this solution is really lightweight and does not that much change system's behavior compared to heavy monitoring approaches). All you would need to do is some copy&paste to duplicate template entries for your various hosts.

      To install this on an Armbian based OMV installation all that's needed is armbianmonitor -r but I would suggest updating the armbianmonitor script with latest commit contents before.

      An overview: cnx-software.com/2016/03/17/rp…spberry-pi-and-orange-pi/
    • Yes, RPi-Monitor seems to be a definite possibility. But actually some DIY using Node-Red with SNMP also might be a contender.

      But I'm not working on it very actively atm.

      I took you up on the suggestion to make sure that the worst case was OK. I improved the air-circulation quite a bit by forcing more air to pass near the HC2s, reducing the leaks. Less space around the HC2s. I use a passive "chimney-effect" cooling system with a 100x1800 mm vertical ventilation pipe, in a book case, through the shelves, as a chimney.

      I had peak HDD temps at 47C degrees before that, during sustained ~1000Mb/s file transfers. Now the HDD temps peak at below 40C. Around 30-35C during typical usage at <=20C ambient. I intend to have some monitoring in place before next summer. The chimney effect runs on the temperature difference.

      In the future I might run my OMV nas1-nas4 with their root filesystems on the EVO860 SSD on the Armbian ctrl, using PXE. I'll do some experimenting at least. Then I'll have full access to the innards of all the HC2s locally from ctrl.
      OMV 4, 5 x ODROID HC2, 2 x 12TB, 2 x 8 TB, 1 x 500 GB SSD, GbE, WiFi mesh
    • Adoby wrote:

      In the future I might run my OMV nas1-nas4 with their root filesystems on the EVO860 SSD on the Armbian ctrl, using PXE. I'll do some experimenting at least. Then I'll have full access to the innards of all the HC2s locally from ctrl
      Hmm... /sys and /proc included?

      Anyway: you know what you're doing so please keep us informed in which direction the little project moves :)
    • Users Online 1

      1 Guest