Web //Blog //Community //OSMC //OSDC
  
 
Hardware Monitoring with the new IPMI Plugin v2Hardware Monitoring with the new IPMI Plugin v2


 

Hardware Monitoring with the New IPMI Plugin v2 (EN)

For many years now, most servers have been equipped with sensors to monitor the rotational speed of fans, temperatures, voltages, power supply redundancy, and more. The IPMI (Intelligent Platform Management Interface) specification allows these sensors to be queried in a standardized way, enabling easy hardware monitoring for heterogeneous server environments too.

 

IPMI offers other advantages for hardware monitoring: Numerical sensors (so-called threshold sensors) like temperature sensors include upper and lower threshold definitions. For stateful sensors (so-called discrete sensors allow up to 15 yes/no states per sensor) IPMI software like FreeIPMI has a configuration included to distinguish states as ok, warning or critical. For a discrete power supply sensor the state 'power supply presence detected' is ok, while the state 'power supply input lost' should trigger a critical alert.

 

IPMI software tools like FreeIPMI read sensor information via a local system interface on the server itself, or remotely via a LAN channel. In the latter case, user names and passwords ensure that only authorized requests are answered.

 

The IPMI Sensor Monitoring Plugin (check_ipmi_sensor) allows monitoring software like Icinga or Nagios to easily monitor all IPMI-compatible hardware. The plugin itself is a bash script. Version 2 of the plugin uses FreeIPMI instead of ipmitool to query sensors, making it now capable of monitoring discrete sensors, too. As IPMI threshold sensors always know whether their state is OK or not, no additional thresholds have to be defined in Icinga. Configuration of the plugin is easy and straightforward. In addition to the actual status monitoring, the plugin extracts performance data for numerical sensors like temperature, fan or voltage sensors. Third party addons like PNP4Nagios use this performance data to draw diagrams over time. Visualizing trends in this way allow for the identification of rising temperature even before the upper critical temperature threshold is reached.

 

I'm sure most of your servers support IPMI. So come, join this talk and start monitoring your hardware! You will get notified about every critical hardware issue like a single failed fan, allowing you to replace it before a second component fails, which could cause a downtown for the system.

 






Werner Fischer

Werner Fisher has been a technology specialist at Thomas Krenn AG since 2005. Is also the author of the check_ipmi_sensor plugins. At the moment he works intensively with hardware monitoring and I/O performance optimization, and in particular with Flash based storage technologies. Beside these, Werner’s interests center on virtualization and high availability. On these areas of expertise he regularly writes for IT publications and speaks at conferences such as Linux Days, OSDC, Profoss, Security Forum Hagenberg etc. At Thomas Krenn, Werner is the editorial manager of the Thomas Krenn Wiki.


<--- back to overview












  • 29 & 30 November 2011
  • Nuremberg - Germany
  • 2 days all about Nagios
  • meet | discuss | experience: Open Source!
  • over 250 international participants
  • stay at the conference hotel and get the most of the event