Uploaded image for project: 'OpenNMS'
  1. OpenNMS
  2. NMS-7106

If a target node is rebooted the RRD/JRB files contains spikes because the sysUpTime check is not working

    XMLWordPrintable

Details

    Description

      Several years ago, the SNMP Collector was refactored to track the sysUpTime change on each collection attempt. That functionality was integrated with another feature which is disabled by default, so the sysUpTime check is not being performed.

      This feature is send a forceRescan if a change is detected on the ifNumber or the sysUpTime, that for some reasons has been disabled.

      The validation of the sysUpTime must be executed no matter if the forceRescan feature is enabled or not for the SNMP Collector, in order to skip the persistence and the threshold processing for counter variables.

      Because this is not happening right now, the counter variables may have big spikes due to counter resets, and the threshold processor may generate invalid alarms.

      Attachments

        1. collectd-configuration.xml
          12 kB
          Aaron Scamehorn
        2. event.png
          55 kB
          Aaron Scamehorn
        3. ifInDiscards.jrb
          37 kB
          Aaron Scamehorn
        4. ifOutDiscards.jrb
          37 kB
          Aaron Scamehorn
        5. jrb.png
          60 kB
          Aaron Scamehorn
        6. lr3prKbz.png
          16 kB
          Aaron Scamehorn
        7. sysuptime.png
          17 kB
          Aaron Scamehorn

        Activity

          People

            agalue Alejandro Galue
            agalue Alejandro Galue
            Votes:
            3 Vote for this issue
            Watchers:
            4 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: