Uploaded image for project: 'Observium'
  1. Observium
  2. OBS-2763

Extreme XOS dBm convert issue

Details

    Description

      Since XOS 22.3.1.4 or some versions below Extreme added the feature to pull dBm values from the ports. Works "ok" so far... only interger values are support crap.

      Observium is supporting this feature but during the convertion of this value it fails.

      Converted DBM value: -7.0E-6 W -> FALSE

      And sensor_unit is wrong. Output is dBm. Attached debug log and a CLI output from the transceiver.

      Attachments

        1. extreme-dbm-convert-issue.txt
          39 kB
        2. extreme-dbm-cli.txt
          1 kB
        3. OBS-2763-DISCOVERY.txt
          66 kB
        4. OBS-2763-POLLER.txt
          227 kB
        5. OBS-2763-DISCOVERY-NEW.txt
          254 kB
        6. Screenshot_2018-09-06 Watts to dBm conversion calculator.png
          Screenshot_2018-09-06 Watts to dBm conversion calculator.png
          11 kB

        Activity

          [OBS-2763] Extreme XOS dBm convert issue
          dklimek Denis Klimek added a comment -

          This sounds more like an SFP+ module bug instead of an issue in the whole firmware of EXOS. Did you try other vendor SFP+ modules?

          dklimek Denis Klimek added a comment - This sounds more like an SFP+ module bug instead of an issue in the whole firmware of EXOS. Did you try other vendor SFP+ modules?
          athompson@merlin.mb.ca Adam Thompson added a comment -

          FYI... in XOS 31.6.1.3 this is still not fixed for all SFP+ modules.  A FiberStore version of an Extreme CW57-10G-80-1570 continues to quantize reported Rx power levels to integer microwatts, thus losing a LOT of precision in the Observium graphs.

          athompson@merlin.mb.ca Adam Thompson added a comment - FYI... in XOS 31.6.1.3 this is still not fixed for all SFP+ modules.  A FiberStore version of an Extreme CW57-10G-80-1570 continues to quantize reported Rx power levels to integer microwatts, thus losing a LOT of precision in the Observium graphs.
          landy Mike Stupalov added a comment -

          great!

          landy Mike Stupalov added a comment - great!
          dklimek Denis Klimek added a comment -

          Extreme released new firmware which contains a bug fix.

          ExtremeXOS 22.7.1.2

          dklimek Denis Klimek added a comment - Extreme released new firmware which contains a bug fix. ExtremeXOS 22.7.1.2
          dklimek Denis Klimek added a comment -

          Update: Extreme confirmed this bug opened a official bug ticket for this. Fetching values via SNMP for "entPhySensorEntry" gives incorrect values. CR#xos0073216

          dklimek Denis Klimek added a comment - Update: Extreme confirmed this bug opened a official bug ticket for this. Fetching values via SNMP for "entPhySensorEntry" gives incorrect values. CR#xos0073216
          dklimek Denis Klimek added a comment -

          We will open a ticket for this issue at Vendor support :/

          dklimek Denis Klimek added a comment - We will open a ticket for this issue at Vendor support :/
          landy Mike Stupalov added a comment -

          ahh, you mean used incorrect scale?

          we exactly use scale value from entPhySensorScale, if here incorrect value.. again write to vendor
          I not see ways how to fix this case inside observium.

          landy Mike Stupalov added a comment - ahh, you mean used incorrect scale? we exactly use scale value from entPhySensorScale , if here incorrect value.. again write to vendor I not see ways how to fix this case inside observium.
          dklimek Denis Klimek added a comment -

          Mhhh... I converted the CLI value 3.39dBm to watt and get 0.0021827299118 Watt. But you're right... you're doing everything right.

          Extreme should return:

          entPhySensorValue.135 = 2184

          or change scale to

          entPhySensorScale.135 = milli

          Do you agree with me?

          dklimek Denis Klimek added a comment - Mhhh... I converted the CLI value 3.39dBm to watt and get 0.0021827299118 Watt. But you're right... you're doing everything right. Extreme should return: entPhySensorValue.135 = 2184 or change scale to entPhySensorScale.135 = milli Do you agree with me?
          landy Mike Stupalov added a comment - 

          entPhysicalDescr.135 = SFP TX Power Sensor
          		 
          entPhysicalDescr.136 = SFP RX Power Sensor
          		 
          entPhySensorValue.135 = 21849
          		 
          entPhySensorValue.136 = -9999
          		 
          entPhySensorType.135 = watts
          		 
          entPhySensorType.136 = watts
          		 
          entPhySensorScale.135 = micro
          		 
          entPhySensorScale.136 = micro


          Converted DBM value: 0.021849 W -> 13.394315646739 dBm
          		 
          Converted DBM value: -0.009999 W -> FALSE

          What incorrect?

          landy Mike Stupalov added a comment - entPhysicalDescr.135 = SFP TX Power Sensor entPhysicalDescr.136 = SFP RX Power Sensor entPhySensorValue.135 = 21849 entPhySensorValue.136 = -9999 entPhySensorType.135 = watts entPhySensorType.136 = watts entPhySensorScale.135 = micro entPhySensorScale.136 = micro Converted DBM value: 0.021849 W -> 13.394315646739 dBm Converted DBM value: -0.009999 W -> FALSE What incorrect?
          dklimek Denis Klimek added a comment -

          Also added a discovery debug.

          dklimek Denis Klimek added a comment - Also added a discovery debug.

          People

            landy Mike Stupalov
            dklimek Denis Klimek
            Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: