Details

    • Sub-task
    • Resolution: Fixed
    • Major
    • None
    • Professional Edition
    • Poller

    Description

      Hello,

      On Wednesday (29.11) we upgraded our Observium instance from 17.4.8487 to 17.11.8930. After that upgrade we experienced great increase of the poller durration times to the devices with Huawei VRP 5.170 OS. Today we have upgraded the Observium to 17.11.8960 but the issue was not resolved. Please view the attached image for reference.

       

      Also we have reverted the Observium to 17.4.8487 version but the the poller duration times was still very higher than before the first upgrade.

       

      Please note that this issue is only with devices with Huawei VRP 5.170 OS. On the other switches with Huawei VRP 5.160 and devices with Mikrotik RouterOS 6.40, Linux, and Windows, the poller duration times are fine.

       

      Could you please advice us how to resolve this issue.

       

      Best regards!

       

      Attachments

        1. 1.3.6.1.4.1.2011.5.25.42.3.1.1.1.log
          438 kB
          Plamen Petkov
        2. br1-tdc.r1-hua-6720.log
          9.86 MB
          Plamen Petkov
        3. poller-duration-chart.png
          46 kB
          Plamen Petkov
        4. poller-duration-time.png
          60 kB
          Plamen Petkov
        5. poller-performance.png
          121 kB
          Plamen Petkov
        6. Q-BRIDGE-MIB.log
          259 kB
          Plamen Petkov

        Activity

          [OBS-2532] Increased pooler duration times to Huawei VRP 5.170 after update

          Issue was fixed in main task (OBS-2620).

          landy Mike Stupalov added a comment - Issue was fixed in main task ( OBS-2620 ).

          Hi Mike,

          Is there any news?

          Best regards

          plamenp Plamen Petkov added a comment - Hi Mike, Is there any news? Best regards

          Hello Mike,

          Is any, hopefully good, news here?

          Best regards!

          plamenp Plamen Petkov added a comment - Hello Mike, Is any, hopefully good, news here? Best regards!

          yah, as I thought, your huawei device report every vlan 1-4094 by Q-BRIDGE-MIB, also if it not exist..

          Ok, I will try to fix by vendor specific MIB.

          landy Mike Stupalov added a comment - yah, as I thought, your huawei device report every vlan 1-4094 by Q-BRIDGE-MIB, also if it not exist.. Ok, I will try to fix by vendor specific MIB.

          Hello Mike,

           

          Please find attached the requested logs. I'm agree with you that the device reports all possible VLANs (also unused and not enabled) and this increase the snmpwalk time. But we didn't have such issue before the upgrade. Is there a option Observium to not trying to gather data from inactive VLANs?

           

          Best regards,

           

          Q-BRIDGE-MIB.log

          1.3.6.1.4.1.2011.5.25.42.3.1.1.1.log

          plamenp Plamen Petkov added a comment - Hello Mike,   Please find attached the requested logs. I'm agree with you that the device reports all possible VLANs (also unused and not enabled) and this increase the snmpwalk time. But we didn't have such issue before the upgrade. Is there a option Observium to not trying to gather data from inactive VLANs?   Best regards,   Q-BRIDGE-MIB.log 1.3.6.1.4.1.2011.5.25.42.3.1.1.1.log

          + one more walk:

          snmpbulkwalk -v2c -c <community> -m HUAWEI-L2VLAN-MIB -M /opt/observium/mibs/rfc:/opt/observium/mibs/net-snmp:/opt/observium/mibs/huawei <device> .1.3.6.1.4.1.2011.5.25.42.3.1.1.1
          

          landy Mike Stupalov added a comment - + one more walk: snmpbulkwalk -v2c -c <community> -m HUAWEI-L2VLAN-MIB -M /opt/observium/mibs/rfc:/opt/observium/mibs/net-snmp:/opt/observium/mibs/huawei <device> .1.3.6.1.4.1.2011.5.25.42.3.1.1.1
          landy Mike Stupalov added a comment - - edited

          plamenp yep, I'm here

          I see your trouble, but this is definitely not related to observium update..
          I not sure what additionally changed in your system and/or on device,
          but trouble is that your device report all possible vlans (also unused and not enabled).

          Please make additional debugs from this device:

          snmpbulkwalk -v2c -c <community> -m Q-BRIDGE-MIB -M /opt/observium/mibs/rfc:/opt/observium/mibs/net-snmp <device> dot1qVlanStaticRowStatus
          

          (replace <community> and <device> with your data)

          I will try make some speedup by this data (if possible)..

          landy Mike Stupalov added a comment - - edited plamenp yep, I'm here I see your trouble, but this is definitely not related to observium update.. I not sure what additionally changed in your system and/or on device, but trouble is that your device report all possible vlans (also unused and not enabled). Please make additional debugs from this device: snmpbulkwalk -v2c -c <community> -m Q-BRIDGE-MIB -M /opt/observium/mibs/rfc:/opt/observium/mibs/net-snmp <device> dot1qVlanStaticRowStatus (replace <community> and <device> with your data) I will try make some speedup by this data (if possible)..
          plamenp Plamen Petkov added a comment - - edited

          Hello Mike,

           

          Any news here?

           

          Best regards!

           

           

          plamenp Plamen Petkov added a comment - - edited Hello Mike,   Any news here?   Best regards!    

          Hello Mike,

          In fact on 01.12 the graphs again become normal and I'm not sure what is the reason. We have not made any change neither on the Observium server nor on the Huawei device. But we saw that the poller duration time is increased for many of our Huawei devices. As reference, please find the attached poller-duration-time.png for the mentioned device. Before the update poller times are near 40 ms while after update such times increased to 200 ms, as the most time is taken by the "ports" module.

           

          Please find the requested files attached.

          br1-tdc.r1-hua-6720.log

          We are looking forward to your feedback.

           

          Best regards!

           

          plamenp Plamen Petkov added a comment - Hello Mike, In fact on 01.12 the graphs again become normal and I'm not sure what is the reason. We have not made any change neither on the Observium server nor on the Huawei device. But we saw that the poller duration time is increased for many of our Huawei devices. As reference, please find the attached poller-duration-time.png for the mentioned device. Before the update poller times are near 40 ms while after update such times increased to 200 ms, as the most time is taken by the "ports" module.   Please find the requested files attached. br1-tdc.r1-hua-6720.log We are looking forward to your feedback.   Best regards!  

          2. additionally pls attach screenshot from Device page -> Performance Data (right menu)

          landy Mike Stupalov added a comment - 2. additionally pls attach screenshot from Device page -> Performance Data (right menu)

          People

            landy Mike Stupalov
            plamenp Plamen Petkov
            Votes:
            1 Vote for this issue
            Watchers:
            3 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: