This discussion has been locked. The information referenced herein may be inaccurate due to age, software updates, or external references.
You can no longer post new replies to this discussion. If you have a similar question you can start a new discussion in this forum.

Unable to poll CPU utilization after upgrading to NPM 9.1

We have recently upgraded to NPM 9.1 from NPM 9.0. Its nice to see various new improvements in the new version.


But today after 48 hours of upgradation we found that the pollers are no longer able to capture the CPU utilization of the devices configured. I have even tried downloading the latest copy of the MIBs from the solarwinds website but the situation is same.


We rebooted all the servers, restarted all the services but the poller is polling everything except the CPU and memory.


Has anybody checked this ???


  • FormerMember
    0 FormerMember

     We made some changes to CPU and memory polling for certain classes of Cisco devices.  Specifically, what kinds of devices are you seeing the problems on?

  • unable to see on cisco 3750 , 7206, 6513 , 3745 , 7604


     


    its on all the devices we have in our environment.


    can you also suggest what i have to do it fix this ??? we cannot keep on sending blank reports to our management.

  • FormerMember
    0 FormerMember in reply to rajivbahl

     Do you have a Support case number?  If not, can you open one today?

  • another issue started with NPM 9.1


    now my poller goes down every night and they do not come up even after restarting the NPM service .


     I have to get the server rebooted for this.


    I think i have to roll back to 9.0 as i can't continue like this .This is impacting my SLA reports and my SLA team and customer is behind me .

  • I Have even tried posting ticket on the Solarwinds support


    It seems link the submit button is not working


  • Issue above works fine with SNMPv2, just not with SNMPv3 
    We use HOST-RESOURCES-MIB and CISCO-PROCESS-MIB for New cisco devices
     
    Adding HOST-RESOURCES-MIB and CISCO-Process-MIB to SNMPv3 VIEWS should resolve the issue of not getting CPU Utilization
      
    snmp-server group GROUP v3 auth read ORIONView
    snmp-server view ViewName View internet included
    snmp-server view ViewName system included
    snmp-server view ViewName interfaces included
    snmp-server view ViewName chassis included
     
    snmp-server view ViewName cisco included
    snmp-server view ViewName ciscoProcessMIB included
     
    Also albe to add in Views by using OIDs 

    snmp-server view ORIONView 1.3.6.1.4.1.9.9.109.1.1.1.1.7 included

    snmp-server view ORIONView 1.3.6.1.4.1.9.9.109.1.1.1.1.8 included

     

    *******************************************   

     

    Pre Orion 9, we used...OID name: cpmCPUTOTAL5min/cpmCPUTotal1min
    OID number: 1.3.6.1.4.1.9.9.109.1.1.1.1.5/4 
      

     

    With 9.1 they started using the following OIDs
    1.3.6.1.4.1.9.9.109.1.1.1.1.7
    1.3.6.1.4.1.9.9.109.1.1.1.1.8

    And the reason is because of this from Cisco:
    https://www.cisco.com/en/US/tech/tk648/tk362/technologies_tech_note09186a0080094a94.shtml 
     
    snmpwalk -Os -c <community string> -v 2c 206.130.126.3
    1.3.6.1.4.1.9.9.109.1.1.1.1.7
    enterprises.9.9.109.1.1.1.1.7.1 = Gauge32: 7
    enterprises.9.9.109.1.1.1.1.7.2 = Gauge32: 25
    enterprises.9.9.109.1.1.1.1.7.3 = Gauge32: 0
     
    So here is the explanation to how we calculate this:

    The device has 3 CPUs. 
    Those three numbers that are returning are individual percentages for each CPU. 
    Prior to 9.1 we were only reporting the CPU utilization for one CPU (7% as an example). 
     
    Now we are fully utilizing the MIB and averaging the utilization of all the reported CPUs and showing that number: 7 + 25 + 0 = 32
     
    Take that a divide by the number of CPUs: 32 / 3 = 10.67 ~ 11% average CPU utilization.
    *******************************************
  • I have this issue too and I guess I am not happy with the result or the explanation.  I did open the case on a 6509 that now in SolarWinds reports a 20% CPU utilization when it is actually (by doing sh proc cpu) running at 65 or 70%.  The average does not give me an acurate picture of how hard my device is working.  I'm in the process of adding additional QoS and ACLs and I would like to know how they affect my CPU utilization.  Without adding custom pollers to pull the information I need, I'm missing a VERY important piece to SolarWinds.  I'm disappointed and personally I would like to see the way CPU is calculated changed back to the way it was in previous versions.