5 Replies Latest reply on Oct 7, 2011 8:03 AM by mavturner

    CBQoS Policy Discovery

    pyro13g

      Can something please be done to make NTA properly discover policy changes and then start collecting properly for the new policies?  Once again a failure getting needed data out of NTA because it failed to pick a policy change up.

        • Re: CBQoS Policy Discovery
          mavturner

          pyro13g,

          We are definitely looking at the several issues around CBQoS. I will make sure this specific scenario is covered. Can you provide a little more detail? Are you just changing the name of the policy? What specific changes are you making to the queues and priorities?

          Thanks,

          Mav

            • Re: CBQoS Policy Discovery
              pyro13g

              Add Class to Policy

              Remove Class from Policy

              Change Class (traffic shaping and policing)

              Those were my recent encounters.

              • Re: CBQoS Policy Discovery
                pyro13g

                Any movement or road map for when it will be fixed?   We were bit by the problem again this morning.

                  • Re: CBQoS Policy Discovery
                    mavturner

                    This is not on the plate for the current "What we are working on". Beyond that I can't comment on when this will be fixed. We acknowledge there is a problem there and can easily reproduce it, but I can't commit to specific road map timelines.

                    Mav

                      • Re: CBQoS Policy Discovery
                        mavturner

                        More information in case other people have problems with this.

                         

                        When a CBQoS policy is changed, those updates (name changes or new policies) will be updated on the next poll of the device. The default poll interval for CBQoS data is 5 minutes. However, if you have only recently added CBQoS to a node, it’s possible it will take much longer (a few hours) to detect that the node now has CBQoS data available. If NTA polls a device for CBQoS, but the device does not respond, after 3 consecutive failures, we change the CBQoS poll interval to 4 hours. This is done to reduce the load on NTA, the end device, and decrease unnecessary traffic (since the device doesn't respond, there is no reason to keep querying those OIDs). If you think this is happening, you can force NTA to retry CBQoS by going to Manage Sources, disable CBQoS for the node, hit submit, wait two minutes, then enable it again in the Manage Sources page. This should create the new job to poll the CBQoS data properly.

                        If you or anyone else is still experiencing this issue, please let us know. We have not been able to duplicate it in our lab so don't have any open bugs to resolve around this.