3 Replies Latest reply on Dec 4, 2013 9:12 AM by ecklerwr1

    NTA 4.0 RC2 notes


      The internal environment we're running NTA in is fairly small (the result of the SolarWinds NetFlow Averages\PDUs Processed per second was ~7). The environment has NPM 6.0, SAM 6.0.0, VNQM 4.0.1, IPAM 4.0, and NTA 3.11 (now 4.0 RC2).


      The installation went smoothly with no issues. We have a separate SQL server used for several other applications shared with SolarWinds although with the size of the environment we could get away with SQL Server Express on the main poller if desired. Due to this, I opted to have the NTA Flow Storage database run off the main poller (no noticeable performance issues from doing this so far but I don't expect any due to the limited flow data received).


      I haven't noticed any issues with the incoming data or viewing historical data. I'll check again once the system has been running for a few days/week of not summarising data.


      I also took the opportunity to upgrade SAM from 6.0.0 to 6.0.1. The SAM installer couldn't shutdown the new NetFlow service so I had to shut it down manually. The installation and configuration wizard proceeded fine after that.




      Polling Completion99.99
      Elements  240 
      Network Node Elements51
      Volume Elements114
      Interface Elements75
      Polling Rate5% of its maximum rate.» Learn more
        • Re: NTA 4.0 RC2 notes

          Apologies, I didn't see the questionnaire!


          I have also upgraded from RC2 to RC3 this morning. No issues so far.

          NTA Engineering Feedback Questionnaire

          1. Installation/Deployment
            1. How would you describe the installation experience in terms of ease of use?
              1. Easy to use but I think it should be made clearer that if you want the Flow Storage installed elsewhere that it should be done first.
            2. NTA 4.0 RC runs a new flow storage database on the Orion main poller machine.  What is your preference, if any, on whether this DB runs on the main poller or on another machine in your network?
              1. Due to our small internal environment I do not mind running the database off the main poller. For some of our customers running large Netflow environments, I definitely prefer offloading this to another server.
            3. Are you currently using or do you plan to involve any 32bit server in your production environment?
              1. No.
          2. Reporting
            1. Are you able to recreate all of your custom Netflow Traffic Analyzer reports in the new web-based reporting UI?
              1. No custom reports. One of our customers has some custom SQL reports but we do not have access to them.
            2. How do you like the new out of the box NTA reports?
              1. Looking good. No issues.
          3. NTA User Interface Updates
            1. NTA Endpoints and Domains resources have changed a bit.  Hostnames and domains are stored historically in flows, and updates to hostnames are made only into newly coming flows since hostname of the endpoint changes.  What is your opinion on this change?
              1. If I'm reading this correctly, I like this change as you will be able to review traffic from an IP endpoint and note the different hostnames associated to the IP over time instead of the traffic being marked against the current hostname/IP relationship.
            2. Endpoints, transmitters and receivers resources now group by hostname instead of IP address so multiple endpoints having same hostname are now shown as one row in the resource. What is your opinion on this change?
              1. Useful - can track hostname traffic regardless of whether they're using their wired or wireless interfaces.
            3. Updates to the IP Groups and Applications and NetFlow sources (disabling interface) settings do not appear immediately into resources, because stored data has to be updated with the new information. What is your opinion on this change? Do you submit changes to IP groups, Applications or NetFlow sources often?
              1. Unknown. We only have 1 NetFlow source device, do not use the IP groups, and do not modify the Applications.
          4. NTA Reliability
            1. Do you notice any gaps in the flow data collected by NTA?
              1. No. Only the expected gap from when the services were down for the install/upgrade.
            2. Have you experienced any crashes or errors generated by NTA or the NTA flow storage?
              1. No.
          5. NTA Performance
            1. How does the NTA 4.0 RC impact other Orion modules you are currently testing?  Do you see a noticeable degradation in NPM, for example?
              1. As by first post in this thread, we had a small issue upgrading SAM to the latest version but easy to workaround. We haven't noticed any performance degradation.
            2. Describe the performance of your Orion main poller.  How do CPU and memory fare in this deployment?  What have you noticed about disk space consumption?
              1. Similar to previous - CPU usage is high once the services come online as the system runs its baseline. Following that it sits on ~5%. Memory usage is fairly constant ~90% with ~500 MB available memory (4 GB total).
            3. What is your opinion on the performance of the resources in NTA 4.0 RC?  Are they comparable to your current (or previous if you upgraded) production environment?
              1. Performance seems slightly faster in NTA. Haven't noticed any difference to the other modules.
          • Re: NTA 4.0 RC2 notes

            Steven, the NTA team is aware of the issue you have reported above, which is related to the FastBit database not stopping quickly enough. This issue is being tracked internally under FB289353. In the meantime, I recommend stopping all SolarWinds Services manually, and waiting for them to come to a complete stop before upgrading or installing any additional Orion modules.

              • Re: NTA 4.0 RC2 notes

                Thanks aLTeReGo... so it takes a while for the FastBit database to stop and the installer wasn't waiting long enough for it before it tried to move on with install I'm guessing.  Manually stopping all the SW services on the NPM/SAM server before running the SAM 6.0.0 to 6.0.1 upgrade installer sounds like the thing to do then.