10 Replies Latest reply on Sep 27, 2017 1:29 PM by rschroeder

    Scheduled Backup Job

    mthomas777777

      Hi,

       

      Does anyone know when running a scheduled backup of devices, what account is being used for this job?

       

      MT

        • Re: Scheduled Backup Job
          jeilers

          I believe the setting you are looking for is Connection Profile. You can see this under NCM Settings -> Connection Profiles.(Mine is in the middle right side of the screen below Security above Advanced on version NCM7.5.1) You can also go into each node and see which Connection Profile it is using. Hope this helps.

          • Re: Scheduled Backup Job
            eevans50

            If you look under Jobs List (Dashboards>Configs>Jobs) you should see who the Author for the job is. Their user-level credentials will be used if you have set the devices up to use them.

             

            I created a separate account in SolarWinds(and ACS) to complete the backups and logged in as them to set up the backup jobs. Any time you edit the job the SolarWinds user you are logged in as will take over as the author.

             

            Jobs List

              • Re: Scheduled Backup Job
                mthomas777777

                Thanks for the details. The problem I am facing is that scheduled backup jobs are failing for me. Previously, everyday, Solarwinds would backup all my devices, however, over the last few days this has stopped working. If I go to the job name and manually start the job it works fine and downloads the configs, so there are no issues with the login account to the devices from Solarwinds. Does anyone have ideas why scheduled backup jobs are failing overnight?

                 

                MT

              • Re: Scheduled Backup Job
                mthomas777777

                Thanks for the details. I have tried with different times but same issue.

                • Re: Scheduled Backup Job
                  jkump

                  I have experienced that same issue.  I run Windows Server 2012 R2 and anytime I have pending windows updates, the job scheduler stops working.  My only resolution at this time is to perform a restart and apply those updates.  Then the scheduler will run fine for another month or so.  I use it as reminder to keep it patched   I have not looked in the KB to see if there is a better solution as it is low hanging fruit in the grand scheme.  If it fails the night before, I manually run in the morning at it works.  Since, I know who and what changes are made if none occurred that day then no worries.  If changes are made during the day, the real-time monitor catches those changes as well.  I have also noticed that if the windows updates are pending sometimes the real-time monitor will fail to immediately grab the configs.  This is in 12.1 so I have not had a window to update to 12.2 yet.  Hope this makes sense and helps.  If there is a better resolution, count me in.

                  • Re: Scheduled Backup Job
                    rschroeder

                    I use a service account created just for Solarwinds to do these jobs.  There are some options to consider when troubleshooting failed backups:

                    • Is the account locked out?
                    • Has someone changed the service account's password in AD but not changed it in NCM (or vice-versa)?
                    • Is something else causing the failure?
                      • Has the size of the configuration files on the devices being backed up has significantly increased?  I saw this happen when we enabled much more port-security and additional features on every port in our chassis switches.  They have 384 ports each, and their configuration files grew so large that they caused NCM to time out during backups.  The solution was to either change the job timeout settings or use a different transport method (e.g.: SCP versus CLI "show run").
                      • Has a server maintenance schedule overlapped with your NCM backup schedule?  Once, when my NCM backups began failing, I found that System Admins had set a schedule for backing up our Solarwinds servers that overlapped with NCM's schedule for backing up switches & routers.  This caused NCM to hang and fail on many backups.  I had the SA's change the backup schedule to be earlier, and I moved NCM's switch/router config backup schedule later, and the problem was resolved.
                    • Diagnose the issue by temporarily turning on Session Traces for NCM and then letting the jobs run and fail as usual.  You can also capture traces of jobs that are manually started, and this can be a real-time tool for you to use that will show you exactly what commands are being sent to the devices, and what their responses are.  Session Traces should not be left enabled--this capture generates a lot of unnecessary overhead and file space, but it's extremely useful for getting to the root of failures.  You can enable and disable traces here:

                     

                    Also note that the above screen shot shows where those session trace files will be stored.  This will happen on each poller in the same file location, so if you have five pollers, you'll have five locations to check for the traces.

                     

                    Remember to turn session trace off once you've got the data you need to troubleshoot & correct the problem, and you've proven the issue is resolved.

                     

                    Swift packets!

                     

                    Rick Schroeder

                      • Re: Scheduled Backup Job
                        christopher.t.jones123

                        I had a question about this particular portion of this post

                         

                        • Has the size of the configuration files on the devices being backed up has significantly increased?  I saw this happen when we enabled much more port-security and additional features on every port in our chassis switches.  They have 384 ports each, and their configuration files grew so large that they caused NCM to time out during backups.  The solution was to either change the job timeout settings or use a different transport method (e.g.: SCP versus CLI "show run").

                         

                        how do i accomplish this?

                          • Re: Scheduled Backup Job
                            rschroeder

                            Here's a path to adjust some of the Config Transfer settings.  Once upon a time I determined through trial and error that I could safely set NCM's jobs to do up to six simultaneous downloads.  Now that I've upgraded to NCM 7.7, I see the Simultaneous Downloads has gone to 25 sessions.  I'm not sure if this accounts for the fact that I've got four pollers supporting the job (~6 downloads per poller?) or not.

                             

                            Also note that if you're using SNMP to transfer configs instead of SCP, you can change the timeout here as well.

                             

                             

                            You can also adjust SNMP timeouts & retries here:

                             

                            1 of 1 people found this helpful
                            • Re: Scheduled Backup Job
                              rschroeder

                              You can set the config transfer protocol for individual devices here:

                               

                              If you wish to change that $(GlobalConfigTransferProtocol) setting, here's where you set it:

                               

                               

                              I'd used SSH2 for everything until my chassis switches' configuration grew very large and started timing out and causing the job to report some failures.  Then I moved them to SCP, and along with changing the time of server anti-virus scanning, server backups, NCM job schedules, and NCM protocol timeouts, the jobs all started completing successfully.

                               

                              The trick was discovering all of the factors impacting the jobs.  I'd been unaware the SysAdmins were backing up my NPM and pollers during the time that I'd selected to run NCM backup jobs, and offsetting the two schedules was very helpful to get jobs completed.

                              1 of 1 people found this helpful