5 Replies Latest reply on Jul 16, 2009 8:53 AM by phoffmann

    New additions are booting existing records

    Rookie

      Hello all,

       

      I have a unique problem.  I am adding a rack full of notebook computers to Landesk.  Each computer has a unique computername, IP, etc..(of course).  The problem is I install the landesk client on a computer, a record is created....then when I go to install the client on computer2, it takes the place of computer1, instead of creating a new record of its own.

       

      These are imaged machines, but they have unique SIDs, computer names, MAC's, etc... Almost all of the computers on our 1000+ computer network are imaged, so that has never been a problem before.  I cant find any information related to his here, or with a google search.

       

      Anyone have any ideas, or have seen this before?

       

      Thanks,

      Scott

        • 1. Re: New additions are booting existing records
          Specialist

          Maybe you also imaged the Agent with his unique ID?

          For the master-Image it is recommended that if already the LANDesk agent is
          present, that the DeviceID fields be deleted.

           

          These can be found in the registry in two locations:

          • HKLM\SOFTWARE\INTEL\LANDesk\Common Api\UniqueID
          • HKLM\SOFTWARE\LANDesk\Common Api\UniqueID

           

          Best Regards

          Troga

          • 2. Re: New additions are booting existing records
            Rookie

            Well, that would be a good idea... but we image all of the machines first.  Then install the landesk client and join them after the fact.   So Landesk isnt part of the image.

             

            -scott

            • 3. Re: New additions are booting existing records
              phoffmann SupportEmployee

              TrogaTrading wrote:

               

              Maybe you also imaged the Agent with his unique ID?

              For the master-Image it is recommended that if already the LANDesk agent is
              present, that the DeviceID fields be deleted.

               

              These can be found in the registry in two locations:

              • HKLM\SOFTWARE\INTEL\LANDesk\Common Api\UniqueID
              • HKLM\SOFTWARE\LANDesk\Common Api\UniqueID

               

              Best Regards

              Troga

               

              I would bet good money on this being EXACTLY the problem . Spot on, Troga

               

              To explain - LANDesk uses its own Unique ID's (similar to the Windows SID's) to uniquely identify devices.

               

              So - say you have a device with ID 1234-1234-1234-1234 ...

               

              ... you then image this device (keeping this information in the registry) ... and distribute it.

               

              You now have all sorts of devices that overwrite each other (all claiming to be the same device) -- though this can be countered with Duplicate Device ID Detection. But yes, you should clean up your image ASAP and - ideally - not include the LDMS agentry on the image in the first place (far better instead to include a command-line to install the agent in the image).

               

              That way, you will always have the most up-to-date agent files (as per your Core), and change management in regards to the LDMS agentry becomes a LOT simpler .

               

              - Paul Hoffmann

              LANDesk EMEA Technical Lead.

              • 4. Re: New additions are booting existing records
                Rookie

                Landesk isnt part of the image.  The client is installed manually on each machine after its imaged.

                 

                 

                • 5. Re: New additions are booting existing records
                  phoffmann SupportEmployee

                  Then there is no way in which clients should be over-writing each other.

                   

                  Another thing that you could be using that makes it look like this is going on is "days to keep inventory scans" being set to a non-0 value. This is a setting for the Inventory Service (so you configure it on the Core's 32-bit Console).

                   

                  If this value is set to "0" - we never delete inventory data automatically.

                   

                  If you set it to any value other than "0" - for instance "10" then after this many days without inventory, we delete devices. We normally recommend to keep this value high (90 or so), to ensure that devices that are "on holiday / in repair" or whatnot, don't get deleted, as this is mainly for devices that are no longer in use and have been scrapped.

                   

                  That said, I've seen plenty of situations where this was set to "2" (or even "1"), meaning that we'd delete shedloads of devices from inventory on Sunday Night's maintenance run.

                   

                  This could give the same impression, potentially.

                   

                  You may want to look into that one .

                   

                  Also, you may want to check out your NT Application Event Log - that may hold some interesting data from us.

                   

                  - Paul Hoffmann

                  LANDesk EMEA Technical Lead