6 Replies Latest reply on May 23, 2008 2:19 AM by Makke

    Houston, we have a problem... we are losing devices!






      We are having unexplained things happening with the devices we have currently listed in All Device.



      We deployed agents to all the devices in our domain through the use of a login script. We monitor the progress of the clients being listed in All Devices. This way, we know how many agents we distributed and how many we still have to go.



      The strange thing is that we "lose" clients listed in All Devices on a regular basis. We can have 800 clients at the end of the workday and have 775 at the start of the next. We have no idea why this is happening. No one other then the LANDesk administrators here has the means to delete the agents from the LDMS, and still we are losing them.



      Has anyone else encountered this problem?






        • 1. Re: Houston, we have a problem... we are losing devices!
          phoffmann SupportEmployee

          Old sock, and the cause is pretty much always one of the two below:


          The question primarily is "do you have your landesk client installed in your image". 99% odds are that the answer is "yes", and as a result you have devices over-writing each other's inventory records (as they all share the same unique identifier - the "UniqueID" which in your case is likely to be "not so unique").




          Two causes are the most common here:


          The 'master-image' contains a LANDesk client where the Device ID's have

          not been deleted, thus several clients share the same ID.



          The customer had "Days to keep inventory scans" set to 90 days, meaning

          that any client which did not send inventory for 90 days would be

          removed by DB Maintenance



          Several things are used to resolve this:



          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



          It is generally recommended to leave "Days to keep Inventory Scans" at

          0 (never delete) - which can be access via the 32-bit console on the

          core only. CONFIGURE -> SERVICES -> in the INVENTORY-tab.


          NOTE - "less" likely to be the cause here, since you lose devices during the course of the day (unless you happen to be running the LANdesk maintenance during daylight hours, as opposed the default 11 pm).


          (3) Duplicate Device removal.


          CONFIGURE -> SERVICES -> in the INVENTORY-tab, click on the

          "DEVICES" and select options as desired. Recommended settings are:


          • BOTH device name and MAC-addresses match

          • Restore old Device IDs.



          In CONFIGURE -> SERVICES -> in the INVENTORY-tab, click on the

          "DEVICE ID's"-button and select options as desired. Recommendations are:


          • Tick "Reject Duplicate Identities"

          • Add "COMPUTER.DEVICE NAME" as attribute

          • Add "COMPUTER.NETWORK.NIC ADDRESS" as attribute

          • Set the value for "Identity Attributes Changed" to "2" .


          Note that (4) is enabled by default in LDMS 8.7. You did not specify what version / service pack you use.


          Paul Hoffmann

          LANDesk EMEA Technical Lead.

          • 2. Re: Houston, we have a problem... we are losing devices!





            First of all thank you for your quick and elaborate reply. Furthermore we use LDMS



            Unfortunately in this case, we fit in the 1% odd that we don't have the agent in our desktop image. We deploy the agent through a login script we serve in Active Directory. Thus our clients do not share the same device ID.



            Furthermore our "Days to keep inventory scans" setting is set to 60 days, however, the deployment of the first agents started only a month ago, so this threshold isn't in effect for another 30 days or so.



            I've checked the Duplicate Device Removal. The settings Device name and MAC-adresses are both enabled however, the Resore old Device ID's isn't. We will enable this setting.



            I've also chacked the Device ID's settings in Configure->Services->Inventory. Only Computer.Device Name and Computer.DeviceID are added. We will add COMPUTER>NETWORK.NIC ADDRESS, as replied. The "Identify attributes Changed"  already has a value of 2.



            As stated above, I will implement your suggestions and will report back shortly.



            If the above should ring other bells with you or other members, please let me know.









            • 3. Re: Houston, we have a problem... we are losing devices!
              phoffmann SupportEmployee

              Hmm - now that IS interesting.


              If you don't have the LANDesk agentry in your image, it shouldn't be a case of stuff over-writing each other.




              I would suggest doing the following (depending on how regularly this happens).


              1 - Stop all Services

              2 - Take a FULL DB Backup of your LANDesk DB.

              3 - On the Core's 32-bit Console go to CONFIGURE -> SERVICES


              4 - In the new Windows go to the Inventory Tab.

              5 - In the Inventory Tab click on the "Advanced Settings" button.

              6 - Search for the value "Store Scans" and enable that (set it to a value of "1").


              7 - This will create a new subdirectory on your Core's
              CoreName\ldmain\ldscan\ directory called "storage". It takes a copy of every scan-file that comes in.

              8 - Let this run for 24 hours ... note down how many devices disappear during this period. (Or "Morning -> Evening" ... a period of time where the problem happens)


              9 - Stop all Services.

              10 - Take again a full DB backup and store the scan-files in a compressed archive.


              11 - Open a support-ticket with LANDesk with the information about the behavior... and be prepared to upload the 2x DB's ("Before" and "After") as well as the scan-files on to our FTP.


              IF the "device-deletions are related in any shape or form to device-ID's we should be able to identify that pretty easily.


              DB Maintenance is the time when the "delete duplicates" occurs, and this doesn't happen during the day ... if it's NOT that, things will get rather interesting...


              Paul Hoffmann

              LANDesk EMEA Technical Lead.

              • 4. Re: Houston, we have a problem... we are losing devices!





                Again many thanks for your quick reply.



                We will examine your suggestion and analyse the impact.



                I'll get back to you about this when we have more to tell.









                • 5. Re: Houston, we have a problem... we are losing devices!
                  phoffmann SupportEmployee

                  No problem - we try to help :).


                  Paul Hoffmann

                  LANDesk EMEA Technical Lead.

                  • 6. Re: Houston, we have a problem... we are losing devices!


                    Checking back in again.



                    After a few weeks of monitoring this issue we can say that it isn't an issue anymore. Although we didn't change anything on our Core Server, we find that no client dissapear from All Devices anymore.



                    I don't think we will ever find what was causing this issue, but we are glad it's gone now.



                    Thanks to everyone who helped us with this issue.