8 Replies Latest reply on Feb 7, 2018 1:21 PM by Jose Santiago

    Need help with a unique situation in CMDB

    Jose Santiago

      We have several CIs in the Computer System class. They all have the same information the only difference is that one has marked as deleted = yes and the other has marked as deleted = no. When we run Normalization the records show as Normalized and Approved as well as ready to merge. When we run reconciliation the records that have marked as deleted = no are not merged into the cmdb. we are populating our CMDB with the data we receive from Discovery (BMC.ADDM).

       

      e.g.

       

      name = test.server.name              name = test.server.name

      serial number = 12345                 serial number = 12345

      category = hardware                    category = hardware

      type = Processing Unit                 type = Processing Unit

      item = Server                               item = Server

      model = xyz                                  model = xyz

      marked as deleted = Yes             marked as deleted = No

       

      So to recap all of the CI information is the same except for the Marked As Deleted field.

       

      Hope this makes sense.

       

      José

        • 1. Re: Need help with a unique situation in CMDB
          Stefan Hall

          Hi Jose,

          I'm not sure I understood the target correctly.

          I assume that MAD=Y is in the ADDM dataset. If not, how or who set it in the asset dataset?

          Please check the precedence rule values for both datasets and/or the sandbox, if you have edited directly in the asset.

           

          Normalized CIs marked with "ready to merge" must be identified in the first step. Please check if both CIs have the same recon id.

          1 of 1 people found this helpful
          • 2. Re: Need help with a unique situation in CMDB
            Jose Santiago

            Hi Stefan, thanks for your response. That is the unique part both set of records(Cis) are in the ADDM dataset. One has MAD=Y and another one has MAD=N. I am still trying to figure out how the records with MAD=Y got set that way in the cmdb.

            Thanks,

            José

             

            “Failure is a part of progress,

            not a final outcome.”

             

            Matthew Kelly

            • 3. Re: Need help with a unique situation in CMDB
              Stefan Hall

              This is normal as long as you don't fix your addm dataset. Afterwards, the new (MAD=N) CI is identified and can be merged.

               

              If both CIs in the ADDM dataset have a recon id, your identification rules don't seem to be correct.

              1 of 1 people found this helpful
              • 4. Re: Need help with a unique situation in CMDB
                Jose Santiago

                The CIs with MAD=N have a recon-id of 0.

                 

                Thanks,

                José

                • 5. Re: Need help with a unique situation in CMDB
                  Stefan Hall

                  "fix your addm" in my last comment I meant "purge your addm", always this autocorrection

                   

                  I'll try to describe the process briefly.

                  1. ADDM finds a server A (reconid 0)
                  2. Recon identifies this server A (reconid != 0)
                  3. Recon merges Server A into ASSET

                    Now you have the same server A in both datasets - ADDM and ASSET.

                  4. ADDM recognises Server A as new Anew server (e. g. moved VM)

                    Now you have two servers A and Anew into ADDM

                  5. a ADDM synchronizes Server A with MAD=Y into ASSET
                    b ADDM tries to identify server Anew and looks in ASSET
                       finds the identical server A in ASSET and tries to use its reconid for server Anew in ADDM

                    Which is not possible because reconid is unique for each dataset. You should find a corresponding error message in your recon logs.
                    You always have to purge your source data. "purge" deletes entries with MAD=Y.

                    Afterwards:
                  6. ADDM tries to identify server Anew and looks in the ASSET (with your identify rules)
                    finds the identical server A, takes over its reconid and merges both data sets.

                   

                  Hope it helps you. Good luck!

                  Stefan

                  1 of 1 people found this helpful
                  • 6. Re: Need help with a unique situation in CMDB
                    Carsten Kowalski

                    Hi Jose,

                     

                    Does your reconciliation Job "ADDM to ASSET" has a purge activity?

                    The reconciliation Job identify and merge a CI from one dataset to the "productive" dataset.

                    If a CI not longer exists in ADDM the sync Job "ADDM to CMDB" sets the flag MaD = yes.

                    After that the CI with flag = yes in the ADDM dataset must be removed after the merge activity.

                     

                    Hope this helps.

                     

                    Kind regards,

                    Carsten

                    • 7. Re: Need help with a unique situation in CMDB
                      Jose Santiago

                      Thanks Stefan, i tried this in our test environment and it worked! Thank you for your responses it is greatly appreciated!

                       

                      José

                      • 8. Re: Need help with a unique situation in CMDB
                        Jose Santiago

                        Thanks Carsten, like I replied to Stefan I tried this in our test environment and it worked so thank you so very much for your reply as well. Your help is greatly appreciated!

                         

                        José