10 Replies Latest reply on Jul 31, 2015 10:54 AM by Santosh Kunnothmannur

    BMA Snapshot Comparision based on XML Elements

      Share This:

      Hi Guys,

       

      I have a doubt in BMA snapshots comparision:

       

      When I compare two snapshots, do I have provision in BMA to provide a report of specific elements in both the XML files?

       

      For ex: Suppose I only need to get the comparision report on JVM heapsizes, JVM Arguments, Classpaths etc, can I get this kind of report intead of whole report?

       

      Please give your suggestions.

       

      Thanks and Regards,

       

      Ashwin

        • 1. Re: BMA Snapshot Comparision based on XML Elements

          3 options:

           

          1. Use the snapshot filters to reduce the size of the snapshots you are comparing.  You cannot filter down to specific attributes or elements at this time.  See the user guide for more info about filtering.

           

          2. Manually edit the xml files to remove unwanted comparisons.  If you want a very specific report generated this is the quickest option

           

          3. Play to BMA's strengths.  Create application specifc xml packages and tokensets.  Import those tokensets into your server profiles and fill out the correct environment specific values.  Now when you want to compare two different environments use the "Push Tokens to Configuration" option to replace token values with tokens in the snapshots then compare them.  This will remove known differences and reduce the amount of false positives that appear in the report.

           

          Using #1 and #3 together allows you to quickly find differences affecting your WebSphere environments.

          • 2. Re: BMA Snapshot Comparision based on XML Elements

            Has the BMA Comparison improved between 2012 and now?

            • 3. Re: BMA Snapshot Comparision based on XML Elements

              Yes, the Eclipse Compare Engine has been upgraded in the 8.7 release.

              • 4. Re: BMA Snapshot Comparision based on XML Elements

                Naresh, what types of improvements are you looking to see? 

                Also what version of BMA are you on?

                 

                There have been changes/modifications in compare since 2012, but if you're looking for something specific please elaborate and I can better answer or see to it that we get your request on our roadmap.

                • 5. Re: BMA Snapshot Comparision based on XML Elements

                  We are looking to compare two different environment/cell /cluster snapshots. When one environment is reported as working and the other doesn't work, we need to be able to compare, push tokens to both and find out any differences to be able to troubleshoot and resolve issues. We need to write our tokensets to eliminate cell name, cluster name etc with context.

                   

                  We looked at the migration wizard as well and what we find is that the migration wizard adds two unique names like ph_cell_name_UAT1cell and ph_cell_name_UAT2cell. Can this aspect be generalized?

                  • 6. Re: BMA Snapshot Comparision based on XML Elements

                    This seems like functionality that is already available within our current release of BMA.  There is the ability to use Smart tokenization which is a combination of using XPATH and RegEx to define a token.  Can you please let me know what version of BMA you are on now?

                    • 7. Re: BMA Snapshot Comparision based on XML Elements

                      BMA 8.7 p3. We already use smart tokenization and define tokens with context using xpath.

                      • 8. Re: BMA Snapshot Comparision based on XML Elements

                        Ok Thanks Naresh,  I'll circle back with the team to make sure you get an appropriate answer.

                        • 9. Re: BMA Snapshot Comparision based on XML Elements

                          We looked at the migration wizard as well and what we find is that the migration wizard adds two unique names like ph_cell_name_UAT1cell and ph_cell_name_UAT2cell. Can this aspect be generalized?

                          Basically, we can generalize token name creation for the cell element and have something like ${ph:cell_name}. But for clusters and other resources (e.g. servers) this is hardly possible to implement, because they need to have different names in order to map to different values.

                           

                          What we can recommend to do: for both environments, you can replace tokens with some generic values, which will have to be common in both environments. For example: tokens "ph_cell_name_UAT1cell" and "ph_cell_name_UAT2cell" can be replaced with some generic value (or token) like "UATcell" or "${ph:cell_name_UAT_cell}". In this case, cell specific elements won't be shown in comparison report.

                          • 10. Re: BMA Snapshot Comparision based on XML Elements
                            Santosh Kunnothmannur

                            The goal is to have the ability to compare 2 environments. The migration wizard is a great tool but fulfills only the purpose of migration. What Naresh is asking is since the Migration wizard tokenizes the middleware specific values ( cell , node cluster , absolute paths etc.) -  Can we extend this to compare environments? Here is a simple use case

                             

                            1. Run full snapshot on Env 1 and Env 2.
                            2. The starting point would be to eliminate middleware specific uniqueness. Run the Migration/Comparison Wizard on these snapshots and eliminate the uniqueness. Example :- all Cell Names become $(ph:cell_name}, all node name become $(ph:node_name}
                            3. Compare the Env1 and Env2 files created by the wizard - This will have all values including the application specific token values that the customer wants.
                            4. At this point, depending on the usecase, the customer can push application tokens to the configuration to reduce the differences to a more readable number.