Best practices for specific implementation scenarios

For specific scenarios, BMC recommends particular settings for Atrium Integrator, Product Catalog, Normalization Engine, and Reconciliation Engine.

Recommendation for scheduling

Importing your data into BMC CMDB, normalizing it, and reconciling it must happen separately, so do not schedule any overlap in these processes. For example, reconciliation will result in errors for a group of related configuration items (CIs) such as a computer system and its components if not all of the CIs and relationships in the group have yet been imported.

Best practices for bulk loads

When you are loading a large amount of data, such as an initial load, into BMC CMDB, use the following guidelines. The objective is to provide a starting point with normalized and reconciled data, using batch jobs. After a bulk load, you can decide how to normalize changes to the data (batch, inline, or continuous normalization).

Best practices for bulk load

Component	Tasks
Atrium Integrator	When transferring a large amount of data, use parallel processing with a different BMC CMDB instance per exchange. For an attribute that is used as a primary key or as a relationship key, create a separate index in `BMC_BaseElement`. Database/External Data Store can have an index on the column that is the Primary Key. Check the log files and confirm that there are no errors, especially data errors.
Product Catalog	Verify that Product Catalog data is current and accurate.
Normalization Engine	Turn on normalization across all datasets into which you are loading bulk data. Create a batch normalization job. Use the batch job after the data is loaded into the import dataset. Then, use the reconciliation batch job to merge the normalized data to a different dataset. Do not modify the default dataset configurations. Note: Do not enable the `Allow new Product Catalog entry` parameter because it could result in duplicate entries for the same product. Do not allow an installer to start normalization jobs because the standard reconciliation Identify and Merge rules are not sufficient to reconcile only normalized configuration items (CIs).
Reconciliation Engine	Create a batch reconciliation job. After the import dataset has been normalized, use the batch reconciliation job to merge the normalized data to a different dataset.

Best practices for incremental loads

After you have loaded initial data, use the following guidelines to normalize and reconcile changes to the data.

Best practices for incremental loads

Component	Tasks
Atrium Integrator	Check the log files and confirm that there are no errors, especially data errors.
Product Catalog	Verify that Product Catalog data is current and accurate.
Normalization Engine	Turn on normalization across all datasets into which you are loading bulk data. Create a continuous normalization job to normalize changes and additions to the import dataset. Do not modify the default dataset configurations. Note: Do not enable the `Allow new Product Catalog entry` parameter because it could result in duplicate entries for the same product.
Reconciliation Engine	Create a continuous reconciliation job. After the initial reconciliation, the continuous job identifies and merges new and changed configuration items (CIs) in the import dataset.

Best practices for CMDB Explorer edits

When modifying configuration items (CIs) in the CMDB Explorer, use the following guidelines. When you promote a CI, it is reconciled using a standard Identify and Merge job.

Best practices for CMDB Explorer edits

Component	Tasks
Atrium Integrator	Not applicable
Product Catalog	Verify that Product Catalog data is current and accurate.
Normalization Engine	Verify that Catalog Mapping aliases are current and accurate. If normalization is turned off across all datasets, do not turn on normalization. Configure the normalization settings for the dataset where the manual edits are made: Normalization Type: Name & CTI lookup Inline Normalization: Enabled Allow Unapproved CI: Disabled Allow new Product Catalog entry: Disabled In the edit form that appears in a failed normalization, you must know which attributes are required and what values are valid. If the Product Catalog has no entries for hardware or other classes, disable normalization for those classes by deleting them from the Class Configuration tab in the Configuration Editor. You can add the class later, if needed.
Reconciliation Engine	Create a reconciliation job for the specific user sandbox to start manually; do not create a schedule or select Continuous for the job.

Best practices for Remedy ITSM suite manual edits

When modifying configuration items (CIs) from the Remedy IT Service Management suite, use the following guidelines.

Best practices for Remedy ITSM suite manual edits

Component	Tasks
Atrium Integrator	Not applicable
Product Catalog	Verify that Product Catalog data is current and accurate.
Normalization Engine	Configure the normalization settings for the specific user sandbox where the manual edits are made: Normalization Type: Name & CTI lookup Inline Normalization: Enabled Allow Unapproved CI: Disabled Allow new Product Catalog entry: Disabled
Reconciliation Engine	You can use the existing reconciliation job for the BMC Remedy ITSM sandbox.