How Merge Replication Tracks and Enumerates Changes

The system tables used for unfiltered and filtered tables contain the following metadata:

• MSmerge_contents contains one row for each row inserted or updated in a published table in the database.

• MSmerge_tombstone contains one row for each row deleted from a published table in the database.

• MSmerge_genhistory contains one row for each generation. A generation is a collection of changes that is delivered to a Publisher or Subscriber. Generations are closed each time the Merge Agent runs; subsequent changes in a database are added to one or more open generations.

The following change tracking process is used for all unfiltered tables:

• When an insert or update occurs on a published table, the MSmerge_ins_<GUID> or MSmerge_upd_<GUID> trigger fires, and a row is inserted into the MSmerge_contents system table. The rowguid column of MSmerge_contents contains the GUID for the inserted or updated row, indicating that the next time synchronization occurs, the corresponding inserted or updated row in the user table should be sent to the Publisher or Subscribers. If subsequent updates occur on a row in a user table, the row in MSmerge_contents is updated to reflect this.

• When a delete occurs on a published table, the MSmerge_del_<GUID> trigger fires, and a row is inserted into the MSmerge_tombstone system table. The rowguid column of MSmerge_tombstone contains the GUID for the deleted row, indicating that the next time synchronization occurs, a delete should be sent to the Publisher or Subscribers for the corresponding deleted row in the user table. If the deleted row is referenced in MSmerge_contents (because it was inserted or updated since the last synchronization), the row is deleted from MSmerge_contents.

In addition to the system tables described in the previous section, three tables in the publication database contain metadata for tracking changes to filtered tables:

• MSmerge_partition_groups contains one row for each partition that is defined in a publication. Partitions can be:

  • Defined explicitly using sp_addmergepartition or the Data Partitions page of the Publication Properties dialog box.

  • Created automatically when a Subscriber synchronizes if the Subscriber requires a partition that does not yet have an entry in MSmerge_partition_groups.

• MSmerge_current_partition_mappings contains one row for each unique combination of rows in MSmerge_contents and MSmerge_partition_groups. For example, if a row in a user table belongs to two partitions, and the row is updated, one row is inserted into MSmerge_contents to reflect the update, and two rows are inserted into MSmerge_current_partition_mappings, to indicate that the updated row belongs to the two partitions.

• MSmerge_past_partition_mappings contains one row for each row that no longer belongs in a given partition. A row moves out of a partition if:

  • The row is deleted. If a row is deleted from a user table, a row is inserted into MSmerge_tombstone and one or more rows are inserted into MSmerge_past_partition_mappings.

  • The value in a column used for filtering has changed. For example, if a parameterized filter is based on the state in which a company is headquartered and the company moves, the row for the company (and related rows in other tables) might move out of one sales person's partition of data into the partition for another sales person. If a row is updated such that it no longer belongs in a partition, a row is inserted or updated in MSmerge_contents and one or more rows are inserted into MSmerge_past_partition_mappings.

Note
If nonoverlapping partitions with one subscription per partition (a value of 3 for the @partition_options parameter of sp_addmergearticle) are used, the system tables MSmerge_current_partition_mappings and MSmerge_past_partition_mappings are not used to track the rows' partition mappings, because each row belongs to only one partition and can be changed at only one Subscriber.

The process described above (in the section "Change Tracking for Unfiltered Tables") for unfiltered tables is also used for filtered tables, with the following additions:

• When an insert occurs on a published table, in addition to data being updated or inserted into MSmerge_contents, a partition mapping is added to MSmerge_current_partition_mappings for each partition that the row belongs to.

• When an update occurs on a published table, in addition to data being updated or inserted into MSmerge_contents, if a partition mapping does not exist in MSmerge_current_partition_mappings for each partition that the row belongs to, one is added. If the update resulted in a row being moved from one partition to another, a row is updated in MSmerge_current_partition_mappings and one is added to MSmerge_past_partition_mappings.

• When a delete occurs on a published table, in addition to a row being inserted into MSmerge_tombstone, a row is deleted from MSmerge_current_partition_mappings and one is added to MSmerge_past_partition_mappings.

When the Merge Agent runs, changes are enumerated using a number of system tables and stored procedures:

• MSmerge_genhistory contains one row for each generation. A generation is a collection of changes that is delivered to a Publisher or Subscriber. Generations are closed each time the Merge Agent runs; subsequent changes in a database are added to one or more open generations.

• sysmergesubscriptions contains information on subscriptions, including a record of the last generations of changes a node has sent and received. In the publication database, this table contains a row for the Publisher and one row for each Subscriber. In a subscription database, this table typically contains a row for the Subscriber and a row for the Publisher.

• MSmerge_generation_partition_mappings is used only for filtered tables, recording whether a given generation contains any changes relevant to a given partition. This table in the publication database contains one row for each unique combination of rows in MSmerge_genhistory and MSmerge_partition_groups.

• sp_MSmakegeneration closes all open generations at the beginning of the enumeration process.

• sp_MSenumchanges enumerates changes for tables (a number of related procedures that have names beginning with sp_MSenumchanges are also used in this process).

• sp_MSgetmetadata determines whether a change from one node should be applied at another node as an insert, update, or delete.

The following process occurs during change enumeration:

1. The system procedure sp_MSmakegeneration is called:

   • For unfiltered and filtered tables, this procedure closes all open generations referenced in MSmerge_genhistory (closed generations have a value of 1 or 2 in the column genstatus).

   • For filtered tables, this procedure populates the system table MSmerge_generation_partition_mappings. If a generation contains one or more changes that are relevant for a partition, a row is inserted into the system table. If a generation does not contain any changes relevant for a given partition, a row is not inserted into MSmerge_generation_partition_mappings, and changes are not enumerated for any Subscribers that receive that partition.

2. The stored procedure sp_MSenumchanges and related procedures are called. These procedures enumerate the changes that have occurred since the last time synchronization occurred:

   1. The procedures first determine the generation at which enumeration starts, based on the columns sentgen (last generation sent) and recgen (last generation received) in the table sysmergesubscriptions.

      For example, when determining which generations' changes must be enumerated for a given Subscriber, the sentgen for the Subscriber (stored in the publication database), and the recgen for the Subscriber (stored in the subscription database) are compared. If the values are the same (which indicates that the last generation sent from the Publisher was successfully received by the Subscriber), changes are enumerated starting with the next generation in MSmerge_genhistory. If the values are not the same, the lower of the two values is used to ensure all required changes are sent.

   2. The procedures then enumerate changes:

      For unfiltered tables, all changes contained in generations after the generation in sentgen or recgen are enumerated: MSmerge_genhistory is joined to MSmerge_contents and MSmerge_tombstone to determine which changes must be sent.

      For filtered tables, MSmerge_generation_partition_mappings is joined to: MSmerge_current_partition_mappings and MSmerge_contents; and MSmerge_past_partition_mappings and MSmerge_tombstone to determine which changes are relevant for the partition that the Subscriber receives.

3. The stored procedure sp_MSgetmetadata is called to determine whether a change should be applied as an insert, update, or delete. At this point, conflict detection and resolution are performed; for more information, see How Merge Replication Detects and Resolves Conflicts.