Is it necessary to update scope_info and scope_config when adding columns to a table in a provisioned database?

Question

I followed the example provided in
https://jtabadero.wordpress.com/2011/04/13/modifying-sync-framework-scope-definition-part-3-workarounds-addingremoving-columns/ to add columns to a table in a provisioned database.  The sample code replaces the stored procedures triggered when the
table is updated with new stored procedures which include the additional columns
However, the sample does not update scope_config. When I am finished, the new columns do not appear in the scope config.  Looking at the SQL code it is immediately obvious why.  Here is an excerpt from the beginning of the query
DROP PROCEDURE [PATIENT_insertmetadata];
DROP PROCEDURE [PATIENT_updatemetadata];
DROP PROCEDURE [PATIENT_deletemetadata];
-- BEGIN Enable Snapshot Isolation on Database 'PatScan' if needed
IF EXISTS (SELECT NAME FROM sys.databases where NAME = N'PatScan' AND [snapshot_isolation_state] = 0)
BEGIN
 ALTER DATABASE [PatScan] SET ALLOW_SNAPSHOT_ISOLATION ON
END
GO
-- END Enable Snapshot Isolation on Database 'PatScan' if needed
-- BEGIN Create Scope Info Table named [scope_info]
IF NOT EXISTS (SELECT t.name FROM sys.tables t JOIN sys.schemas s ON s.schema_id = t.schema_id WHERE t.name = N'scope_info' AND s.name = N'dbo')
BEGIN
 CREATE TABLE [scope_info] ([scope_local_id] int IDENTITY(1,1) NOT NULL, [scope_id] uniqueidentifier DEFAULT NEWID() NOT NULL, [sync_scope_name] nvarchar(100) NOT NULL, [scope_sync_knowledge] varbinary(max) NULL, [scope_tombstone_cleanup_knowledge] varbinary(max)
NULL, [scope_timestamp] timestamp NULL, [scope_config_id] uniqueidentifier NULL, [scope_restore_count] int DEFAULT 0 NOT NULL, [scope_user_comment] nvarchar(max) NULL)
 ALTER TABLE [scope_info] ADD CONSTRAINT [PK_scope_info] PRIMARY KEY ([sync_scope_name])
END
GO
-- END Create Scope Info Table named [scope_info]
-- BEGIN Create Scope Config Table named [scope_config]
IF NOT EXISTS (SELECT t.name FROM sys.tables t JOIN sys.schemas s ON s.schema_id = t.schema_id WHERE t.name = N'scope_config' AND s.name = N'dbo')
BEGIN
 CREATE TABLE [scope_config] ([config_id] uniqueidentifier NOT NULL, [config_data] xml NOT NULL, [scope_status] char NULL)
 ALTER TABLE [scope_config] ADD CONSTRAINT [PK_scope_config] PRIMARY KEY ([config_id])
END
GO
-- END Create Scope Config Table named [scope_config]
In an already provisioned database, the scope_info and scope_config tables already exist.  Hence the portions of the script to create the scope_info and scope_config tables are skipped.  I tried modifying the query to drop these tables.  When
I do so, ne tables are created but they only define the modified table.
Is it necessary to update scope_info and scope_config?  Are there are side effects (such as deprovisioning) if I drop and then recreate these tables? Can I just update the existing scope_config, replacing the portion of the XML which defines the table
being modified? Do I have to make changes in scope_info?
Thanks
Howard Weiss
Howard P. Weiss

Answer 1

I followed the example provided in
https://jtabadero.wordpress.com/2011/04/13/modifying-sync-framework-scope-definition-part-3-workarounds-addingremoving-columns/ to add columns to a table in a provisioned database.  The sample code replaces the stored procedures triggered when the
table is updated with new stored procedures which include the additional columns
However, the sample does not update scope_config. When I am finished, the new columns do not appear in the scope config.  Looking at the SQL code it is immediately obvious why.  Here is an excerpt from the beginning of the query
DROP PROCEDURE [PATIENT_insertmetadata];
DROP PROCEDURE [PATIENT_updatemetadata];
DROP PROCEDURE [PATIENT_deletemetadata];
-- BEGIN Enable Snapshot Isolation on Database 'PatScan' if needed
IF EXISTS (SELECT NAME FROM sys.databases where NAME = N'PatScan' AND [snapshot_isolation_state] = 0)
BEGIN
 ALTER DATABASE [PatScan] SET ALLOW_SNAPSHOT_ISOLATION ON
END
GO
-- END Enable Snapshot Isolation on Database 'PatScan' if needed
-- BEGIN Create Scope Info Table named [scope_info]
IF NOT EXISTS (SELECT t.name FROM sys.tables t JOIN sys.schemas s ON s.schema_id = t.schema_id WHERE t.name = N'scope_info' AND s.name = N'dbo')
BEGIN
 CREATE TABLE [scope_info] ([scope_local_id] int IDENTITY(1,1) NOT NULL, [scope_id] uniqueidentifier DEFAULT NEWID() NOT NULL, [sync_scope_name] nvarchar(100) NOT NULL, [scope_sync_knowledge] varbinary(max) NULL, [scope_tombstone_cleanup_knowledge] varbinary(max)
NULL, [scope_timestamp] timestamp NULL, [scope_config_id] uniqueidentifier NULL, [scope_restore_count] int DEFAULT 0 NOT NULL, [scope_user_comment] nvarchar(max) NULL)
 ALTER TABLE [scope_info] ADD CONSTRAINT [PK_scope_info] PRIMARY KEY ([sync_scope_name])
END
GO
-- END Create Scope Info Table named [scope_info]
-- BEGIN Create Scope Config Table named [scope_config]
IF NOT EXISTS (SELECT t.name FROM sys.tables t JOIN sys.schemas s ON s.schema_id = t.schema_id WHERE t.name = N'scope_config' AND s.name = N'dbo')
BEGIN
 CREATE TABLE [scope_config] ([config_id] uniqueidentifier NOT NULL, [config_data] xml NOT NULL, [scope_status] char NULL)
 ALTER TABLE [scope_config] ADD CONSTRAINT [PK_scope_config] PRIMARY KEY ([config_id])
END
GO
-- END Create Scope Config Table named [scope_config]
In an already provisioned database, the scope_info and scope_config tables already exist.  Hence the portions of the script to create the scope_info and scope_config tables are skipped.  I tried modifying the query to drop these tables.  When
I do so, ne tables are created but they only define the modified table.
Is it necessary to update scope_info and scope_config?  Are there are side effects (such as deprovisioning) if I drop and then recreate these tables? Can I just update the existing scope_config, replacing the portion of the XML which defines the table
being modified? Do I have to make changes in scope_info?
Thanks
Howard Weiss
Howard P. Weiss