Help regarding Flow of internal Sync Flow

[arjunbharara]

Which stored procedure is responsible for populating data in the tracking table where the updated_scope_id is set to 0000-0000-0000-0000? This is important as null values are already being handled in triggers.

AND
which procedure actually selects the changes for Sync.

[arjunbharara]

I have a multi-scope setup, and it is generating similar sync_scope_id values for different scopes, which is causing issues. Could you please specify what might be the reason behind the creation of similar sync_scope_id values?

[Mimetis]

0000-0000-0000-0000 is the server scope, always

[arjunbharara]

f 0000-0000-0000-0000 is always the server scope, then what is the role of sync_scope_id? How does it function in the overall synchronization process if it ultimately resolves to 0000-0000-0000-0000?

[Mimetis]

sync_scope_id is uniquely identifying each client with a random GUID generated on the first sync
sync_scope_id is always Guid.Empty (0000-......) for server side

[arjunbharara]

I have found that the _changes stored procedure is called to fetch changes from both sides. It takes two parameters: timestamp and sync_scope_id. On the client side, the sync_scope_id is sent as null.

Is there any way to intercept or modify this behavior to replace null with a normal ID or a standard GUID?

[Mimetis]

If you look at the GetChanges method, you have this WHERE clause:

WHERE (
	[side].[timestamp] > @sync_min_timestamp
	AND ([side].[update_scope_id] <> @sync_scope_id OR [side].[update_scope_id] IS NULL)
)

@sync_min_timestamp is the inferior limit from which you want the changes.
Usually, this value comes from the scope_info_client table

@sync_scope_id is the client_id from which one you dont want the rows.

Explanation:

Each time a database is updating a row locally, the update_scope_id is set to NULL:

• NULL means that this row has been updated locally.

ON the client side, when you try to get changes for a particular client ( for example ClientId = 56FGH-DDD....), you want:

• ALL the rows updated locally (update_scope_id == NULL)
• NOT the rows where the update_scope_id has been updated to the correct value by the server (when server receive a NULL value, it replaces it with the correct client_id), so @sync_scope_id != 56FGH-DDD....

ON the server side, you want the rows updated by the server (NULL) and all the others clients , to be sent to a particular client (56FGH-DDD....):

• ALL the rows updated locally (update_scope_id == null)
• ALL the rows that are different from the client we are targetting @sync_scope_id != 56FGH-DDD....
• NOT the rows updated by the server with a valid client value value ( @sync_scope_id != 56FGH-DDD....)

If you want to see how it's working you can play with the GetChanges() method you have on LocalOrchestrator and RemoteOrchestrator.

That being said, I don't know what you want to achieve by changing the @sync_scope_id value, but you may have a better view with this explanation