Ability to re-run migration and keep data in sync

Aug 26, 2015 at 5:30 AM

I am working on a migration project for a very old and big application/website with bunch of large databases. Source databases are hosted on a SQL server cluster in a private data center and the plan is to move everything to Azure Sql Dbs.

Using this tool, I've successfully done the first run of migration which is a proof of concept and the next run is supposed to be for go-live. Since there is a large amount of data and a long down-time being unacceptable by the client - I was thinking about different options. One of them being this:
  1. Do the first sync using the migration wizard: source Sql cluster to destination Azure Sql
  2. Apply the required schema changes from the migration in the other direction: e.g. source Azure Sql (or the tool) to the Sql cluster (not sure what the quickest way to do this is)
  3. Setup some sort of data transfer between the two to keep data in sync (not sure how to do this - need community opinion)
  4. Considering the data is already being synced - just do the switch when ready instead of running migrations once again
My questions are:
  • Is the above possible to achieve? If so - please provide additional details on step 2 and 3
  • Are there any easier alternatives?
Aug 28, 2015 at 1:10 PM
Edited Sep 2, 2015 at 1:33 PM

A lot of people use SQLAzureMW for the first datasync and then move over to SSIS or Azure Data Factory to keep things up to date.

Sep 2, 2015 at 8:58 AM
Thank you very much - I will give it a go!