Hello re:Post community,
I would appreciate input on options to move databases residing in multiple RDS MySQL instances to a single RDS MySQL instance as detailed below with minimal (a few minutes) downtime.
Current state
- Existing RDS MySQL 5.7 instance, ARIA, with DBs ap, aq
- Existing RDS MySQL 5.7 instance, ARIB, with DBs bx, by
Desired future state
- New (and only) RDS MySQL 5.7 instance, ARIC, with DBs ap, aq, bx, by
- Appropriately sized instance with GP3 storage
This re:Post article describes a method to arrive at future state using "standard" DMS (with endpoints). This method requires a fair amount of additional work including setting extra connection attributes, dumping and restoring DDL prior to data migration, recreating indexes etc.
Using a MySQL-compatible database as a source for AWS DMS has this recommendation:
- "But for a homogeneous migration, where you are migrating from a MySQL database to a MySQL database, we recommend that you use a homogeneous data migrations migration project. homogeneous data migrations uses native database tools to provide an improved data migration performance and accuracy when compared to AWS DMS."
However, the new'ish (June 2023) homogeneous data migration capabilities in DMS appear to not allow for such RDS instance consolidation based on the following:
- There is no field to provide a (source) database name when creating a data provider with the MySQL engine type.
- Creating a data migration in AWS DMS states
- "For data migrations of the Full load and Full load and change data capture (CDC) type, AWS DMS deletes all data, tables, and other database objects on your target database. Make sure you have a backup of your target database."
Please comment if any additional information would be of help with sharing your input.
Thanks!
A few updates after working some more with Homogeneous Data Migrations (HDM):
Given the above, arriving at the desired state using HDMs could be possible, albeit with higher downtime, by performing full loads followed by taking downtime and doing CDCs sequentially for each of the four source databases.
The community's feedback and suggestions would be welcome.
Thanks!