Hello all :) SymmetricDS' blog talks about RedShift as target-only without going into details.
Updates can be problematic in RedShift, performance-wise.
Contexte is: A one-way replication from a MySQL instance to RedShift has already been set up and databases are in synch. Rows in MySQL then gets updated.
So, the question is:
- Will the line get updated in RedShift, with the JDBC way, with the S3-bulk-load way? Does the S3-bulk even work with updates?
Additional non necessary questions:
- Performance-wise, is JDBC still ok for 1m updates a day?
- Is Vaccuum performed by SymmetricDS? Do I need to schedule it?
- If the S3-bulk-update way exists, how does it work internally? Does it loads as a sort of log table that is then propagated to the RedShift target table?
EDIT: I'd love to go parse the source code, but the feature is not yet available (last update of publicly open source files date before the feature)