Common MySQL Source Setup Mistakes

May 2026

MySQL source setup looks straightforward until CDC enters the picture and the binlog settings stop being optional background details.

Short Answer

The most common MySQL source mistakes are relying on an imprecise connection string, being vague about which tables should be ingested, and turning on CDC without verifying the binlog prerequisites. Skippr can read MySQL tables in batch mode with a normal connection string, but CDC specifically requires binlog_format = ROW, binlog_row_image = FULL, and a replication user with REPLICATION SLAVE and REPLICATION CLIENT privileges.

That is where teams often lose time. A batch read can succeed with a standard app user, then CDC fails because the source is emitting statement-oriented binlog events or incomplete row images. Another common problem is assuming restart safety exists automatically even though the real resume point is the stored binlog filename and position, which only helps if binlog retention covers the outage window.

Why Teams Struggle with This

MySQL source setup has two layers: ordinary table access and binlog-backed mutation capture. Many teams validate the first and forget the second, which is why their first CDC test reveals missing replication privileges or source settings that were never part of app-database onboarding.

• A valid MySQL connection string does not imply the user has the replication privileges CDC needs.
• CDC depends on ROW binlog format and FULL row images, not just on binlog being enabled in some form.
• Omitting table scope during evaluation can make the first run noisier and harder to verify than necessary.
• Restart safety depends on binlog retention covering downtime, because Skippr resumes from stored filename and position.

How Skippr Handles It

Skippr makes the split between ordinary extraction and CDC explicit. The connector docs explain the connection-string and table-selection surface, while the CDC source docs spell out the replication configuration, row-event types, and stored resume position.

That clarity matters during rollout. Teams can inspect whether they are truly ready for WRITE_ROWS, UPDATE_ROWS, and DELETE_ROWS handling instead of treating CDC as a checkbox on top of an unrelated batch connector.

• Simple MySQL batch extraction through a standard connection string and optional table list.
• CDC support through binlog replication with stored filename and position resume points.
• Explicit prerequisites for row-based binlog format and full row images.
• A clean path from MySQL row events to downstream final-state destinations.

What the First Useful Version Looks Like

A useful first version is a read-capable connection string plus a short list of important tables so the raw landing shape is easy to verify.

If CDC is the real target, check binlog settings and replication privileges before building downstream models, because those source invariants decide whether updates and deletes are usable.