DataStream fails to start with unsupported version error on MariaDB 10.6

Google Cloud Data Analytics
1

Hello,

I’m currently using Google Datastream to replicate data from a MariaDB 10.6 instance into BigQuery. According to the official documentation, MariaDB 10.6 should be supported, and all prerequisites are met (binary logging configured, connection tested, etc.).

However, when I attempt to create a new stream, Datastream fails to start with an error message indicating that my MariaDB version is not supported.

What’s puzzling is:

  • I have other existing Datastream pipelines pointing to the same MariaDB 10.6 database that are working without issues.
  • I also have streams from MariaDB 10.4 that work fine.
  • When I force the stream to start (even though its status shows as “failed”), the data flows normally into BigQuery.
  • There are no connection issues, and no data loss has been observed.

This seems like either a version detection bug or a regression in the latest Datastream version checks.

Is anyone else experiencing this behavior? Are there any workarounds or configuration flags to skip the version validation, or is this something that needs escalation to the Datastream engineering team?

Any insight would be appreciated.

Thanks,
Iván

2

Hi @ivan_rey,

Welcome to the Google Developer Program forums!

This appears to be a bug in the version validation logic during stream setup, not a problem with actual data replication, which explains why your existing streams and forced starts continue working as expected. You can try any of the following

  • Proceed with forced start if replication works reliably, but monitor replication lag and errors closely.

  • Verify the output of SELECT VERSION(); on your MariaDB instance—Datastream may be misinterpreting certain version suffixes. Suffixes such as -MariaDB-log or operating system-specific identifiers may cause compatibility validation to fail. If the version string includes such variations, it may be helpful to place a proxy in front of the database to normalize the response.

  • Contact Google Cloud Support with full logs and metadata to help resolve what may be a recent regression in Datastream’s version validation mechanism.

  • Google Cloud is actively updating Datastream compatibility and validation logic, so I suggest tracking release notes for any fix announcements.

Was this helpful? If so, please accept this answer as “Solution”. If you need additional assistance, reply here within 2 business days and I’ll be happy to help.