<@U04S8PLJWMT> I have Airbyte 2.0.1 and a connecti...
# replication-ask-ai
a
@kapa.ai I have Airbyte 2.0.1 and a connection between Source MySQL (3.51.5) and Destination BigQuery (3.0.16). After a series of successful syncs, I have a series of syncs that are still successful but that do not import and extract any rows. The only way to fix this is to refresh the entire connection. Why does this happen?
k
I’m sorry, but the knowledge sources don’t contain a clear, explicit root cause for this exact behavior on Airbyte 2.0.1 with MySQL 3.51.5 → BigQuery 3.0.16. I can only point to similar, possibly related issues and patterns. From the available information: - There is an open issue where the MySQL source sync finishes as “successful” but no data is synced. In that case, the job is marked successful, no rows are emitted, and neither “successful” nor “failed” notifications are triggered. [[MySQL success bug](https://github.com/airbytehq/airbyte/issues/68104)] - Another user reports the same issue with MySQL connector 3.51.5, i.e., syncs show success but are not actually successful. [[MySQL success bug](https://github.com/airbytehq/airbyte/issues/68104)] - There are multiple forum threads where: - Syncs suddenly start failing or behaving incorrectly after some time, and users temporarily fix it by resetting/refreshing the connection or splitting data into multiple connections. [[Suddenly failing syncs](https://discuss.airbyte.io/t/2656); [MySQL truncation issue](https://discuss.airbyte.io/t/2032)] - A MySQL→BigQuery pipeline behaves incorrectly in incremental mode but works in full refresh, again pointing to bugs in the MySQL + GCS/BigQuery incremental path. [[MySQL + BigQuery incremental](https://discuss.airbyte.io/t/298)] Putting this together, the most likely explanation (based on the docs and issues, not on speculation) is: - There is a known bug in the MySQL source connector where a sync can be marked as successful even though no data is actually read/emitted, and this affects recent 3.x versions (including 3.51.5). [[MySQL success bug](https://github.com/airbytehq/airbyte/issues/68104)] - Refreshing/resetting the connection effectively “rewinds” the state and forces a full re-read, which is why your data starts flowing again, similar to how a Refresh Sync is described to clear cursor/state and re-import data. [[Refreshing data](https://docs.airbyte.com/platform/operator-guides/refreshes)] However, the knowledge sources do not provide: - A precise technical root cause (e.g., specific state bug, CDC edge case, or catalog issue) for your exact version combination. - A documented permanent fix beyond upgrading connectors / Airbyte or tracking the open GitHub issue. Given what we have, the actionable steps that are supported by the sources are: 1. Treat this as the known “MySQL shows success but isn’t successful” bug and follow that GitHub issue for updates or fixes. [[MySQL success bug](https://github.com/airbytehq/airbyte/issues/68104)] 2. Use Refresh/Reset as a workaround when the issue appears, since this is consistent with how others temporarily recover from state-related MySQL issues. [[Refreshing data](https://docs.airbyte.com/platform/operator-guides/refreshes); [Suddenly failing syncs](https://discuss.airbyte.io/t/2656)] 3. If possible, try newer connector versions once a fix is released in that issue, as several threads suggest upgrading connectors when facing MySQL→BigQuery anomalies. [[Suddenly failing syncs](https://discuss.airbyte.io/t/2656); [MySQL to BigQuery fails](https://discuss.airbyte.io/t/2146)] If you can share (outside this chat) a link to your job logs in a GitHub issue, the Airbyte team typically asks for that to further diagnose similar MySQL/BigQuery incremental problems.
a
Hello, just reverting here as I found the root cause for me. It was due to a long-running uncommitted transaction in my DB, that kept writing to the binlog but prevented Airbyte from moving forward from that transaction. It was fixed once the transaction was committed, and we reset the pipeline data. I hope this can be helpful to anyone!