2026-06-09 035443 source ERROR Jun 09, 2026 65443 AM com.github.shyiko.mysql.binlog.BinaryLogClient connect 2026-06-09 035443 source ERROR : Connected to 10.25.41.8:3306 at mysql-bin.049323/15676253 (sid:6135, cid:404454) 2026-06-09 035444 source ERROR Jun 09, 2026 65444 AM com.github.shyiko.mysql.binlog.BinaryLogClient$5 run 2026-06-09 035444 source ERROR : threadExecutor is shut down, terminating keepalive thread

The knowledge sources do not contain enough information to give a definitive diagnosis for your specific log snippet alone. The log shows: 1. A successful connection to your MySQL instance at

10.25.41.8:3306

(binlog position

mysql-bin.049323/15676253

) 2. Almost immediately after, the

threadExecutor is shut down, terminating keepalive thread

— indicating the connection was torn down very quickly This pattern appears in several related MySQL CDC issues. Based on the knowledge sources, here are the most likely causes and things to check: Possible Causes 1. Binlog deserialization error (most common) In similar cases, the keepalive thread shutdown is preceded or followed by a Debezium binlog deserialization error (e.g.,

Failed to deserialize data of EventHeaderV4

or

EOFException

). Check your full logs for errors immediately after this snippet. [CDC binlog issue] [append-only sync issue] 2. Binlog position no longer available If the saved binlog offset has been purged from the server, Airbyte may fail to resume CDC streaming. Look for a log message like: >

Saved offset no longer present on the server, Airbyte is going to trigger a sync from scratch

To prevent this, increase binlog retention on your MySQL server:

-- For standard MySQL
SET GLOBAL binlog_expire_logs_seconds = 604800; -- 7 days

[MySQL troubleshooting] Recommended Steps • Check the full logs around and after this timestamp for any

ERROR

or stack trace entries — the keepalive shutdown is usually a symptom, not the root cause. • Verify binlog retention is set to at least 24–168 hours on your MySQL server. • Check for deserialization errors — if present, a reset of the CDC state (triggering a new full refresh) may be needed. If you can share more of the surrounding log output (especially any errors after the keepalive shutdown), a more precise diagnosis would be possible. 👍 0 button 👎 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.