Running the binlog_streamer from MariaDB seems to be a little bit different.
an error seems to occur:
PANI[0001] tried to advance to a zero log position: mysql-bin.047756 0 string tag=source_binlog_streamer
I tracked it down to the following events:
replication.TableMapEvent, replication.GenericEvent, *replication.QueryEvent
Which is documented to happen when a replica is unable to understand MariaDB's protocol:
"Dummy event replacing event type X that slave cannot handle."
Running the binlog_streamer from MariaDB seems to be a little bit different.
an error seems to occur:
I tracked it down to the following events: replication.TableMapEvent, replication.GenericEvent, *replication.QueryEvent
Which is documented to happen when a replica is unable to understand MariaDB's protocol: "Dummy event replacing event type X that slave cannot handle."
To me it looks like it comes from the following events: ANNOTATE_ROWS_EVENT (https://mariadb.com/kb/en/annotate_rows_event/) As this is what I sniffed from the traffic:
and the plain text "BEGIN" which both gets shoved in the replication.QueryEvent
Further more the also state that the TABLE_MAP_EVENT is used for row-based binlog: https://mariadb.com/kb/en/table_map_event/
Currently debugging for more data, will come with a PR that works with my current MariaDB: 10.3.28-MariaDB-1:10.3.28+maria~focal-log