log_reader internal worker waits at: issue_resync_and_wait and blocks log reader
because of fixed size internal queue and the copy thread waits at wait_for_state:
│ LOG_DEBUG(LOG_PG_LOG_MGR, "Wait for state syncing, db: {}", _db_id);
│ // ensure the pipeline was stalled before we complete
│ _internal_state.wait_for_state(STATE_SYNCING);
Also when the parent table and child table are resynced in parallel, there are times these tables' ddl go out of order causing parent table to lose partition.
cc: @garth @craig @Arun