Hi Jenni,
Thanks for your help. I add clean_run => true in conf file and run logstash again, no more error occurs.
So I think: in my issue, sql_last_value stored in last_run_metadata_path is not the numeric value or datetime value, after clean_run => true reseted the wrong value to 0 or 1970-01-01 , the thread goes on and data be indexed successfully, am I right?