FDW Foreign Table Access: strange LOG message

[Albrecht Dreß <albrecht.dress@xxxxxxxx>]

Hi all,

I use Postgres 10.3 on a Debian Stretch system with foreign tables, and noticed strange LOG messages when accessing them.

The data base setup is basically

---8<------------------------------------------------------------------------------------------
CREATE EXTENSION IF NOT EXISTS postgres_fdw WITH SCHEMA public;
CREATE SERVER ext_server FOREIGN DATA WRAPPER postgres_fdw OPTIONS (
    dbname 'ext_db', host 'localhost', updatable 'false');
CREATE FOREIGN TABLE public.ext_table (
    […]
) SERVER ext_server OPTIONS (schema_name 'public', table_name 'some_table', updatable 'false');
---8<------------------------------------------------------------------------------------------

Now I use the following trivial Python2 (psycopg2) code to access the tables:

---8<------------------------------------------------------------------------------------------
import psycopg2
conn = psycopg2.connect(dbname='my_db')
with conn.cursor() as csr:
   csr.execute("SELECT * FROM […] LIMIT 1")
   csr.fetchone()
conn.close()
---8<------------------------------------------------------------------------------------------

When I access a “local” table of my_db in the SELECT statement, there is no LOG message.  However, reading from ext_table works just fine, but the conn.close() statement above triggers the log message

---8<------------------------------------------------------------------------------------------
[time stamp/pid] user@my_db LOG:  could not receive data from client: Connection reset by peer
---8<------------------------------------------------------------------------------------------

Did I miss something in the setup here?  Or is there any other way to get rid of the message (which fills ~95% of my logs)?

Thanks in advance,
Albrecht.

Attachment: pgp2eZdZd5aP2.pgp
Description: PGP signature