python - Psycopg2 db connection hangs on lost network connection -

Problem Description

I use psycopg2 to connect to your PostgreSQL database on a remote host I am I open a connection and wait for the requests, then for each request, I run a query on connection and return data.

But when the network connection is lost after connection, the next DB query is already hanging and I

details:

• This less
• I can not use the asynchronous connection because I need a transaction
• Python 2.6
• psycopg 2.2.1
• Debian Linux 6 and 7, 64bit
• Postgresclines 8.4 and 9.1

What do I need

I need reliable way to detect some unsuccessful connections before running a query, so my program will not hang, or cursor.exact (..) make way failed to raise an exception on connection

example :.

  import psycopg2 import time Conn = psycopg2.connect ( "host = ' DBS 'dBNAME =' foo 'user =' joe 'password =' ​​x '") time.sleep (10) # I manually switch off VPN during this sleep Vendor am .. cu = conn.cursor () cu.execute ( 'SELECT 1') # & lt; -  

What I have tried (and what has not worked):

• Before importing Psycopg2, I said:

    import socket .setdefaulttimeout (10)  
   
   TCP timeout "globally" TCP timeout set to  psycopg.connection  socket: 
   
    .. Conn = psycopg2.connect (... S = socket.fromfd (conn.fileno (), socket.AF_INET, socket.SOCK_STREAM) s.settimeout (5).  
   
   
    To enable the keepalive for  psycopg.connection  socket: 
   
    ... Conn = psycopg2.connect (... s = socket.fromfd (conn.fileno (), socket.AF_INE T Soketksok_strym) Sksettimut (5) Sksetsocopt (Soketksol_soket, Soketkso_keeplive 1) Sksetsocopt (Soketkipproto_tkp, Soketktkp_keapidle 1) Sksetsocopt (Soketkipproto_tkp, socket. TCP_KEEPINT VL, 3) s.setsockopt (socket.IPPROTO_TCP, socket.TCP_KEEPCNT, 5) ...