connect function timeout not working with SQL Server using SSPI or documentation incorrect about pyodbc HOT 6 CLOSED

Hi dkleehammer,

First of all, thanks for an awesome project.

I'm having this problem with both 3.0.6 and the latest beta (3.0.7-beta03). If I set an incorrect server name in the connect string it tries to connect forever even though I've added the timeout flag to the connect method. Keyboard interrupts don't work until maybe a minute in or so.

I'm on Mac OS X 10.8.2 (64-bit) and the database I'm trying to connect to is a SQL Server 2008 R2 x64 (10.50.1617.0).

I can see that you closed this ticket. Was this fixed?

from pyodbc.

jmagnusson,

Actually, this project is ran by my brother, mkleehammer, but I'm sure he'll check this one. :D

I was incorrect about the keywords in the connection string. The timeout is soley used for query timeouts. We are investigating why the odbc connection timeout isn't being handled properly. We believe we are on the right track and will update pyodbc with the fix as quickly as possible.

dk

from pyodbc.

Did this ever get fixed?
Currently I have a firewall blocking outbound database connections. I've only just found this out after some debugging - unfortunately the timeout keyword seems to be being ignored and it hangs forever. On non firewalled host it works fine but I need me application to raise an error not just hang.

CONN_STRING = 'DSN={};DATABASE={};UID={};PWD={}'.format(SERVER,DATABASE,UID,PWD) conn = pyodbc.connect(CONN_STRING,timeout=15)

from pyodbc.

Is this issue fixed?

from pyodbc.

I don't believe so

On Wed, 6 Apr 2016, 09:13 Sreejith Kesavan, [email protected]
wrote:

Is this issue fixed?

—
You are receiving this because you commented.
Reply to this email directly or view it on GitHub
#11 (comment)

Sent from Android

from pyodbc.

Just discovered this same issue on Windows Server 2019 using ODBC Driver 17 for SQL Server. I have a canary that is checking health of services using pyodbc and hit was hanging on this when port is not responding.

Work around is to do a port scan of 1433 before trying to connect.

Here is a helper for scanning ports:

# (bPortOpen, iErrno) = scanPort('<ip>', <port>)
def scanPort(host:str, port:int, timeout:int=4) -> tuple(bool, int):
    errno = 0 # see TCP socket error codes https://gist.github.com/gabrielfalcao/4216897
    sock = None
    try:
        sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
        sock.settimeout(timeout)
        errno = sock.connect_ex((host, port))
    except socket.timeout:
        errno=-1
    except socket.gaierror:
        errno=112
    except socket.error:
        errno=111
    try:
        sock.close()
    except:
        None
    return ((errno == 0), errno)

from pyodbc.