Thread: ODBC problem with local PostgreSQL 8.1 DBMS
Hi together, I have a problem with connections to a PostgreSQL 8.1 system. My locally running application does about a few hundred SELECTs. During these SELECTs at a random point one SELECT (not even the same) fails always. The failing appears as some sort of Windows exception which I cannot details of because my C++ application catches this exception only by 'catch (...)'. Any other possible exception class is caught before and therefore not candidate for this special exception. Another thing is that the selects run really long but maybe this is just a problem of the last (failing) SELECT and a timeout. I am not sure about this. The strange thing about this is that the problem only occurrs when connecting to the DBMS from the same (local) machine. It does NOT occur when connecting to the DBMS from remote. So my idea was that maybe ODBC is doing something wrong. Any idea how to trace this? Thanks in advance! Heiko Nardmann
> I have a problem with connections to a PostgreSQL 8.1 system. My locally > running application does about a few hundred SELECTs. During these SELECTs at > a random point one SELECT (not even the same) fails always. The failing > appears as some sort of Windows exception which I cannot details of because > my C++ application catches this exception only by 'catch (...)'. Any other > possible exception class is caught before and therefore not candidate for > this special exception. What psqlodbc version are you using? What is the connection string? Could you provide us the common log output? Regards, Luf P.S. I'm going to be on holiday next week. I hope someone else will help you.
Am Donnerstag, 1. März 2007 17:01 schrieb Ludek Finstrle: > > I have a problem with connections to a PostgreSQL 8.1 system. My locally > > running application does about a few hundred SELECTs. During these > > SELECTs at a random point one SELECT (not even the same) fails always. > > The failing appears as some sort of Windows exception which I cannot > > details of because my C++ application catches this exception only by > > 'catch (...)'. Any other possible exception class is caught before and > > therefore not candidate for this special exception. > > What psqlodbc version are you using? What is the connection string? > Could you provide us the common log output? > > Regards, > > Luf > > P.S. I'm going to be on holiday next week. I hope someone else will help > you. In the meanwhile we solved the problem. The problem was that the ODBC definition had 'Debug' option set to '1' which seems to create some overhead which leads to either timing issues or too much data. I'm using OTL for accessing ODBC - maybe it is an issue there. -- Heiko Nardmann