On 7/30/2010 1:44 AM, Matthew Wakeling wrote:
> On Thu, 29 Jul 2010, David Kerr wrote:
>> Getting protocols out of sync sort of sounds to me like a network glitch
>> of some sort to me. does that sound possible?
>
> No. TCP has pretty robust mechanisms for preventing that sort of error.
>
> Matthew
>
Ok, yes that's consistent with what we found yesterday where we were
able to reproduce it on system that was being used by other people
without problem.
we've got a working theory that the program opens up a connection to
DB1 and DB2. In DB1 we're running a very long query (15 minutes) and
after some sort of timeout(or something) the program hangs leaving the
idle in TX in DB2.
Thanks
Dave