Or at least I think that is why we turn autocommit OFF. Our coding standards clearly said that we should always start a transaction explicitly. Alas, this is not very good in production environments, as if you are not aware of the behaviour or just oblivious , you can cause blocking situations. If I follow your steps I do indeed see that the transaction does not prevent the store being cleaned up. If you really want an answer, opening a case is a better approach, but if they think it is not a bug, the case may cost you an arm and a leg. Monday, March 27, 3: Remove From My Forums.
|Date Added:||26 September 2010|
|File Size:||62.86 Mb|
|Operating Systems:||Windows NT/2000/XP/2003/2003/7/8/10 MacOS 10/X|
|Price:||Free* [*Free Regsitration Required]|
ODBC – who is using it? | Uniface Deployment | Forum
Edited by cbreemer Tuesday, March 21, 2: Thursday, March 30, 9: In both cases the transaction will have begun by the time you run the first query. Alas, this is not very good in production environments, as if you are not aware of the behaviour or just obliviousyou can unifcae blocking situations.
As I pointed out, in the day I did live in a world with no autocommit, I still preferred to start the transactions explicitly. Microsoft Technology Center – Dallas My blog.
This hasn’t caused problems for our customers so far, but one customer uses snapshot isolation on their DB and are observing massive bloat of the tempdb, apparently caused by the fact that each process always has a transaction open. Hope someone can help me – I could not find the answer to my query anywhere. Monday, March 27, 9: Sunday, April 9, 1: All fine, this is as expected and documented.
Since you commit them, you might as well start them. Monday, March 27, 3: But there is two way to turn it off: Sunday, April 9, 8: Sunday, April 9, 6: It does happen that you get response from Microsoft on Connect, but there is no guarantee.
Help us improve MSDN.
Which is more or less what you suggest, right? Why is that obvious?
Thanks unifwce the tip Erland,l I’ll keep that possibility in mind. The content you requested has been removed. In that case autocommit would be ignored.
But we may well need to study in more depth the relation between autocommit and implicit transactions.
I have not posted in these forums before.
SQL Support on ODBC
Monday, May 1, 3: We want to control the transactions from Uniface, so obviously autocommit is turned off. It’s really appreciated, very useful information. I agree it would be better to start the transactions unifaxe, but unfortunately that is not how our drivers work.