Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.server -> Re: Row values changing back to old values randomly???
Randy Harris wrote:
> "DG" <info_at_no-spam-e-resitve.com> wrote in message > news:wmKme.12391$F6.2626011_at_news.siol.net... > > [snip...] > >>>>>>levels. Is it possible that the database is configured to "READ
>>>>>So back to the technical side. I though about transaction isolation
> > is >>>>But I think 2 is pretty likely.
>>>>>really showing data which hasn't been committed yet. Later the session
>>>>>disconnects and the data is converted back.
>>>>>
>>>>>Thanks, David
>>>>
>>>>There are truly only three possibilities.
>>>>
>>>>1. There is no commit and you don't realize it.
>>>>
>>>>2. The commit is failing and your understanding of the Oracle
>>>>transaction model is so weak you don't understand how to interpret
>>>>what you are seeing or your code doesn't trap for the failure.
>>>>
>>>>3. The commit is succeeding and you have a problem with caching
>>>>on the application server or in your application.
>>>
>>>
>>>4. The TOAD is looking from another session that started a transaction
>>>before the app server started it's update transaction, after looking
>>>from a session that started after the update transaction committed.
>>>
> > > David, I've been reading this thread with interest. I read this NG to learn > and find there is a much value in learning from the experiences of others. > Thanks for posting the solution you've uncovered, it adds to the value for > learners, such as myself. > > I do quite a lot of work with ODBC myself, mostly using ADO. From what I > know of it, I find it difficult to accept that the driver is somehow caching > that data, then triggering a revert. Rolling the transaction back after > some sort of timeout seems much more plausible, though this explanation is, > of course, contradicted by what you see with TOAD. > > You've obviously found a "fix" for the problem and I have no more reasonable > explanation to offer. I still, however, find this quite puzzling. >
Hello, sorry for bugging you all. After a joint session with the customer's DBA we finally got to looking at the logs. And it turned out that there was a third process which was replicating tables from another view and which was periodically overwriting all the data the users stored in our web app. So, it was only a coincidence that the replication didn't happen when we installed our "fix".
Best regards,
David Received on Tue May 31 2005 - 07:11:51 CDT