RE: Grabbing sequence values blocks in consecutive order: Need a guaranteed method.

From: Amar Kumar Padhi <amar.padhi_at_gmail.com>
Date: Wed, 25 Mar 2009 21:04:55 +0400 (GST)
Message-ID: <32310419.2111238000882872.JavaMail.seven_at_aomfe4p1>

Hmmm... Guys thanks for your feedback. Much appreciated. I have received such feedback in the past and I have identified that simple checks put in the right place with this logic just makes it foolproof. The old note I gave link to is an elementary form of how to use the concept, and not the complete design. I have used this in quite a few heavy load designs and never found any issues, some are running for years now.

Places where I could not implement this is where the application_info call is already in use. For instance, oracle apps already uses it to register session information (had to resort to using a generic lock holding table for such designs).  

I do agree that locking is also a good option, but this did result in some serious bottlenecks on different client runtime.

And yes, I haven't tried this on RAC. This will come up for me soon, so will keep this list posted.

thanks!
Amar
Www.amar-Padhi.com

-original message-

 RE: Grabbing sequence values blocks in consecutive order: Need a guaranteed method.

--

http://www.freelists.org/webpage/oracle-l Received on Wed Mar 25 2009 - 12:04:55 CDT