Re: Understanding Terracotta caching
Date: Fri, 22 Jan 2010 10:48:41 -0700
Message-ID: <3c5f7821001220948p13f9dbfflcf89029bbcbf4dea_at_mail.gmail.com>
I'm already making a list of questions for the training. I've also been gathering more information on the tables I know about. I have already set up a test database specifically for this project. Unfortunately, the amount of data is not high due to disk limitations. I've asked for additional disk so we can do a valid test. I've been pullilng the most frequently used sql for the tables the DA has told me about so we can make sure those are run during the testing.
Any other suggestions? We have never been able to do a real load test, but I've got a few ideas around that too. Just need to flesh them out.
To Job's last response - some of our code is using Hibernate, but the bulk is not. The DA is aware that code will have to be rewritten for the caching. That's why he's opposing my idea of caching the table that has a read/write ratio of millions:one. No one has discussed with me how they intend to maintain data coherence. That's one of my questions for the training. Some very good information. Since I was never involved in the discussions, I have no idea if they've discussed any of this or not. Yesterday, I talked to my boss and said I wanted more involvement since it has a direct correlation to the activity against the database. He's going to look into it. He isn't the VP in charge of the project so he can't automatically make it happen. I still think the people driving the project are missing the boat by not making the application more efficient. Think we should look at every possible option we can to improve performance and scalability. But then I'm just the DBA--the person you blame when performance is bad. Makes me smile when I can prove it was the crappy code on top of the horrible design.
--
Sandy
Transzap, Inc.
--
http://www.freelists.org/webpage/oracle-l
Received on Fri Jan 22 2010 - 11:48:41 CST