Last day of the event and I think I finally know where Moscone( North, South, West), the Marriott, and Yerba Buena Center for the Arts are located for the different sessions.
I will post yesterday’s events as time permits.
Wednesday was a good day at Oracle Open World; some hands-on, good sessions on performance tuning, and one-on-one with Joel Kallman, (Director, Software Development for Oracle Application Express).
The hands-on session was an Introduction to Oracle Data Integrator Enterprise Edition. At a high level ODE is a tool used for extract, transform, and load (ETL) of data, any type of data. Flat files, comma delimited, XML, other database platforms. I even noticed my old friend Informix on the list of sources. The lab was taking XML data sources, joining the data, and inserting it into an Oracle database. Like most of the Oracle tools these days, OED is Java based so it will run on Linux, Apple and Winblows. These one hour hands-on sessions are a good test of the product for me. I figure if I can get through the lesson then the software must be fairly simple to operate out of the box. Granted we only scratch the surface of the products. This is one of those products that we need, I know we could use it, but we don’t know we need it. It’s a cycle… More information on OED can be found here:
I had three sessions related to tuning and all were very informative.
1. Performance Tuning with DB Time
2. SQL Tuning Roundtable with Oracle Gurus
3. Tuning the Oracle Grid
We’ve been tuning with DB Time and did not even realize it. We used it fix performance issue with ETAS and CARMA. It’s built into Grid Control. I remember studying for the 10g OCP exam that DB Time was a key metric to pay attention to because it was new in 10g and the fundamental measurement of Oracle performance. This is one that requires the slides to help explain. But, basically, when you take all the idle wait time out of a histogram, it compresses everything and you end up with the graph on the performance page of Grid for a database. I need to look at the real-time SQL monitoring of Grid. They showed a number of charts with it and from what I understand; the real-time SQL monitoring can be used to “watch” SQL run. For example, if a long query is running and it is loading data, a person should be able to see where it is in the load by using real-time SQL monitoring and then be able to make some estimates of when it will be finished… This was mentioned is both sessions 1 and 2.
One question for the round table was about best practices for implementing profile changes into a production environment and passing it through the change control processes, (ITIL:-) no, SDL. The response was comparing an update statistics in production without having to test first. It should not be a problem to implement a profile directly into production. In 11g multiple profiles can be staged and switched out if needed. New in 11g is SQL Plan Management (SPM) which uses the baseline logic to implement the plan management process.. Here is a URL to some additional information for performance tuning of 11g.
I went to the demo grounds in one of the exhibit halls looking specifically to find someone that could answer my question on best practices for architecture of Apex. There are three different ways to get Apex up and running:
- Using Apache with mod_plsql fronted Apex
- Have it all inside the database ( came out in 11g )
- Use Oracle application server ( OC4J or the Web Logic version)
The demo grounds are miles of stand up posts with a divider which will allow up to 4 stations. There must have been a few hundred of these stations. I got my map out to try and narrow down where the Apex hands-on area was and found that it was in Moscone North, Database area. Great, found it… But when I got there I see a sea of these demo things just under the database section. And, you have to walk around each one to see what area they are focusing on. Oh, and two rows so you can get dizzy very easy walking around these things. But I found one on Apex, don’t remember what it was but I knew I was close. Kept walking around and found one that may be of some help. So I asked the person at the station my question and she took me to the Oracle documentation that outlines the three methodologies to get Apex up and running. Great, but I was looking for best practices.. She mentioned that I talk to the director of Apex development and he just happened to walk by so she grabbed him.
His name was Joel Kallman and was very helpful. With the new version of Apex, 4.0 they introduced an Apex listener which eliminates the limitations of mod_plsql and Apache 1.3. Starting in 10g, Oracle released their HTTP server which is now based on Apache 2.0. This is good because we are not tied to the Apache 1.3 platform with the mod_plsq. Joel told me that Apex.oracle.com was based on the Apache 2.0 HTTP server which will allow for connection pooling using the Apex listener. Here is a URL to his blog one this. http://joelkallman.blogspot.com/2008/01/oracle-http-server-apache-20-and.html
Still need to do some reading on this but it sound promising to just use Apache 2.0, ( aka Oracle HTTP server) as the architecture.
Comments