ALL POSTS

MERRY CHRISTMAS from Pythian’s EBS Teams, Oracle Style ;-)

Pythian’s friendly Oracle e-Business Suite DBAs would like to congratulate all of our friends, clients, Pythian colleagues, prospects, new team members whom we haven’t yet met (join us!), all database geeks, and blog readers for successfully completing 2012! Have a very peaceful Christmas and the happiest year ever in 2013! May all of your pagers be quiet, and your year-end maintenance be incident-free! ;) Here comes our Christmas card to you :)

Revealing the Hints Behind the Automatic SQL Tuning Advisor

A client recently supplied a list of 50+ SQL IDs that should receive SQL profiles, and I’ve been working with Gwen Shapira to review the list. Further discussion showed that this list had come from the Automatic SQL Tuning feature, installed by default in Oracle 11g. The report includes a list of recommended SQL profiles ordered by “Maximum Benefit”, and in our case it included several hundred statements. The expected workflow, as far as I can gather, is to see the recommendations, look at the before- and after- execution plans, and accept the recommendations. Before blindly accepting recommendations, though, I like to see what exact changes are being proposed. They aren’t listed anywhere in the report, and require some extra work to uncover. The first step is to get the automatic SQL tuning advisor report. (

Changing SID on a RAC Environment

Just a short note documenting a procedure that isn’t done frequently, and I didn’t see described anywhere else. Somewhere last month a customer asked for my help in changing the SIDs on one of his RAC databases to match the new corporate standard. The database name matched the standard so we could leave that alone, but the SID needed to be changed. Here’s what we did to change it

Temporary Tablespaces: to BIGFILE or Not to BIGFILE

The environment was an Exadata environment undergoing pre-production stress testing. We used Real Application Testing to take a highly-concurrent OLTP workload, and replayed the workload with the synchronization parameter set to FALSE, effectively increasing concurrency beyond the original test system. AWR showed a large volume of buffer busy activity.

Page 48 of 256« First...102030...4647484950...607080...Last »