Franck's Oracle blog
Blog of Franck Pachot, Consultant at dbi services
I'm a Senior Consultant, and Oracle Technology Leader at dbi services (Switzerland).
Certified DBA (OCM 11g, OCP 12c, Performance Tuning Expert, Exadata Implementation) I cover all database areas: architecture, data modeling, database design, tuning, operation, and training.
My preferred area is troubleshooting oracle and performance tuning, especially when I acheive to enable an efficient collaboration between the developers and the
As an Oracle Ace, I participate in the Oracle Community in forums, blogs, articles and presentation. You can follow my activity on this blog: and my twitter account: @FranckPachot
If you want a statement to be hard parsed on its next execution, you can flush the shared pool, but you don't want all the cursors to be hard parsed. Here is how to flush only one statement, illustrated with the case where it can be useful.
During the performance training, here is how I introduce Adaptive Cursor Sharing, here is how I show the bind variable peeking problem that is well known by everyone that was DBA at the times of 9iR2 upgrades.
This is a question that starts to be raised quite frequently. Oracle released the multitenant architecture 1.5 year ago. And now says that the previous architecture - known as non-CDB - is deprecated. What does it mean? Do we have to go to the CDB architecture even if we dont want to use/pay multi-tenant?
This month we started to give our Oracle Tuning Workshop. And with a new workshop comes new questions. We advise to give to the optimizer the most accurate statsistics that we can have. That suggests that WORKLOAD statistics are better than NOWORKLOAD ones because they gather the average number of blocks read in multiblock read, rather than using default values. But then, the question is: which time period do you choose to gather workload statistics, and with which interval duration?
In europe we have accents and non US7ASCII characters. We need special characterset. I'm not talking about Unicode here that solves all the problems. If you have a Java application, you have no problem: it's Unicode. You can store all characters in one multi-byte characterset. But for other applications, on Windows, you have 2 possible charactersets for Western Europe WE8MSWIN1252 and WE8PC850. WE8MSWIN1252 is the one that is set by default in the registry, but is it the right one?
This is a second part of the previous post about metadata link. I've shown how a sharing=metadata function becomes a sharing=none function when it is changed in the pdb - i.e when not having the same DDL, not having a different signature.
Here is another experimentation doing the opposite: change the function in root and see what happens in the pdb. Again playing with internals in order to understand the 'upgrade by unplug-plug' feature available in 12c multi-tenant (and single-tenant).
Of course, playing with internals is fun ;)
But it's not only that. We will implement CDB databases (multi-tenant, and single-tenant as I think non-CDB will be deprecated in future versions). We will operate pluggable databases (clone, upgrade, downgrade,...), encounter issues, etc. Before doing that on critical environments, I need to be confident about the multitenant implementation. I'ts new, it's undocumented, and - at the first sight - it looks like a 'hack' in the way the dictionary is working for years.
calendar today tomorrow message ----------------- ------ --------- -------------- Arabic Hijrah 1436 1436 English Hijrah 1436 1436 Gregorian 2014 2015 Happy New Year Japanese Imperial 0026 0027 Happy New Year Persian 1393 1393 ROC Official 0103 0104 Happy New Year Thai Buddha 2557 2558 Happy New Year