Posts tagged ‘12c’

August 12, 2013

Oracle 12c – Multi-Tenant Databases — #STEM, #Oracle

by C. Michael Eliasz-Solomon


Oracle 12c

Oracle’s newest database (version 12c) has many new features, the discussion of which are too big for a single blog article (or even a series of blogs). The substantial high-level bulleted list of new features is in the 12c New Features Manual . But the concepts and low level SQL language details show a much larger change then you might perceive.

Multitenant Database

The new paradigm shift, Multitenant Databases, will stop DBAs pretty quick, particularly in Windows where the installer creates a Container DB. Previous to 12c, all databases were Non-Container DBs. With 12c you can create a Non-Container DB or a Container DB. The Container DB can have zero, one, or more Pluggable DBs within. A Non-Container DB can never have a Pluggable DB. So that becomes an upfront database creation decision.

You can and you should read the Oracle Intro to Multitenant Databases .

I first relaized the Oracle Installer had created a container database for me when I went through the normal process of creating a database user using the same old techniques I always did and received an Oracle Error: ORA-65096. WHAM, I slammed right into the new paradigm without me even knowing it existed. The error description and the necessary action introduced to another part of the Multitenant Database paradigm: Comon User vs. Local User. That quickly led to Containers. Of course, with any new features, comes an array of new database schema tables like, v$pdbs for example. You will also probably use a new Sql*Plus command a LOT: SHOW CON_NAME to know what container (root or pluggable database) you are connected to. Some DBA commands must be done in the root container (CDB$ROOT). Your pluggable databases (in Windows) will be by default: PDB$SEED and PDBORCL. Every container database has precisely one seed pluggable database from which all pluggable databases are created from.

This paradigm shift will be  seriously disorienting feeling to long time DBAs, especially if were not aware this was coming. Fortunately, there are many DBA bloggers out there sharing their 12c experiences. They were a help for me to gather the necessary jargon. But it was not until I discovered that Oracle had created a tutorial on Multitenant Databases and I had spent an hour or two playing with the tutorial on my newly created sandbox database (on Windows) which was by default a Container DB. This tutorial is an excellent way to jump start your understanding of the new paradigm.

By the way, I think either the default should be a NON-CONTAINER DB (so you are backwards compatible) or the Oracle Intsaller needs to make it clear that a CONTAINER DB will require new DBA processes (i.e. a learning curve) and give you an OVERT option to create a NON-CONTAINER DB for backwards compatibility.


Read the Oracle Introduction to Multitenant Databases to understand the concepts. Then immediately work your way through the tutorial in a test database that is a Container DB. Ultimately, I think Container DBs are the way to go. I think this is what you want to do to implement a CLOUD or in a Virtualized Environment.

August 8, 2013

Wordless Wednesday — #Oracle, #12c, #STEM, #GEEK

by C. Michael Eliasz-Solomon

Oracle 12c installed . Getting my #GEEK on this week.



August 7, 2013

Oracle v 12c … vs. Greenplum MPP — #STEM, #Oracle, #Greenplum, #BigData

by C. Michael Eliasz-Solomon

Studying up on Oracle v. 12c. As usual, there are many new features to recommend migrating or deploying to the new version of Oracle. Last blog, I talked about just a few: ILM, ADO, HEAT_MAP and how these buzz-worthy acronyms were related to compression inside the database. Before,  I get into today’s topic, I wanted to talk about  a bit more about the Automatic Data Optimization (ADO).

I failed to make clear yesterday, that the ADO, automatically relocates your cold data or compresses your data as it ages through its Lifecycle automatically. That is the magic. You define the policies and the database will relocate or compress a segment(s) or a row(s) to save space or to clear space on more expensive hard disk, by relocating to slower/less accessible disk storage media. Pretty nifty idea.

By the way, you may be wondering … 8i, 9i, 10g, 11g, 12c what is the pattern or meaning of these major release versions from Oracle.?  Well, “8i / 9i” were from the era, when Oracle was the “Internet” database (you know  like iPhone, or i-<Anything>). Then “10g / 11g” were to be the “Grid”. Grid never really achieved buzz-worthy status. Now we have “12c”. It should not surprise you that we are now in the “Cloud” era. So Oracle’s letters are for: Internet, Grid, and Cloud . Now you know.

That Cloud and yesterday’s ADO  will figure in today’s blog too. You see, I was recently asked about Greenplum. Could I use it? As is my wont, I took a step back and studied the question. Here is my answer.



MPP platform

MPP – RAC(aka Oracle parallel server)

Full SQL (Postgres)

Full SQL (Oracle, ANSI)


Compression since 11g, ADO/ILM 12c

B-Tree / BitMap Indexes

B-Tree / BitMap Indexes


JDBC/ODBC/OLE/Pro*C (etc.)

Parallel Query Optimizer

Parallel Query Optimizer

External Tables

External Tables

GreenPlum HD (HDFS)

External Tables using an HDFS

I believe that as an Oracle expert (28+ years from v2.0-11g inclusive), that I could effectively use Greenplum on a project. If you look at the above chart, I think you will see what I am about to explain.

Green is an MPP platform. Very nice acrhitecture. Oracle can sit on top of any architecture (MPP, SMP, or any cluster or Highly Available or Fault-Tolerant Failover set of servers) you can setup.

Both use FULL SQL.  That means ANSI compliance and with enhancements (POSTGRES for Greenplum and ORACLE, uh, for Oracle).

B-Tree and Bit Map Indexes for both — yawn old hat. Parallel Query Optimizer – been there, seen that for a while.

Greenplum has JDBC/ODBC/OLE interfaces. Oracle has those too, plus a full complement of Pro*C (or many other languages) embedded pre-compiled 3GL languages. Oracle is well supported by Shell Scripts like PHP or PERL that have their interfaces to Oracle. Slight advantage to Oracle. But the point is, Oracle professionals have done this for more than a decade.

External Tables too are a feature in both databases.  GreenPlum HD uses the External Table to provide HDFS access in GreenPlum via SQL or other in-database features. Now I had not previously thought to try and use HDFS with Oracle. But the External Table is precisely the feature I would use. Can it be done? A look at Oracle’s documentation answers that:


CREATE TABLE [schema.]table
   ( column datatype, ... )
                        DEFAULT DIRECTORY directory
                        ACCESS PARAMETERS
                            ( PREPROCESSOR HDFS_BIN_PATH:hdfs_stream access_parameters
                        ... )
                        LOCATION (file1,file2...)

So I recommend that companies fell free to utilize Oracle consultants on Greenplum databases. There is an awful lot of overlap that the Oracle specialist can leverage from his/her background and transfer to the Greenplum database.

Of course, for companies without Greenplum, it looks like you can use many of the same features already in Oracle including using HDFS filesystems with External Tables.

So get to that BigData, your friendly Oracle expert can help you.

August 6, 2013

This Jester Has Been Consulting the Oracle — #STEM, #ILM, #ADO, #Oracle

by C. Michael Eliasz-Solomon

Dateline 06 Aug 2013 — 

If you are the same age as Stanczyk, then when you see the acronym, ILM, you probably think of George Lucas’ Industrial Light & Magic. But this article is about the Oracle of Larry Ellison. Oracle has released its latest version of its database, 12c, on June 25th, 2013.

So the ILM, of this blog is about Information Lifecycle Management. I thought you might need a buzz-word upgrade too — hence this blog. In the latest 12c, Oracle is advancing its ILM paradigm to make Automatic Data Optimization (ADO) a differentiator in Data / Databases. You see data storage is eating the planet or at least the IT budgets of many large companies. That Big Data has to live somewhere and the costs to house that data is very significant. Ergo, Oracle is giving you a way to  Tier your data storage  amongst differing costs media (hi to low) and using differing levels of compression, depending on your data’s lifecycle. Hence ILM.


Source:  Oracle Documentation

The idea is that data ages from very active, to less active, to historical, to archival. You ideally would want to place the most active data on the fastest, most reliable, … most costly hardware. Likewise, as the data ages, it would be preferable to place on less costly storage devices or in a more compressed state to save space and costs. How can you do that effectively and without a large staff of IT professionals?  This is where the ADO comes in.

Using your familiar create table or alter table commands you can add an ILM policy to compress or relocate your data. Oracle provides segment level or  even row level granularity for these policies. How do you know what data is active vs inactive? Oracle has implemented a HEAT_MAP facility for detecting data usage. HEAT_MAP is a db parameter. Set it on in your init.ora file or via an alter session command in sql*plus (to do it on a session basis instead of database wide.


You can check on things via:


There is even a PL_SQL stored package:  DBMS_HEAT_MAP.

So this is a quick update on ILM, ADO, and HEAT_MAP in Oracle 12c database. Go to the Oracle yourself and see what you can get on this new technology.

Tags: , , , ,
Stem Cellular

Science and technology improving health outcomes

Steve Szabados Genealogy

Genealogy Columnist for the Polish American Journal and Author

From Shepherds and Shoemakers

Sharing musings, insights, resources and strategies as I discover my family history.

Find Lost Russian & Ukrainian Family

Uncovering the secrets of finding family and records in the former USSR

Historia pamięcią pisana

Historia wsi Święty Józef na Pokuciu

The Dystopian Nation of City-State

A cruel, futuristic vision created by science fiction authors James Courtney and Kaisy Wilkerson-Mills. ©2013-2016. All Rights Reserved. All writings available through Amazon.

What's Past is Prologue

Adventures in genealogy

The Family Kalamazoo

A genealogical site devoted to the history of the DeKorn and Zuidweg families of Kalamazoo and the Mulder family of Caledonia

Interesting Literature

A Library of Literary Interestingness

Globe Drifting

Global issues, travel, photography & fashion. Drifting across the globe; the world is my oyster, my oyster through a lens.

Oracle Scratchpad

Just another Oracle weblog

toledo's kuschwantz

a Polish kid and her family from Toledo

Author Michael Charton

Home of Author Michael Charton


A Journey through History in Search of a Vanished Family

The Blog

The latest news on and the WordPress community.

%d bloggers like this: