Emre Baransel - Oracle Blog

I'm reading the "Expert Oracle Database Architecture" book of Thomas Kyte, which is a must read for everyone who is serious on being an Oracle DBA.

Jonathan Lewis already said: "Frankly, if every DBA and developer in the world were made to work carefully through Tom Kyte’s book, I’d probably have to start offering consultancy services to SQL Server users because the number of clients needing Oracle consultancy would drop dramatically." in the foreword of the book. Still, i wanted to share some paragraphs from the "Data Files" part of the book. Following are taken out of only three pages of the book. I'm sharing this to show how the book is full of information and encourage you to immediately buy one and read.

Segments
==============================
You will find many CREATE statements that create multisegment objects. The confusion lies in the fact that a single CREATE statement may ultimately create objects that consist of zero, one, or more segments! For example, CREATE TABLE T ( x int primary key, y clob ) will create four segments: one for the TABLE T, one for the index that will be created in support of the primary key, and two for the CLOB (one segment for the CLOB is the LOB index and the other segment is the LOB data itself). On the other hand, CREATE TABLE T ( x int, y date ) cluster MY_CLUSTER will create zero segments (the cluster is the segment in this case).

Extents
==============================
Extents vary in size from one Oracle data block to 2GB. 11g Release 2 has introduced the concept of a “deferred” segment—a segment that will not immediately allocate an extent-, so in that release and going forward, a segment might defer allocating its initial extent until data is inserted into it.

Blocks
==============================
Here’s a little-known fact: the default block size for a database does not have to be a power of two. Powers of two are just a commonly used convention. You can, in fact, create a database with a 5KB, 7KB, or nKB block size, where n is between 2KB and 32KB.

Most blocks, regardless of their size, have the same general format, which looks something like:

 ------------------
|Header           | -> type of block (table block, index block, and so on), 
|                  |   transaction information when relevant regarding active and 
|                 |   past transactions on the block; and the address (location) 
|                 |   of the block on the disk.
|------------------|   The next two block components are found on the HEAP-                 |                  |   organized tables.      
|Table Directory   | -> The table directory, if present, contains information  
|                 |   about the tables that store rows in this block
|------------------|
|Row Directory     | -> The row directory contains information describing the 
|                 |   rows that are to be found on the block.
|------------------|
|                  | -> The remaining 2 pieces of the block are straightforward: 
|Free Space        |   there may be free space on a block, and
|                  |   then there will generally be used space that is 
|                  |   currently storing data.
|------------------|
|                  |
|Data              |
|                  |
|------------------|
|Tail              |
 ------------------


Exceptions to this format include LOB segment blocks and hybrid columnar compressed blocks in Exadata storage, for example, but the vast majority of blocks in your database will resemble the format in Figure