Oracle components and architecture

1. User-related processes

User-related processes include user processes and server processes.

A user process can be connected to only one server process

However, in some Oracle configurations, multiple user processes can share the same server process.

The PAG is also created. Each server process has a PGA

2. Logical Memory Structure, collectively referred to as an oracle instance

An oracle instance is composed of an oracle main memory structure, called SGA (system global area), and server background processes.

When a user accesses data in the database, the server process is responsible for communicating with the SGA memory block.

An oracle instance includes three components:

1) required components:

Shared Pool: caches SQL statements recently used by database users;

Database cache: information recently used by database users;

Redo log: stores the transaction information restored to the target.

Optional internal components:

Java pool: cache the recently used java targets when the java Virtual Machine is used.

Large pool: Data Change and big operations, such as recovery management, backup and recovery activities, and Shared Server Components

Streams pool: cache queue request information related to the database. When the oracle Advanced queue option is enabled

SGA uses algorithms to manage memory at least.

2) background process

There are many background processes, each of which executes a special task to help manage oracle instances. Five background processes are mandatory, including:

1. The system monitors processes, implements disaster recovery of processes, merges free space of databases, and manages memory space for sorting;

2. process monitoring process to clear database users that fail;

3. The database write process writes database blocks from the database cache area of SGA to database files on the disk;

4. The log writing process writes the recovery information from the REDO log cache in SGA to the online redo log file on the disk;

5. The Checkpoint Process updates the database file immediately after the checkpoint event.

3. The physical file architecture is collectively referred to as a database

A database consists of physical files that reside on the disk drive of the host server. These files can be called data files, control files, and log files. Other files are also related to databases,

But it does not belong to a database, such as a password file and database file, a configuration file (pfile, spfile), or an archived redo log file ).

I. Control Files

Is a key part of the database. They store important information which is not provided elsewhere. Including:

• Database Name;
• Name, location, and size of the data file and redo log file;
• The information used to restore the database when the disk fails or the user is wrong.

When the database is created, the control file www.bkjia.com has been created according to the location described in the control_files parameter of the parameter file. The loss of the control file will negatively affect the ability to restore the database.

Most database products provide multiple database files and store them in multiple locations concurrently. When necessary, CKPT dynamically updates the control file of each database and maintains the consistency of the control file content.

You can use the following example to view detailed information such as the location and name of the control file:

1. SQL>Select Name FromV $ controlfile;
3. NAME
4. --------------------------------------------------------------------------------
5. /U01/app/oracle/oradata/nihao/control01.ctl
6. /U01/app/oracle/flash_recovery_area/nihao/control02.ctl
8. SQL>

1 2 Next Page