12 Tuning the System Global Area

This chapter describes how to tune the System Global Area (SGA). If you are using automatic memory management to manage the database memory on your system, then there is no need to tune the SGA as described in this chapter.

This chapter contains the following topics:

12.1 Using Automatic Shared Memory Management

Automatic shared memory management simplifies the configuration of the SGA by automatically distributing the memory in the SGA for the following memory pools:

  • Database buffer cache (default pool)

  • Shared pool

  • Large pool

  • Java pool

  • Streams pool

Automatic shared memory management is controlled by the SGA_TARGET parameter. Changes in the value of the SGA_TARGET parameter automatically resize these memory pools. If these memory pools are set to nonzero values, then automatic shared memory management uses these values as minimum levels. Oracle recommends that you set the minimum values based on the minimum amount of memory an application component requires to function properly.

The following memory caches are manually-sized components and are not controlled by automatic shared memory management:

  • Redo log buffer

    The redo log buffer is sized using the LOG_BUFFER initialization parameter, as described in "Configuring the Redo Log Buffer".

  • Other buffer caches (such as KEEP, RECYCLE, and other nondefault block size)

    The KEEP pool is sized using the DB_KEEP_CACHE_SIZE initialization parameter, as described in "Configuring the KEEP Pool".

    The RECYCLE pool is sized using the DB_RECYCLE_CACHE_SIZE initialization parameter, as described in "Configuring the RECYCLE Pool".

  • Fixed SGA and other internal allocations

    Fixed SGA and other internal allocations are sized using the DB_nK_CACHE_SIZE initialization parameter.

The memory allocated to these memory caches is deducted from the value of the SGA_TARGET parameter when automatic shared memory management computes the values of the automatically-tuned memory pools.

The following sections describe how to access and set the value of the SGA_TARGET parameter:

See Also:

12.1.1 User Interfaces for Setting the SGA_TARGET Parameter

This section describes the user interfaces for setting the value of the SGA_TARGET parameter.

This section contains the following topics:

12.1.1.1 Setting the SGA_TARGET Parameter in Oracle Enterprise Manager Cloud Control

You can change the value of the SGA_TARGET parameter in Oracle Enterprise Manager Cloud Control (Cloud Control) by accessing the SGA Size Advisor from the Memory Parameters SGA page.

12.1.1.2 Setting the SGA_TARGET Parameter in the Command-Line Interface

You can change the value of the SGA_TARGET parameter in the command-line interface by querying the V$SGA_TARGET_ADVICE view and using the ALTER SYSTEM command.

12.1.2 Setting the SGA_TARGET Parameter

This section describes how to enable and disable automatic shared memory management by setting the value of the SGA_TARGET parameter.

This section contains the following topics:

12.1.2.1 Enabling Automatic Shared Memory Management

To enable automatic shared memory management, set the following initialization parameters:

  • STATISTICS_LEVEL to TYPICAL or ALL

  • SGA_TARGET to a nonzero value

    The SGA_TARGET parameter can be set to a value that is less than or equal to the value of the SGA_MAX_SIZE initialization parameter. Set the value of the SGA_TARGET parameter to the amount of memory that you intend to dedicate to the SGA.

12.1.2.2 Disabling Automatic Shared Memory Management

To disable automatic shared memory management, set the value of the SGA_TARGET parameter dynamically to 0 at instance startup.

This disables automatic shared memory management and the current auto-tuned sizes will be used for each memory pool. If necessary, you can manually resize each memory pool, as described in "Sizing the SGA Components Manually".

12.2 Sizing the SGA Components Manually

If the system is not using automatic memory management or automatic shared memory management, then you must manually configure the sizes of the following SGA components:

  • Database buffer cache

    The database buffer cache is sized using the DB_CACHE_SIZE initialization parameter, as described in "Configuring the Database Buffer Cache".

  • Shared pool

    The shared pool is sized using the SHARED_POOL_SIZE initialization parameter, as described in "Configuring the Shared Pool".

  • Large pool

    The large pool is sized using the LARGE_POOL_SIZE initialization parameter, as described in "Configuring the Large Pool".

  • Java pool

    The Java pool is sized using the JAVA_POOL_SIZE initialization parameter.

  • Streams pool

    The Streams pool is sized using the STREAMS_POOL_SIZE initialization parameter.

  • IM column store

    The IM column store is sized using the INMEMORY_SIZE initialization parameter.

The values for these parameters are also dynamically configurable using the ALTER SYSTEM statement.

Before configuring the sizes of these SGA components, take the following considerations into account:

See Also:

12.2.1 SGA Sizing Unit

Memory for the buffer cache, shared pool, large pool, and Java pool is allocated in units of granules. If the SGA size is less than 1 GB, then the granule size is 4MB. If the SGA size is greater than 1 GB, the granule size changes to 16MB. The granule size is calculated and fixed when the database instance starts up. The size does not change during the lifetime of the instance.

To view the granule size that is currently being used for the SGA, use the V$SGA_DYNAMIC_COMPONENTS view. The same granule size is used for all dynamic components in the SGA.

12.2.2 Maximum Size of the SGA

The maximum amount of memory usable by the database instance is determined at instance startup by the value of the SGA_MAX_SIZE initialization parameter. You can expand the total SGA size to a value equal to the SGA_MAX_SIZE parameter. The value of the SGA_MAX_SIZE parameter defaults to the aggregate setting of all the SGA components.

If the value of the SGA_MAX_SIZE parameter is not set, then decrease the size of one cache and reallocate that memory to another cache if necessary. Alternatively, you can set the value of the SGA_MAX_SIZE parameter to be larger than the sum of all of the SGA components, such as the buffer cache and the shared pool. Doing so enables you to dynamically increase a cache size without having to decrease the size of another cache.

Note:

The value of the SGA_MAX_SIZE parameter cannot be dynamically resized.

12.2.3 Application Considerations

When configuring memory, size the memory caches appropriately based on the application's needs. Conversely, tuning the application's use of the memory caches can greatly reduce resource requirements. Efficient use of the memory caches also reduces the load on related resources, such as latches, CPU, and the I/O system.

For optimal performance, consider the following:

  • Design the cache to use the operating system and database resources in the most efficient manner.

  • Allocate memory to Oracle Database memory structures to best reflect the needs of the application.

  • If changes or additions are made to an existing application, resize Oracle Database memory structures to meet the needs of the modified application.

  • If the application uses Java, investigate whether the default configuration for the Java pool needs to be modified.

See Also:

Oracle Database Java Developer's Guide for information about Java memory usage

12.2.4 Operating System Memory Use

For most operating systems, it is important to consider the following when configuring memory:

See Also:

Your operating system hardware and software documentation, and the Oracle documentation specific to your operating system, for more information on tuning operating system memory usage

12.2.4.1 Reduce Paging

Paging occurs when an operating system transfers memory-resident pages to disk solely to load new pages into memory. Many operating systems page to accommodate large amounts of information that do not fit into real memory. On most operating systems, paging reduces performance.

To determine whether significant paging is occurring on the host system, use operating system utilities to examine the operating system. If significant paging is occurring, then the total system memory may not be large enough to hold the memory caches for which memory is allocated. Consider either increasing the total memory on the system, or decreasing the amount of memory allocated.

12.2.4.2 Fit the SGA into Main Memory

Because the purpose of the SGA is to store data in memory for fast access, the SGA should reside in the main memory. If pages of the SGA are swapped to disk, then the data is no longer quickly accessible. On most operating systems, the disadvantage of paging significantly outweighs the advantage of a large SGA.

This section contains the following topics:

12.2.4.2.1 Viewing SGA Memory Allocation

To view how much memory is allocated to the SGA and each of its internal structures, use the SHOW SGA statement in SQL*Plus as shown in the following example:

SQL> SHOW SGA

The output of this statement might look like the following:

Total System Global Area  840205000 bytes
Fixed Size                   279240 bytes
Variable Size             520093696 bytes
Database Buffers          318767104 bytes
Redo Buffers                1064960 bytes
12.2.4.2.2 Locking the SGA into Physical Memory

To prevent the SGA from being paged out, consider locking the SGA into physical memory by enabling the LOCK_SGA parameter. The database does not use the MEMORY_TARGET and MEMORY_MAX_TARGET parameters when the LOCK_SGA parameter is enabled.

12.2.4.3 Allow Adequate Memory to Individual Users

When sizing the SGA, ensure that you allow enough memory for the individual server processes and any other programs running on the system.

12.2.5 Iteration During Configuration

Configuring memory allocation involves distributing available memory to Oracle Database memory structures, depending on the needs of the application. The distribution of memory to Oracle Database structures can affect the amount of physical I/O necessary for Oracle Database to operate properly. Having a proper initial memory configuration provides an indication of whether the I/O system is effectively configured.

After the initial pass through the memory configuration process, it may be necessary to repeat the steps of memory allocation. Subsequent passes enable you to make adjustments to earlier steps, based on changes in subsequent steps. For example, decreasing the size of the buffer cache enables you to increase the size of another memory structure, such as the shared pool.

12.3 Monitoring Shared Memory Management

Table 12-1 lists the views that provide information about SGA resize operations.

Table 12-1 Shared Memory Management Views

View Description

V$SGA_CURRENT_RESIZE_OPS

Displays information about SGA resize operations that are currently in progress.

V$SGA_RESIZE_OPS

Displays information about the last 800 completed SGA resize operations. This does not include operations that are currently in progress.

V$SGA_DYNAMIC_COMPONENTS

Displays information about the dynamic components in the SGA. This view summarizes information of all completed SGA resize operations that occurred after instance startup.

V$SGA_DYNAMIC_FREE_MEMORY

Displays information about the amount of SGA memory available for future dynamic SGA resize operations.

See Also:

Oracle Database Reference for information about these views

12.4 Improving Query Performance with the In-Memory Column Store

The In-Memory Column Store (IM column store) is an optional portion of the system global area (SGA) that stores copies of tables, partitions, and other database objects in columnar format, and this columnar data is optimized for rapid scans. As the IM column store stores database objects in memory, Oracle Database can perform scans, queries, joins, and aggregates on that data much faster as compared to performing these operations on a data that is stored on a disk.

Note:

  • The IM column store and database buffer cache store the same data, but in different formats. The IM column store does not replace the row-based storage in the database buffer cache, but supplements it for achieving better query performance.

  • The IM column store is available starting with Oracle Database 12c Release 1 (12.1.0.2).

See Also:

Oracle Database In-Memory Guide for more information about the IM column store

12.5 Enabling High Performance Data Streaming with the Memoptimized Rowstore

The Memoptimized Rowstore enables high performance data streaming for applications, such as Internet of Things (IoT).

This section contains the following topics:

12.5.1 About the Memoptimized Rowstore

The Memoptimized Rowstore enables high performance data streaming for applications, such as Internet of Things (IoT) applications, which typically stream small amounts of data in single-row inserts from a large number of clients simultaneously and also query data for clients at a very high frequency.

The Memoptimized Rowstore provides the following functionality:

  • Fast ingest

    Fast ingest optimizes the processing of high-frequency, single-row data inserts into a database. Fast ingest uses the large pool for buffering the inserts before writing them to disk, so as to improve data insert performance.

  • Fast lookup

    Fast lookup enables fast retrieval of data from a database for high-frequency queries. Fast lookup uses a separate memory area in the SGA called the memoptimize pool for buffering the data queried from tables, so as to improve query performance.

    Note:

    For using fast lookup, you must allocate appropriate memory size to the memoptimize pool using the MEMOPTIMIZE_POOL_SIZE initialization parameter.

12.5.2 Using Fast Ingest

Fast ingest optimizes the processing of high-frequency, single-row data inserts into database from applications, such as Internet of Things (IoT) applications.

Fast ingest uses the MEMOPTIMIZE_WRITE hint to insert data into tables specified as MEMOPTIMIZE FOR WRITE. The database temporarily buffers these inserts in the large pool and automatically commits the changes at the time of writing these buffered inserts to disk. The changes cannot be rolled back.

The inserts using fast ingest are also known as deferred inserts, because they are initially buffered in the large pool and later written to disk asynchronously by background processes.

Steps for using fast ingest for inserting data into a table

The following are the steps for using fast ingest for inserting data into a table:

  1. Enable a table for fast ingest

    This task is achieved by specifying the MEMOPTIMIZE FOR WRITE clause in the CREATE TABLE or ALTER TABLE statement.

    See "Enabling a Table for Fast Ingest" for more information.

  2. Specify a hint for using fast ingest for an insert operation

    This task is achieved by specifying the MEMOPTIMIZE_WRITE hint in the INSERT statement.

    See "Specifying a Hint for Using Fast Ingest for Data Inserts" for more information.

Limitations for using fast ingest

Tables with the following characteristics cannot use fast ingest:

  • Tables with:

    • disk compression
    • in-memory compression
    • column default vales
    • encryption
    • functional indexes
    • domain indexes
    • bitmap indexes
    • bitmap join indexes
    • ref types
    • varray types
    • OID$ types
    • sub-partition stats
    • unused columns
    • virtual columns
    • LOBs
    • triggers
    • binary columns
    • foreign keys
    • row archival
    • invisible columns
  • Temporary tables

  • Nested tables

  • Index organized tables

  • External tables

  • Materialized views with on-demand refresh

The following are some additional considerations for fast ingest:

  • Because fast ingest buffers data in the large pool, there is a possibility of data loss in the event of a system failure. To avoid data loss, a client must keep a local copy of the data after performing inserts, so that it can replay the inserts in the event of a system failure before the data is written to disk. A client can use the DBMS_MEMOPTIMIZE package subprograms to track the durability of the inserts. After inserts are written to disk, a client can destroy its local copy of the inserted data.

  • Queries do not read data from the large pool, hence data inserted using fast ingest cannot be queried until it is written to disk.

  • Parent-child transactions must be synchronized to avoid errors. For example, foreign key inserts and updates of rows inserted into the large pool can return errors, if the parent data is not yet written to disk.

  • Index operations are supported by fast ingest similar to the regular inserts. However, for fast ingest, database performs index operations while writing data to disk, and not while writing data into the large pool.

Note:

A table can be configured for using both fast ingest and fast lookup.

12.5.2.1 Enabling a Table for Fast Ingest

You can enable a table for fast ingest by specifying the MEMOPTIMIZE FOR WRITE clause in the CREATE TABLE or ALTER TABLE statement.

To enable a table for fast ingest:

  1. In SQL*Plus, log in to the database as a user with ALTER TABLE privileges.

  2. Run the CREATE TABLE or ALTER TABLE statement with the MEMOPTIMIZE FOR WRITE clause.

    The following example creates a new table test_fast_ingest and enables it for fast ingest:

    CREATE TABLE test_fast_ingest (
         id        NUMBER(5) PRIMARY KEY,
         test_col  VARCHAR2(15))
      MEMOPTIMIZE FOR WRITE;

    The following example enables the existing table hr.employees for fast ingest:

    ALTER TABLE hr.employees MEMOPTIMIZE FOR WRITE;
12.5.2.2 Specifying a Hint for Using Fast Ingest for Data Inserts

You can use fast ingest for data inserts by specifying the MEMOPTIMIZE_WRITE hint in INSERT statements.

Prerequisites

This task assumes that a table is already enabled for fast ingest.

To use fast ingest for data inserts:

  1. In SQL*Plus, log in to the database as a user with the privileges to insert data into tables.

  2. Run the INSERT statement with the MEMOPTIMIZE_WRITE hint for a table that is already enabled for fast ingest.

    For example:

    INSERT  /*+ MEMOPTIMIZE_WRITE */ INTO test_fast_ingest VALUES (1,'test');
12.5.2.3 Disabling a Table for Fast Ingest

You can disable a table for fast ingest by specifying the NO MEMOPTIMIZE FOR WRITE clause in the ALTER TABLE statement.

To disable a table for fast ingest:

  1. In SQL*Plus, log in to the database as a user with the ALTER TABLE privileges.

  2. Run the ALTER TABLE statement with the NO MEMOPTIMIZE FOR WRITE clause.

    The following example disables the table hr.employees for fast ingest:

    ALTER TABLE hr.employees NO MEMOPTIMIZE FOR WRITE;
12.5.2.4 Managing Fast Ingest Data in the Large Pool

You can view the fast ingest data in the large pool using the V$MEMOPTIMIZE_WRITE_AREA view. You can also view and control the fast ingest data in the large pool using the subprograms of the packages DBMS_MEMOPTIMIZE and DBMS_MEMOPTIMIZE_ADMIN.

Overview of the V$MEMOPTIMIZE_WRITE_AREA view

The V$MEMOPTIMIZE_WRITE_AREA view provides the following information about the memory usage and data inserts in the large pool by fast ingest:

  • Total amount of memory allocated for fast ingest data in the large pool

  • Total amount of memory currently used by fast ingest data in the large pool

  • Total amount of memory currently free for storing fast ingest data in the large pool

  • Number of fast ingest insert operations for which data is still in the large pool and is yet to be written to disk

  • Number of clients currently using fast ingest for inserting data into the database

See Also:

Oracle Database Reference for information about the V$MEMOPTIMIZE_WRITE_AREA view

Overview of the DBMS_MEMOPTIMIZE package subprograms

You can use the following subprograms of the DBMS_MEMOPTIMIZE package to view and control the fast ingest data in the large pool:

Subprogram Description

GET_APPLY_HWM_SEQID

Returns the low high-water mark (low HWM) of sequence numbers of data records that are successfully written to disk by all the sessions.

GET_WRITE_HWM_SEQID

Returns the high-water mark (HWM) sequence number of the data record that is written to the large pool for the current session.

WRITE_END

Flushes all the fast ingest data from the large pool to disk for the current session.

See Also:

Oracle Database PL/SQL Packages and Types Reference for information about the DBMS_MEMOPTIMIZE package

Overview of the DBMS_MEMOPTIMIZE_ADMIN package subprograms

You can use the following subprograms of the DBMS_MEMOPTIMIZE_ADMIN package to control the fast ingest data in the large pool:

Subprogram Description

WRITES_FLUSH

Flushes all the fast ingest data from the large pool to disk for all the sessions.

See Also:

Oracle Database PL/SQL Packages and Types Reference for information about the DBMS_MEMOPTIMIZE_ADMIN package

12.5.3 Using Fast Lookup

Fast lookup enables fast data retrieval from database tables for applications, such as Internet of Things (IoT) applications.

Fast lookup uses a hash index that is stored in the SGA buffer area called memoptimize pool to provide fast access to blocks of tables permanently pinned in the buffer cache, thus avoiding disk I/O and improving query performance.

Steps for using fast lookup for a table

The following are the steps for using fast lookup for a table:

  1. Enable the memoptimize pool

    This task is achieved by setting the MEMOPTIMIZE_POOL_SIZE initialization parameter to a non-zero value.

    See "Enabling the Memoptimize Pool" for more information.

  2. Enable a table for fast lookup

    This task is achieved by specifying the MEMOPTIMIZE FOR READ clause in the CREATE TABLE or ALTER TABLE statement.

    See "Enabling a Table for Fast Lookup" for more information.

Limitations for using fast lookup

The following are the limitations for using fast lookup:

  • Tables enabled for fast lookup cannot be compressed.

  • Tables enabled for fast lookup must have a primary key constraint enabled.

Note:

A table can be configured for using both fast ingest and fast lookup.

See Also:

12.5.3.1 Enabling the Memoptimize Pool

You must enable the memoptimize pool before using fast lookup. The memoptimize pool resides in the SGA, and stores the data and hash index for the tables that are enabled for fast lookup.

Prerequisites

This task assumes that the COMPATIBLE initialization parameter is set to 18.0.0 or higher.

To enable the memoptimize pool:

  1. In SQL*Plus, log in to the database as a user with administrative privileges.

  2. Set the MEMOPTIMIZE_POOL_SIZE initialization parameter to a non-zero value. The minimum setting is 100 MB. When you set this initialization parameter in a server parameter file (SPFILE) using the ALTER SYSTEM statement, you must specify SCOPE=SPFILE.

    For example, the following statement sets the memoptimize pool size to 10 GB:

    ALTER SYSTEM SET MEMOPTIMIZE_POOL_SIZE = 10G SCOPE=SPFILE;
  3. Restart the database for the change to take effect.

Example: Enabling the Memoptimize Pool

Assume that the MEMOPTIMIZE_POOL_SIZE initialization parameter is initially set to 0. The following example enables the memoptimize pool by setting the MEMOPTIMIZE_POOL_SIZE to 10 GB:

SQL> SHOW PARAMETER MEMOPTIMIZE_POOL_SIZE

NAME                   TYPE         VALUE
---------------------  -----------  -----
memoptimize_pool_size  big integer  0

SQL> ALTER SYSTEM SET MEMOPTIMIZE_POOL_SIZE=10G SCOPE=SPFILE;

System altered.

SQL> SHUTDOWN IMMEDIATE
Database closed.
Database dismounted.
ORACLE instance shut down.

SQL> STARTUP
ORACLE instance started.

Total System Global Area 1.1832E+10 bytes
Fixed Size                  9010864 bytes
Variable Size            1.1799E+10 bytes
Database Buffers           16777216 bytes
Redo Buffers                7766016 bytes
Database mounted.
Database opened.

SQL> SHOW PARAMETER MEMOPTIMIZE_POOL_SIZE

NAME                   TYPE         VALUE
---------------------  -----------  -----
memoptimize_pool_size  big integer  10G

Note:

The MEMOPTIMIZE_POOL_SIZE value does count toward SGA_TARGET, but the database does not grow and shrink the memoptimize pool automatically. For example, if SGA_TARGET is 10 GB, and if MEMOPTIMIZE_POOL_SIZE is 1 GB, then a total of 9 GB is available for SGA memory other than the memoptimize pool.

See Also:

12.5.3.2 Enabling a Table for Fast Lookup

You can enable a table for fast lookup by specifying the MEMOPTIMIZE FOR READ clause in the CREATE TABLE or ALTER TABLE statement.

Prerequisites

This task assumes that the memoptimize pool is enabled.

To enable a table for fast lookup:

  1. In SQL*Plus, log in to the database as a user with ALTER TABLE privileges.

  2. Run the CREATE TABLE or ALTER TABLE statement with the MEMOPTIMIZE FOR READ clause for the table that needs to be enabled for fast lookup.

    The following example creates a new table test_fast_lookup and enables it for fast lookup:

    CREATE TABLE test_fast_lookup (
         id        NUMBER(5) PRIMARY KEY,
         test_col  VARCHAR2(15))
      MEMOPTIMIZE FOR READ;

    The following example enables the existing table hr.employees for fast lookup:

     ALTER TABLE hr.employees MEMOPTIMIZE FOR READ;
12.5.3.3 Disabling a Table for Fast Lookup

You can disable a table for fast lookup by specifying the NO MEMOPTIMIZE FOR READ clause in the ALTER TABLE statement.

Prerequisites

This task assumes that a table is already enabled for fast lookup.

To disable a table for fast lookup:

  1. In SQL*Plus, log in to the database as a user with the ALTER TABLE privileges.

  2. Run the ALTER TABLE statement with the NO MEMOPTIMIZE FOR READ clause for the table that needs to be disabled for fast lookup.

    The following example disables the hr.employees table for fast lookup:

    ALTER TABLE hr.employees NO MEMOPTIMIZE FOR READ;
    
12.5.3.4 Managing Fast Lookup Data in the Memoptimize Pool

The memoptimize pool stores the data (fast lookup data) of all the tables that are enabled for fast lookup. You can explicitly delete or populate fast lookup data for a table in the memoptimize pool using the DBMS_MEMOPTIMIZE package subprograms.

Overview of the DBMS_MEMOPTIMIZE package subprograms

The following are the DBMS_MEMOPTIMIZE package subprograms that can be used to delete or populate fast lookup data for a table in the memoptimize pool:

Subprogram Description

DROP_OBJECT

Removes fast lookup data for a table from the memoptimize pool.

POPULATE

Populates fast lookup data for a table in the memoptimize pool.

See Also: