The large pool is used to allocate large memory to handle the larger memory of the shared pool and is a standby pool to ease Oracle's pressure on shared pool and PGA memory use
The main objects used with large pool are:
Multithreaded Server MTS: Allocating UGA in the large_pool of the SGA
Parallel query Parallel exection for statements: used as a message buffer between processes
Recovery Manager Rman: Used as a disk IO buffer at backup
To view the default setting value of the Large_pool parameter, 0 should indicate that the ASMM is turned on
Show Parameter Large_pool_size
View ASMM actual allocation size during Large_pool current run
Select Pool,sum (bytes) from V$sgastat group by pool;
Java Pool
Oracle has added Java support to the kernel, which is designed for Java development and application, and it is not necessary to change the default size of the buffer without Java programs
To view the default setting value of the Java_pool parameter, 0 should indicate that the ASMM is turned on
Show Parameter Java_pool_size
View ASMM actual allocation size during Java_pool current run
Select Pool,sum (bytes) from V$sgastat group by pool;
Streams_buffer
Buffering for Convection replication
View Size
Show parameter streams_pool_size;
Redo_log_buffer
Any modifications to the database are recorded sequentially in the buffer, and then the LGWR process writes to the Redolog file on the disk in batches based on conditional change information to conserve disk IO, which does not participate in ASMM dynamic management and is not automatically resized
View Redo_log_buffer Size
Show parameter Log_buffer;
Select Name,bytes from V$sgastat where name= ' Log_buffer ';
Manually modify Redo_log_buffer Size
Alter system set log_buffer=6000000 Scope=spfile;
Oracle Process:
Background process
Server process
User process
Background process:
A background process is an Oracle program that manages the reading, writing, recovery, and monitoring of a database, primarily through background processes and user processes, and by data exchange between background processes and user processes
To view a background process that is already running
$ps-ef |grep Ora_
Sql>select name,description from v$bgprocess where paddr!= ' 00 ';
DBWN (Database write process)
Data buffer data is written to the data file, is responsible for data buffer management of a background process. The default number is 1, up to 10.
Data buffer Fast write disk, you want to write each data cache fast content to the data file in the specified fast, because these designations are usually discontinuous, so write these fast is called discrete write, because the head needs to repeatedly locate the data file within the specified fast, and the head repeated positioning is time-consuming, so the discrete write is relatively slow
The role of DBWN:
The LRU algorithm manages the data buffer, writes the modified dirty buffer data to the data file, maintains clean of the data buffer so that the user process can always find enough free buffers
Optimizes disk IO reading and writing by delaying write data
Lgwr
Writes log data from the log buffer to the Redolog log file group, where the contents of each log buffer block are written to a specified speed and are written to each fast in the log file, so they are called sequential writes and are faster in sequential writes.
The role of LGWR ·:
Manage • Log buffers, write database changes to log files to maintain data consistency and provide a basis for recovery after data loss
Optimizing Disk IO reading and writing by delaying write logging
About transactions
' One transaction generates one or more redo records, and a redo record contains multiple change vectors, which, when restored with log data, take the smallest unit