session_max_open_files:
說明: 指定可在任一給定會話中打開的 BFILE 的最大數量
值范圍:
默認值:
parallel_execution_message_size:
說明: 指定並行執行 (並行查詢
值范圍:
默認值: 如果 PARALLEL_AUTOMATIC_TUNING 為 FALSE
Paralle_min_percent:
說明: 指定並行執行要求的線程的最小百分比
值范圍:
默認值:
Parallel_automatic_tuning:
說明: 如果設置為 TRUE
值范圍: TRUE | FALSE
默認值: FALSE
parallel_threads_per_cpu:
說明: 說明一個 CPU 在並行執行過程中可處理的進程或線程的數量
值范圍: 任何非零值
默認值: 根據操作系統而定 (通常為
parallel_broadcast_enabled:
說明 : 通過使用一個散列聯接或合並聯接
值范圍: TRUE | FALSE
默認值 : FALSE
parallel_adaptive_multi_user:
說明: 啟用或禁用一個自適應算法
值范圍: TRUE | FALSE
默認值: 如果 PARALLEL_AUTOMATIC_TUNING = TRUE
parallel_max_servers:
說明: 指定一個例程的並行執行服務器或並行恢復進程的最大數量
值范圍:
默認值: 由 CPU_COUNT
parallel_min_servers
說明: 指定為並行執行啟動例程後
值范圍:
默認值:
log_archive_dest_state_
說明: 指定相應的歸檔日志目標參數 (僅 LOG_ARCHIVE_DEST_
值范圍: ENABLE | DEFER
默認值: ENABLE
log_archive_dest_state_
說明: 指定相應的歸檔日志目標參數 (僅 LOG_ARCHIVE_DEST_
值范圍: ENABLE | DEFER
默認值: ENABLE
log_archive_dest_state_
說明: 指定相應的歸檔日志目標參數 (僅 LOG_ARCHIVE_DEST_
值范圍: ENABLE | DEFER
默認值: ENABLE
log_archive_dest_state_
說明: 標識特定日志歸檔目標的最近的用戶定義狀態
值范圍: ENABLE
log_archive_dest_state_
說明: 標識特定日志歸檔目標的最近的用戶定義狀態
值范圍: ENABLE
log_archive_dest_state_
說明: 標識特定日志歸檔目標的最近的用戶定義狀態
值范圍: ENABLE
alter table emp parallel (degree
select degree from user_tables where table_name =
select count(*) from emp;
alter table emp noparallel;
SELECT /*+ PARALLEL(emp
FROM emp;
一
大表掃描
創建大的索引
基於分區的索引掃描
大量的插入
拷貝操作
二
SMP
足夠的I/O帶寬
足夠的內存來支持消耗內存的進程(排序
三
在DSS系統中
銀行電信行業的月終年終結算
基於許多大表的即時查詢
數據庫的備份恢復
四
當該參數置為TRUE時
五
系統硬件配置
CPUS =
Main Memory =
Disk =
Users =
DBA 做如下設置:
PARALLEL_AUTOMATIC_TUNING = TRUE
SHARED_POOL_SIZE =
TRANSACTIONS =系統缺省值
Oracle自動調整以下參數的值
PARALLEL_MAX_SERVERS =
PARALLEL_ADAPTIVE_MULTI_USER = TRUE
PARALLEL_THREADS_PER_CPU =
PROCESSES =
SESSIONS =
TRANSACTIONS =
LARGE_POOL_SIZE =
Parameter Settings for DOP and the Adaptive Multi
The DBA parallelizes every table having more than
similar to the following:
ALTER TABLE employee PARALLEL;
In this example
CPUs is
TRUE
time of the query
系統硬件配置
CPUS =
Main Memory =
Disk =
Users =
The DBA makes the following settings:
n PARALLEL_AUTOMATIC_TUNING = TRUE
n PARALLEL_ADAPTIVE_MULTI_USER = FALSE
n PARALLEL_THREADS_PER_CPU =
n SHARED_POOL_SIZE =
The DBA also sets other parameters unrelated to parallelism
responds by automatically adjusting the following parameter settings:
n PROCESSES =
n SESSIONS =
n TRANSACTIONS =
n PARALLEL_MAX_SERVERS =
n LARGE_POOL_SIZE =
Parameter Settings for DOP and the Adaptive Multi
The DBA parallelizes some tables in the data warehouse while creating other views
for special users:
ALTER TABLE sales PARALLEL;
CREATE VIEW invoice_parallel AS SELECT /*+ PARALLEL(P) */ * FROM invoices P;
The DBA allows the system to use the PARALLEL_THREADS_PER_CPU setting of
processes while more complex queries use
系統硬件配置
Uses MPP Architecture (Massively Parallel Processing)
Main Memory =
Disk =
Users =
The DBA uses manual parallel tuning by setting the following:
n PARALLEL_AUTOMATIC_TUNING = FALSE
n PARALLEL_THREADS_PER_CPU =
n PARALLEL_MAX_SERVERS =
n SHARED_POOL_SIZE =
n PARALLEL_SERVER_INSTANCES =
n PARALLEL_SERVER = TRUE
n PROCESSES =
n SESSIONS =
n TRANSACTIONS =
The DBA also sets other parameters unrelated to parallel execution
PARALLEL_AUTOMATIC_TUNING is set to FALSE
execution buffers from the SHARED_POOL
Parameter Settings for DOP and the Adaptive Multi
The DBA parallelizes tables in the data warehouse by explicitly setting the DOP
using syntax similar to the following:
ALTER TABLE department
ALTER TABLE department
CREATE VIEW current_sales AS SELECT /*+ PARALLEL(P
In this example
the DBA has manually set all parallel execution parameters
系統硬件配置
CPUS =
Main Memory
Disk =
Users =
The DBA makes the following settings:
n PARALLEL_AUTOMATIC_TUNING = TRUE
n PARALLEL_MAX_SERVERS =
n PARALLEL_MIN_SERVER =
n LARGE_POOL_SIZE =
n SHARED_POOL_SIZE =
n PROCESSES =
n SESSIONS =
n TRANSACTIONS =
Parameter Settings for DOP and the Adaptive Multi
The DBA has carefully evaluated which users and tables require parallelism and
has set the values according to their requirements
mentioned in the earlier examples
command during peak user hours to enable the adaptive DOP algorithms:
ALTER SYSTEM SET PARALLEL_ADAPTIVE_MULTI_USER = TRUE;
During off hours when batch processing is about to begin
adaptive processing by issuing the command:
ALTER SYSTEM SET PARALLEL_ADAPTIVE_MULTI_USER = FALSE;
From:http://tw.wingwit.com/Article/program/Oracle/201311/18583.html