ORACLE啟動過程簡析
ORACLE資料庫啟動時，經歷了三個過程：(用命名如下）
startup nomount;
alter database mount; alter database open;
當然資料庫關閉時也是經歷了三個相反的過程：
alter database close;
alter database dismount;
shutdown;（shutdown 後面跟了四個引數：normal;immediate;transactional;abort)
四種方式關閉資料庫的比較：
(NO YES)
分析第一個過程startup nomount:
這個過程資料庫首先到引數檔案（pfile/spfile)中讀取資料庫的設定，建立例項.
資料庫所在的作業系統版本：
[oracle@localhost ~]$ lsb_release -a
LSB Version: :core-3.1-ia32:core-3.1-noarch:graphics-3.1-ia32:graphics-3.1-noarch
Distributor ID: EnterpriseEnterpriseServer
Description: Enterprise Linux Enterprise Linux Server release 5.5 (Carthage)
Release: 5.5
Codename: Carthage
BANNER
--------------------------------------------------------------------------------
Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - Production
因為spfile是二進位制檔案，不能直接讀取，在linux中，可以用命令String轉儲出來：
[oracle@localhost dbs]$ strings spfileorcl3939.ora
orcl3939.__db_cache_size=54525952
orcl3939.__java_pool_size=4194304
orcl3939.__large_pool_size=4194304
orcl3939.__oracle_base='/u01/app/oracle'#ORACLE_BASE set from environment
orcl3939.__pga_aggregate_target=171966464
orcl3939.__sga_target=251658240
orcl3939.__shared_io_pool_size=0
orcl3939.__shared_pool_size=176160768
orcl3939.__streams_pool_size=4194304
*.audit_file_dest='/u01/app/oracle/admin/orcl3939/adump'
*.audit_trail='db'
*.compatible='11.2.0.0.0'
*.control_files='/u01/app/o
racle/oradata/orcl3939/control01.ctl','/u01/app/oracle/flash_recovery_area/orcl3939/control02.ctl','/u01/app/oracle/oradata/orcl3939/control03.ctl'
*.db_block_size=8192
*.db_domain='localdomain'
*.db_name='orcl3939'
*.db_recovery_file_dest='/u01/app/oracle/flash_recovery_area'
*.db_recovery_file_dest_size=4039114752
*.diagnostic_dest='/u01/app/oracle'
*.dispatchers='(PROTOCOL=TCP) (SERVICE=orcl3939XDB)'
*.instance_name='ORCL3939'
*.local_listener='(ADDRESS=(PROTOCOL=TCP)(HOST = local
host.localdomain)(PORT = 1521))'
*.memory_target=423624704
*.open_cursors=300
*.processes=150
*.remote_login_passwordfile='EXCLUSIVE'
*.service_names='a,b,c,d'
*.trace_enabled=TRUE
*.undo_tablespace='UNDOTBS1'
spfile檔案中你可以看到資料庫在nomount時做了些什麼，根據引數檔案的內容，建立了instance,分配了相應的記憶體區域，啟動了相應的後臺程式。
我們再看告警日誌檔案（alert_.log）：讀取了引數檔案，啟動了例項
Starting up:
Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options.
Using parameter settings in server-side spfile /u01/app/oracle/product/11.2.0/dbhome_1/dbs/spfileorcl3939.ora
System parameters with non-default values:
processes = 150
memory_target = 404M
control_files = "/u01/app/oracle/oradata/orcl3939/control01.ctl"
control_files = "/u01/app/oracle/flash_recovery_area/orcl3939/control02.ctl"
control_files = "/u01/app/oracle/oradata/orcl3939/control03.ctl"
db_block_size = 8192
compatible = "11.2.0.0.0"
db_recovery_file_dest = "/u01/app/oracle/flash_recovery_area"
db_recovery_file_dest_size= 3852M
undo_tablespace = "UNDOTBS1"
remote_login_passwordfile= "EXCLUSIVE"
db_domain = "localdomain"
instance_name = "ORCL3939"
service_names = "a,b,c,d"
dispatchers = "(PROTOCOL=TCP) (SERVICE=orcl3939XDB)"
local_listener = "(ADDRESS=(PROTOCOL=TCP)(HOST = localhost.localdomain)(PORT = 1521))"
audit_file_dest = "/u01/app/oracle/admin/orcl3939/adump"
audit_trail = "DB"
db_name = "orcl3939"
open_cursors = 300
diagnostic_dest = "/u01/app/oracle"
trace_enabled = TRUE
Thu Apr 02 14:59:41 2015
PMON started with pid=2, OS id=5989
Thu Apr 02 14:59:41 2015
VKTM started with pid=3, OS id=5991 at elevated priority
VKTM running at (10)millisec precision with DBRM quantum (100)ms
Thu Apr 02 14:59:41 2015
GEN0 started with pid=4, OS id=5995
Thu Apr 02 14:59:41 2015
DIAG started with pid=5, OS id=5997
Thu Apr 02 14:59:41 2015
DBRM started with pid=6, OS id=5999
Thu Apr 02 14:59:41 2015
PSP0 started with pid=7, OS id=6001
Thu Apr 02 14:59:41 2015
DIA0 started with pid=8, OS id=6003
Thu Apr 02 14:59:41 2015
MMAN started with pid=9, OS id=6005
Thu Apr 02 14:59:41 2015
DBW0 started with pid=10, OS id=6007
Thu Apr 02 14:59:41 2015
LGWR started with pid=11, OS id=6009
Thu Apr 02 14:59:41 2015
CKPT started with pid=12, OS id=6011
Thu Apr 02 14:59:41 2015
SMON started with pid=13, OS id=6013
Thu Apr 02 14:59:41 2015
RECO started with pid=14, OS id=6015
Thu Apr 02 14:59:41 2015
MMON started with pid=15, OS id=6017
Thu Apr 02 14:59:41 2015
MMNL started with pid=16, OS id=6019
starting up 1 dispatcher(s) for network address '(ADDRESS=(PARTIAL=YES)(PROTOCOL=TCP))'...
starting up 1 shared server(s) ...
ORACLE_BASE from environment = /u01/app/oracle
資料庫根據引數建立例項之後，後臺程式依次啟動，注意上面輸出中包含了PID資訊以及OS ID兩個資訊，PID代表該程式在資料庫內部的識別符號編號，而OS ID則代表該程式在作業系統上的程式編號。
我們可以透過oracle中的動態檢視v$process，可以把後臺程式和作業系統的程式想關聯起來：
SQL> select addr,pid,spid,username,program from v$process;
ADDR PID SPID USERNAME PROGRAM
ADDR PID SPID USERNAME PROGRAM
-------- ---------- ------------------------------------------------------------------------ --------------------------------------------- ------------------------------------------------------------------------------------------------------------------------------------------------
393B9444 1 PSEUDO
393B9F1C 2 5989 oracle oracle@localhost.localdomain (PMON)
393BA9F4 3 5991 oracle oracle@localhost.localdomain (VKTM)
393BB4CC 4 5995 oracle oracle@localhost.localdomain (GEN0)
393BBFA4 5 5997 oracle oracle@localhost.localdomain (DIAG)
393BCA7C 6 5999 oracle oracle@localhost.localdomain (DBRM)
393BD554 7 6001 oracle oracle@localhost.localdomain (PSP0)
393BE02C 8 6003 oracle oracle@localhost.localdomain (DIA0)
393BEB04 9 6005 oracle oracle@localhost.localdomain (MMAN)
393BF5DC 10 6007 oracle oracle@localhost.localdomain (DBW0)
393C00B4 11 6009 oracle oracle@localhost.localdomain (LGWR)
393C0B8C 12 6011 oracle oracle@localhost.localdomain (CKPT)
393C1664 13 6013 oracle oracle@localhost.localdomain (SMON)
393C213C 14 6015 oracle oracle@localhost.localdomain (RECO)
393C2C14 15 6017 oracle oracle@localhost.localdomain (MMON)
393C36EC 16 6019 oracle oracle@localhost.localdomain (MMNL)
393C41C4 17 6021 oracle oracle@localhost.localdomain (D000)
393C4C9C 18 6023 oracle oracle@localhost.localdomain (S000)
393C5774 19 6757 oracle oracle@localhost.localdomain (TNS V1-V3)
已選擇19行。
分析第二個過程mount:
告警日誌中：
alter database mount
Thu Apr 02 15:33:03 2015
Successful mount of redo thread 1, with mount id 3864558315
Database mounted in Exclusive Mode
Lost write protection disabled
Completed: alter database mount
有引數檔案中，找到了control_file的位置並鎖定控制檔案：
SQL> show parameter control_files;
NAME TYPE VALUE
------------------------------------ --------------------------------- ------------------------------
control_files string /u01/app/oracle/oradata/orcl39
39/control01.ctl, /u01/app/ora
cle/flash_recovery_area/orcl39
39/control02.ctl, /u01/app/ora
cle/oradata/orcl3939/control03
.ctl
這三個控制檔案的大小一樣，3個控制檔案最好放在不同的物理磁碟上，往控制檔案中寫資訊的時候併發同時寫，所以3個控制檔案的內容是相同的，但是讀取的時候，只讀取第一個，如果3個控制檔案有一個出錯了，oracle就不能啟動了。在實際的生產工程中，不建議放在同一磁碟上，這樣不利於資料庫遇到磁碟介質損壞的恢復。
控制檔案中包含了聯機重做日誌檔案和資料檔案的位置。
分析第三個過程open:
由於控制檔案中記錄了資料檔案，日誌檔案的位置，檢查點資訊等重要的資訊，在open階段時，資料庫根據控制檔案中記錄的這些資訊找到這些檔案，然後進行檢查點及完整性檢查。如果沒有問題可以啟動資料庫，如果存在不一致或者檔案丟失則需要恢復資料庫。關於資料庫的一致性性檢查在這裡不做闡述。
在這三個過程中，每個過程可以查些什麼動態效能檢視（動態效能檢視是在資料庫啟動時自動建立）:
nomount:
只是啟動了例項，啟動例項的資訊主要來自引數檔案，引數檔案中記錄的資訊可以查詢，可以查：v$parameter,v$spparameter,v$sga,v$sgastat,v$bh,v$instance,v$option,v$version,v$process,v$session
mount:
此時控制檔案被讀取，和控制檔案相關的檢視可以查詢，這要有：v$thread,v$controlfile,v$database,v$datafile,v$logfile,v$datafile_header
open:
open之後，所有的動態效能檢視都可以查詢。
startup nomount;
alter database mount; alter database open;
當然資料庫關閉時也是經歷了三個相反的過程：
alter database close;
alter database dismount;
shutdown;（shutdown 後面跟了四個引數：normal;immediate;transactional;abort)
四種方式關閉資料庫的比較：
(NO YES)
| 關閉方式 | A | I | T | N |
|
允許新的連線 |
N | N | N | N |
| 等待活動會話終止 | N | N | N | Y |
| 等待活動事務終止 | N | N |
|
Y |
| 強制進行checkpoint,關閉所有檔案 | N | Y | Y | Y |
分析第一個過程startup nomount:
這個過程資料庫首先到引數檔案（pfile/spfile)中讀取資料庫的設定，建立例項.
資料庫所在的作業系統版本：
[oracle@localhost ~]$ lsb_release -a
LSB Version: :core-3.1-ia32:core-3.1-noarch:graphics-3.1-ia32:graphics-3.1-noarch
Distributor ID: EnterpriseEnterpriseServer
Description: Enterprise Linux Enterprise Linux Server release 5.5 (Carthage)
Release: 5.5
Codename: Carthage
資料庫版本：
SQL> SELECT * FROM v$version where rownum=1;BANNER
--------------------------------------------------------------------------------
Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - Production
因為spfile是二進位制檔案，不能直接讀取，在linux中，可以用命令String轉儲出來：
[oracle@localhost dbs]$ strings spfileorcl3939.ora
orcl3939.__db_cache_size=54525952
orcl3939.__java_pool_size=4194304
orcl3939.__large_pool_size=4194304
orcl3939.__oracle_base='/u01/app/oracle'#ORACLE_BASE set from environment
orcl3939.__pga_aggregate_target=171966464
orcl3939.__sga_target=251658240
orcl3939.__shared_io_pool_size=0
orcl3939.__shared_pool_size=176160768
orcl3939.__streams_pool_size=4194304
*.audit_file_dest='/u01/app/oracle/admin/orcl3939/adump'
*.audit_trail='db'
*.compatible='11.2.0.0.0'
*.control_files='/u01/app/o
racle/oradata/orcl3939/control01.ctl','/u01/app/oracle/flash_recovery_area/orcl3939/control02.ctl','/u01/app/oracle/oradata/orcl3939/control03.ctl'
*.db_block_size=8192
*.db_domain='localdomain'
*.db_name='orcl3939'
*.db_recovery_file_dest='/u01/app/oracle/flash_recovery_area'
*.db_recovery_file_dest_size=4039114752
*.diagnostic_dest='/u01/app/oracle'
*.dispatchers='(PROTOCOL=TCP) (SERVICE=orcl3939XDB)'
*.instance_name='ORCL3939'
*.local_listener='(ADDRESS=(PROTOCOL=TCP)(HOST = local
host.localdomain)(PORT = 1521))'
*.memory_target=423624704
*.open_cursors=300
*.processes=150
*.remote_login_passwordfile='EXCLUSIVE'
*.service_names='a,b,c,d'
*.trace_enabled=TRUE
*.undo_tablespace='UNDOTBS1'
我們再看告警日誌檔案（alert_.log）：讀取了引數檔案，啟動了例項
Starting up:
Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options.
Using parameter settings in server-side spfile /u01/app/oracle/product/11.2.0/dbhome_1/dbs/spfileorcl3939.ora
System parameters with non-default values:
processes = 150
memory_target = 404M
control_files = "/u01/app/oracle/oradata/orcl3939/control01.ctl"
control_files = "/u01/app/oracle/flash_recovery_area/orcl3939/control02.ctl"
control_files = "/u01/app/oracle/oradata/orcl3939/control03.ctl"
db_block_size = 8192
compatible = "11.2.0.0.0"
db_recovery_file_dest = "/u01/app/oracle/flash_recovery_area"
db_recovery_file_dest_size= 3852M
undo_tablespace = "UNDOTBS1"
remote_login_passwordfile= "EXCLUSIVE"
db_domain = "localdomain"
instance_name = "ORCL3939"
service_names = "a,b,c,d"
dispatchers = "(PROTOCOL=TCP) (SERVICE=orcl3939XDB)"
local_listener = "(ADDRESS=(PROTOCOL=TCP)(HOST = localhost.localdomain)(PORT = 1521))"
audit_file_dest = "/u01/app/oracle/admin/orcl3939/adump"
audit_trail = "DB"
db_name = "orcl3939"
open_cursors = 300
diagnostic_dest = "/u01/app/oracle"
trace_enabled = TRUE
Thu Apr 02 14:59:41 2015
PMON started with pid=2, OS id=5989
Thu Apr 02 14:59:41 2015
VKTM started with pid=3, OS id=5991 at elevated priority
VKTM running at (10)millisec precision with DBRM quantum (100)ms
Thu Apr 02 14:59:41 2015
GEN0 started with pid=4, OS id=5995
Thu Apr 02 14:59:41 2015
DIAG started with pid=5, OS id=5997
Thu Apr 02 14:59:41 2015
DBRM started with pid=6, OS id=5999
Thu Apr 02 14:59:41 2015
PSP0 started with pid=7, OS id=6001
Thu Apr 02 14:59:41 2015
DIA0 started with pid=8, OS id=6003
Thu Apr 02 14:59:41 2015
MMAN started with pid=9, OS id=6005
Thu Apr 02 14:59:41 2015
DBW0 started with pid=10, OS id=6007
Thu Apr 02 14:59:41 2015
LGWR started with pid=11, OS id=6009
Thu Apr 02 14:59:41 2015
CKPT started with pid=12, OS id=6011
Thu Apr 02 14:59:41 2015
SMON started with pid=13, OS id=6013
Thu Apr 02 14:59:41 2015
RECO started with pid=14, OS id=6015
Thu Apr 02 14:59:41 2015
MMON started with pid=15, OS id=6017
Thu Apr 02 14:59:41 2015
MMNL started with pid=16, OS id=6019
starting up 1 dispatcher(s) for network address '(ADDRESS=(PARTIAL=YES)(PROTOCOL=TCP))'...
starting up 1 shared server(s) ...
ORACLE_BASE from environment = /u01/app/oracle
我們可以透過oracle中的動態檢視v$process，可以把後臺程式和作業系統的程式想關聯起來：
SQL> select addr,pid,spid,username,program from v$process;
ADDR PID SPID USERNAME PROGRAM
ADDR PID SPID USERNAME PROGRAM
-------- ---------- ------------------------------------------------------------------------ --------------------------------------------- ------------------------------------------------------------------------------------------------------------------------------------------------
393B9444 1 PSEUDO
393B9F1C 2 5989 oracle oracle@localhost.localdomain (PMON)
393BA9F4 3 5991 oracle oracle@localhost.localdomain (VKTM)
393BB4CC 4 5995 oracle oracle@localhost.localdomain (GEN0)
393BBFA4 5 5997 oracle oracle@localhost.localdomain (DIAG)
393BCA7C 6 5999 oracle oracle@localhost.localdomain (DBRM)
393BD554 7 6001 oracle oracle@localhost.localdomain (PSP0)
393BE02C 8 6003 oracle oracle@localhost.localdomain (DIA0)
393BEB04 9 6005 oracle oracle@localhost.localdomain (MMAN)
393BF5DC 10 6007 oracle oracle@localhost.localdomain (DBW0)
393C00B4 11 6009 oracle oracle@localhost.localdomain (LGWR)
393C0B8C 12 6011 oracle oracle@localhost.localdomain (CKPT)
393C1664 13 6013 oracle oracle@localhost.localdomain (SMON)
393C213C 14 6015 oracle oracle@localhost.localdomain (RECO)
393C2C14 15 6017 oracle oracle@localhost.localdomain (MMON)
393C36EC 16 6019 oracle oracle@localhost.localdomain (MMNL)
393C41C4 17 6021 oracle oracle@localhost.localdomain (D000)
393C4C9C 18 6023 oracle oracle@localhost.localdomain (S000)
393C5774 19 6757 oracle oracle@localhost.localdomain (TNS V1-V3)
已選擇19行。
分析第二個過程mount:
告警日誌中：
alter database mount
Thu Apr 02 15:33:03 2015
Successful mount of redo thread 1, with mount id 3864558315
Database mounted in Exclusive Mode
Lost write protection disabled
Completed: alter database mount
有引數檔案中，找到了control_file的位置並鎖定控制檔案：
SQL> show parameter control_files;
NAME TYPE VALUE
------------------------------------ --------------------------------- ------------------------------
control_files string /u01/app/oracle/oradata/orcl39
39/control01.ctl, /u01/app/ora
cle/flash_recovery_area/orcl39
39/control02.ctl, /u01/app/ora
cle/oradata/orcl3939/control03
.ctl
控制檔案中包含了聯機重做日誌檔案和資料檔案的位置。
分析第三個過程open:
由於控制檔案中記錄了資料檔案，日誌檔案的位置，檢查點資訊等重要的資訊，在open階段時，資料庫根據控制檔案中記錄的這些資訊找到這些檔案，然後進行檢查點及完整性檢查。如果沒有問題可以啟動資料庫，如果存在不一致或者檔案丟失則需要恢復資料庫。關於資料庫的一致性性檢查在這裡不做闡述。
在這三個過程中，每個過程可以查些什麼動態效能檢視（動態效能檢視是在資料庫啟動時自動建立）:
nomount:
只是啟動了例項，啟動例項的資訊主要來自引數檔案，引數檔案中記錄的資訊可以查詢，可以查：v$parameter,v$spparameter,v$sga,v$sgastat,v$bh,v$instance,v$option,v$version,v$process,v$session
mount:
此時控制檔案被讀取，和控制檔案相關的檢視可以查詢，這要有：v$thread,v$controlfile,v$database,v$datafile,v$logfile,v$datafile_header
open:
open之後，所有的動態效能檢視都可以查詢。
來自 “ ITPUB部落格 ” ，連結：http://blog.itpub.net/30496894/viewspace-1806105/，如需轉載，請註明出處，否則將追究法律責任。
相關文章
- ORACLE啟動過程淺析Oracle
- Android 啟動過程簡析（一）之 init 程式Android
- Oracle的啟動過程Oracle
- 淺析Android Activity的啟動過程Android
- 淺析Spring Framework框架容器啟動過程SpringFramework框架
- Linux啟動過程簡介Linux
- Oracle啟動的三個過程Oracle
- 深入理解ORACLE啟動過程Oracle
- Oracle資料庫啟動過程Oracle資料庫
- oracle的內部啟動過程Oracle
- ORACLE關閉過程淺析Oracle
- Flutter啟動流程簡析Flutter
- HMaster啟動流程簡析AST
- Oracle-解析啟動的全過程Oracle
- Tomcat啟動流程簡析Tomcat
- MySQL資料庫啟動過程的簡述MySql資料庫
- Windows 啟動過程Windows
- App 啟動過程（含 Activity 啟動過程） | 安卓 offer 收割基APP安卓
- React Native 啟動流程簡析React Native
- oracle資料庫啟動過程大揭祕Oracle資料庫
- Angular的啟動過程Angular
- main的啟動過程AI
- Nginx的啟動過程Nginx
- Service啟動過程分析
- Activity啟動過程分析
- linux啟動過程Linux
- 【LINUX】啟動過程Linux
- RecyclerView 之 Adapter 的簡化過程淺析ViewAPT
- Windows啟動過程(MBR引導過程分析)Windows
- Oracle 資料庫例項啟動關閉過程Oracle資料庫
- 使用10046 來了解oracle的啟動過程Oracle
- 原始碼簡析Spring-Integration執行過程原始碼Spring
- SpringMVC請求處理過程原始碼簡析SpringMVC原始碼
- Android App啟動過程AndroidAPP
- 計算機啟動過程計算機
- app的啟動過程（三）APP
- Liferay 啟動過程分析
- Spring Boot 啟動過程Spring Boot