透过alert日志看Oracle Startup的三个阶段

        alert日志记录oracle实例生命周期中发生的重要事件，例如数据库启动、关闭、日志切换等，方便程序员和管理员发现问题。其中，数据库启动分为3个阶段，即nomount，mount和open。那么，在这三个阶段alert日志会记录哪些信息呢。下面我们就按照 nomount，mount和open的顺序逐步启动数据库。

nomount

通过startup nomount命令将数据库启动到nomount状态，此时，数据库实例的状态是STARTED。如下所示：

点击(此处)折叠或打开

1. SQL>
   
2. SQL> startup nomount
   
3. ORACLE instance started.
   
4. 
   
5. Total System Global Area 941600768 bytes
   
6. Fixed Size 1348860 bytes
   
7. Variable Size 515902212 bytes
   
8. Database Buffers 419430400 bytes
   
9. Redo Buffers 4919296 bytes
   
10. SQL>
    
11. SQL> select status from v$instance;
    
12. 
    
13. STATUS
    
14. ------------
    
15. STARTED
    
16. 
    
17. SQL>

接下来，查看alert日志的输出内容如下：

点击(此处)折叠或打开

1. Sat May 09 22:14:16 2015
   
2. Starting ORACLE instance (normal)
   
3. LICENSE_MAX_SESSION = 0
   
4. LICENSE_SESSIONS_WARNING = 0
   
5. Picked latch-free SCN scheme 2
   
6. Using LOG_ARCHIVE_DEST_1 parameter default value as /u01/app/oracle/product/11.2.0/dbhome_1/dbs/arch
   
7. Autotune of undo retention is turned on.
   
8. IMODE=BR
   
9. ILAT =27
   
10. LICENSE_MAX_USERS = 0
    
11. SYS auditing is disabled
    
12. Starting up:
    
13. Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - Production
    
14. With the Partitioning, OLAP, Data Mining and Real Application Testing options.
    
15. ORACLE_HOME = /u01/app/oracle/product/11.2.0/dbhome_1
    
16. System name:    Linux
    
17. Node name:    enmoedu1.example.com
    
18. Release:    2.6.18-164.el5PAE
    
19. Version:    #1 SMP Thu Sep 3 02:28:20 EDT 2009
    
20. Machine:    i686
    
21. VM name:    VMWare Version: 6
    
22. Using parameter settings in server-side spfile /u01/app/oracle/product/11.2.0/dbhome_1/dbs/spfileHOEGH.ora
    
23. System parameters with non-default values:
    
24.   processes = 150
    
25.   memory_target = 900M
    
26.   memory_max_target = 900M
    
27.   control_files = \"/u01/app/oracle/oradata/HOEGH/control01.ctl\"
    
28.   control_files = \"/u01/app/oracle/oradata/HOEGH/control02.ctl\"
    
29.   db_block_size = 8192
    
30.   compatible = \"11.2.0.0.0\"
    
31.   undo_tablespace = \"UNDOTBS1\"
    
32.   remote_login_passwordfile= \"EXCLUSIVE\"
    
33.   db_domain = \"\"
    
34.   dispatchers = \"(PROTOCOL=TCP) (SERVICE=HOEGHXDB)\"
    
35.   audit_file_dest = \"/u01/app/oracle/admin/HOEGH/adump\"
    
36.   audit_trail = \"DB\"
    
37.   db_name = \"HOEGH\"
    
38.   open_cursors = 300
    
39.   diagnostic_dest = \"/u01/app/oracle\"
    
40. Sat May 09 22:14:16 2015
    
41. PMON started with pid=2, OS id=5054
    
42. Sat May 09 22:14:16 2015
    
43. PSP0 started with pid=3, OS id=5056
    
44. Sat May 09 22:14:17 2015
    
45. VKTM started with pid=4, OS id=5063 at elevated priority
    
46. VKTM running at (1)millisec precision with DBRM quantum (100)ms
    
47. Sat May 09 22:14:17 2015
    
48. GEN0 started with pid=5, OS id=5067
    
49. Sat May 09 22:14:17 2015
    
50. DIAG started with pid=6, OS id=5069
    
51. Sat May 09 22:14:17 2015
    
52. DBRM started with pid=7, OS id=5071
    
53. Sat May 09 22:14:17 2015
    
54. DIA0 started with pid=8, OS id=5073
    
55. Sat May 09 22:14:17 2015
    
56. MMAN started with pid=9, OS id=5075
    
57. Sat May 09 22:14:17 2015
    
58. DBW0 started with pid=10, OS id=5077
    
59. Sat May 09 22:14:17 2015
    
60. LGWR started with pid=11, OS id=5079
    
61. Sat May 09 22:14:17 2015
    
62. CKPT started with pid=12, OS id=5081
    
63. Sat May 09 22:14:17 2015
    
64. SMON started with pid=13, OS id=5083
    
65. Sat May 09 22:14:17 2015
    
66. RECO started with pid=14, OS id=5085
    
67. Sat May 09 22:14:17 2015
    
68. MMON started with pid=15, OS id=5087
    
69. Sat May 09 22:14:17 2015
    
70. MMNL started with pid=16, OS id=5089
    
71. starting up 1 dispatcher(s) for network address \'(ADDRESS=(PARTIAL=YES)(PROTOCOL=TCP))\'...
    
72. starting up 1 shared server(s) ...
    
73. ORACLE_BASE from environment = /u01/app/oracle

由上可以看出，在nomount阶段，Oracle首先寻找参数文件(pfile/spfile)，然后根据参数文件中的设置（如内存分配等设置），创建实例（INSTANCE），分配内存，启动后台进程。Nomount的过程也就是启动数据库实例的过程。这个过程在后台是启动Oracle可执行程序的过程，Windows上是oracle.exe文件的初始化，在Unix/Linux上是oracle可执行文件的初始化。

mount

接下来，通过alter database mount;命令将数据库启动到mount状态，此时，数据库实例的状态是 MOUNTED 。如下所示：

点击(此处)折叠或打开

1. SQL>
   
2. SQL> alter database mount;
   
3. 
   
4. Database altered.
   
5. 
   
6. SQL>
   
7. SQL> select status from v$instance;
   
8. 
   
9. STATUS
   
10. ------------
    
11. MOUNTED
    
12. 
    
13. SQL>

查看alert日志的输出内容如下：

点击(此处)折叠或打开

1. Sat May 09 22:15:30 2015
   
2. alter database mount
   
3. Sat May 09 22:15:34 2015
   
4. Successful mount of redo thread 1, with mount id 2100626370
   
5. Database mounted in Exclusive Mode
   
6. Lost write protection disabled
   
7. Completed: alter database mount

mount数据库的过程是读参数文件中描述的控制文件,校验控制文件的正确性,将控制文件的内容读入到内存,mount是挂接的意思,是操作系统中的概念。一旦mount之后,就是将一个没有意义的实例和一个数据库发生了联系。

open

通过alter database open;命令将数据库启动到open状态，此时，数据库实例的状态是OPEN。如下所示：

点击(此处)折叠或打开

1. SQL>
   
2. SQL> alter database open;
   
3. 
   
4. Database altered.
   
5. 
   
6. SQL>
   
7. SQL> select status from v$instance;
   
8. 
   
9. STATUS
   
10. ------------
    
11. OPEN
    
12. 
    
13. SQL>

查看alert日志的输出内容如下：

点击(此处)折叠或打开

1. Sat May 09 22:16:34 2015
   
2. alter database open
   
3. Beginning crash recovery of 1 threads
   
4.  parallel recovery started with 2 processes
   
5. Started redo scan
   
6. Completed redo scan
   
7.  read 12 KB redo, 9 data blocks need recovery
   
8. Started redo application at
   
9.  Thread 1: logseq 7, block 118
   
10. Recovery of Online Redo Log: Thread 1 Group 1 Seq 7 Reading mem 0
    
11.   Mem# 0: /u01/app/oracle/oradata/HOEGH/redo01.log
    
12. Completed redo application of 0.00MB
    
13. Completed crash recovery at
    
14.  Thread 1: logseq 7, block 143, scn 918639
    
15.  9 data blocks read, 9 data blocks written, 12 redo k-bytes read
    
16. Sat May 09 22:16:34 2015
    
17. Thread 1 advanced to log sequence 8 (thread open)
    
18. Thread 1 opened at log sequence 8
    
19.   Current log# 2 seq# 8 mem# 0: /u01/app/oracle/oradata/HOEGH/redo02.log
    
20. Successful open of redo thread 1
    
21. MTTR advisory is disabled because FAST_START_MTTR_TARGET is not set
    
22. Sat May 09 22:16:35 2015
    
23. SMON: enabling cache recovery
    
24. [5095] Successfully onlined Undo Tablespace 2.
    
25. Undo initialization finished serial:0 start:4294562150 end:4294562220 diff:70 (0 seconds)
    
26. Verifying file header compatibility for 11g tablespace encryption..
    
27. Verifying 11g file header compatibility for tablespace encryption completed
    
28. SMON: enabling tx recovery
    
29. Database Characterset is AL32UTF8
    
30. No Resource Manager plan active
    
31. replication_dependency_tracking turned off (no async multimaster replication found)
    
32. Starting background process QMNC
    
33. Sat May 09 22:16:35 2015
    
34. QMNC started with pid=22, OS id=5274
    
35. Completed: alter database open
    
36. Sat May 09 22:16:36 2015
    
37. Starting background process CJQ0
    
38. Sat May 09 22:16:36 2015
    
39. CJQ0 started with pid=24, OS id=5296

这一阶段数据库读取 控制文件中描述的数据文件 并 验证数据文件的一致性，如果不一致，使用日志文件将数据库文件恢复到一致的状态。 数据库open后,普通用户才可以访问数据库。

~~~~~~~ the end~~~~~~~~~

hoegh

2015.05.09

      

来自 “ ITPUB博客 ” ，链接：http://blog.itpub.net/30162081/viewspace-1631808/，如需转载，请注明出处，否则将追究法律责任。

转载于:http://blog.itpub.net/30162081/viewspace-1631808/