实力恢复可以细分为instance recovery和crash recovery。

crash recovery的产生：

如果数据库异常关闭，数据库就没机会修改V$THREAD.STATUS状态了。因此，数据库在MOUNT状态通过检查这个状态，就可以判断使句库是否需要进行crash recover。（仅限于单INSTANCE）

单INSTANCE环境：

SQL> select status from v$thread;

STATUS
------
OPEN

SQL> shutdown abort;
ORACLE instance shut down.
SQL> startup mount;
ORACLE instance started.

Total System Global Area 1073741824 bytes
Fixed Size                  2089368 bytes
Variable Size             373296744 bytes
Database Buffers          687865856 bytes
Redo Buffers               10489856 bytes
Database mounted.
SQL> select status from v$thread;

STATUS
------
OPEN

如果是RAC环境，结果却不同。（原因还未知）

SQL> SELECT status FROM V$THREAD;

STATUS
------
OPEN
OPEN

SQL> shutdown abort;
ORACLE instance shut down.
SQL> startup mount
ORACLE instance started.

Total System Global Area 184549376 bytes
Fixed Size                  2019448 bytes
Variable Size             146804616 bytes
Database Buffers           33554432 bytes
Redo Buffers                2170880 bytes
Database mounted.
SQL> SELECT status FROM V$THREAD;

STATUS
------
CLOSED
OPEN

参考：http://www.itpub.net/thread-1722733-1-1.html

ITPub博客

实列恢复之Crash recovery的产生

最新文章