You are in the middle of a transaction and very crucial data has been modified. Because of a hardware failure, the instance has shut down before synchronizing all the database files. Which two statements are true? (Choose two.) A) On startup, SMON coordinates instance recovery. B) On startup, CKPT coordinates instance recovery. C) On startup, use RMAN to perform instance recovery. D) Uncommitted changes will be rolled back after the database is opened. E) On startup, perform media recovery and then instance recovery. F) On startup, all the files will be synchronized and you get both committed and uncommitted data.
Answer: A,D 当实例失败后,SMON首先找到最后SCN记录的Redo Log File,寻找最后一个打入的Checkpoint。Oracle会从这个checkpoint position处开始应用日志文件进行“前滚”来重演整个操作过程直到实例失败的那一刻,而后假如数据库发现有些事务没有进行提交,则会从Undo表空间中寻找到旧版本SCN的数据块信息,来进行SGA中Buffer Cache数据块恢复。“回滚”未提交的事务。
|