undo_retention简单定义,就是最多数据的最少保留时间。AUM模式下,undo_retention参数用于事务commit后undo数据保留的时间。单位为秒。这是个no guarantee的限制。也就是,若空间足够,他只是个‘花瓶’;当可用空间不足而又有事务需要回滚空间,则这些数据依然会被覆盖。这个行为可能会导致ORA-01555错误,这些数据被记忆的时间可用v$undostat里面的字段TUNED_UNDORETENTION来查询。
很多时候,我们希望undo数据能够被留存,而不是被覆盖。那么在10g,oracle对undo增加了guarantee控制,也就是,用户可以指定undo表空间必须满足undo_retention的限制。
alter tablespace undotbs1 retention guarantee|noguarantee;
通过设置期望的保留时间,修改undo表空间属性,就可以使undo表空间运行在guarantee模式。下面我们用一个实验体验一下,noguarantee和guarantee有什么区别:
sys@ORCL> select tablespace_name,contents,retention from dba_tablespaces where tablespace_name like '%UNDOTBS%'; TABLESPACE_NAME CONTENTS RETENTION ------------------------------ --------- ----------- UNDOTBS1 UNDO NOGUARANTEE sys@ORCL> alter system set undo_retention=800; System altered. sys@ORCL> alter tablespace undotbs1 retention guarantee; Tablespace altered. sys@ORCL> select tablespace_name,contents,retention from dba_tablespaces where tablespace_name like '%UNDOTBS%'; TABLESPACE_NAME CONTENTS RETENTION ------------------------------ --------- ----------- UNDOTBS1 UNDO GUARANTEE 将undo表空间自动扩展属性取消 sys@ORCL> select file_name,tablespace_name,bytes/1024/1024 m from dba_data_files where tablespace_name like '%UNDOTBS%'; FILE_NAME TABLESPACE_NAME M ------------------------------------------------------------------------ ------------------------------ ---------- /u01/app/oracle/oradata/ORCL/datafile/o1_mf_undotbs1_8050fkc6_.dbf UNDOTBS1 30 sys@ORCL> alter database datafile '/u01/app/oracle/oradata/ORCL/datafile/o1_mf_undotbs1_8050fkc6_.dbf' autoextend off; Database altered. 尝试循环小批量删除数据,在guarantee设置下,很快会出现ORA-30036错误: hr@ORCL> select count(*) from t; COUNT(*) ---------- 1462140 hr@ORCL> begin 2 for i in 1..1000 3 loop 4 delete from t where rownum<1001; 5 commit; 6 end loop; 7 end; 7 / begin * ERROR at line 1: ORA-30036: unable to extend segment by 8 in undo tablespace 'UNDOTBS1' ORA-06512: at line 4 hr@ORCL> select count(*) from t; COUNT(*) ---------- 14620000 而在修改了undo表空间retention属性后,删除可以顺利完成 hr@ORCL> begin 2 for i in 1..1000 3 loop 4 delete from t where rownum<1001; 5 commit; 6 end loop; 7 end; 8 / PL/SQL procedure successfully completed.
undo_retention的设置可尽责于闪回功能。对于列的类型为LOB的,自动undo数据的记忆是不支持LOBs,必须设置undo_retention的值来尽力挽留之。oracle在undo segment header上创建一个retention table用于记录相关undo存储的提交时间,从而实现其保留策略。
最优的undo_retention如何计算?
模糊计算可以借助:
如果你的事务隔离级别是 serializable或read only ,你可以让undo_retention稍微长于最长运行的事务;
如果你的事务隔离级别是read commited,你可以让undo_retention稍微长于最长运行的语句(DML)。
精确计算可以借助公式:
我们需要选一段比较有代表性的时间段来测试,或者去多段,然后加权平均
SELECT d.undo_size/(1024*1024) "ACTUAL UNDO SIZE [MByte]", SUBSTR(e.value,1,25) "UNDO RETENTION [Sec]", ROUND((d.undo_size / (to_number(f.value) * g.undo_block_per_sec)))"OPTIMAL UNDO RETENTION [Sec]" FROM ( SELECT SUM(a.bytes) undo_size FROM v$datafile a, v$tablespace b, dba_tablespaces c WHERE c.contents = 'UNDO' AND c.status = 'ONLINE' AND b.name = c.tablespace_name AND a.ts# = b.ts# ) d, v$parameter e, v$parameter f, ( SELECT MAX(undoblks/((end_time-begin_time)*3600*24))undo_block_per_sec FROM v$undostat ) g WHERE e.name = 'undo_retention' AND f.name = 'db_block_size' ACTUAL UNDO SIZE [MByte] UNDO RETENTION [Sec] OPTIMAL UNDO RETENTION [Sec] 30 900 40421