关于SCN的理解，请指正

cenwell

cool

courser001

好

h1yn

scn号是唯一的，如果第一scn号是1000的话，那么在出现第二个
scn号的时候，如果是1004，这就意味着在这个期间数据库做了某些内部的操作（例如buffer 块的清除）,
scn是唯一的，但可能不连贯
除非对数据库重新建立，否则scn号永远不会被重置为0

topgunroot

get latest SCN (depends on your oracle version :-p)

select dbms_flashback.get_system_change_number from dual;

rchsh

SQL> select checkpoint_change#  from v$database;

CHECKPOINT_CHANGE#
------------------
           1069220

SQL> select dbms_flashback.get_system_change_number from dual;

GET_SYSTEM_CHANGE_NUMBER
------------------------
                 1070544

为什么这2个查到的SCN不一样呢

Yong Huang

最初由 rchsh 发布
[B]SQL> select checkpoint_change#  from v$database;

CHECKPOINT_CHANGE#
------------------
           1069220

SQL> select dbms_flashback.get_system_change_number from dual;

GET_SYSTEM_CHANGE_NUMBER
------------------------
                 1070544

为什么这2个查到的SCN不一样呢 [/B]

They're two different things. Checkpoint SCN should always lag behind the latest SCN in the database. If you alter system checkpoint and immediately run your two queries, they should differ less, but still checkpoint SCN is smaller.

Yong Huang

localhosts

最初由 Yong Huang 发布
[B]

They're two different things. Checkpoint SCN should always lag behind the latest SCN in the database. If you alter system checkpoint and immediately run your two queries, they should differ less, but still checkpoint SCN is smaller.

Yong Huang [/B]

[php]
sys@DBAP01> select max(ktuxescnw*power(2,32)+ktuxescnb) from x$ktuxe;

MAX(KTUXESCNW*POWER(2,32)+KTUX
------------------------------
                      52211024

已用时间:  00: 00: 00.00
sys@DBAP01> alter system checkpoint;

系统已更改。

已用时间:  00: 00: 00.06
sys@DBAP01> select CHECKPOINT_CHANGE# from v$database;

CHECKPOINT_CHANGE#
------------------
          52211055

已用时间:  00: 00: 00.00
sys@DBAP01> select max(ktuxescnw*power(2,32)+ktuxescnb) from x$ktuxe;

MAX(KTUXESCNW*POWER(2,32)+KTUX
------------------------------
                      52211053

已用时间:  00: 00: 00.00
sys@DBAP01>
[/php]

biti_rainy

最初由 localhosts 发布
[B]

[php]
sys@DBAP01> select max(ktuxescnw*power(2,32)+ktuxescnb) from x$ktuxe;

MAX(KTUXESCNW*POWER(2,32)+KTUX
------------------------------
                      52211024

已用时间:  00: 00: 00.00
sys@DBAP01> alter system checkpoint;

系统已更改。

已用时间:  00: 00: 00.06
sys@DBAP01> select CHECKPOINT_CHANGE# from v$database;

CHECKPOINT_CHANGE#
------------------
          52211055

已用时间:  00: 00: 00.00
sys@DBAP01> select max(ktuxescnw*power(2,32)+ktuxescnb) from x$ktuxe;

MAX(KTUXESCNW*POWER(2,32)+KTUX
------------------------------
                      52211053

已用时间:  00: 00: 00.00
sys@DBAP01>
[/php] [/B]

x$ktuxe 计算出来的是已经结束的最新的事务的commit scn ，所以可小于当前系统scn。 检查点 scn 自然也小于当前系统scn。 但是 检查点scn 和 x$ktuxe 计算出来的大小却倚赖于 系统状况了。

Yong Huang

最初由 biti_rainy 发布
[B]
x$ktuxe 计算出来的是已经结束的最新的事务的commit scn ，所以可小于当前系统scn。 检查点 scn 自然也小于当前系统scn。 但是 检查点scn 和 x$ktuxe 计算出来的大小却倚赖于 系统状况了。 [/B]

You're saying both the x$ktuxe SCN and the checkpoint SCN are smaller than the "current system SCN", whatever that is, so there's no relationship between these two (x$ktuxe SCN and checkpoint SCN). Correct? The question becomes, what is exactly that "current system SCN"? Is there any way to find that?

Yong Huang

biti_rainy

最初由 Yong Huang 发布
[B]

You're saying both the x$ktuxe SCN and the checkpoint SCN are smaller than the "current system SCN", whatever that is, so there's no relationship between these two (x$ktuxe SCN and checkpoint SCN). Correct? The question becomes, what is exactly that "current system SCN"? Is there any way to find that?

Yong Huang [/B]

current  scn 是 系统当前所产生的最大 scn ，可能是当前未结束事务所产生的scn。 在9i 的 dbms_flashback.get_system_number 可以得到这个值，这个值应该是大于等于  x$ktuxe SCN  (这个view 记录的是 当前数据库结束事务的最大scn)