undo segment behavior

FAQ – Automatic Undo Management (AUM) / System Managed Undo (SMU) (Doc ID 461480.1) 

What is the Fast Ramp-up Routine ?

The concept of Fast Ramp-Up indicates the following:

"In previous versions, when the instance was restarted or when you switched undo tablespaces, the ramp-up time for the system to grow the number of online undo segments could be several minutes.  Generally, this time delay was unacceptable.

After an instance startup or a switch of undo tablespaces in Oracle Database 10g, the database decides how many segments to place online, based on existing data stored in the the AWR.

The algorithm to compute the # of undo segments to online for "Fast Ramp Up" is:

Has the instance been up for more than 7 days?

    Yes- use max(maxconcurrency) from v$undostat

     No- Is this he first time this function is called? 

        Yes - Is there select_workload_repository function (SWRF) snaphot data? 

                     No- online minimal # of undo segs

                     Yes- attempt to get the max(maxconcurrency) from wrh$_undostat for the last 7 days. 

                          If we cannot find that information, attempt to get the max(rbs cnt) from wrh$_rollstat

                                  for the last 7 days.  

                  Store the value in internal variable.

          No- use an internal variable.

猜你喜欢

转载自buralin.iteye.com/blog/2041261