High US - contention enqueue / rowcache lock while trying to online a NEW undo segment (Doc ID 1577753.1)

Last updated on AUGUST 05, 2016

Applies to:

Oracle Database - Enterprise Edition - Version 10.2.0.4 to 11.1.0.7 [Release 10.2 to 11.1]
Information in this document applies to any platform.
***Checked for relevance on 06-Aug-2016***

Symptoms

High US - contention enqueue , latch: row cache objects and rowcache locks may be seen while trying to online a new undo segment.

When transaction concurrency increases sharply (e.g. following a connection storm or high DML activity), Automatic UNDO Management (AUM) may need to bring online existing undo segments, or create additional undo segments.

The process of creating additional undo segments or onlining existing undo segments appears to be slower in 11.1.0.6 than in 10.2.0.3 .

The Test Case Done for the Bug Investigation creates 1782 undo segments in first run on freshly created db and onlines 1782 undo segments in second run .

The Following are the Test Case results , which shows slight increase in the Database Time , slight increase in US enqueue while onlining ,

and Huge increase in latch: row cache objects wait time in both scenarios (creating or onlining undo segments)

 
Wait Event time create undo segments10.2.0.311.1.0.6
DB time 5538 6375
DB CPU 164 151
US - contention 4636 2587
row cache objects latch 96 3434
row cache lock wait time 569 169
latch free wait time 58 118

 

 
Wait Event time onlining undo segments10.2.0.311.1.0.6
DB time 2019 2190
DB CPU 58 53
US - contention 1141 1532
row cache objects latch 98 226
row cache lock 667 313
latch free 28 178

 

Note: All times are in seconds.

 

Cause

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms