Job Seekers   Employers
dbapool: Authors | Submissions | Contact Us
   Forgot password? | Sign up
  Home   Discussion Forum   Articles   Interview Questions   FAQs   Scripts   Rewards   Analyzer   White Papers   Blog   Certification   Downloads   Tools
 

Topic: Performance & Tuning >> Hit Ratio of Database Buffer Cache

Disclaimer: The purpose of all dbapool forums including OCP and Other Oracle Certification forums is to help each other with specific issues but not to share dump and copyrighted exam content, materials or intellectual property. You may review the entire Oracle Certification Program Candidate Agreement online Here.



  


 Title: Hit Ratio of Database Buffer Cache
 Navi  Posted: Jun 13, 2007 08:09:00 AM

 Total Post: 60
 Joined: Jun, 2007






 Hi,

Anyone please answer this question
Why 99%+ Database Buffer Cache hit ratio is not OK?

Regards,
Naveen

 Ratnaker
Posted: Jul 03, 2007 02:23:36 AM  

 Total Post: 136
 Joined: Apr, 2007






 
hi Naveen,

This is the percentage of requests
for a particular block which are satisfied within the cache without the need for physical IO.
Although historically known as one of the most important statistics, the buffer cache hit ratio
can sometimes be misleading. A high (e.g. 99%) cache hit ratio normally indicates the cache
is adequately sized, however this may not be the case in all circumstances. For example,
frequently executed SQL statements which repeatedly refer to a small number of buffers via
indexed lookups can skew the buffer gets statistic. When these blocks are read, they are
placed at the most recently used (MRU) end of the buffer cache; iterative access to these
blocks can artificially inflate the cache hit ratio. This makes tuning the buffer cache a
challenging activity.
On some sites, it is possible to identify a too small buffer cache by the appearance of the
‘write complete waits’ event, which indicates that hot blocks (i.e. blocks which are still being
modified) are aging out of the cache while they are still needed; check the Wait events
section for evidence of this event.
Alternatively, a lower buffer cache hit ratio does not necessarily mean the cache is too small;
it may be that (potentially valid) full table scans are artificially reducing what is otherwise a
good hit ratio.

 Time Zone: EDT

  




Forum Rules & Description


Who Can Read The Forum? Any registered user or guest
Who Can Post New Topics? Any registered user
Who Can Post Replies? Any registered user




 








Get FREE Magazines

Top 10 Forum User

  Murtuja Khokhar814
  Mohammed Taj694
  Jayanta Sur479
  Vigyan Kaushik386
  positive fanatic361
  Gitesh Trivedi322
  Gopu Gopi239
  neeraj sharma228
  Ramesh Jois226
  snehalatha p169






oracle Mag



  About Us Advertise Terms of Use Privacy Newsletters Contact Us    

Home   Discussion Forum   FAQs  Articles  Jobs   Newsletters  Directory  Downloads 

Our Premium Sponsor
Confio Software