Re: row cache objects latch and shared_pool_size

From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Sun, 6 Apr 2003 11:31:17 +0100
Message-ID: <b6ovju$qr8$1$8300dec7@news.demon.co.uk>

I won't add to Howard's response. But taking a different tack I'd like to point out that the row cache objects latch is a parent latch (in your version of Oracle) with about 22 child latches that correspond to the different "parameter" parent types in the v$rowcache view.

If you are certain that there is significant contention for this latch you should check whether it is restricted to particular elements of the rowcache. This may give you some clues.

Example: if you are using pipelined functions heavily, then there are three entries in v$rowcache that will get hammered - off the top of my head, these are:

    dc_users
    dc_object_ids
    dc_global_oids

If your problem happens to be is in this area, then sizing of pools is irrelevant.

(In fact, noting your other post about reloads reports in v$sql, there is a hint that perhaps your problem is "shared pool too small" - excessive reloads would lead to heavy thrashing on the 'commoner' bits of the row cache)

--
Regards

Jonathan Lewis
http://www.jlcomp.demon.co.uk

  The educated person is not the person
  who can answer the questions, but the
  person who can question the answers -- T. Schick Jr


One-day tutorials:
http://www.jlcomp.demon.co.uk/tutorial.html

____UK_______April 8th
____UK_______April 22nd
____Denmark__May 21-23rd
____USA_(FL)_May 2nd

Three-day seminar:
see http://www.jlcomp.demon.co.uk/seminar.html
____UK_(Manchester)_May
____Estonia___June (provisional)
____USA_(CA, TX)_August

The Co-operative Oracle Users' FAQ
http://www.jlcomp.demon.co.uk/faq/ind_faq.html


"NetComrade" <andreyNSPAM_at_bookexchange.net> wrote in message
news:3e89cf77.3113443725_at_nyc.news.speakeasy.net...


> Row cache objects latch:

> ........................

>

>         In order to reduce contention for this latch, we need to

>         tune the data dictionary cache. In Oracle7 this basically

> means

>         increasing the size of the shared pool (SHARED_POOL_SIZE)

>

>

> Could it be the other way around? Could the shared_pool_size be too

> large? how can one tell?

> .......

> We use Oracle 8.1.7.4 on Solaris 2.7 boxes

> remove NSPAM to email

Received on Sun Apr 06 2003 - 05:31:17 CDT