Re: Checking the rebuildability of an index

From: Connor McDonald <hamcdc_at_yahoo.co.uk>
Date: Tue, 06 Aug 2002 01:09:38 -0800
Message-ID: <F001.004ABE50.20020806010938@fatcity.com>

I'll throw a contentious comment...I would say that using the figures in index_stats to rebuild indexes is a waste of time. Consider

1. an index that has 30% deleted space:

This could be due to everyday usage patterns. Do you really want to rebuild it - since by tomorrow (or the next day) it will be back to 30% deleted space.

Alternatively, the 30% might be due to a one-off archive of data. If that deleted space is *entire* blocks that have been freed up, then maybe you do not want to rebuild since those blocks will be re-used anyway. Then again, if you don't plan on doing any more dml, maybe you do want to rebuild it to speed up its use in queries. Then again, rebuilding it could pack more into a block and introduce contention. Then again, rebuilding it could spread things out and make spanning parts of the index actually slower

b) an index may have *no* deleted space

You might want to rebuild with a very low pctfree if there is no more dml to be issued against it. Then again, you might want to rebuild with a higher pctfree to avoid contention on some of the blocks

Except in DSS or partition-style environments where drop/rebuild is mandated for load/space/etc reasons, I would say there are very few times when the decision on whether to rebuild an index can be made solely on the stats or some other automated decision.

hth
connor

• "Jesse, Rich" <Rich.Jesse_at_qtiworld.com> wrote: > Hi all,
  >
  > On our ERP DB, 8.1.7.2.0 on HP/UX, I'm finally able
  > to implement an
  > automated index rebuild job. I've got two
  > questions:
  >
  > First, I've planned to use the gauge of b*-tree
  > HEIGHT and "DEL_LF_ROWS /
  > LF_ROWS) > 20%" from the SYS.INDEX_STATS view to
  > determine what indexes to
  > rebuild. I've read in "Gaja 101" that the leaf
  > block density should also be
  > considered. Is this LF_ROWS / LF_BLKS (from
  > INDEX_STATS)? If so, are there
  > any guidelines as to what the density should be? Or
  > is it only the
  > percentage decrease over time that matters? And if
  > it is, what percentage
  > decrease over what period of time should be
  > considered? This would seem to
  > be very installation-dependant, but I'm just hoping
  > for general guidelines
  > for a starting point so I'm not rebuilding our
  > largest indexes weekly if
  > it's not needed.
  >
  > Second, how does one determine if there's enough
  > room to REBUILD ONLINE an
  > index? Other than comparing MAX(BYTES) to the
  > INITIAL clause of the index,
  > that is. I've looked in Metalink, "Gaja 101", and
  > several websites, but
  > nothing jumped out at me.
  >
  > TIA!
  >
  > Rich Jesse System/Database
  > Administrator
  > Rich.Jesse_at_qtiworld.com Quad/Tech
  > International, Sussex, WI USA
  > --
  > Please see the official ORACLE-L FAQ:
  > http://www.orafaq.com
  > --
  > Author: Jesse, Rich
  > INET: Rich.Jesse_at_qtiworld.com
  >
  > Fat City Network Services -- (858) 538-5051 FAX:
  > (858) 538-5051
  > San Diego, California -- Public Internet
  > access / Mailing Lists
  >
  

  ---

  
  > To REMOVE yourself from this mailing list, send an
  > E-Mail message
  > to: ListGuru_at_fatcity.com (note EXACT spelling of
  > 'ListGuru') and in
  > the message BODY, include a line containing: UNSUB
  > ORACLE-L
  > (or the name of mailing list you want to be removed
  > from). You may
  > also send the HELP command for other information
  > (like subscribing).
  

  ---

Connor McDonald
http://www.oracledba.co.uk
http://www.oaktable.net

"Remember amateurs built the ark - Professionals built the Titanic"

---

Do You Yahoo!?
Everything you'll ever need on one web page from News and Sport to Email and Music Charts http://uk.my.yahoo.com

-- 
Please see the official ORACLE-L FAQ: http://www.orafaq.com
-- 
Author: =?iso-8859-1?q?Connor=20McDonald?=
  INET: hamcdc_at_yahoo.co.uk

Fat City Network Services    -- (858) 538-5051  FAX: (858) 538-5051
San Diego, California        -- Public Internet access / Mailing Lists
--------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).

Received on Tue Aug 06 2002 - 04:09:38 CDT