Re: db file scattered/sequential read wait

From: Daniel Fink <Daniel.Fink_at_Sun.COM>
Date: Wed, 16 Jun 2004 12:34:11 -0600
Message-id: <40D092A3.1050504@sun.com>

Wolfgang,

Excellent word of caution. I use a global temp table that I load up with the output from dba_segments, slap an index on it and perform the query. The whole process takes a fraction of what it takes to apply this type of predicate to the dba_segments query.

Regards,
Daniel Fink

Wolfgang Breitling wrote:
> Just a word of caution. You don't want to run this query every time you want to
> find out the object for a db file read. That alone could bring your system to
> its knees, especially if you have LMTs - and if not, why not? (those who still
> manage Oracle 7 and lower systems are excused).
> I run something like that in the wee hours of the morning when there is spare
> capacity and load the response into a table indexed by file# and block#. This
> way I don't go after the catalog tables and the response is immediate. Sure, I
> risk missing the occasional block (hasn't happened yet) if the object went into
> a new extent after I ran my report.
>
> Quoting "Mark W. Farnham" <mwf_at_rsiz.com>:
>
>

>>I think you want a minus one in the second term.
>>
>>Let's see, if block_id = 1 and you read 8 blocks, then 1...9, is *not* what
>>you want, so yeah, (block_id+blocks-1)
>>would tighten up that predicate.
>>
>>mwf
>>

>
>

Please see the official ORACLE-L FAQ: http://www.orafaq.com

To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line.

--
Archives are at http://www.freelists.org/archives/oracle-l/
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------

Received on Wed Jun 16 2004 - 13:31:30 CDT