]]
Reply To: [EMAIL PROTECTED]
Sent: Friday, October 25, 2002 9:13 PM
To: Multiple recipients of list ORACLE-L
Subject: Re: can clustering help INSERTS ?
Rahul,
It does no good to speculate; let's work with facts...
What wait-events are occurring in the sessions running
DENNIS WILLIAMS wrote:
Stephane
You mentioned each additional index costs about 2.5 times the cost of
inserting into a non-indexed table. I just wanted to point out that Kevin
Loney has done some performance tests involving the number of indexes. I
don't know if he has published these
Naveen -
He provided figures, and they are on the handout that is somewhere in my
office. Maybe I'll run across it someday, or even better, maybe he'll
publish his results. Here are a few more details that I posted to this list
earlier.
Kevin Loney (author of Oracle DBA Handbook) has
Gorman[SMTP:[EMAIL PROTECTED]]
Reply To: [EMAIL PROTECTED]
Sent: Friday, October 25, 2002 9:13 PM
To: Multiple recipients of list ORACLE-L
Subject: Re: can clustering help INSERTS ?
Rahul,
It does no good to speculate; let's work with facts...
What wait
Subject: Re: can clustering help INSERTS ?
Rahul,
It does no good to speculate; let's work with facts...
What wait-events are occurring in the sessions running the
INSERTs?
If
you
can locate the sessions in the V$SESSION view, then use the value
I believe Kevin gave that presentation at OpenWorld -- either last year
or the year before. His paper is available for download on the TUSC
site, as he works for TUSC.
--- DENNIS WILLIAMS [EMAIL PROTECTED] wrote:
Naveen -
He provided figures, and they are on the handout that is somewhere
Rahul
--
From: Jared Still[SMTP:[EMAIL PROTECTED]]
Sent: Sunday, October 27, 2002 8:27 AM
To: [EMAIL PROTECTED]; Rahul
Subject: Re: can clustering help INSERTS ?
*no* waits?
How is this possible?
Is intantaneous computing now a reality?
Sorry
:[EMAIL PROTECTED]]
Reply To: [EMAIL PROTECTED]
Sent: Friday, October 25, 2002 9:13 PM
To: Multiple recipients of list ORACLE-L
Subject: Re: can clustering help INSERTS ?
Rahul,
It does no good to speculate; let's work with facts...
What wait-events are occurring
--
From: Tim Gorman[SMTP:[EMAIL PROTECTED]]
Reply To: [EMAIL PROTECTED]
Sent: Friday, October 25, 2002 9:13 PM
To: Multiple recipients of list ORACLE-L
Subject: Re: can clustering help INSERTS ?
Rahul,
It does no good to speculate; let's work
Dennis,
Did he publish any figures? I mean, it seems common-sense that adding the
first index will hurt but adding 11th index to a table won't hurt that much.
As stephane pointed out, cost of an index is 2.5 times more than the cost of
insert in a non-indexed table.
So assuming cost is 1,
]
Sent: Friday, October 25, 2002 9:13 PM
To: Multiple recipients of list ORACLE-L
Subject: Re: can clustering help INSERTS ?
Rahul,
It does no good to speculate; let's work with facts...
What wait-events are occurring in the sessions running the INSERTs? If
you
can locate
events while the process
is
running..
there are no indexes on the tables
--
From: Tim Gorman[SMTP:[EMAIL PROTECTED]]
Reply To: [EMAIL PROTECTED]
Sent: Friday, October 25, 2002 9:13 PM
To: Multiple recipients of list ORACLE-L
Subject: Re: can clustering help INSERTS
on the tables
--
From: Tim Gorman[SMTP:[EMAIL PROTECTED]]
Reply To: [EMAIL PROTECTED]
Sent: Friday, October 25, 2002 9:13 PM
To: Multiple recipients of list ORACLE-L
Subject:Re: can clustering help INSERTS ?
Rahul,
It does no good to speculate
List,
i have two heavily inserted tables, the structures are same.
currently these tables reside on separate disks, can i increase the
performance
of inserts if i create these tables in a cluster ? as a cluster would force
the rows of both the tables
to be physically close on the disk !
Rahul wrote:
List,
i have two heavily inserted tables, the structures are same.
currently these tables reside on separate disks, can i increase the
performance
of inserts if i create these tables in a cluster ? as a cluster would force
the rows of both the tables
to be physically close on
Rahul,
It does no good to speculate; let's work with facts...
What wait-events are occurring in the sessions running the INSERTs? If you
can locate the sessions in the V$SESSION view, then use the value in the
column SID to locate associated rows in the V$SESSION_EVENT view, sorting by
the
Stephane
You mentioned each additional index costs about 2.5 times the cost of
inserting into a non-indexed table. I just wanted to point out that Kevin
Loney has done some performance tests involving the number of indexes. I
don't know if he has published these anywhere. In a nutshell, the
17 matches
Mail list logo