Index tablespace has the most wasted space and yes, there are several
indices in those tablespaces. I need to keep it available or I'd just
drop and recreate the entire index.
good to know I wasn't hallucinating!
--- DENNIS WILLIAMS <[EMAIL PROTECTED]> wrote:
> Rachel - Right you are, as "Stop D
Hi!
>- Are you trying to keep the index available to users while you're
> rebuilding? I'm assuming this is the reason you are looking at rebuilding
> the index twice. Or is it because rebuilding an index probably won't cause
a
> large sort?
Sort is still needed, even when rebuilding. It's jus
Rachel - Right you are, as "Stop Defragmenting . . . " points out, there are
several types of fragmentation.
- Is it your table or your index that contains wasted space?
- Are there multiple objects in each tablespace or just a single object?
- Are you trying to keep the index available t
uld say "ah well doesn't cost much lets buy
> > another
> > bookcase".
> >
> > In summary Niall's 2nd rule states that "data always goes in but
> > never
> > comes out". It's parkinsons law for databases
> >
>
t; comes out". It's parkinsons law for databases
>
> Niall
>
> > -----Original Message-
> > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
> > Behalf Of Stephen Lee
> > Sent: 13 June 2003 18:45
> > To: Multiple recipients of list ORACLE-L
&g
es that "data always goes in but never
comes out". It's parkinsons law for databases
Niall
> -Original Message-
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
> Behalf Of Stephen Lee
> Sent: 13 June 2003 18:45
> To: Multiple recipients of list ORACLE-L
>
I think those guys work here now :-)
-Original Message-
Sent: Friday, June 13, 2003 1:45 PM
To: Multiple recipients of list ORACLE-L
That's one thing good about the databases here. Tablespace
fragmentation is
rarely a problem. Most of the database here are a Database Roach Motel:
"D
That's one thing good about the databases here. Tablespace fragmentation is
rarely a problem. Most of the database here are a Database Roach Motel:
"Data checks in. It doesn't check out." Somehow, the data purge part of
the application -- that they intended to put in "one of these days" -- nev
of list ORACLE-L
there was a debate on here 2 weeks ago where it was concluded that until you
get to thousands of extents it just doesnt matter how many you have.
>
> From: "VIVEK_SHARMA" <[EMAIL PROTECTED]>
> Date: 2003/06/13 Fri AM 06:39:36 EDT
> To: Multiple rec
Title: RE: RE: Fragmentation ?
Depends ... who you ask ...
If you ask Microsoft
1. you are fragmented if you have at-least _one_ non-windows server in your corporation
2. Your thinking is fragmented if you are even _considering_ LINUX
If you ask SCO
1. You are fragmented if you use
t ORACLE-L <[EMAIL PROTECTED]>
> Subject: RE: Fragmentation ?
>
> Dennis , List
>
> What may be the OTHER forms of fragmentation ?
>
> What Number of Extents may be considered Critical warranting RE-Organization for
> Manually Sized Objects existing in LMTs ?
>
Dennis , List
What may be the OTHER forms of fragmentation ?
What Number of Extents may be considered Critical warranting RE-Organization for
Manually Sized Objects existing in LMTs ?
Thanks for the great paper . Had read it previously though .
Thanks
-Original Message-
Sent: Wedne
Vivek
Make sure you've read "How to Stop Defragmenting and Start Living"
http://metalink.oracle.com/cgi-bin/cr/getfile_cr.cgi?239049
The authors point out that uniform extents stop fragmentation at the
tablespace level. However they point out that there are other forms of
fragmentation.
Dennis
Seema,
If you COPY the datafiles at the OS level then you are getting an
exact image of the datafile. The same for the RMAN copy command. If you
use SQL "insert ,, select " type of copying data from one server to
another then the data will fill the extents and eliminate fragmentation.
It will not
Hi Seema,
define what you mean when you say 6 are fragmented.
John
[EMAIL PROTECTED] wrote:
> Hi
> If one tablespace has 10 tables and 6 are fragmented table in that
> particular tablespace.If I use COPY command to move data from one
> server to another then is COPy command bring data with f
Shaibal is right on target. If you're going to go to the trouble to
defragment then you may as well switch to uniform extent sizing so you don't
have the problem in the future.
If you have the room you can do it in stages. First create an index
tablespace and rebuild all your indexes in that.
Title: Fragmentation of data dictionary
Helmut,
X$
tables are NEVER stored in the disk. THey are just memory structures in the SGA
and the contents are zeroed (are reset) when you shutdown the database. THey
will not cause data dictionary fragmentation.
Best Regards,K GopalakrishnanBan
Seema,
From your posting, it is not clear to me about the content of the mentioned tablespace. If it contains all of one users schema you can just take export for that particular schema. The step when you are recreating the tablespace, create it with pct increase 0 and do create initial and next e
|
| To: Multiple recipients of list ORACLE-L|
| <[EMAIL PROTECTED]> |
| cc: (bcc: Mike Hately/ETECH) |
| Subject: Re: Fragment
st thought it was funny that the recommendation was
> now going back to the original extent recommendations :)
>
> Rachel
>
> >From: [EMAIL PROTECTED]
> >Reply-To: [EMAIL PROTECTED]
> >To: Multiple recipients of list ORACLE-L <[EMAIL PROTECTED]>
> >Subj
TECTED]
>To: Multiple recipients of list ORACLE-L <[EMAIL PROTECTED]>
>Subject: Re: Fragmentation & Locally Managed Tablespaces
>Date: Thu, 28 Jun 2001 04:35:35 -0800
>
>
>
>Rachel,
>That maximum isn't a hard limit. Oracle themselves impose no limit of this
>kind.
Paul Drake [SMTP:[EMAIL PROTECTED]]
> > Sent: Thursday, June 28, 2001 1:46 PM
> > To: Multiple recipients of list ORACLE-L
> > Subject:Re: Fragmentation & Locally Managed
> Tablespaces
> >
> > VIVEK_SHARMA wrote:
> > >
> > > Is Fragment
Multiple recipients of list ORACLE-L|
| <[EMAIL PROTECTED]> |
| cc: (bcc: Mike Hately/ETECH) |
| Subject: Re: Fragment
so we are going back to the Oracle specified maximum extents for a
particular blocksize?
>From: Paul Drake <[EMAIL PROTECTED]>
>Reply-To: [EMAIL PROTECTED]
>To: Multiple recipients of list ORACLE-L <[EMAIL PROTECTED]>
>Subject: Re: Fragmentation & Locally Manage
Thanks indeed for the Article's Excerpt , Paul
Our db_block_size=8K
> -Original Message-
> From: Paul Drake [SMTP:[EMAIL PROTECTED]]
> Sent: Thursday, June 28, 2001 1:46 PM
> To: Multiple recipients of list ORACLE-L
> Subject: Re: Fragmentation & Loc
VIVEK_SHARMA wrote:
>
> Is Fragmentation of Objects meaningless in Locally managed Tablespaces
> Assuming EXTENTS is 3,000 for Some of the Objects
>
> Or Do the Objects need to be DE-Fragmented using exp/imp ?
>
Vivek,
funny, I was just looking into this tonight.
what is your block size?
I
3000 extents seems a lot to me.
I usually kept the number of extents for a segment
below the number of blocks kept in the extent map, the
extent map depends on the OS block size. For example
if the block size is 8K you should not have more than
504 extents.
--- VIVEK_SHARMA <[EMAIL PROTECTED
27 matches
Mail list logo