Matthias Hoys wrote:
> "DA Morgan" <damorgan_at_psoug.org> wrote in message
> news:1117116008.895087_at_yasure...
>
>>Matthias Hoys wrote:
>>
>>>"DA Morgan" <damorgan_at_psoug.org> wrote in message
>>>news:1117062970.747334_at_yasure...
>>>
>>>
>>>>Matthias Hoys wrote:
>>>>
>>>>
>>>>
>>>>>I meant that there are a lot of "holes" in our tablespace map because of
>>>>>the large differences between the initial/next extent size of all the
>>>>>objects. I don't know if we will see a performance gain after the
>>>>>migration but we will definitely win diskspace (30-40%) by "compacting"
>>>>>the data.
>>>>
>>>>The subject says 10g. That means you are using LMT. So before you go
>>>>fixing "holes" and claiming that initial and next extent sizes are
>>>>different ... explain how you have done whatever it is you've done.
>>>>What is your storage strategy? Because the above statements just don't
>>>>compute.
>>>>--
>>>
>>>
>>>No no, I mean in our CURRENT db which is Oracle 8i with DM tablespaces.
>>>So I suspect we will gain diskspace by moving to LMT on 10g.
>>
>>Not if you load the data properly. As I asked ... what is your storage
>>strategy?
>>--
>
>
> Well it's a database I "inherited", and there was no storage strategy at all
> ... There were about 20 developers each using their own extent specs ...
> 80K, 128K, 64K, different sizes for initial/next extents, some
> exports/imports done with COMPRESS=Y ...
Then I would suggest that you stop work immediately, convene in a
conference room, and agree on a storage strategy. Anarchy is rarely
an effective solution unless you are being paid by the hour and have
on hand an adequate supply of pain relievers.
--
Daniel A. Morgan
Relational theory is not something that is simply a nice-to-have.
http://www.psoug.org
damorgan_at_x.washington.edu
(replace x with u to respond)
Received on Thu May 26 2005 - 14:53:45 CDT
