joel garry wrote:
> DA Morgan wrote:
>> joel garry wrote:
>>> DA Morgan wrote:
>>>> Frank van Bortel wrote:
>>>>> DA Morgan schreef:
>>>>>> That is about as horrible a definition as I can imagine. I can't think
>>>>>> of a single justification for either the 64K extents or a pctincrease
>>>>>> that is anything except 1.
>>>>>>
>>>>> And what would be wrong with 64kB extents?
>>>>>
>>>>> I'm off to the Chimay :D Sayonara!
>>>> The problem with 64K extents is one Howard Rogers ranted about many
>>>> years ago here at c.d.o. The fact that it doesn't match any operating
>>>> system. Or at least not Windows or any variant of UNIX or Linux.
>>>>
>>> Block size isn't extents, is it?
>> You are correct I wasn't reading closely though I suspect that may still
>> be what the OP intended. Either way that too is a bad idea. If a table
>> is going to grow to 1GB ... do you really want to do this 64K at a time?
>> I sure don't even without dictionary managed tablespaces.
>
> I guess this must be "Let's hit dizwell day."
> http://www.dizwell.com/prod/node/66
>
> jg
> --
> @home.com is bogus.
> And Sept. 19 is http://www.talklikeapirateday.com/wordpress/
I agree with Howard ... but it isn't about the number of extents to
which I refer. Rather why update that bitmap a zillion times when a few
will do. Keep in mind that while that extent is being added, update
bitmap, update seg$, update a bunch of view, you are gaining nothing but
wait.
--
Daniel Morgan
Puget Sound Oracle Users Group
Received on Fri Sep 08 2006 - 20:28:18 CDT
