This is the same documentation that says SYSTEM can be created locally
managed, which is only true of 9i release 2. I wouldn't therefore pay it too
much attention on this issue; it is simply wrong if it implies 8K is the
only permitted block size on NT. I do 16K every time.
Regards
HJR
--
-----------------------------------------------
Resources for Oracle : http://www.hjrdba.com
===============================
"Ed Stevens" <spamdump_at_nospam.noway.nohow> wrote in message
news:3cadd7b0.92777827_at_ausnews.austin.ibm.com...
> Restating my original post, with clarification . . . .
>
> My systems are Ora 8.0.5 or Ora 8.1.7 on NT
>
>
> In looking for information on an entirely unrelated Oracle issue, I came
across
> the following statement in the Oracle8i Administrator's Guide
> Release 2 (8.1.6) for Windows NT, Chapter 1, "Oracle8i Differences between
> Windows NT and UNIX"
>
> <quote regarding the NT implementation>
> The Oracle block size is 8K. The maximum number of blocks per data file is
4
> million. The maximum number of data files per database depends on block
size.
> </quote>
>
> Hmm. "The Oracle block size _IS_ 8K. " (emphasis mine). This would
> indicate to me that the Oracle block size in NT is a fixed value,
rendering the
> init parameter DB_BLOCK_SIZE useless. Based on earlier readings (much,
MUCH
> earlier, back at 7.3 when we were charged with bringing up our first
Oracle db
> without even
> knowing how to spell Oracle) we have that parm set to 4096 in all of our
> databases.
>
> Hmm. Then, after asserting that the block size *is* 8k, it goes on to say
"The
> . . . number . . . depends on block size.", implying that block size is
*not* a
> fixed value.
>
> I guess it really does matter what the meaning of "is" is . . . .
>
>
> Comments?
Received on Fri Apr 05 2002 - 11:47:06 CST
