Re: Enable 32K Block in 8K Block DB

From: Daniel Morgan <damorgan_at_x.washington.edu>
Date: Sun, 28 Mar 2004 23:25:35 -0800
Message-ID: <1080545112.770029@yasure>

Howard J. Rogers wrote:
> "Daniel Morgan" <damorgan_at_x.washington.edu> wrote in message
> news:1080537417.493976_at_yasure...
>
>

>>>HJR
>>
>>I agree with one caveat. IIRC AIX 5L allows for other block sizes
>>up to 256K. You must check with your SA to determine the block size
>>before installing Oracle.
>>
>>Reference: http://hpcf.nersc.gov/vendor_docs/ibm/gpfs/am3admst09.html
>>
>>So while I believe you and Steve are correct on the Oracle basics. It is
>>possible that the original 8K block size is correct Ok: Highly unlikely
>>but possible.

>
>
> No, it's a valuable caveat which I've mentioned before, and which it doesn't
> do any harm to mention again. The rule that says Linux and AIX = 4K and
> HP-UX and Solaris = 8K depends *entirely* on accepting the defaults when
> creating file systems. It is possible to change the *o/s block size* at the
> time of creating the file system, at which point you have another degree of
> freedom as it were.
>
> But I would caution against using a 256K o/s block size (if such a thing
> were possible) because it isn't possible in Oracle (64K being the biggest
> one I ever heard of, and then only for some MVS, iirc -32K being the
> "normal" maximum). And if you can't persuade Oracle to use 256K blocks, then
> you've just broken Adams' first law.
>
> Regards
> HJR
Agreed. Here's a link to the 256K block size with 5L.

Search for "block size" in the document.

The block size change is required by 5L to achieve large file systems so it may be that an SA makes that choice without any understanding of the needs of Oracle that will later be laid down on it by a DBA.

I'm not advocating larger block sizes. Only reminding everyone that they may exist and that their UNIX SA could have created them wihtout their knowledge.

-- 

Daniel Morgan
http://www.outreach.washington.edu/ext/certificates/oad/oad_crs.asp
http://www.outreach.washington.edu/ext/certificates/aoa/aoa_crs.asp
damorgan_at_x.washington.edu
(replace 'x' with a 'u' to reply)

Received on Mon Mar 29 2004 - 01:25:35 CST