cancel
Showing results for 
Search instead for 
Did you mean: 

Oracle block size

SOLVED
Go to solution
Tim Medford
Valued Contributor

Oracle block size

We are installing Oracle 8.1.7 on a new HP L3000 server running HPUX 11.0.

We plan to use an 8K blocksize in Oracle (ie db_block_size = 8192).

I want to set the OS up to use 8k blocks as well. Is this simply the vps_pagesize kernel parameter or is there more to it?

Thanks, Tim
4 REPLIES
Bruce Regittko_1
Esteemed Contributor

Re: Oracle block size

Hi,

vps_pagesize sets the minimum page size for virtual memory, not disk I/O. I believe you set Oracle's extent size from within Oracle. Check out the following threads for more on Oracle block size:

http://forums.itrc.hp.com/cm/QuestionAnswer/1,1150,0xc9eb6c96588ad4118fef0090279cd0f9,00.html

http://my1.itrc.hp.com/cm/QuestionAnswer/1,1150,0xc9eb6c96588ad4118fef0090279cd0f9,00.html

http://forums.itrc.hp.com/cm/QuestionAnswer/1,1150,0x9bf26c96588ad4118fef0090279cd0f9,00.html

--Bruce

www.stratech.com/training
Dave Wherry
Esteemed Contributor

Re: Oracle block size

You'll want to set the block size when you create the file systems for your database:
newfs -F vxfs -b 8192 -o largefiles /dev/vg99/lvol99
Steven Sim Kok Leong
Honored Contributor

Re: Oracle block size

Hi,

The default block size for HFS is 8 kb. If you are intending to use HFS, you do not need to specify the block size explicitly ie. -b option is not required to be specified.

However, HFS is very much slower than JFS in database I/O operations such as tablespace creation. If you are using an Oracle block size of 8 kb, in general, you can improve I/O performance (especially in tablespace creation) by specifying an OS filesystem block size of 1 kb instead of 8 kb.

Hope this helps. Regards.

Steven Sim
Brainbench MVP for Unix Admin
http://www.brainbench.com
Email: steven@beepz.com. Homepage: https://www.beepz.com
Dan Hetzel
Honored Contributor
Solution

Re: Oracle block size

Hi Tim,

Adjusting the filesystem block size to match Oracle block size should give you the best possible database performance.

It's at the time of filesystem creation that you'll set the pagesize:
# newfs -F vxfs -b 8192 -o largefiles /dev/vg??/lvol?

vps_pagesize is unrelated to the disk I/O pagesize as it is the minimum pagesize for virtual memory.

Best regards,

Dan
Everybody knows at least one thing worth sharing -- mailto:dan.hetzel@wildcroft.com