Re: data file sizes

There is the issue that on Unix File System, inode locking takes place - readers block writers and writers block readers albeit fairly briefly.

Consequently for high-activity systems, especially if there is a fair proportion of writing, you spread the collisions by having a higher number of smaller files.

The effect is usually not terribly significant. However when you start running Parallel Execution slaves, which leaves multiple processes reading and writing to the temporary tablespace to do a large parallel hash join, say, then you may notice a difference. If you make any special cases, at all, make the temporary tablespace a collection of N smaller files, where N is around 2 x your preferred degree of parallelism

--
Jonathan Lewis
Yet another Oracle-related web site:  http://www.jlcomp.demon.co.uk

Practical Oracle 8i:  Building Efficient Databases

Publishers:  Addison-Wesley
See a first review at:
http://www.ixora.com.au/resources/index.htm#practical_8i
More reviews at: http://www.jlcomp.demon.co.uk/book_rev.html



gdas_at_my-deja.com wrote in message <93nehn$mbc$1_at_nnrp1.deja.com>...


>Hi,

>

>Are there any performance considerations to take into account when

>sizing index and data tablespace datafiles?  Do fewer but larger

>datafiles have any performance benefits to More but smaller datafiles?

>Does the fact that the files are on RAID 5 change the answer.

>

>I've seen some posts here that support more but smaller files due to

>increased manageability, but I was curious if there were any

>performance implications.

>

>Thanks,

>Gavin

>

>

>Sent via Deja.com

>http://www.deja.com/