Re: ORACLE on Linux - IO bottleneck

Wyvern wrote:
>
> I didnīt know what was an IO scheduler. Iīve been looking for and
> what Iīve seen is that kernel 2.6 implements one but kernel 2.4
> doesnīt, isnīt it?
>

In the "old" times of 2.4 kernel it was called I/O elevator.

>
> OK. Reads should not be secuential in a OLTP database, Iīm wrong? So,
> I suppose reads are not secuential as you say ...
>

That's can be true at RDBMS level but the I/O at OS level is different. Oracle asks for blocks of 8K so for the OS (on raw) they are 16 contiguos 512 byte reads or writes that could be merged.

Personally I would try the same workload on a filesystem (on this configuration with and without direct I/O).

> hdparm says readahead in that device (and in all other) is 120. I
> didnīt know what this parameter meant but Iīve been reading about and
> what Iīve understood is that for non-secuential reads readahead should
> be smaller than in a secuentially big file reading, isnīt it?
>

On most OLTP it can be disabled reducing io wait (and maybe service time). Of course, test before implementing it!

Regards

-- 
Fabrizio Magni

fabrizio.magni_at_mycontinent.com

replace mycontinent with europe