Re: Slower performance after enabling async io on Oracle Linux
From: Mladen Gogala <gogala.mladen_at_gmail.com>
Date: Wed, 16 May 2012 12:41:32 +0000 (UTC)
Message-ID: <pan.2012.05.16.12.41.32_at_gmail.com>
On Wed, 16 May 2012 01:32:51 -0700, Matthias Hoys wrote:
>
> So, you see, turning off async io drastically improves performance in my
> situation (409 versus 716 seconds). I'm going to leave it off for now,
> and maybe create a service request on MOS to see if they can help. I
> couldn't find a lot on OTN.
>
> Matthias Hoys
Date: Wed, 16 May 2012 12:41:32 +0000 (UTC)
Message-ID: <pan.2012.05.16.12.41.32_at_gmail.com>
On Wed, 16 May 2012 01:32:51 -0700, Matthias Hoys wrote:
>
> So, you see, turning off async io drastically improves performance in my
> situation (409 versus 716 seconds). I'm going to leave it off for now,
> and maybe create a service request on MOS to see if they can help. I
> couldn't find a lot on OTN.
>
> Matthias Hoys
Yup, with the "async" option, you have time waited for sequential read as 242.64 and for scattered read as 4.93 while without async the time for single block IO is 84.01 and for scattered read 7.05. The worst problem is for single block IO, while scattered read is even slightly faster than the synchronous version. If I am allowed to guess, the problem is caching. You may want to turn off direct IO and just leave async IO and to set the read ahead on the block devices, with blockdev --setra. As I have previously said, async I/O on Linux leaves a lot to be desired, especially when the network is involved.
-- http://mgogala.byethost5.comReceived on Wed May 16 2012 - 07:41:32 CDT