Re: troubleshooting I/O performance bottleneck

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Mon, Apr 27, 2009 at 10:35:58PM +0200, Emmanuel Dreyfus wrote:
> [...]
> Namei         Sys-cache     Proc-cache     26 ioapic1 pin 0        ndcpy
>     Calls     hits    %     hits     %        ioapic0 pin 16       fltcp
>         6        6  100                   809 ioapic0 pin 19       zfod
>                                               vcpu0 ioapic0        cow
> Disks:   md0   cd0   wd0   wd1 raid0      150 vcpu0 clock       64 fmin
>  seeks                                        vcpu0 xenbus      85 ftarg
>  xfers               417   429   423          vcpu0 xbd3.3         itarg
>  bytes               13M   13M   13M          vcpu0 xvif3.0   3851 wired
>  %busy              89.0  58.0  98.0          vcpu0 xbd13.3        pdfre

Writes to disks are almost as fast a with domU: 13MB/s here while it was
11MB/s from the domU.

> [...]
> 
> > It could also come from partition layouts; modern disks are known to be 
> > faster
> > when writting at the beggining of the drive than at the end.
> 
> That would explain a 200% hit?

probably not that much, but it's noticeable.
Did you have something else doing I/O when you did your test, like
a parity rewrite ? systat shows raid0 almost 100% busy in both cases, 
and for almost the same data transfer rate.

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--