[Top] [All Lists]

Re: nfs/local performance with software raid5 and xfs and SMP

To: Simon Matter <simon.matter@xxxxxxxxxxxxxxxx>
Subject: Re: nfs/local performance with software raid5 and xfs and SMP
From: Seth Mos <knuffie@xxxxxxxxx>
Date: Thu, 19 Jul 2001 11:45:26 +0200
Cc: Tru Huynh <tru@xxxxxxxxxx>, "linux-xfs@xxxxxxxxxxx" <linux-xfs@xxxxxxxxxxx>
In-reply-to: <3B56A79F.EA8658D2@xxxxxxxxxxxxxxxx>
References: <3B55E49E.5B4B6CEF@xxxxxxxxxx> <>
Sender: owner-linux-xfs@xxxxxxxxxxx
At 11:25 19-7-2001 +0200, Simon Matter wrote:
Seth Mos schrieb:
> >4) try SMP+raid5 on local IDE HD to rule out 3ware issue
> >- I just have a single HD on the primary IDE controller
> >   I can only repartion it by destroying the scratch partition.
> >   Will 3 logical partitions be ok for a minimal raid5 device?
> Yes that will do

But performance will be bad anyway with all partitions on the
same physikal disk.

That does _not_ matter. It is neccesary to see if the problem lies with the 3ware controller or not. This is not a matter of performance. Read the thread before giving hints.

> >5) change the chunk-size from 64 to 256/512.
> >- I guess that is fine tuning :)

I was playing around with chunk-size but it didn't really change
something for me.

It does if you have large files. The disks can read bigger chunks at once and improve throughput at the cost of latency. But with a 450GB you will probably want enough speed so you can actually fill it within 2 days.

Every program has two purposes one for which
it was written and another for which it wasn't
I use the last kind.

<Prev in Thread] Current Thread [Next in Thread>