xfs
[Top] [All Lists]

Re: XFS write speed drop

To: Helge Hafting <helge.hafting@xxxxxxxxxxxxx>
Subject: Re: XFS write speed drop
From: Jan Engelhardt <jengelh@xxxxxxxxxxxxxxx>
Date: Fri, 26 May 2006 10:51:19 +0200 (MEST)
Cc: Nathan Scott <nathans@xxxxxxx>, linux-xfs@xxxxxxxxxxx, Linux Kernel Mailing List <linux-kernel@xxxxxxxxxxxxxxx>
In-reply-to: <4476AC3B.5030704@aitel.hist.no>
References: <Pine.LNX.4.61.0605190047430.23455@yvahk01.tjqt.qr> <20060522105326.A212600@wobbly.melbourne.sgi.com> <Pine.LNX.4.61.0605221308290.11108@yvahk01.tjqt.qr> <20060523084309.A239136@wobbly.melbourne.sgi.com> <Pine.LNX.4.61.0605231517330.25086@yvahk01.tjqt.qr> <4476AC3B.5030704@aitel.hist.no>
Sender: xfs-bounce@xxxxxxxxxxx
>> Well as mentioned, -o nobarrier solved it, and that's it. I do not
>> actually need barriers (or an UPS, to poke on another thread), because
>> power failures are rather rare in Germany.
>
> Then there are mistakes like someone stepping on the power
> cord, pulling out the wrong one, drilling holes in the wall,
> there are kernel crashes, there is lightning,
> there is the possibility that some
> component inside the box just breaks.
>
A mathematician has fear of flying because there could possibly be a 
terrorist bomb inside the plane. What does he do?
Taking an bomb with himself, because the probablity that there are two 
bombs inside a plane is lower than one.

What I am wanting to say is that the factors you mentioned just do not happen
here, at this one - home - box.


Jan Engelhardt
-- 


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