[Top] [All Lists]

Re: Q: Filesystem block sizes available?

Subject: Re: Q: Filesystem block sizes available?
From: "D. Stimits" <stimits@xxxxxxxxxx>
Date: Tue, 13 Nov 2001 09:50:45 -0700
Cc: XFS Mailing list <linux-xfs@xxxxxxxxxxx>
References: <Pine.LNX.4.30.0111131556430.1160-100000@xxxxxxxxxxxxxxxxx> <3BF14697.F90270CD@xxxxxxxxxx> <20011113101659.E9701@xxxxxxx>
Reply-to: stimits@xxxxxxxxxx
Sender: owner-linux-xfs@xxxxxxxxxxx
Nathan Straz wrote:
> On Tue, Nov 13, 2001 at 09:13:11AM -0700, D. Stimits wrote:
> > FYI, I've managed to get just over 30 MB/sec sustained (read; write
> > isn't too far off; both XFS and ext2 do well with average size files) on
> > a good 10K rpm U160 scsi drive (64 bit 66 MHz bus) if conditions are
> > right...no raid at all. With 5 drives on RAID 0 and your hardware specs,
> > can't imagine not getting far better than you get now. It sure sounds
> > like something is wrong.
> Why?  IIRC, RAID0 is just concatenated disks.  You're not going to get a
> speed boost out of it.  Not until RAID1 (mirrored) will you start seeing
> a speed boost.

I do not refer to logical append of disks as RAID0. RAID0 should be
stripped, and very good performance, splitting the load over all the
disks evenly (with no redundancy if anything fails). RAID5 will probably
make write performance worse, and read performance much better. RAID
definitions seem to have all kinds of different terminology depending on
who you talk to, but the linux software RAID docs I've seen always refer
to RAID0 as striped and high performance, not as merely a logical

D. Stimits, stimits@xxxxxxxxxx

> --
> Nate Straz                                              nstraz@xxxxxxx
> sgi, inc                                           http://www.sgi.com/
> Linux Test Project                                  http://ltp.sf.net/

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