Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 11 Aug 2000 18:37:01 -0400
From:      Mike Tancsa <mike@sentex.net>
To:        Mike Smith <msmith@FreeBSD.ORG>
Cc:        hardware@FreeBSD.ORG
Subject:   Re: 3ware and RAID0 vs RAID 1 
Message-ID:  <4.2.2.20000811183002.0456fb88@mail.sentex.net>
In-Reply-To: <200008112235.PAA00819@mass.osd.bsdi.com>
References:  <Your message of "Fri, 11 Aug 2000 10:07:00 EDT." <4.3.2.7.0.20000811092428.00da32d0@marble.sentex.ca>

next in thread | previous in thread | raw e-mail | index | archive | help
At 03:35 PM 8/11/2000 -0700, Mike Smith wrote:
> > >Ok, thanks.  It *definitely* shouldn't "just hang" at all.
> >
> > OK, some more quick testing... I was wrong about the hanging, I just did
> > not wait long enough.
>
>That's a relief.  I take it that you didn't look at the disks?


Yes, sorry to get you worried needlessly.  I did blast the disks 
individually and performance was as expected identical.  And as RAID0, it 
all works quite well both reading *and* writing...

> >
> > Doing a 500MB write on a regular drive on this machine takes about 3min,
> > 30seconds. On the RAID1, it takes 19 min ??
>
>Can you be more specific about "a 500M write"?  Are you referring to the
>bonnie run?  I wouldn't expect it to be 5x as slow.  Note that the
>rewrite case suggests that your test is still too small; I typically use
>1 or 10GB on a 128MB machine.


Yes, the times I reference are for bonnie to complete a
bonnie -s 500

The thing I dont understand is why the RAID0 writes would be very fast, but 
RAID1 5 times slower.  But then again, I really dont know anything about 
hardware so I suppose it can make perfect sense :-)



> > Do you see similar results with your controllers and disks in RAID1 ?
>
>Bonnie results on these cards tend to suggest that they're fairly slow
>writers.  It's 3ware also observed this, but commented that with two
>writers they found more bandwidth.  I suspect that the write overhead may
>be fairly high, leading to a lot of dead time while writing (the read
>overhead is also an issue I guess).  I haven't done a lot of timing tests,
>as I'm typically more concerned with functionality and reliability.

Understood.  Is there anyone else out there who can confirm these results ? 
It (for me anyways) makes this card pretty useless for RAID1. A shame if 
its the case.

         ---Mike
--------------------------------------------------------------------
Mike Tancsa,                          	          tel +1 519 651 3400
Network Administration,     			  mike@sentex.net
Sentex Communications                 		  www.sentex.net
Cambridge, Ontario Canada			  www.sentex.net/mike



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hardware" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4.2.2.20000811183002.0456fb88>