Date: Mon, 11 Sep 2000 21:59:39 +0300 (EEST) From: Esko Petteri Matinsola <esko@kapi.net> To: hackers@freebsd.org Subject: Vinum RAID5 rawio performance Message-ID: <Pine.BSF.4.10.10009112158340.19651-100000@sputnik>
next in thread | raw e-mail | index | archive | help
On Sat, 9 Sep 2000, Greg Lehey wrote:
> Please don't use bonnie. It's not measuring what you think it is.
OK, I used rawio as you suggested.
> I don't know how you'd find anything at vinumvm.org which tells you
> what to expect from bonnie.
No I did not. I found that I should be expecting about 25% of the write
performance of striping than with raid5.
> Well, using master and slave together isn't fine in my book.
I ordered additional IDE-card three weeks ago, but haven't recieved
it so I thought I should get started with these two on the motherboard.
So, I tested with rawio.
Here's rawio -a -p 8 /dev/ad0e:
Random read Sequential read Random write Sequential write
ID K/sec /sec K/sec /sec K/sec /sec K/sec /sec
ad0e 1355.1 85 5335.0 326 1530.4 95 6416.9 392
With the the following configuration of vinum:
drive drive0 device /dev/ad0e
drive drive1 device /dev/ad1e
drive drive2 device /dev/ad2e
drive drive3 device /dev/ad3e
volume raid0
plex org striped 256k
sd length 512m drive drive0
sd length 512m drive drive1
sd length 512m drive drive2
sd length 512m drive drive3
I ran rawio -a -p 8 /dev/vinum/raid0 and got:
Random read Sequential read Random write Sequential write
ID K/sec /sec K/sec /sec K/sec /sec K/sec /sec
aid0 3593.0 221 17807.7 1087 11551.0 711 17503.5 1068
And with the the following configuration of vinum:
drive drive0 device /dev/ad0e
drive drive1 device /dev/ad1e
drive drive2 device /dev/ad2e
drive drive3 device /dev/ad3e
volume raid10
plex org striped 256k
sd length 512m drive drive0
sd length 512m drive drive1
plex org striped 256k
sd length 512m drive drive2
sd length 512m drive drive3
I ran rawio -a -p 8 /dev/vinum/raid10 and got:
Random read Sequential read Random write Sequential write
ID K/sec /sec K/sec /sec K/sec /sec K/sec /sec
aid10 3727.4 230 20024.5 1222 5727.5 358 10073.2 615
And with the the following configuration of vinum:
drive drive0 device /dev/ad0e
drive drive1 device /dev/ad1e
drive drive2 device /dev/ad2e
drive drive3 device /dev/ad3e
volume raid5
plex org raid5 256k
sd length 512m drive drive0
sd length 512m drive drive1
sd length 512m drive drive2
sd length 512m drive drive3
I ran rawio -a -p 8 /dev/vinum/raid5 and got:
Random read Sequential read Random write Sequential write
ID K/sec /sec K/sec /sec K/sec /sec K/sec /sec
aid5 3569.9 221 19001.7 1160 1169.8 73 1105.1 67
And with the the following configuration of vinum:
drive drive0 device /dev/ad0e
drive drive1 device /dev/ad1e
drive drive2 device /dev/ad2e
drive drive3 device /dev/ad3e
volume raid0
plex org striped 256k
sd length 512m drive drive0
# sd length 512m drive drive1
sd length 512m drive drive2
# sd length 512m drive drive3
I ran rawio -a -p 8 /dev/vinum/raid0 and got:
Random read Sequential read Random write Sequential write
ID K/sec /sec K/sec /sec K/sec /sec K/sec /sec
aid0 3756.6 233 16859.6 1029 5972.7 368 17213.3 1051
And with the the following configuration of vinum:
drive drive0 device /dev/ad0e
drive drive1 device /dev/ad1e
drive drive2 device /dev/ad2e
drive drive3 device /dev/ad3e
volume raid0
plex org striped 256k
sd length 512m drive drive0
sd length 512m drive drive1
# sd length 512m drive drive2
# sd length 512m drive drive3
I ran rawio -a -p 8 /dev/vinum/raid0 and got:
Random read Sequential read Random write Sequential write
ID K/sec /sec K/sec /sec K/sec /sec K/sec /sec
aid0 3756.6 233 16859.6 1029 5972.7 368 17213.3 1051
So, is this just what I should get or is there something I could do to
that raid5 write performance ? Based on these figures, it's nowhere near
25% what I expect. I expect to get somewhat better performance when I get
that IDE-card when I can put every drive as master on its own channel.
Please comment.
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.10.10009112158340.19651-100000>