On Dec 29, 2009, at 11:26 AM, Brad wrote:

@relling
"For small, random read IOPS the performance of a single, top-level
vdev is
performance = performance of a disk * (N / (N - P))
             133 * 12/(12-1)=
             133 * 12/11

where,
N = number of disks in the vdev
P = number of parity devices in the vdev"

performance of a disk => Is this a rough estimate of the disk's IOP?


"For example, using 5 disks @ 100 IOPS we get something like:
2-disk mirror: 200 IOPS
4+1 raidz: 125 IOPS
3+2 raidz2: 167 IOPS
2+3 raidz3: 250 IOPS"

So if the rated iops on our disks is @133 iops
133 * 12/(12-1) = 145

11+1 raidz: 145 IOPS?????

If that's the rate for a 11+1 raidz vdev, then why is iostat showing
about 700 combined IOPS (reads/writes) per disk?

Because the model is for small, random read IOPS
over the full size of the disk. What you are seeing is
caching and seek optimization at work (a good thing).
But, AFAIK,  there are no decent performance models
which take caching into account. In most cases, storage
is sized based on empirical studies.
 -- richard

r/s w/s Mr/s Mw/s wait actv wsvc_t asvc_t %w %b device
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 c0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 c0t0d0
1402.2 7805.3 2.7 36.2 0.2 54.9 0.0 6.0 0 940 c1
10.8 1.0 0.1 0.0 0.0 0.1 0.0 7.0 0 7 c1t0d0
117.1 640.7 0.2 1.8 0.0 4.5 0.0 5.9 1 76 c1t1d0
116.9 638.2 0.2 1.7 0.0 4.6 0.0 6.1 1 78 c1t2d0
116.4 639.1 0.2 1.8 0.0 4.6 0.0 6.0 1 78 c1t3d0
116.6 638.1 0.2 1.7 0.0 4.6 0.0 6.1 1 77 c1t4d0
113.2 638.0 0.2 1.8 0.0 4.6 0.0 6.1 1 77 c1t5d0
116.6 635.3 0.2 1.7 0.0 4.5 0.0 6.0 1 76 c1t6d0
116.2 637.8 0.2 1.8 0.0 4.7 0.0 6.2 1 79 c1t7d0
115.3 636.7 0.2 1.8 0.0 4.4 0.0 5.8 1 77 c1t8d0
115.4 637.8 0.2 1.8 0.0 4.5 0.0 5.9 1 77 c1t9d0
114.8 635.0 0.2 1.8 0.0 4.3 0.0 5.7 1 76 c1t10d0
114.9 639.9 0.2 1.8 0.0 4.7 0.0 6.2 1 78 c1t11d0
115.1 638.7 0.2 1.8 0.0 4.4 0.0 5.9 1 77 c1t12d0
1.6 140.0 0.0 15.1 0.0 0.6 0.0 4.4 0 8 c1t13d0
1.3 9.1 0.0 0.1 0.0 0.0 0.0 1.0 0 0 c1t14d0
--
This message posted from opensolaris.org
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss

_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss

Reply via email to