gstripe performance scaling with many disks

[Ivan Voras]

Eric Anderson wrote:

> Maybe this should be tried with md* devices just to remove the network
> component from this mix.

Network was introduced here to do bandwidth limiting so to get
predictable results, but you're right - network could, at least in
theory, be the issue here.

I remember trying to do such a benchmark with USB sticks, which also
have high latency, and getting similar scaling problems because the
requests were split to small blocks.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 250 bytes
Desc: OpenPGP digital signature
Url : http://lists.freebsd.org/pipermail/freebsd-geom/attachments/20061230/c3205ffe/signature.pgp