On Thu, Jan 29, 2009 at 06:05,  <[email protected]> wrote:
> 1) Does anybody have a similar configuration running? Are there some special
> configuration "tricks", i.e. jumbo frames (how to setup on the solaris,
> cisco and ESX/HBA side).
Have you disabled Nagle's algorithm?  This can lead to significant
performance improvements with iSCSI.  What storage is backing your
targets (e.g., a zpool made of one disk?  a dozen disks?  7k disks or
15k?)? What hardware is running Solaris?  What does local disk
performance look like according to bonnie++ (or something similar)?

> 2) How can I measure the perfomance of such a configuration in a objective,
> comparable way?
Run your expected workload against it.  Anything else is mostly for
brag rights; it doesn't matter if you can read really fast
sequentially if your application doesn't do that.

Will
_______________________________________________
storage-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/storage-discuss

Reply via email to