Date: Thu, 22 Dec 2011 16:59:56 +0100 From: Ivan Voras <ivoras@freebsd.org> To: freebsd-geom@freebsd.org Subject: Re: RFC: GEOM MULTIPATH Rewrite Message-ID: <jcvk5s$jbu$1@dough.gmane.org> In-Reply-To: <7bf62b8af857803e03363b81ead54484@feld.me> References: <7bf62b8af857803e03363b81ead54484@feld.me>
next in thread | previous in thread | raw e-mail | index | archive | help
On 22/12/2011 03:05, Mark Felder wrote: > more benchmarks and I might switch the scheduler to 4BSD because this > box will be doing nothing but I/O and network traffic and my research > indicates ULE is not optimal in that type of a workload. Thanks for testing the new gmultipath patches! I just want to say that you should test 4BSD/ULE very carefully before making any conclusions because ULE fails in very, very specific edge cases which are not very common. I've done this 4BSD/ULE test just for laughs on an network IO-intensive workload on a 24-core machine and the performance with 4BSD was laughably slow (which is expected as it is of a very old design).
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?jcvk5s$jbu$1>