Date: Fri, 29 Dec 2006 15:39:33 +0100 From: Ivan Voras <ivoras@fer.hr> To: freebsd-geom@freebsd.org Subject: Re: gstripe performance scaling with many disks Message-ID: <en39bb$lq$1@sea.gmane.org> In-Reply-To: <20061229142638.45844.qmail@web30315.mail.mud.yahoo.com> References: <en380a$tl7$1@sea.gmane.org> <20061229142638.45844.qmail@web30315.mail.mud.yahoo.com>
next in thread | previous in thread | raw e-mail | index | archive | help
[-- Attachment #1 --] R. B. Riddick wrote: > --- Ivan Voras <ivoras@fer.hr> wrote: >> Vasil Dimov wrote: >>> 17 drives: >>> dT: 0.563s w: 0.500s >>> L(q) ops/s r/s kBps ms/r w/s kBps ms/w %busy Name >>> 5 185 178 848 18.9 0 0 0.0 88.8| ggate100 >> Ok, so ops/s is increasing but it's not conclusive. Does anyone know how >> to get KB/t stats for synthetic geom devices? >> > I would say: > transaction size = (volume per second) / (operations per second) > E. g.: > 848KBps/(178r/s) = 4.76KB/op > or > 1028KBps/(116r/s) = 8.86KB/op Well "doh", nowadays I don't know how to calculate anymore :) Hmmm, it's strange - it's 2x lower than it should be. For 8 drives, each should do 16 KB/s but they do 8 KB/s. Maybe something's chopping the requests at 64 KB each? [-- Attachment #2 --] -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.4 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFFlSirldnAQVacBcgRApn/AJ4wsi/Ot3cfgXAurLI9b+rx0D8XFACbBG12 O5D9ZDdEAq7nfGKsOIpnQAE= =4zvK -----END PGP SIGNATURE-----
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?en39bb$lq$1>