Date: Mon, 26 Nov 2007 09:59:27 -0500 From: Chris Marlatt <cmarlatt@rxsec.com> To: =?UTF-8?B?6Z+T5a625qiZIEJpbGwgSGFja2Vy?= <askbill@conducive.net> Cc: freebsd-stable@freebsd.org Subject: Re: SAS5IR performance issue with Dell 860 Message-ID: <474ADF4F.7090303@rxsec.com> In-Reply-To: <474988C0.4070909@conducive.net> References: <47496BE1.8080206@modula.no> <474988C0.4070909@conducive.net>
next in thread | previous in thread | raw e-mail | index | archive | help
韓家標 Bill Hacker wrote: > > You haven't indicated what drives are on that controller, or how (RAID?) > arranged.,, what sort of on-drive or on-controller cahce and policy. > > Nor how you measured the '..performs better', which a single <anything> > can often do compared to several of the possible RAID configurations. > > Otherwise, those are not actually 'bad' numbers for sustained I/O - > especially if anything-at-all is going on at the same time that needs to > intervene and move the r/w heads 'elsewhere' - however briefly. > > More info? > > Bill > The issue has already been confirmed and a "fix"/hack committed so there's little point busting his balls for lack of information in his email. It seems like he was only requesting any updated information on the issue, not a full diagnosis. The Dell PE-860 only comes with 2 drives, so its going to be independent drives, raid 1, or raid 0. Frankly if independent drives or raid 0 are performing that poorly as well it would seem the issue is even worse than previously indicated. As for the cache policy, seeing as that's the fix, your question seems a little odd. Best regards, Chris
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?474ADF4F.7090303>