Date: Tue, 4 Feb 2003 17:21:47 -0800 (PST) From: Julian Elischer <julian@elischer.org> To: Poul-Henning Kamp <phk@freebsd.org> Cc: current@freebsd.org Subject: Re: Preview: GEOMs statistics code. Message-ID: <Pine.BSF.4.21.0302041709240.84482-100000@InterJet.elischer.org> In-Reply-To: <25779.1044395087@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 4 Feb 2003, Poul-Henning Kamp wrote: > > > > Now, discussion time: > > The timestamps cost something to make, and my plan was to only collect > them while a monitoring program is running. (Is this a good idea ?) > probably it is a good idea though it does lead to the possibility of a heisenbug in some corner cases.. "It doesn't have the problem when I look at the stats" Turning on and off the stamping could lead to corner cases, but I'm sure that overall it's worthwhile assuming all methods of gathering stats can notify the kernel that they are doing so.. > In difference from the devstat framework which measures how big a > percentage of the time a drive has one or more outstanding requests, > I think that measuring the responstime is a much more useful metric. > (comments, input, references welcome) The devstat figure is probably more useful in working out what the total capacity of a device is and what percentage of that figure is being used. Command Latencies are however also of great interest so I can see a point for having both sets of numbers. > Comments, ideas, suggestions. Progress is good. and I agree with the way this is going.. > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.21.0302041709240.84482-100000>