Date: Fri, 12 Feb 1999 13:09:49 +1030 From: Greg Lehey <grog@lemis.com> To: Brad Knowles <blk@skynet.be> Cc: freebsd-questions@FreeBSD.ORG Subject: Re: Software RAID-5 comparison: FreeBSD 3.0 vs. Solaris 2.6/x86... Message-ID: <19990212130949.I491@lemis.com> In-Reply-To: <19990211094929.028342@relay.skynet.be>; from Brad Knowles on Thu, Feb 11, 1999 at 09:49:29AM %2B0100 References: <19990211094011.J71962@freebie.lemis.com> <19990211094929.028342@relay.skynet.be>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thursday, 11 February 1999 at 9:49:29 +0100, Brad Knowles wrote: > On Thu, Feb 11, 1999, Greg Lehey <grog@lemis.com> wrote: > >>> the DiskSuite tools let me grow the filesystem while it's online >>> (and it doesn't have to be blown away and re-newfs'ed), >> >> That's in the pipeline, but it's not there yet. > > It's also not strictly necessasry. I don't anticipate adding any > more disks to this system any time soon, but it would be nice if the > feature were there. Yes, it's definitely on the "to-do" list. >>> and there are some pretty cool performance monitoring tools >>> available from Rich Petit (SymbEL 3.x) and Adrian Crockcraft (author >>> of the book _Sun Performance and Tuning: Java and the Internet_). >> >> And we don't have much in that area either. Vinum maintains a number >> of statistics, but that's about all. > > Actually, I was thinking primarily of system performance tuning > tools, not just filesystem tools. In addition to wanting to know how > long the wait queue is for I/O commands to a particular disk or volume, > how many I/O operations I'm getting per second, etc... I'd also like to > know things like how much RAM is allocated to the kernel, how much is > shared between some processes, how much is private for some processes, > how much is allocated to buffer cache, etc.... Most of this stuff is outside Vinum. Some of it is there, and I'm trying to find a way to measure Vinum throughput (average I/O time and standard deviation, for example, per Vinum object). >> What I've been seeing with Vinum is that it loses about 2-5% >> performance on concatenated volumes, and gains up to 45% on a two-way >> striped volume. > > That's concatenated and striped, but what about RAID-5? Assume the same as striped for reading RAID-5. Writing is approximately 60% of the speed of reading, unless a drive dies, in which case it gets a little faster. But I'm not going to publish any ``exact'' numbers now, because they're not as exact as they look. Greg -- See complete headers for address, home page and phone numbers finger grog@lemis.com for PGP public key To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990212130949.I491>