Date: Sat, 16 Feb 2008 11:37:21 +0000 From: Thomas Hurst <tom.hurst@clara.net> To: Dag-Erling Sm?rgrav <des@des.no> Cc: Ian FREISLICH <ianf@clue.co.za>, Ken Smith <kensmith@cse.Buffalo.EDU>, freebsd-current@freebsd.org Subject: Re: FreeBSD 7.0-RC2 Available Message-ID: <20080216113721.GA55702@voi.aagh.net> In-Reply-To: <86r6fdx0tf.fsf@ds4.des.no> References: <rea-fbsd@codelabs.ru> <HNc5KTwAOlChZd/l%2BN1vPPiVFRE@3SQePivZkcJXerr/O1l2SLu1NoU> <E1JPD3j-0000wk-6u@clue.co.za> <msJHdNXRu5fGqwNIwehl3Qsvvmg@L/B2HsSNkA3O1ZRIaMxnTL95W%2Bo> <86r6fdx0tf.fsf@ds4.des.no>
next in thread | previous in thread | raw e-mail | index | archive | help
* Dag-Erling Sm?rgrav (des@des.no) wrote:
> Not cost-effective? What is the "street price" of 16 GB disk space
> these days? About the same as a couple of Big Macs?
That's roughly half of a common 36G SCSI drive, and still a fairly
significant chunk of a 73G one. Granted, you probably don't get all
that many high-memory systems with just one or two dinky disks.
For us, our systems with the most memory have little need for much
storage or IO; they have a pair of mirrored 73G SAS drives and 20G of
memory; they currently run with 2G of swap, which if they ever have to
use, will make them useless. Blowing 30% of available local storage on
swap doesn't really make sense; we're much more likely to have 20G
application cores than kernel ones.
Speaking of, it'd be really nice if you could interrupt the generation
of coredumps; big ones take a while. Perhaps the dump loop could check
kern.coredump every few thousand pages and exit early if it's 0?
--
Thomas 'Freaky' Hurst
http://hur.st/
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20080216113721.GA55702>