Date: Mon, 19 Sep 2011 12:07:00 -0700 (PDT) From: Jason Usher <jusher71@yahoo.com> To: freebsd-fs@freebsd.org Subject: Re: ZFS obn FreeBSD hardware model for 48 or 96 sata3 paths... Message-ID: <1316459220.35419.YahooMailClassic@web121209.mail.ne1.yahoo.com> In-Reply-To: <alpine.GSO.2.01.1109171451500.1973@freddy.simplesystems.org>
next in thread | previous in thread | raw e-mail | index | archive | help
=0A=0A--- On Sat, 9/17/11, Bob Friesenhahn <bfriesen@simple.dallas.tx.us> w= rote:=0A=0A> 150KB is a relatively small file size given that the=0A> defau= lt zfs blocksize is 128KB.=A0 With so many files you=0A> should definitely = max out RAM first before using SSDs as a=0A> l2arc.=A0 It is important to r= ecognize that the ARC cache=0A> is not populated until data has been read.= =A0 The cache=0A> does not help unless the data has been accessed several= =0A> times. You will want to make sure that all metada and=0A> directories = are cached in RAM.=A0 Depending on how the=0A> files are used/accessed you = might even want to intentionally=0A> disable caching of file data.=0A=0A=0A= How does one make sure that all metadata and directories are cached in RAM?= Just run a 'find' on the filesystem, or a 'du' during the least busy time= of day ? Or is there a more elegant, or more direct way to read all of th= at in ?=0A=0AFurther, if this (small files, lots of them) dataset benefits = a lot from having the metadata and dirs read in, how can I KEEP that data i= n the cache, but not cache the file data (as you suggest, above) ?=0A=0ACan= I explicitly cache metadata/dirs in RAM, and cache file data in L2ARC ?=0A= =0A=0A=0A> Are the writes expected to be synchronous writes, or are=0A> the= y asynchronous?=A0 Are the writes expected to be=0A> primarily sequential (= e.g. whole file), or is data=0A> accessed/updated in place?=0A=0A=0AIt's a = mix, I'm afraid.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1316459220.35419.YahooMailClassic>