Date: Sun, 08 Aug 2010 19:55:40 +0200 From: Antonio Vieiro <antonio@antonioshome.net> To: freebsd-questions@freebsd.org Subject: Re: File System Performance on FreeBSD Message-ID: <4C5EEF9C.30804@antonioshome.net> In-Reply-To: <20100808182211.000029f0@unknown> References: <4C5EB94F.509@gmail.com> <4C5EE5CA.2090706@potentialtech.com> <20100808182211.000029f0@unknown>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi, I heard that Linux filesystems were not reliable because of some bad way of doing caching or something like that. For a study on Linux FS reliability see [1] by Toshiba guys. It seems Linux was upset on this about one year ago [2]. Quoting: "Torvalds, for one, didn't seem too excited about the delayed synchronization. He writes on the mailing list, "Doesn't at least ext4 default to the insane model of 'data is less important than metadata, and it doesn't get journalled'? And ext3 with 'data=writeback' does the same, no? Both of which are -- as far as I can tell -- total brain damage." I don't mind if a filesystem is very fast: I want it to be reliable first. I wonder if that Phoronix test suite checks for reliability first or not. Cheers, Antonio [1] elinux.org/images/2/26/Evaluation_of_Data_Reliability-ELC2010.pdf [2] http://www.linux-magazine.com/Online/News/Linus-Torvalds-Upset-over-Ext3-and-Ext4 On 08/08/2010 19:22, Bruce Cran wrote: > On Sun, 08 Aug 2010 13:13:46 -0400 > Bill Moran<wmoran@potentialtech.com> wrote: > >> To someone technical who might be looking to investigate the results >> with an eye toward fixing them, it's useless. > > Anyone can download the Phoronix Test Suite though, so it should be > fairly easy to check if the results are valid at least. >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4C5EEF9C.30804>