Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!newsroom.utas.edu.au!munnari.OZ.AU!spool.mu.edu!howland.reston.ans.net!newsfeed.internetmci.com!news.kei.com!nntp.coast.net!swidir.switch.ch!scsing.switch.ch!news.dfn.de!gina.zfn.uni-bremen.de!marvin.pc-labor.uni-bremen.de!news.uni-stuttgart.de!news.rhrz.uni-bonn.de!RRZ.Uni-Koeln.DE!zpr.uni-koeln.de!se From: se@ZPR.Uni-Koeln.DE (Stefan Esser) Newsgroups: comp.unix.bsd.freebsd.misc,comp.os.linux.development.system Subject: Re: The better (more suitable)Unix?? FreeBSD or Linux Date: 17 Feb 1996 19:44:06 GMT Organization: Institute for Mathematics, University of Cologne, Germany Lines: 60 Sender: se@Sysiphos (Stefan Esser) Message-ID: <4g5b66$q6k@news.rrz.uni-koeln.de> References: <4er9hp$5ng@orb.direct.ca> <jlemonDMtpAz.3rJ@netcom.com> <4g0l6o$gcl@park.uvsc.edu> <4g2213$e3f@cebaf4.cebaf.gov> <strenDMxJ0H.L2s@netcom.com> NNTP-Posting-Host: sysiphos.mi.uni-koeln.de To: stren@netcom.com (Sam Trenholme) Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:14529 comp.os.linux.development.system:18248 In article <strenDMxJ0H.L2s@netcom.com>, stren@netcom.com (Sam Trenholme) writes: |> 3) The fact that FreeBSD, though stuck to a 4k block size (yuck, from a |> news feed perspective-- 512-byte blocks would not be too small, but |> ext2's default 1024 is acceptable), seems to have 1k "fregments". DOes |> this mean 4 1k postings will only take up one block? Hmm, seems you got something confused ... The BSD FFS uses a block size of 8KB by default, but can also work with 4KB or 16KB (and any higher power of two in principle). But it achieves a file size granularity of one eigth a block, by using a bit mask to specify which of eight possible fragments are containing data for this file in the last block. Since a (non zero size) file needs at least one fragment to store it's data, you will be able to fit 4 1K postings into a 4KB block. But it would even be possible to have 8 512byte control files in those 4KB :) |> If not, I think not supporting a 1k block size, just so that a 4 gig file |> can be referenced with only dual-indirection, is silly. Since I deal with Well, as I said above, you can have a 512byte fragment size (but are limited to 1GB file size using dual-indirection, then). With 1KB you actually get 4GB with dual-indirection, but the current file size limit (with triple- indirection) is 1000GB, and could easily be raised to 8000GB, in case anybody has an urgent need for it :) |> 64meg+ files so rarely, I would rather have the smaller block size, and |> have less slack on my news archive, than have better huge-file |> performance. Smaller than 512byte would require smaller sector sizes and would COST capacity ... |> Of course, with FreeBSD adding ext2 support, hopefully this issue will |> become moot. Ideally, FreeBSD's ext2 will be done right (whatever that |> means), and not just there to run the occasional Linux file system it |> comes across. Why should it? FFS is far advanced, compared to EXT2FS, IMHO. Concepts for save and fast meta data updates have been developed and will no doubt be integrated into the FreeBSD file system soon. Regards, STefan -- Stefan Esser, Zentrum fuer Paralleles Rechnen Tel: +49 221 4706021 Universitaet zu Koeln, Weyertal 80, 50931 Koeln FAX: +49 221 4705160 ============================================================================== http://www.zpr.uni-koeln.de/~se <se@ZPR.Uni-Koeln.DE>