Return to BSD News archive
Newsgroups: comp.unix.bsd.freebsd.misc,comp.os.linux.development.system Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!news.mel.connect.com.au!munnari.OZ.AU!news.ecn.uoknor.edu!paladin.american.edu!gatech!newsfeed.internetmci.com!howland.reston.ans.net!ix.netcom.com!netcom.com!jlemon From: jlemon@netcom.com (Jonathan Lemon) Subject: Re: The better (more suitable)Unix?? FreeBSD or Linux Message-ID: <jlemonDMtpAz.3rJ@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) References: <4er9hp$5ng@orb.direct.ca> <4fo1tu$n31@news.jf.intel.com> <DMrCE4.3HF@pe1chl.ampr.org> <4ftjt9$fjs@park.uvsc.edu> Date: Thu, 15 Feb 1996 15:20:59 GMT Lines: 46 Sender: jlemon@netcom22.netcom.com Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:14334 comp.os.linux.development.system:17979 In article <4ftjt9$fjs@park.uvsc.edu>, Terry Lambert <terry@lambert.org> wrote: > >Let us also agree that the UFS use of synchronus updates is as >a simple method of ensuring write ordering without violating >patents or rewriting a lot of code. > >Let us discuss the desirability of order metadata writes. > > >Posit: ordered metadata writes insure the determisitic > recoverability of changes to file system structure, > >Conclusion: ordered metadata writea are desirable. > >] The claim (usually) isn't that async metadata is better in >] this respect, but it is not worse and it is faster. Hence >] i > >Since the recoverability of file system structure is unrelated >to the recovery of data within files as container objects, your >claim about "not worse" applies only to files as container >objects, and is not applicable to the discussion of write >ordering or the use of syncronous I/O as a mechanism used in >the implementation of write ordering. It is apparent to me that the filesystem safety is dependent on ordered metadata updates. FFS uses sync writes as the _implementation_ for ordered updates. 'sync' in this sentence, meaning "write this to disk right NOW". However, if metadata is written in the _same_ order, but async, I would believe that the recoverability of the filesystem is preserved. 'async' in this context, meaning "write this to disk when convenient". However, this introduces the problem of data dependencies (eg: a datablock write may have to be delayed until the metadata is flushed first) and would probably be difficult to implement, as Terry alluded to above. In this sense, "sync vs async" is a the wrong argument; it should really be "ordered vs unordered" instead. My question is, does ext2fs preserve ordering, even though it writes metadata async? If not, then it is arguably more unsafe, even from the user's point of view. (eg: if you have to re-run 'mkfs' on a disk to fix consistency problems, then you lose more data than just truncating a couple of files, no?) -- Jonathan