Date: Wed, 28 Apr 1999 22:38:03 -0500 (EST) From: Alfred Perlstein <bright@rush.net> To: Greg Lehey <grog@lemis.com> Cc: "Paul L. Poynter" <ppoynter@rorke.com>, freebsd-questions@FreeBSD.ORG Subject: Re: BSD Software RAID 0 and 1? Journaling file system? Message-ID: <Pine.BSF.3.96.990428222303.10204k-100000@cygnus.rush.net> In-Reply-To: <19990429112005.I46511@freebie.lemis.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 29 Apr 1999, Greg Lehey wrote: > [Format recovered--see http://www.lemis.com/email/email-format.html] > > On Wednesday, 28 April 1999 at 13:35:14 -0500, Alfred Perlstein wrote: > > On Wed, 28 Apr 1999, Paul L. Poynter wrote: > > > >> Will I have the ability to mirror my boot drives with bsd? > > Yes. > > >> Is the filesystem "journaling" like IRIX XFS? > > 4.4BSD includes a journaling file system, but it hasn't been used > much, and I suspect that it's no longer completely reliable. This > doesn't seem to be much of an issue for FreeBSD systems. It's not in the base distribution anymore (if you are refering to LFS) > > >> (if the filesystem needs to be fsck'd and the filesystem is a very > >> large raid, this may take several hours - don't want this) > > Our goal is to avoid crashing :-) > > > unfortunatly at this time i'm afraid FreeBSD doesn't offer this > > functionality, we DO have raid support, just not fast fsck because > > of logging. > > You may have misunderstood the question. When a journaling/logging > file system crashes, you don't need to run fsck: you replay the log. > It makes up for the faster fsck by slower performance all the time. > This is probably why nobody has been too concerned about it in the > past. Er.... when you have a logging filesystem you can have an in place log, since the superblock points to the last sucessful checkpoint you just need to examine all partial segements after the last sucessful checkpoint verifying thier completeness. I don't think you have to sacrifice speed at all as segments written out can contain both metadata and data blocks, in fact 4.4 BSD's LFS is superior at writing isn't it? > > btw, softupdates is _almost_ metadata logging, anyone care to > > describe what it would take to make fsck and softupdates co-operate > > to check point a filesystem? > > Not I. All you would have to do is some sort on disk pointer to all the data operations modified in the last softupdates "burst" -Alfred To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.96.990428222303.10204k-100000>