SUMMARY: fsck checks filesystems twice

From: Ron Stanonik (stanonik@nprdc.navy.mil)
Date: Mon Dec 23 1991 - 19:03:18 CST


The "problem":
        If fsck finds a problem in /usr, it doesn't exit immediately after
        repairing the problem, rather it checks all of the other filesystems.
        Consequently, all of the filesystems get checked twice; ie, again
        during the reboot after repairing /usr. (Yes, this is a rare;
        for example, autoreboot due to a power glitch.)

From: simon@Aus.Sun.COM (Simon Woodhead - Technical Consultant)
        4.1.2 includes the QuickCheck utility which keeps an eye on the
        file system state *all the time* (well, between /usr/etc/updates),
        and sets a flag for each filesystem indicating that it is stable.

From: Mike Raffety <miker@sbcoc.com>
        However, how often does fsck ever need to fix
        anything in /usr? We generally make the mount read-only, so fsck
        doesn't even look at it, and things go even faster. Even if it were
        read-write, I'd say that the problem is awfully rare (once in a hundred
        boots, since 99 are clean halts or shutdowns).

From: Charles <mcgrew@porthos.rutgers.edu>
        We do /usr as pass 1 as well...

From: mdl@cypress.com (J. Matt Landrum)
        I think it's probably a mistake on sun's part.
        As and aside...
        If you aren't going to 4.1.2 soon, I recommend you get
        quick check (from Sun - $500 and worth every penney).
        One of my 4/280s with 4 892 disks takes 10 minutes
        for a full fsck (it used to take 45+). I think
        this will be bundled in future O/S releases.

From: Jim Guyton <guyton%condor@rand.org>
        If you see this enough to bother you, then something is wrong.

Thanks also to
From: etnibsd!vsh@uunet.UU.NET (Steve Harris)
From: shipley@kizmiaz.tcs.com
From: erueg@cfgauss.uni-math.gwdg.de (Eckhard Rueggeberg)

Ron Stanonik
stanonik@nprdc.navy.mil



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:21 CDT