Subject: Re: LFS troubles from around 20020405
To: None <>
From: gabriel rosenkoetter <>
List: current-users
Date: 06/04/2002 00:29:58
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Mon, Jun 03, 2002 at 09:16:58PM -0700, Konrad Schroder wrote:
> If this is the problem I think it is, I think if you upgrade lfs_cleanerd
> to later than 20020426 or so, it won't do this any more.

So that 20020430 snapshot I haven't bothered to install might be a
good idea, then? ;^>

Incidentally, my first impulse to make use of lfs_cleanerd's -q flag
and cron is a really bad idea in this situation, if anyone else is
in it (rather than one lfs_cleanerd chewing your disk, you get a
linearly increasing number).

> If the cleaner is not running, a "sync" followed by fsck_lfs should tell
> you if your ifile is corrupted, even without unmounting your fs.  (If it
> is messed up fsck_lfs may not be able to help you yet.)

Hrm. Found me a bunch of unassigned inodes, which I blithely let it
add back to the free list. Let's see if I lose anything. (Like I
said, it's sort of immaterial if I do; there's nothing here that ten
minutes with abcde wouldn't get me back.)

In any case, fsck_lfs seems to have modified my file system without
borking all over it, which is a pleasant change from the last time I
tried it (admittedly, ages ago).

> When you say "LFS" I am listening, though I don't always have time :^/


> It might be worthwhile trying one of the 1.6_BETA snapshots...there is a
> lingering race somewhere that causes my test box to fall over about once
> every twelve hours in reassignbuf() (this is with QUEUEDEBUG, ymmv) but it
> should be significantly more stable than an early April snapshot.

The next build I do for this system will probably be 1.6; I'd sort
of like to have it on -release rather than -current being as it's my
firewall and all...

gabriel rosenkoetter

Content-Type: application/pgp-signature
Content-Disposition: inline

Version: GnuPG v1.0.7 (NetBSD)