bin/118300: fsck -y failed to cure UFS2 volume with lost '.' entry in one directory

Vladimir Grebenschikov vova at fbsd.ru
Wed Nov 28 04:17:42 PST 2007


On Wed, 2007-11-28 at 12:53 +0100, Remko Lodder wrote:
> Vladimir Grebenschikov wrote:
> > On Wed, 2007-11-28 at 12:34 +0100, Remko Lodder wrote:
> >> Vladimir Grebenschikov wrote:
> >>> On Wed, 2007-11-28 at 10:55 +0000, remko at FreeBSD.org wrote:
> >>>> Synopsis: fsck -y failed to cure UFS2 volume with lost '.' entry in one directory
> >>> It is very concrete problem report - 
> >>> "fsck -y failed to cure UFS2 volume with lost '.' entry in one
> >>> directory"
> >>>
> >>> I have UFS2 volume, and fsck -y reports that it cures it, but actually
> >>> it has inconsistencies after that.
> >>>
> >>> Do you think that it is not a problem ? 
> >>>
> >>> Please reopen PR.
> >>>
> >> Hello,
> >>
> >> No I will not reopen the ticket till more information is available that
> >> shows that this is indeed a bug that should be fixed. It could just be
> >> a combination of problems on your local end which are not actually a bug
> >> in the code or something.
> > 
> > What else information do you want to see here ?  I'll provide.
> > 
> 
> It should be reproducable on other hardware as well if this is a real
> bug, 

I pretty sure, that everybody who will mount my volume will see the same
problem.

> so perhaps you can printout more information on the /compat/linux
> directory which is giving you problems, are the directories all there
> that should 'support' the problematic missing ``.`` ? etc.

Looks like we are speaking on different languages.

Once again: 

1. only directory that has a problem
is /compat/linux/usr/lib/locale/af_ZA/
2. IMHO contents of all other part of /compat/linux is irrelevant here,
but if anybody think that it is relevant I may provide
find /compat/linux/ output or any else.
3. for my knowledge I've provide enough information about broken
directory and its entries. I am happy to provide more, if anybody points
me what to provide.

> I am not a FS expert, so hence the reasoning to poke the fs mailinglist
> since there are people with much more clue then me over there.  If they
> agree this is a real problem and should be fixed, I will surely reopen
> the ticket and include usefull information in it so that we can keep
> track of the PR etc. For the moment this still does not seem like a PR
> to me but perhaps a local value or something that messed up which might
> just not be a bug but a local problem

>From my point it is clearly malfunction of fsck - it failed to fix
volume.

> (that does not mean that is not a problem for you of course).
For me it is rather simple to fix problem by hands (with fsdb - as I
suggested in PR or just reformat volume with newfs).

But my main intention was - to improve quality of fsck by reporting
problem and providing debug/verify/etc cycle.

If such FS problem happens for me it may happens for everybody else,
with more or less probability.
 
> > I've write to fs at freebsd.org already with no any feedback for three
> > days:
> > 
> 
> People are busy so it would be unfair to have them reply within a
> certain timeframe right? You could just follow-up on it and reply
> 'did someone have the time to look at this already and/or what else
> can I do to provide more valuable information?'
> 
> Someone will reply eventually and an occassional friendly reminder
> can be greatly appreciated :)
> 
> Cheers
> remko
> 
-- 
Vladimir B. Grebenschikov
vova at fbsd.ru


More information about the freebsd-bugs mailing list