Re: 2.6.1 + XFS wierdness

From: Andi Kleen (ak_at_suse.de)
Date: 01/24/04

  • Next message: Dave Jones: "Re: Illegal instruction with gl"
    To: mru@kth.se (Måns Rullgård)
    Date:	24 Jan 2004 16:45:10 +0100
    
    

    mru@kth.se (Måns Rullgård) writes:

    > Igor <ivi@runbox.com> writes:
    >
    > > Ok, as advised I'm reporting what happened to my system:
    > > I run Kernel 2.6.1 with XFS on a laptop, I forgot to send it to "sleep"
    > > and battery died, so there was unclean unmount (This is, what I
    > > believe was the cause),
    > > at some point after I restarted my system many of the files couldn't
    > > be executed:
    > > "binary file can't be executed reported", However the system was
    > > functional and I could boot it.
    > > So I hexopened some of the problematic files and found that although
    > > the size of the file is maintained, there was no data, every byte was
    > > replaced by 0, I guess it was lucking reference on a hard drive or
    > > maybe something else. The reason I think the root of the problem is
    > > filesystem + kernel because the "corrupted" files have nothing in
    > > common, e.g:
    > > /usr/bin/file
    > > /etc/init.d/cron
    > > /usr/bin/lynx
    > > and that only happened when I updated kernel to 2.6.1
    >
    > See http://oss.sgi.com/projects/xfs/faq.html#nulls

    I don't think his description fits the FAQ.
    The XFS 0 problem should only happen to files that have been written
    shortly before the crash.
    Zeroing/destroying random files that haven't been written looks more like a
    bug (either in XFS or in a driver)

    -Andi
     
    -
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/


  • Next message: Dave Jones: "Re: Illegal instruction with gl"

    Relevant Pages

    • Re: LVM2
      ... XFS is an SGI project. ... >>Disk size: What are you doing with it. ... >>>Please read the FAQ at http://www.tux.org/lkml/ ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: printk and long long
      ... > Måns Rullgård wrote: ... >> can be either long or long long depending on machine? ... send the line "unsubscribe linux-kernel" in ... Please read the FAQ at http://www.tux.org/lkml/ ...
      (Linux-Kernel)
    • Re: (0 == foo), rather than (foo == 0)
      ... mru@kth.se (Måns Rullgård) writes: ... >> Hi Linus. ... send the line "unsubscribe linux-kernel" in ... Please read the FAQ at http://www.tux.org/lkml/ ...
      (Linux-Kernel)
    • Re: [BUG] Serious scheduler starvation
      ... On Sunday 17 August 2003 14:11, Måns Rullgård wrote: ... Look in top's "PRI" column, which is where you see the effects of the dynamic ... send the line "unsubscribe linux-kernel" in ... Please read the FAQ at http://www.tux.org/lkml/ ...
      (Linux-Kernel)
    • Re: Is there a "make hole" (truncate in middle) syscall?
      ... On Thu, 4 Dec 2003, Rob Landley wrote: ... Using XFS XFS_IOC_GETBMAPX you might ... send the line "unsubscribe linux-kernel" in ... Please read the FAQ at http://www.tux.org/lkml/ ...
      (Linux-Kernel)