fsck dumps core

[Ian Lepore]

On Mon, 2014-02-24 at 19:52 +0400, Dmitry Sivachenko wrote:
> On 24 февр. 2014 г., at 19:50, Slawa Olhovchenkov <slw at zxy.spb.ru> wrote:
> 
> > On Mon, Feb 24, 2014 at 06:27:15PM +0400, Dmitry Sivachenko wrote:
> > 
> >> Forgot to mention that it is ~90TB partition and this server has 24GB of RAM.
> >>> Is there any way to complete fsck to get this drive working?
> > 
> > Try to add some swap space.
> 
> 
> I have 48GB of swap.
> 
> It dumps core almost immediately on start, not that it is running out of swap.

It looks to me like it's trying to allocate an insanely large amount of
memory, perhaps in response to some insane value it found while trying
to walk the filesystem data.  I had something like that happen back in
the 90s, and I was eventually able to track it down to garbage written
into one inode block.  I manually zero'd that block with dd and lost a
few files in the process, but was able to recover most of the
filesystem.

I'm not recommending that you start zeroing out blocks on your disk at
this point, just mentioning one possibility of how much difficulty you
may face.  fsck does an amazing job, but it can't cope with every
possible type of insanity it may encounter.

-- Ian