[TAG] segmentation fault in aptitude

[Kapil Hari Paranjape]

Hello,

On Wed, Apr 06, 2005 at 10:18:12PM -0400, Benjamin A. Okopnik wrote:
> On Thu, Apr 07, 2005 at 08:03:04AM +0530, Kapil Hari Paranjape wrote:
> > On Tue, Apr 05, 2005 at 10:22:01PM +0530, J.Bakshi wrote:
> > > write(1, "\rBuilding Dependency Tree... 67%"..., 33) = 33
> > > --- SIGSEGV (Segmentation fault) @ 0 (0) ---
> > > +++ killed by SIGSEGV +++
> > 
> > Another couple of possiblities:
> > 
> > 1. Not enough swap space (the OOM killer?!).
> 
> "VM space" is, I think, a bit more precise - but the essential point is
> about the same.

OOPs killer :-) You're right of course.

> > 2. Swap space is buggy. Run the commands
> > 	swapoff <name of swap device>
> > 	mkswap -c <name of swap device>
> > 	swapon <name of swap device>
> 
> Interesting. I've seen a swap partition that decided to refuse to mount
> for no obvious reason - and the above was the way I'd fixed it
> (superblock corruption, maybe?), but I've never seen it mount and not
> work. [shrug] What the heck - seems like it would be worth trying.

Actually, what I had in mind was HDD corruption of part of the swap space.
which is why I put in the "-c".

We had this happen here once. The disk just acquired a huge number of
bad blocks rather suddenly. So we restored from backup and still found
there were problems. Someone (grrr!) had done bad block checks on the
file-system partitions but forgotten (felt lazy more like!) to do a bad
block check on the swap partition.

Regards,

Kapil.

--