Re: sarge dist-upgrade "Package is in a very bad inconsistent state"
>From Carlos Sousa on Monday, 2004-09-06 at 22:38:14 +0100:
> On Mon, 6 Sep 2004 23:18:25 +0200 Conrad Newton wrote:
> [...]
> > It is not an answer to your question, and maybe it is not even
> > relevant to your problem, but I am having similar problems
> > here with tetex-base. When I try to correct the problem,
> > I get a system crash or even trigger a reboot.
> >
> > The system is a mixture of testing and unstable (pinned to
> > testing, but with several major applications upgraded).
> > I have had numerous instabilities lately, and thought
> > that I might improve the situation by upgrading to
> > unstable (my unstable box works just fine), but instead
> > things look worse than ever . . .
> >
> > For example, I was just going to try the following
> >
> > apt-get -f install
> > dpkg --configure -a
> >
> > but already at the first step, I get a reboot!!!????!!!
> > What could be causing this?
>
> Sure smells a lot like hardware trouble (memory?). Better be
> ready with your backups...
I did a memtest check today, and the results were the same as two years
ago: failure with test #5, but otherwise OK. What this means is that I
cannot compile kernels or glibc on the box (I first became aware of this
problem because I tried to install gentoo), but otherwise the box has been
working fine for the last six years, with various Linux distributions,
and without any file corruption.
I suspect the root problem affecting memtest is not the memory, but the
processor itself, as there are reported problems with the AMD K6-300 MHz.
The number of reported errors varies every time I run memtest, and it
fails on the same test with either of the two sticks of memory, although
both sticks were bought from different suppliers at different times.
Could it be that the newer kernel is probing some part of memory or the
processor that was unexplored before? I'm using 2.6.8. I'm wondering
what other explanations could account for the observed behavior.
My worst nightmare would be a bad hard disk, but this one (a 60 GB Seagate)
has been running smoothly now for three years, so I have no reason
to be suspicious.
Conrad
Reply to: