[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: system instablity



Dan hursh wrote:
> 
> Hi all,
> 
>     I'm asking here because I'm hoping someone might have a nicer
> diagnosis for what I'm seeing.  I had been running a debian hamm
> installation that I had upgraded with no trouble until I had a disk
> failure.  I verified it was the hard drive that was hosed.
> 
>     Anyhow, I recent bought hamm on CD and installed again on a different
> drive.  I've had a few odd problems here and there but nothing unusual
> until the past couple of days.  I have been seeing dpkg die with odd perl
> syntax errors, seg-faults, etc.  Other programs such as netscape 4.05,
> emacs, smailconfig, and kfm (and many other I can't remember) have been
> dieing with illegal operations, segs, and bus errors.
> 
>   The problems will be repeatable for a short time, and then the program
> will behave, or exhibit a new and interesting problem.  I know linux and
> the software I'm using is more stable than this.  I guess I'm wondering if
> anyone know if there may have been problems with the official cd image?  I
> would image that the problems I'm see would have been fixed if it were the
> software.  I guess another question would be, does anyone have any idea
> which hardware is most likely to cause this without crashing the whole OS?
> Judging by the randomness of the error, I'm guessing not the CPU.  Maybe a
> bad spot in memory?
> 
>     I'm running hamm pretty much out of the box still.  I have not
> compiled my own kernel yet.  I'm not using any new hardware.  This has all
> work within the last month.  If anyone has any idea's, I'd appreciate it.
> 
> Thanks,
> Dan Hursh
> hursh@infonet.isl.net
> 


	Thanks Dan, I thought it was just me.  I reported transient errors with
apt/dpkg.  Do the errors start with 'General Protection:  000000' or
'General failure:  0000' or something like that?  The only person who
responded to that part of my message, insisted the problem was elsewhere
and not Debian, but I can't figure out any other explanation other than
corrupted data files used by apt/dpkg or some similar problem.
	In my original post, I believed dselect/apt/dpkg were being pushed
beyond their capabilities.  Dpkg is being swamped by packages, now more
than 2700.  However, the errors, from my point of view, are far more
common when running dselect with apt as the access method.
	This Friday I tried an upgrade using dselect/apt.  I had a lot of stuff
on hold, so decided to let dselect upgrade all of them (~30 packages). 
The download was a long one, but uneventfull.  When installing though it
started installing the latest netbase package of slink and
crashed...hard.  I first got the same kind of 'General failure:   00000'
message from apt/dpkg then I got the kernel message that began with
'Aiyee' or something like that.  It was the first time I've seen a hard
kernel crash outside of the X win system.
	The wierd thing, and this is why this problem is so frustrating, is
that I rebooted, let fsck fix the inevitable filesystem errors then
installed the same packages manually with 'dpkg -i' and everything
installed sucessfully.
	Like you this is something I've only seen recently, I don't remember
having these kinds of problems back with Deb 1.3.1.


-- 
Ed C.


Reply to: