[TriLUG] Debian Sarge is dying somehow

[Scott G. Hall]

On 12-31-2004 12:52:56 EST, Brian Henning <brian at strutmasters.com> wrote:
> Hi Y'all,
> [In trying to work with rdesktop to access IE6 via a link], things went to
> heck.  The dialog for creating the link never really disposed of itself.
> It ceased to respond, and the entire system suddenly bogged down to a 
> nearly 100% unresponsive crawl (no mouse response, etc).  Couldn't even
> switch virtual terminals.  I walked away for a bit and returned to a text
> screen chock full of out-of-memory -related error text (which had filled
> the entire scrollback buffer). On reboot, the system reaches "Starting
> system logging daemon: syslogd" and hangs.  There is some hard disk
> activity now and then, and eventually, if it is left to its own devices,
> it will again fill the buffer with out-of-memory messages.  Before this
> machine was loaded with Linux, it was a 98 machine that had started to
> bluescreen uncontrollably.  I was hoping it was a software issue, but this
> is starting to make me think maybe it's hardware-related. Does anyone have
> some input based on what I've just described?  Seen it before and know what
> is likely the cause, and/or how to fix it?
> 
> Thanks, everyone. ~Brian

I would have to agree -- it does sound like some ailing hardware issue.
Quick standard troubleshooting steps:
1) Remove the cover, then make sure the system has been cleaned out of
    dust.  A lot of dust and dirt can cause intermittent shorts and the
    like with humidity, restrict air flow to cool components, cause heat
    sinks to become less efficient and effective, and bog down fans.
2) Make sure that every card, memory SIMM, device cable and connector is
    fully seated.  You might want pull each card and connector and to spray
    some contact cleaner before reseating.  Even if a system isn't moved,
    over time temperature and humidity changes can cause a connector or
    card to dislodge or shift.  Even just a minor fraction of an inch is
    enough to allow a thin layer of dirt to accumulate and cause contact
    problems.  Don't overlook anything -- depending on the motherboard and
    BIOS even a slightly loose panel button or speaker cable can cause a
    lot of I/O interrupts and overwhelm any operating system.
3) Make sure the heat sink (and any aux. fan attached) is properly seated
    on the CPU.  More than once I have had a CPU fan go out on me, or for
    some reason the heat sink dislodge itself (usually due to me forcing
    extra cables into contorted routes...).
4) Make sure cables are routed to allow for sufficient air flow.  Heat is
    your enemy.  Some components that were more tolerant of higher
    temperatures when new may now have problems, and need to be cooled
    better.  You can buy extra fans for cooling your drive bay, and your
    card area if you have a lot of drives or cards installed.  You can also
    buy new cables that allow for more air circulation.
5) You may be experiencing aging components -- such as memory SIMMS, video,
    sound chips and I/O chips, even the CPU -- that once were in specs on
    the borderline, but now are out of specs.  I have seen bad memory SIMMs
    create just the problems you are experiencing -- but I have also seen
    overheated ones do it too, and that can be fixed by better cooling.
6) You may also have a bad harddisk.  Over time harddisks do wear out, and
    if they have been subjected to overheating for much of their life, will
    wear out fast than intended.  I figure 2 to 4 years on a good harddisk
    run everyday.

P.S.  This is a good spring cleaning list for all systems out there --
    working properly or not.  I know of a good after-hours company that
    makes house calls and will do this for you :) ...  call me at 624-5973
    for details.

-- 
Scott G. Hall
Raleigh, NC, USA
ScottGHall at BellSouth.Net