On Sat, 2002-10-26 at 18:36, Jack Coates wrote: > Hi, > > I have a machine (naturally a fairly important one) which I built and > which is fairly unstable -- give it ten to twenty days, and it will > suffer a video lock up or a disk error that forces a reboot. > > FIC AZ11 with a VIA KT133 chipset. > Athlon 900 MHz > 768MB PC100 RAM > Voodoo3 3000 AGP card > Maxtor 92049U6 (20G 7200 RPM disk), etx3fs on all partitions. > Polaroid CD-RW IDE3212 > (2) Lite-On Communications Inc LNE100TX PCI tulip NICs > > The system runs MDK 9.0 and ran 8.2 before that. I keep an eye on it > with lm_sensors, and it doesn't go over 60C, so I've tentatively ruled > CPU heat out as a cause. I've also run memtest86 several times for runs > of up to 12 hours, with no RAM problems detected. > > today's reboot was caused by a corrupt wtmp that prevented logins. fsck > fixed it. > > At this point, I'm suspecting the motherboard's IDE controller, the hard > disk itself, the video card, the RAM, gremlins :-) I'm just not sure > where to go next in troubleshooting the system, and I don't want to > replace everything one part at a time. > > any ideas for narrowing this down? > -- > Jack Coates > Monkeynoodle: A Scientific Venture...
Power Supply? "THG continues to receive hundreds of e-mails regarding the causes of sudden and sometimes frequent "crashes" of users' computers and computer systems. As these computer crashes are random and unpredictable, users are losing important data and other critical information as a result. These issues of "stability" can cost hours of use time while the system is continued to be tinkered with in vain in an effort to resolve the issue. When this occurs, ambitious users are inclined to search for the cause in the processor settings, system memory or on the motherboard, since the Front Side Bus, CPU core voltage and GPU of the graphics card are often pushed to the limit to achieve maximum performance. This results in the user resetting all of the system parameters to default values - but to no avail. The computer continues to crash, eventually so frequently that it is no longer possible to continue using it." One possible cause of unreliability that is often overlooked by users is the power supply that is installed in their computer. Unless the computer was custom built and the user recently had the power supply upgraded, chances are that it may contain a less-than-capable power supply. Users who stand to fare the worst are those who have purchased their computer from a computer super store and/or discount retailer. These systems are generally fitted with cheaply made, low-cost power supplies, which often can sustain damage even under minimal loads. The power supply is something that many users overlook when they upgrade their PCs. Since upgraded components often require more power, a larger, more capable power supply should be a standard replacement when significant upgrades have been made to a computer. Only the lab test tells the truth - even if the maximum load is specified as 300 watts, the power supply often "gives up the ghost" long before reaching this load. Some manufacturers of power supplies apparently assume that computer users will never need maximum power from their power supply. Full Load and Overload - Power Supply Units Pushed to the Limits http://www6.tomshardware.com/howto/02q4/021021/index.html Power supplies should be added to the troubleshooting list when tracking down gremlins. Gabriel
signature.asc
Description: This is a digitally signed message part
