Problem with hardware, or is it...?
Here's a good one for you all:
380W Enermax EG365AX-G-FMA Manual Fan/Temp Control
Gigabyte GA-K8NS Pro nForce3 Pro 250, ATX, DDR400, Gbit Lan, USB 2.0, AC'97 Audio, SATA, Raid
AMD Athlon 64 2800 32/64Bit CPU S754 Newcastle Retail inc Heat Sink Fan
512 Mb (2 x 256 Mb) Corsair TwinX, DDR, PC3200C2PT, Cas 2
200Gb Maxtor Plus 10 (7200rpm,8MB) SATA NCQ
Asus Black 16x DVD-ROM + 48x CD-ROM IDE Retail
Pioneer DVR-108 Black x16 Dual Layer DVD Writer, UK OEM
128Mb Asus 9200SETD (D-Sub, DVI, TV-Out)
~2 years old.
I've since replaced PSU, Graphics card and HDD. See below.
Using Debian I get semi-random lockups/reboots. Any other distro which uses Xorg locks up almost instantly on starting X. We're talking hard locked. No amount of ctrl and alt and anything will help. Num lock unresponsive. Reset button time.
With Xfree86 I only saw lockups on very hot days when playing videos. This led me to believe the graphics card was at fault so it was promptly replaced with an Nvidia Geforce 6200 AGP card with correct drivers. This made no difference. The system was useable, just keep it cool and all was well.
I recently (last week) had a HDD failure. Fair enough I thought, it's 2 years old, so I replaced the HDD with a shiny new Samsung spinpoint 250Gb and at the same time treated myself to a new 550W PSU in case the 380 was too weedy.
These installed I was expecting the problem to go away, not get worse. I get lockups now so frequently that the system is unusable. I have, however, found that running knoppix from a live CD the machine does not crash, ever.
Feeling around inside, nothing is hot. I have 9 8cm fans around the case plus CPU and PSU fans.
The RAM is slightly warm to the touch, not hot. Same goes for the northbridge, GPU and anything else I can reach to feel inside. The CPU reports a scorching (!) 32 celsius, so probably not him then. Incidentally, system load wasn't a problem. I can crank mencoder hard encoding movies and watch the CPU temp rise a few degrees to about 45 and nothing went wrong. Playing videos though inevitably causes a bork.
Strangely though pointing a large fan into the open case seems to help with stability.
'It's your memory' I hear you cry. I thought so too. I ran a memtest86 and found no errors. I ran it twice more removing 1 of the 2 sticks each time. No errors. I tried running the system on one stick at a time (each seperately). Still crashes.
So, here's my question:
What, the *&£(ing hell, really hates Xorg and can overheat enough to cause crashes without *feeling* warm? The only things I haven't replaced are the mobo, processor, RAM and associated heatsinks and the optical drives.
There is of course nothing in the logs as they don't get to flush, and this happens with a fresh install as mush as any other, and seems independent of the kernel used (every one since 2.6.10)
The End. Thanks for your time.
Answers on a postcard