The hardware part of this machine is as follows
box: Arctic Cooling Silentium T1 with psu. (was: PC6070BSCANDV2 black miditower from Lian Li.)
power supply: Seasonic 350W, active PFC (was: MIST 500W (revision 3))
motherboard: M2A-VM HDMI from Asus
memory: 4 x 2GB DDR2 PC6400 800 MHz DIMMs - 8 GB total. (was: 2 kits of 2 GB (2 x 1 GB) DDR2 PC6400 800 MHz Ultra 4-4-4-12, GX22GB6400UDC from Geil. Thats 4 DIMMs and 4 GB total). (max memory is 8 GB)
cpu: AMD, Phenom 9500 2.2 GHz Boxed 4MB Cache Socket AM2+
cpu cooler: Noctua, NH-U12F, 4 dual heat pipes, 120 mm 17dB(A), socket775/754/939/940/AM2
ad0: 152627MB <WDC WD1600AVJB-63J5A0 01.03E01> at ata0-master UDMA133
ad4: 953869MB <Seagate ST31000528AS CC49> at ata2-master UDMA100 SATA 3Gb/s
ad6: 953869MB <Seagate ST31000340AS SD1A> at ata3-master UDMA100 SATA 3Gb/s
ad8: 1907729MB <SAMSUNG HD204UI 1AQ10001> at ata4-master UDMA100 SATA 3Gb/s
(was: hard drive: Samsung SpinPoint T166 500 GB SATA2 16MB 7200RPM, HD501LJ)
(was: hard drive 2: Samsung SpinPoint 320 GB, HD322HJ)
(was: optical drive: NEC AD-5170 IDE, DVD+R/+RW/DVD-R/-RW (Dual layer))
video: radeonhd - ATI - Radeon X1200 Series
sound: snd_hda - ATI - IXP SB600 High Definition Audio Controller
network: re - Realtek Semiconductor - RTL8168/8111 PCI-E Gigabit Ethernet NIC
bios: version 5001 (Release Date: 02/04/2010)
misc: boot menu - F8.
2013-02-13: this morning the machine was unresponsive - it answered on ping, but local console didn't respond, and I couldn't ssh into it. Powered it off, and tried a reboot. Didn't work. Further testing this evening reveals that the machine works if "cold" (has been powered off for a long time), but starts rebooting all the time after a while. The HW monitor in the BIOS claims that voltages are within specifications (within 5% of specified value). Bad capacitors on the motherboard perhaps?
2013-02-12: I took out the old motherboard (MSI RS480M2-IL) of the "quiet" machine and installed this motherboard instead. Installed the new memory, the machine now has 8 GB RAM. I also took out the floppy drive (defective) and the CD writer (doesn't work anymore). The CPU cooler almost touches the side panel of the cabinet. ad4 and ad6 switched places too. Hmm, it seems like ad6 is jumpered for 1.5 GBps operation. From dmesg:
ad6: 953869MB <Seagate ST31000340AS SD1A> at ata3-master UDMA100 SATA 1.5Gb/s
Ok, I need to open up the machine and remove that jumper. Ok, done. Now ad6 is in shape:
root@kg-quiet# dmesg | grep ad6
ad6: 953869MB <Seagate ST31000340AS SD1A> at ata3-master UDMA100 SATA 3Gb/s
2013-02-12: I fetched the package from PS.no at the post office today.
2013-02-10: I ordered a new set of memory (from PS.no) for this motherboard. 4 x 2 GB DDR2 800 MHz PC6400 DIMMS for NOK 773.10 including shipping. That will give me 8 GB, which is max for this board.
2012-10-12: Testing with a new PSU in a breadboard setup; the motherboard works. Hmm, OK.
2012-10-10: Testing with the new motherboard revealed that the PSU was malfunctioning, so I need to test this motherboard too.
2012-10-08: I ordered a replacement (motherboard + cpu), the replacement is a Asus M5A78L-M / USB3 motherboard, with an AMD FX-6100 cpu.
2012-10-07: It looks like the motherboard died just now. The green LED (SB_PWR) on it is lit, but when I press the power button, the fans just spin up for a second, then stops again. I can also see a few capacitors with bulging tops on the motherboard. A bit of searching on the bad caps forum reveals that this board has capacitors from a known "bad" vendor. I'll just replace it, I think.
2012-07-21: slice 4 - the machine hung (totally, no response from keyboard) and had to be reset (reset switch). No info on console, and no info in logs afterwards.
2012-03-13: slice 4 - the machine hung (totally, no response from keyboard) and had to be reset (reset switch). No info in logs afterwards.
2012-02-09: slice 4 - the server had a kernel panic again (MCA again).
2011-12-17: slice 4 - the server had a kernel panic again. This time it hung in core dump routine, while dumping pages to the disk, thus no core dump. But it was a MCA interrupt again. Checked for a newer BIOS, but version 5001 is still the newest one.
2011-11-28: slice 4 - the server had a kernel panic again. This time it hung in core dump routine, while dumping pages to the disk, thus no core dump. But it was a MCA interrupt again.
2011-07-13: slice 4 - the server had a kernel panic again. This time it hung in core dump routine, while dumping pages to the disk, thus no core dump. But it was a MCA interrupt again.
2011-01-28: Upgrading bios again, to version 5001 (Release Date: 02/04/2010) this time. I tried a new approach this time; I just replaced one of the Geil DIMMs (the one farthest away from the cpu fan) with my Kingston DIMM and that seems to have worked equally well. Ah - much easier; now I dont have to remove the fan anymore. Then again; in the BIOS, DDR2 Voltage Control was set to "+2.10V", so maybe Asus has fixed their bios. Note: for some reason, the usb memory stick was detected as drive C:, I thought it was going to be A:, oh well. I also added a second hard drive (Samsung SpinPoint 320G, HD322HJ), so I could do more testing.
2011-01-05: The machine has had some crashes and reboots, so I decided to test it; the machine has been running memtest86+ (version 4.10) for about 27 hours now; 19 passes, no errors detected.
2010-09-16: new beta BIOS available, version 5001. Not tested yet.
2009-02-13: new BIOS available, version 2302. Not tested yet.
2008-11-10: new BIOS available, version 2201. Not tested yet.
2008-09-18: upgraded the bios to version 2101. I had to perform the same procedure as before; temporarily swap installed memory with one Kingston 512MB DIMM which works with 1.8V, and set the memory voltage to 2.1V. See the 2008-01-11 and 2007-12-22 entries for details.
2008-09-12: new BIOS available, version 2101. Not tested yet.
2008-08-18: new BIOS available, version 2001. Not tested yet.
2008-07-11: new BIOS available, version 1803. Not tested yet.
2008-06-05: Today I found out that there is a new BIOS available, version 1801. I haven't tried it yet.
2008-01-11: BIOS upgraded to version 1604 (Release Date: 12/19/2007). Asus calls this a beta bios. After the upgrade, I couldn't boot the machine, not even with just one DIMM. The Geil memory I have requires 2.1V, and apparently the upgrade resets the DDR2 voltage to "auto" (which must be 1.8V?). Fixed by removing all four DIMMS, and (temporarily) using a Kingston 512MB DIMM (which requires only 1.8V) to change the setting in BIOS.
2007-12-30: I ran memtest86+ on the machine. It found one error. Further testing found that one of the four DIMMs was bad.
2007-12-30: BIOS: I changed the Memory Clock Frequency from Auto to DDR2 800. It doesn't seem to effect anything.
2007-12-22: BIOS version: 1603 (ASUS M2A-VM HDMI ACPI BIOS Revision 1603. Release Date: 11/30/2007).
2007-12-22: the machine would not boot properly with all 4 DIMMs installed. It either failed, or froze. I had to remove two DIMMs, then adjust DDR2 Voltage Control in BIOS (Advanced, JumperFree Configuration) from "auto" to "+2.10V". I should have known - there was a warning note with the memory modules about using the proper voltage.
2007-12-22: labels for the power and reset button cables was reversed (the power button connector was labeled reset ...). Easy - I just switched the plugs when connecting them to the motherboard.
2007-12-22: The fan of the cpu cooler sits just above the DIMM closest to the cpu. You have to remove the fan if you want to remove the DIMM. Luckily, the fan is attached with two springs, so it is easy to remove it.