I have got a strange problem with slow NTFS crash on a new computer (assembled by me) after adding new memory module to it a day later after it was assembled.
Configuration: mobo: asrock am2nf2; Sempron AM2 3000+; 2xNCP 512Mb 667Mhz (old one and new problematic module) ; Seagate 80gb; Samsung 160Gb; New FSP 400W PowerSupply.
I noticed the problem only on the next day when the system did not boot.
First thing I did – executed memtest86 1.65+ for about 11 hours and it made about 20 passes without any errors after I turned the computer off. Also I tried windiag - also no errors.
I booted from recovery live CD and tried to check NTFS with chkdsk /f
It found many inconsistencies, index problems, etc. It tried to fix some chains but finally said it cannot continue and stopped. (Now I understand because its work was not stored correct on the hdd. Also it read incorrect information).
I tried to copy some important files to other computer on the network but most of them were damaged (due to MFT crash I guess)
Note: previously I used this Samsung drive for about 1.5 year without any problems. Checked it about every 3 months with chkdsk /f and everything was fine always. I was sure it is not a problem with HDD but to be sure I moved this HDD to another healthy Linux computer and tried to format it using mkfs –t ext3. Just after formatting done I tried to check the file system using fsck –f /dev/hda4 and it was fine. Then I moved HDD to problematic computer, booted from Gentoo 2006.1 mini install CD and found that fsck –f /dev/hda4 produced many errors, it was an IDE Samsung drive. To be finally sure it is not an HDD problem I purchased a new 401Gb SATA2 Samsung HDD and found it had the same problems on my mkfs/fsck test. On other healthy computer it passed mkfs/fsck without any problems. I tried to change linux kernel versions: 2.6.17; 2.6.18 it did not help. I tried to boot from Windows Live CD and format Samsung HDD on a problematic system. It refused to do this. I was very confused because other Seagate 80 Gb HDD worked fine on the problematic computer. It passed my mkfs/fsck test. So I understood it is problem with at least all Samsung (both SATA and IDE) drives and at least not with old Seagate drives. I tried to change the mobo and purchased new Gigabyte S3G-M55Plus and did not see any change with my problem. Later in forums I read that other people had similar problems on nForce chipsets with Maxtor, Hitachi, etc drives except Seagate. Also I was confused because other healthy system ran on am2nf3 motherboard purchased 2 weeks earlier than my second am2nf3 mobo with the same memory NCP.
Finally I decided to check this with each memory module separately and found that my mkfs/fsck test passed fine with two memory modules I purchased a few weeks ago in one shop and did not pass with the last module which was purchased in a different shop. I checked this on 3 mobo, two power supply modules, changed IDE and SATA cables many times. And the most confusing part was that all 3 modules passed all tests like: Memtest 86, WinDiag and MPrime (linux analog of Prime 95). So people if there are strange problems on your nForce(3,4,..) computers you may want to add a simple mkfs/fsck test to you collection of tests.
Happy End: Today I changed the module to similar NCP in the shop where previous one was purchased and it works fine so far. I had a backup of all the information and restored it.