molnarcs's Journal: my puter or my OS borked

...I think. Left it building KDE 3.2.2 for the night, when returning saw an error message, saying something about hd timeouts. Reboot, single user mode, fsck. While doing /dev/ad0s1f (my /usr)I saw the same timeout/resetting devices message. After a couple of yes, fix it sessions, fsck exited, saying filesystem is still dirty. Damn! This is not the first time it happens btw (it happened on another machine, running fbsd 5.x).

The error codes were different each time, but 51 was most frequent. So the question is: is this a hw quirk or a sw bug (ffs/ata code in 5.2 branch?). If it happened once, I would have thought the former, but now, that two puters have borked displaying the same errors, I'm suspicious.

I mounted /usr with -f, so I could archive my important data on another hd - sometimes saw the same resetting devices crap. I can't tell how much I'm frustrated with this. I wasn't planning for a reinstall (and I'm not sure that would help either. Good thing I did 'make packages' for large ones like OpenOffice - at least I don't have to build them to have athlon and O2 optimized packs.

What really bothers is is that I don't know what to do PR-wise. The other puter began degrading after multiple reboots (power went out). FSCK always repaired inconsistencies ... for a time. Then I just saw those damn error messages, even though fsck reported filesystem to be clean at boottime(first I didn't even think that there is a connection) - ran manually, it found inconsistencies. But now, that it happens to another puter, I don't know what to think anymore. The problem is, this is not an easily reproducible bug. OS install went without problems on both machines, I didn't even have to do the disable acpi or dma trick. Doh!

I never reply to these, but.

[Coniptor]

I never reply to these but I wanted to know.
I had the same problem and determined in my case when it happend that for some reason my system (Redhat) was booting up and checking with ext3 but when finished mounting I was using ext2.
If you run fsck on a ext2 mounted drive which has ext3 data you will loose alot of data.
At least I did when this happended to me about 2 to 2 and a half years ago.
At the time I solved by recompiling my kernel with ext3 in the kernel along with ext2. I think it was probably a scrip

Re:I never reply to these, but.

[molnarcs]

Thanks for your comment :) - The question is: can software damage my hd? My only reason to be suspicious is that it happened twice on different machines (with different chipsets and processors), but on the same OS. Filesystem is UFS2 (native to FreeBSD 5.x), and I ask this because since I discovered the problem, sometimes my hd makes strange noises. One or twice, when switching on my puter, it acted as if it was switching on and off (the hd) - it was a really scary noise.

Also, after backing up, I reinstall

Re:I never reply to these, but.

[Coniptor]

I have one question. Assuming your using ide drives are they Maxtor at all? I have had a few maxtor drives die on me strangly exhibiting the same problems you describe and a friend of mine has as well. It could be the hardware if they're maxtor or even western digital. I haven't bought any new drives recently but my friend and I are looking to buy seagate as we've both had problems with wd's prior to maxtor's.
I had a maxtor drive in a dual boot machine between linux and winxp. winxp was primary on the mothe