What the heck is wrong with Gentoo

[rkfsm]

Several weeks ago, my Gentoo machine died. It was like someone had picked 30% of my files at random and deleted them. I rebuilt again from scratch on a new machine. New motherboard, CPU and, most importantly, new HDD's. I made sure I used all new passwords. Well, now it has happened again, only worse. This time about 75% of the files are gone. Family photos, 20GB of music, work documents and my web site are all gone. Most of it has been backed up, but still.... WHY?

The damage is spread between two hard drives; a 40GB Maxtor and a 120GB Seagate.

Can someone tell me what I can do to analyze what's going on so that I can stop it?

RK

[Veldrin]

you mentioned new hdds: only for the system, or for you data too? 40 and 120 GB does not really sound new... how old are those 2?

install smartmontools. check the disk themselves, and run a long test...

[schachti]

Which file system did you use?
_________________
Never argue with an idiot. He brings you down to his level, then beats you with experience.

How-To: Daten verschlüsselt auf DVD speichern.

[rkfsm]

/dev/hda1 /boot ext2 (OK)
/dev/hda3 / reiserfs (damaged)
/dev/hdb2 /usr/portage reiserfs (OK)
/dev/hdb3 /pub jfs (damaged)

hda2 and hdb1 are both swap partitions.

Every directory had damage done. /boot was deleted entirely, /home and /root still existed but were empty. /bin and /sbin had only a few files in each.

The system log has been corrupted.

RK

[Habbit]

Well, I don't know of a "rampage mode" kernel patch being included in the gentoo-sources patches, so if you tell us the new machine is a new installation with no relationship to the old one (new hardware, software & system configs), I can only assume the damage comes from outside. The main causes that spring to mind are power failures (usually here some people will scream reiserfs is to blame, but a rock-hard jfs partition has also been damaged it seems) and human attack. You mentioned you ran a website. Can you check any logs that haven't been compromised? (you mentioned the syslog was corrupted, maybe by an attacker, so you could try with your ISP's) Have you put proper security measures? (new passwords are not enough, the Hardened profile would do much better for a server)
_________________

[MostAwesomeDude]

I would gladly blame reiserfs, but something else is going on here.

First off, were the disks dirty at any point, or did they always mount clean? Are the disks really new, especially the 40GB one? (A quick search is telling me that there are Maxtor is not making new 40GB drives.) What does badblocks have to say?
_________________
Don't believe the "n00b" under my name.

[rkfsm]

The Maxtor is fairly old and came from an unused MS box, but smartctl shows no errors. The Seagate was given to me at Christmas.

I have not had any dirty partitions for a long time. However, after my system crashed, when I was getting errors saying things like the init could not be found, I had to do a hard reset and when I ran fsck from a livecd to look at the damage, I had dozens of screens of text about missing links and metadata.

It's all moot now.

I have reformated and am installing hardened with SSP, PIE, PAX and Selinux on XFS (root, home and pub) and EXT2 (boot and usr/portage) partitions.

RK

[padoor]

i have a feeling new version of boost has something to do with this problem.
i have no proof to support
but the followings happened.
i crashed my very nicely working system by installing kde-4 which failed to install withglxchooseview error.
then i run a depclean and kill my whole system with missing gcc among others.
reinstall is what i decided to do and started new installation
before emerging kde4 as my desktop i updated the system only with xorg installed.
expat problem was small as small world only present.
then after emerging boost i found libstdc++ missing
revdep rebuild installed gcc-3.3,4 and stdc problem solved.
then i find that the boost wants to emerge again this time gcc-4.1.2 went missing.
it took 6 tries to emerge boost latest version.
many times it tried to kill init,reboot,some files missing etc.
tis can be your reason for trouble.
_________________
reach out a little bit more to catch it (DON'T BELIEVE the advocate part under my user name)

[likewhoa]

[gkmac]

[Clad in Sky]

Hello there,
I got a similar problem with a hdd. I ran the smartctl -t long command.
How do I get the results of that test? Shell says:

[psycco]

smartctl --all /dev/hda (Prints all SMART information)

[jcat]

[cwr]

I've seen something like this, tho' not with Gentoo, when the CMOS setting went out
to lunch and corrupted the memory timings. Everything appeared to work, and every
file and directory written was corrupted. See if your memory timings are what you
think they are.

Good luck - Will