The filesystem disappear

antiqui.populi · 10-26-2006, 12:36 PM

Hi. Sorry about my english.
I'm running Red Hat 8, with

Code:

uname -a
Linux 2.4.18-14 #1 Wed Sep 4 13:35:50 EDT 2002 i686 i686 i386 GNU/Linux

When I reboot, two partitions of the hard drive /dev/sdc (that's is, /dev/sdc1, /dev/sdc2) are mounted without a problem, at /mnt/sdc1 and /mnt/sdc2. I run a program which read & write files in this partitions. This program takes to run ~1 day. So I let the program running, I go back to my house. The next day, when I come back to work, the entire filesystem on /mnt/sdc1 and /mnt/sdc2 had disappeard. It's like the HD is empty.That is, if I make

Code:

[root@INTA01 boot]# ls -l /mnt/sdc1
total 0

The program that I was running don't appear if I make a ps -A

The partition is still mounted (if I try to mount it again, it says that is already mounted). If I umount and try to mount it again, I get

Code:

[root@INTA01 boot]# mount -t ext3 /dev/sdc1 /mnt/sdc1/
mount: wrong fs type, bad option, bad superblock on /dev/sdc1,
       or too many mounted file systems

If I make fsck, I get

Code:

[root@INTA01 boot]# /sbin/fsck.ext3 /dev/sdc1
e2fsck 1.27 (8-Mar-2002)
Couldn't find ext2 superblock, trying backup blocks...
/sbin/fsck.ext3: Bad magic number in super-block while trying to open /dev/sdc1

The superblock could not be read or does not describe a correct ext2
filesystem.  If the device is valid and it really contains an ext2
filesystem (and not swap or ufs or something else), then the superblock
is corrupt, and you might try running e2fsck with an alternate superblock:
    e2fsck -b 8193 <device>

So one would think that the superblock is corrupted. I tryed using alternatives superblocks, and they are also broken.

Now, the WEIRD THING, is that if I reboot, everything goes back to normal: the partitions are well mounted, I can see the files stored in them. But, if I repeat the sequence (let this program running->return the next day) the filesystem disappeard again.

Any clues? If you need more information, just tell me

Thanks you very much

fordeck · 10-27-2006, 08:14 AM

What do you see in your logs (i.e. /var/log/messages)?
What is this program you're running do?

blackhole54 · 10-27-2006, 08:26 AM

Is this program running as root?
Is this computer connected to a network?

antiqui.populi · 10-27-2006, 08:36 AM

Thank for the replies.

As you can guess, I'm totally newbee, so I don't know exactly where to look when I get errors.
So, this is my /var/log/messages

Code:

Oct 26 05:09:11 INTA01 kernel: 08:21: rw=0, want=714690832, limit=71681998
Oct 26 05:09:11 INTA01 kernel: attempt to access beyond end of device
Oct 26 05:09:11 INTA01 kernel: 08:21: rw=0, want=714866740, limit=71681998
Oct 26 05:09:11 INTA01 kernel: attempt to access beyond end of device
Oct 26 05:09:11 INTA01 kernel: 08:21: rw=0, want=715128884, limit=71681998
Oct 26 05:09:11 INTA01 kernel: attempt to access beyond end of device
Oct 26 05:09:11 INTA01 kernel: 08:21: rw=0, want=715391028, limit=71681998
Oct 26 05:09:11 INTA01 kernel: attempt to access beyond end of device
...
...
...

Oct 26 06:09:38 INTA01 kernel: EXT3-fs error (device sd(8,33)): ext3_add_entry: bad entry in directory #360449: rec_len is smaller than minimal - offset=0, inode=2147549472, rec_len=5, name_len=12
Oct 26 06:09:48 INTA01 kernel: EXT3-fs error (device sd(8,33)): ext3_readdir: bad entry in directory #2: rec_len is smaller than minimal - offset=0, inode=131616, rec_len=0, name_len=12

Of course, this looks ugly. It's look like the filesystem was bad partitionated, but I think I use all the defaults options for mkfs.ext3, and the HD is brand new
But maybe I had make a mistake, I'll check again

The program I'm running just read and write files, in a very usual way. It runs fine in others HD.

blackhole54 · 10-28-2006, 01:29 AM

About a year ago I got some error messages in a RH 8.0 system that (IIRC) were similar to yours. The logs are long gone. I never knew what caused them and I never had the entire contents of a partition "disappear." But it happened twice within about a week, and IIRC I had to repair with e2fsck. I used the situation as motivation to upgrade to the (then) current vanilla 2.4 kernel, which I had been planning on doing anyway. I have not had problems since. Of course, I have no way of knowing whether I would have continued to have problems had I not upgraded. I am just giving you this info to use as you wish.

If you wish to do what I did, I will give you a word of caution. After I upgraded the kernel, I had a problem that when I logged in from the keyboard as normal user I would get some error (no problem when logging in as root). You can read about it in this post. After learning what I put in that post I didn't try to finish solving the problem since I usually run this machine headless and therefore do not have a problem with that error message. Also, if you don't use unicode, I don't think that message is anything more than a nuisance.