bad ext2fs corruption in 2.2.16

From: Sean Harding (
Date: Sun Aug 27 2000 - 21:25:50 EST

I'm experiencing a problem on my laptop that is seemingly resulting in
rapid, massive ext2fs corruption on an IDE drive. The problem seems to be
correlated to use of "cdparanoia" to rip audio data from audio CDs to the
hard drive. I'm running kernel 2.2.16 on a Dell Latitude CPi A366T. 366 MHz
Pentium II, 128MB of RAM, 6GB hard drive.

IDE controller and disk info:

PIIX4: IDE controller on PCI bus 00 dev 39
PIIX4: not 100% native mode: will probe irqs later
    ide0: BM-DMA at 0x0860-0x0867, BIOS settings: hda:DMA, hdb:pio
    ide1: BM-DMA at 0x0868-0x086f, BIOS settings: hdc:DMA, hdd:pio
hda: IBM-DADA-26480, ATA DISK drive
ide2: ports already in use, skipping probe
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
ide1 at 0x170-0x177,0x376 on irq 15
hda: IBM-DADA-26480, 6194MB w/460kB Cache, CHS=789/255/63, UDMA
hdc: ATAPI 24X CD-ROM drive, 128kB Cache
Uniform CD-ROM driver Revision: 3.09
Partition check:
 hda: hda1 hda2 < hda5 hda6 >

This system has been in service since June of '99 without any major
disk-related problems (obviously some kernel upgrades). A few weeks ago, I
ran 'cdparanoia' and had processes start crashing randomly, so I rebooted
and fsck found lots of icky stuff. I was a little concerned, but after that
it went for weeks without any more problems (including several manual fscks
to be sure there wasn't more corruption creeping in). Today I ran
'cdparanoia' to rip some audio and almost instantly programs started
crashing again. Virtually everything I ran crashed with a seg fault. One of
the directories I was working in lost all of its entries, including "." and
"..". When the system came back up and fsck ran, there were lots and lots of
problems. Dozoens of "duplicate/bad block(s) in inode n:", "unattached
inode", "free blocks count wrong for group #n", etc. At one point there was
a message which scrolled by so quickly I couldn't read it with hundreds or
tousands of negative numbers (I'm assuming they were supposed to be block or
inode numbers), followd by "fix <y>?" I had at least 10 files in /usr/local
become corrupted. What was originally a tcl script was replaced with
apparently random binary data. What was previously a shared library was
replaced with data that begins:

sse: ") ("signature" "Unterschrift: ") ("opening" "Anrede: ") ("closing" "Schlu\
"s: ")] 15])
#@58 Insert ENVIRONMENT and prompt for recipient and address.

/lost+found has some interesting files in it:

eris /lost+found
9% ls -al
total 4353701167
c---rw--wx 1 16650 25963 32, 53 Oct 6 2023 #1162
c--xrw--wx 1 25963 29541 32, 102 Apr 2 2029 #1163
c---rwx-w- 1 29793 25710 10, 32 Apr 3 1987 #1165
c--xrw-r-- 1 28532 8294 48, 32 Oct 6 2023 #1167
br-xr-s-wx 1 28274 15717 54, 51 Aug 15 2027 #1168
b--Sr-S--- 1 28009 14900 101, 110 Mar 24 1987 #813
c---rw--w- 1 13360 22597 118, 101 Sep 19 2028 #814
drwxr-xr-x 2 root root 12288 Aug 27 18:54 ./
drwxr-xr-x 22 root root 1024 Jan 1 2000 ../

I don't have strong evidence that cdparanoia is causing this. It just has
been an awfully strange coincidence that both times the problem has
happened, it has been immediately after I started cdparanoia. And I have
seen absolutely no other disk-related problems.

Any ideas on what's going on here? Ideas for more troubleshooting and



Sean Harding        |"Well, contempt loves the silence             | It thrives in the dark"
                                        | --Natalie Merchant
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at

This archive was generated by hypermail 2b29 : Thu Aug 31 2000 - 21:00:20 EST