Crash with hard disk errors (ATA exception)

Nock441 · 2013-12-27 02:29:41

My Arch just crashed displaying in a TTY errors like this:

[   24.229347] ata2.00: exception Emask 0x0 SAct 0x1 SErr 0x0 action 0x0
[   24.229397] ata2.00: irq_stat 0x40000008
[   24.229426] ata2.00: failed command: READ FPDMA QUEUED
[   24.229463] ata2.00: cmd 60/08:00:70:3e:c3/00:00:03:00:00/40 tag 0 ncq 4096 in
         res 51/40:08:70:3e:c3/00:00:03:00:00/40 Emask 0x409 (media error) <F>
[   24.229552] ata2.00: status: { DRDY ERR }
[   24.229579] ata2.00: error: { UNC }
[   24.234145] ata2.00: configured for UDMA/133
[   24.234168] sd 1:0:0:0: [sda] Unhandled sense code
[   24.234171] sd 1:0:0:0: [sda]  
[   24.234173] Result: hostbyte=0x00 driverbyte=0x08
[   24.234176] sd 1:0:0:0: [sda]  
[   24.234177] Sense Key : 0x3 [current] [descriptor]
[   24.234181] Descriptor sense data with sense descriptors (in hex):
[   24.234183]         72 03 11 04 00 00 00 0c 00 0a 80 00 00 00 00 00 
[   24.234192]         03 c3 3e 70 
[   24.234196] sd 1:0:0:0: [sda]  
[   24.234198] ASC=0x11 ASCQ=0x4
[   24.234200] sd 1:0:0:0: [sda] CDB: 
[   24.234202] cdb[0]=0x28: 28 00 03 c3 3e 70 00 00 08 00
[   24.234210] end_request: I/O error, dev sda, sector 63127152
[   24.234261] Buffer I/O error on device dm-1, logical block 7864270
[   24.234321] ata2: EH complete

I shutdown the laptop by pressing the power button. On the reboot, a lot of these errors displayed but the system booted successfully. Here is the dmsg : http://pastebin.com/FAEAXwwY

It is a 8 months old MSI GT70 laptop which never suffered from mechanical shocks in any way in the past.

Should I get worried about this and start making — full — backups?

okubax · 2013-12-27 02:38:33

Yes. Similar issue here

*On another forum, someone claimed to have 'fixed' it by first unmounting the affected partition/drive and running "e2fsck -c -c" on it

Last edited by okubax (2013-12-27 02:42:19)

Nock441 · 2013-12-27 03:37:19

The system does not boot anymore. Here is the boot log: http://imgur.com/cUsvmYl

I cannot e2fsck the /dev/sda2 partition since it is a LUKS volume that contains a LVM PV and not an ext fs.

EDIT: I unplugged the HDD and booted on the Arch CD but I get this error: http://imgur.com/pahBZWw

Last edited by Nock441 (2013-12-27 04:06:23)

R00KIE · 2013-12-27 15:22:38

Get a new disk, at least as big as that one, and do a full image before you try anything else.
Those errors seems to be a very strong warning to start making full backups.

Next you can try to boot with a livecd and recover most data and maybe try to fsck the partiton(s) if you can still unlock the luks volume and enable the lvm pv and get access to the lv inside.

Nock441 · 2013-12-27 17:58:05

The problem seems fixed now, at least I hope so.

If the live CD did not booted this isn't because of the hard drive but of my NVIDIA Optimus config, adding “acpi=off” as kernel parameter fixed that.

In the live CD I was able to open the LUKS volume, activate the LVM VG and mount the LV without any errors or visible problems (such as lags).

Then I launched a “e2fsck -c -c” on the root LV (after unmouting it, of course); it reported 9 errors (9/0/0). I answered “yes” to all questions of e2fsck.

I had the idea to regenerate the initramfs with mkinitcpio as I remembered I have canceled the last mkinitcpio during the process, maybe it generated a corrupted ramfs?

After rebooting, everything worked normally. But I'm afraid I only have a reprieve, I will still clone my disk while it is still readable.

Last edited by Nock441 (2013-12-27 17:59:12)

Arch Linux

#1 2013-12-27 02:29:41

Crash with hard disk errors (ATA exception)

#2 2013-12-27 02:38:33

Re: Crash with hard disk errors (ATA exception)

#3 2013-12-27 03:37:19

Re: Crash with hard disk errors (ATA exception)

#4 2013-12-27 15:22:38

Re: Crash with hard disk errors (ATA exception)

#5 2013-12-27 17:58:05

Re: Crash with hard disk errors (ATA exception)

Board footer