dead disk on Centos 6

timl · 12-27-2016, 05:18 PM

Hi,

this morning my server (Centos 6) was spewing out messages about a disk problem. Mistake #1 - I thought a reboot would fix it so I did not write down the error. In fact I could not boot the system as the boot loader is on this disk. So I fell back to a linux mint rescue CD I use. This revealed that the "other" disk - the backup disk - is intact, while the disk with the operating system is not. I swapped over cables to see whether this was the problem but no joy. Similarly, I tried different ports on the motherboard to no avail.

The problem is that while the "important" stuff is backed up...There are a lot of "doodles" on the dead disk which have mounted up over the years and I would really like to recover these. A simple command from the rescue CD

Quote:

sudo mount /dev/sdb1 /mnt/aaa/ -t ext4
mount: special device /dev/sdb1 does not exist
[32]mint@mint ~ $ sudo mount /dev/sda1 /mnt/aaa/ -t ext4
mint@mint ~ $ ls /mnt/aaa/
backup lost+found tim

and a suggestion I picked up from the net

Quote:

[1]mint@mint ~ $ sudo sfdisk -R /dev/sdb
BLKRRPART: Input/output error

I also ran gparted and this reported an error reading this disk and it's display just shows the content of the disk as "unallocated".

So I can simply stick in a new disk and recover all my backups but I would like to look at the contents of the dead disk. Any suggestions or troubleshooting hints?

Cheers

syg00 · 12-27-2016, 05:26 PM

Doesn't sound promising.
dmesg should tell you the extent of the errors and whether a device node was created for it at boot (Mint liveCD). What does "lsblk -f" return ?.

wpeckham · 12-27-2016, 05:27 PM

Quote:

Originally Posted by timl

Hi,

this morning my server (Centos 6) was spewing out messages about a disk problem. Mistake #1 - I thought a reboot would fix it so I did not write down the error. In fact I could not boot the system as the boot loader is on this disk. So I fell back to a linux mint rescue CD I use. This revealed that the "other" disk - the backup disk - is intact, while the disk with the operating system is not. I swapped over cables to see whether this was the problem but no joy. Similarly, I tried different ports on the motherboard to no avail.

The problem is that while the "important" stuff is backed up...There are a lot of "doodles" on the dead disk which have mounted up over the years and I would really like to recover these. A simple command from the rescue CD

and a suggestion I picked up from the net

I also ran gparted and this reported an error reading this disk and it's display just shows the content of the disk as "unallocated".

So I can simply stick in a new disk and recover all my backups but I would like to look at the contents of the dead disk. Any suggestions or troubleshooting hints?

Cheers

If sections of that disk are truly dead, there is a good reason. Depending upon the reason, you might be able to get access to some files, or you may get nothing at all.
Going forward, you need to back up EVERYTHING (but only about as often as it changes), and get that backup OFF SITE. Another disk on the same server, or in the next room, is not really good enough if your data matters.

Now to the present: I would load up a live-cd image with photorec and give it a couple of days (more or less) to work on that disk. If you can get anything at all, it will tell you.

If you have the option, I would try to image the disk first as a special backup. As it works, the problem may get much worse and prevent access to the things that are currently working, so preserve what you can.

timl · 12-27-2016, 05:39 PM

The first snippet is this:

Quote:

NAME FSTYPE LABEL MOUNTPOINT
sda
└─sda1
sdb
sr0 iso9660 LXFDVD184 /lib/live/mount/fromiso
loop0 /lib/live/mount/medium
loop1 /lib/live/mount/rootfs/filesystem.squashfs

I got this from dmesg (more further) down

Quote:

[Tue Dec 27 23:07:23 2016] sdb: unable to read partition table

Which says it all. I'll have a dig around for restoreing partition tables

dmesg is quite long so I have pasted in the first reference to sdb

Quote:

[Tue Dec 27 23:07:23 2016] ata5.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0
[Tue Dec 27 23:07:23 2016] ata5.00: irq_stat 0x40000001
[Tue Dec 27 23:07:23 2016] ata5.00: failed command: READ DMA
[Tue Dec 27 23:07:23 2016] ata5.00: cmd c8/00:02:02:00:00/00:00:00:00:00/e0 tag 0 dma 1024 in
[Tue Dec 27 23:07:23 2016] res 51/04:02:02:00:00/00:00:00:00:00/e0 Emask 0x1 (device error)
[Tue Dec 27 23:07:23 2016] ata5.00: status: { DRDY ERR }
[Tue Dec 27 23:07:23 2016] ata5.00: error: { ABRT }
[Tue Dec 27 23:07:23 2016] ata5.00: configured for UDMA/133
[Tue Dec 27 23:07:23 2016] ata5: EH complete
[Tue Dec 27 23:07:23 2016] ata5.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0
[Tue Dec 27 23:07:23 2016] ata5.00: irq_stat 0x40000001
[Tue Dec 27 23:07:23 2016] ata5.00: failed command: READ DMA
[Tue Dec 27 23:07:23 2016] ata5.00: cmd c8/00:02:02:00:00/00:00:00:00:00/e0 tag 0 dma 1024 in
[Tue Dec 27 23:07:23 2016] res 51/04:02:02:00:00/00:00:00:00:00/e0 Emask 0x1 (device error)
[Tue Dec 27 23:07:23 2016] ata5.00: status: { DRDY ERR }
[Tue Dec 27 23:07:23 2016] ata5.00: error: { ABRT }
[Tue Dec 27 23:07:23 2016] ata5.00: configured for UDMA/133
[Tue Dec 27 23:07:23 2016] ata5.00: failed command: READ DMA
[Tue Dec 27 23:07:23 2016] ata5.00: cmd c8/00:04:04:00:00/00:00:00:00:00/e0 tag 0 dma 2048 in
[Tue Dec 27 23:07:23 2016] res 51/04:04:04:00:00/00:00:00:00:00/e0 Emask 0x1 (device error)
[Tue Dec 27 23:07:23 2016] ata5.00: status: { DRDY ERR }
[Tue Dec 27 23:07:23 2016] ata5.00: error: { ABRT }
[Tue Dec 27 23:07:23 2016] ata5.00: configured for UDMA/133
[Tue Dec 27 23:07:23 2016] sd 4:0:0:0: [sdb]
[Tue Dec 27 23:07:23 2016] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
[Tue Dec 27 23:07:23 2016] sd 4:0:0:0: [sdb]
[Tue Dec 27 23:07:23 2016] Sense Key : Aborted Command [current] [descriptor]
[Tue Dec 27 23:07:23 2016] Descriptor sense data with sense descriptors (in hex):
[Tue Dec 27 23:07:23 2016] 72 0b 00 00 00 00 00 0c 00 0a 80 00 00 00 00 00
[Tue Dec 27 23:07:23 2016] 00 00 00 04
[Tue Dec 27 23:07:23 2016] sd 4:0:0:0: [sdb]
[Tue Dec 27 23:07:23 2016] Add. Sense: No additional sense information
[Tue Dec 27 23:07:23 2016] sd 4:0:0:0: [sdb] CDB:
[Tue Dec 27 23:07:23 2016] Read(10): 28 00 00 00 00 04 00 00 04 00
[Tue Dec 27 23:07:23 2016] ata5: EH complete
[Tue Dec 27 23:07:23 2016] sdb: unable to read partition table
[Tue Dec 27 23:07:23 2016] sd 4:0:0:0: [sdb] CDB:
[Tue Dec 27 23:07:23 2016] Read(10): 28 00 00 00 00 02 00 00 02 00
[Tue Dec 27 23:07:23 2016] end_request: I/O error, dev sdb, sector 2
[Tue Dec 27 23:07:23 2016] Buffer I/O error on device sdb, logical block 1
[Tue Dec 27 23:07:23 2016] ata5: EH complete
[Tue Dec 27 23:07:23 2016] ata5.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0
[Tue Dec 27 23:07:23 2016] ata5.00: irq_stat 0x40000001
[Tue Dec 27 23:07:23 2016] ata5.00: failed command: READ DMA
[Tue Dec 27 23:07:23 2016] ata5.00: cmd c8/00:04:04:00:00/00:00:00:00:00/e0 tag 0 dma 2048 in
[Tue Dec 27 23:07:23 2016] res 51/04:04:04:00:00/00:00:00:00:00/e0 Emask 0x1 (device error)
[Tue Dec 27 23:07:23 2016] ata5.00: status: { DRDY ERR }
[Tue Dec 27 23:07:23 2016] ata5.00: error: { ABRT }
[Tue Dec 27 23:07:23 2016] ata5.00: configured for UDMA/133
[Tue Dec 27 23:07:23 2016] ata5: EH complete
[Tue Dec 27 23:07:23 2016] ata5.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0
[Tue Dec 27 23:07:23 2016] ata5.00: irq_stat 0x40000001
[Tue Dec 27 23:07:23 2016] ata5.00: failed command: READ DMA
[Tue Dec 27 23:07:23 2016] ata5.00: cmd c8/00:04:04:00:00/00:00:00:00:00/e0 tag 0 dma 2048 in

rknichols · 12-27-2016, 05:59 PM

It's bad that the report is "device error" and not "medium error". That suggests that the problem is in the drive's controller, not the disk surface, and that you won't be able to read anything at all from that disk. Professional data recovery would stand a good chance of recovering everything from that disk, but that's going to be expensive. How much are your "doodles" worth to you?

stanvan · 12-27-2016, 06:03 PM

I used ddrescue a long while back to recover data from a dead USB stick. So long ago, in fact, that I can't offer any better advice than just the name of this tool. Sorry. But wishing you luck to recover your data.

timl · 12-27-2016, 08:47 PM

thanks for all the feedback. I was really testing the water to see whether the disk was recoverable and it sounds like it isn't easily so I'll ponder my next move. I have a spare drive available so I might set the system back up again and plug in that disk later.

The content is mostly home directories from other computers I used to back up to that server. Also fstab and smb.conf would have been good as reference points but it is not the end of the world.

Cheers