[SOLVED] ata error on HD, but which one?

madeye · 2015-01-25 09:25:47

Hi there,

My server has started to give me some ata errors. They seem to be consistent, and appears all the time. A little excerpt can be seen below:

[Wed Jan 14 19:42:47 2015] ata20.00: status: { DRDY }
[Wed Jan 14 19:42:47 2015] ata20.00: failed command: WRITE FPDMA QUEUED
[Wed Jan 14 19:42:47 2015] ata20.00: cmd 61/00:a0:00:44:2a/04:00:20:01:00/40 tag 20 ncq 524288 out
         res 40/00:c4:58:08:00/00:00:20:01:00/40 Emask 0x10 (ATA bus error)
[Wed Jan 14 19:42:47 2015] ata20.00: status: { DRDY }
[Wed Jan 14 19:42:47 2015] ata20.00: failed command: WRITE FPDMA QUEUED
[Wed Jan 14 19:42:47 2015] ata20.00: cmd 61/00:a8:00:48:2a/04:00:20:01:00/40 tag 21 ncq 524288 out
         res 40/00:c4:58:08:00/00:00:20:01:00/40 Emask 0x10 (ATA bus error)
[Wed Jan 14 19:42:47 2015] ata20.00: status: { DRDY }
[Wed Jan 14 19:42:47 2015] ata20.00: failed command: WRITE FPDMA QUEUED
[Wed Jan 14 19:42:47 2015] ata20.00: cmd 61/00:b0:00:4c:2a/04:00:20:01:00/40 tag 22 ncq 524288 out
         res 40/00:c4:58:08:00/00:00:20:01:00/40 Emask 0x10 (ATA bus error)
[Wed Jan 14 19:42:47 2015] ata20.00: status: { DRDY }
[Wed Jan 14 19:42:47 2015] ata20.00: failed command: WRITE FPDMA QUEUED
[Wed Jan 14 19:42:47 2015] ata20.00: cmd 61/00:b8:00:50:2a/04:00:20:01:00/40 tag 23 ncq 524288 out
         res 40/00:c4:58:08:00/00:00:20:01:00/40 Emask 0x10 (ATA bus error)
[Wed Jan 14 19:42:47 2015] ata20.00: status: { DRDY }
[Wed Jan 14 19:42:47 2015] ata20.00: failed command: READ FPDMA QUEUED
[Wed Jan 14 19:42:47 2015] ata20.00: cmd 60/08:c0:58:08:00/00:00:20:01:00/40 tag 24 ncq 4096 in
         res 40/00:c4:58:08:00/00:00:20:01:00/40 Emask 0x10 (ATA bus error)
[Wed Jan 14 19:42:47 2015] ata20.00: status: { DRDY }
[Wed Jan 14 19:42:47 2015] ata20: hard resetting link
[Wed Jan 14 19:42:48 2015] ata20: SATA link up 3.0 Gbps (SStatus 123 SControl 320)
[Wed Jan 14 19:42:48 2015] ata20.00: configured for UDMA/133
[Wed Jan 14 19:42:48 2015] ata20: EH complete

I am afraid one of the drives are going bad. I think I know which drive it is, but smartctl does not show any logged errors on the drive.
Now my main question is:
Is there a way to see which drive is actually connected to ata20:00? That way I can concentrate my efforts on this specific drive instead of fumbling around in the dark.

From seaching on the forum and google, I found other people having similar errors. But none of them had an explanation of the connection between ataXX:00 and a disk.

Last edited by madeye (2015-01-25 17:11:04)

nstgc · 2015-01-25 14:33:16

I had a question with the answer you are looking for: https://bbs.archlinux.org/viewtopic.php?id=192172

[edit] Hmm, this isn't quite the same, you want the physical disk.

First:

djgera wrote:

Hi, According to errors message, the problem is in the link layer.
Identify what disk is connected to ata15.00 (ls -l /sys/block/sd*), and replace the cable.

Now that will give you block device. You can then use something like smartctl to get the serial number of the drive with "smartctl -a /dev/sdX" where sdX is the problem device.

Last edited by nstgc (2015-01-25 14:35:10)

madeye · 2015-01-25 17:10:33

Cool. Thanks. Now I know for sure what drive it is. Now to find the reason.
And unfortunately it was not the disk I was suspecting.

ls -l /sys/block/sd*|grep ata20

gives me

lrwxrwxrwx 1 root root 0 Jan  2 13:12 /sys/block/sdg -> ../devices/pci0000:00/0000:00:1c.7/0000:09:00.0/ata20/host19/target19:0:0/19:0:0:0/block/sdg

So I'll mark this as solved. My question was answered perfectly!

Arch Linux

#1 2015-01-25 09:25:47

[SOLVED] ata error on HD, but which one?

#2 2015-01-25 14:33:16

Re: [SOLVED] ata error on HD, but which one?

#3 2015-01-25 17:10:33

Re: [SOLVED] ata error on HD, but which one?

Board footer