[SOLVED] ata Exception: Interface fatal error

[Commifreak]

SOLVED
Check your cables/hard drives if you get this error!


===========================================================================

Hey

Just testing latest beta on my System:

Asus C60M1-I (http://www.asus.com/Motherboards/AMD_CPU_on_Board/C60M1I) with AMD FCH A50M (Hudson M1) Chipset
4 x WD20EARX (all working! tested with WD-Diagnostics)

Raid 1 for /swap (for all devices)
Raid 1 for /boot (for all devices)
Raid 5 for LVM

LVM for / and /home partitions.

Sata-Type: AHCI

The first boot shows following errors (all other boots shows them, too):

Code: [Select]

Jul  1 00:26:32 server kernel: [  136.120728] ata3.00: exception Emask 0x10 SAct 0x2 SErr 0x400000 action 0x6 frozen
Jul  1 00:26:32 server kernel: [  136.120739] ata3.00: irq_stat 0x08000000, interface fatal error
Jul  1 00:26:32 server kernel: [  136.120747] ata3: SError: { Handshk }
Jul  1 00:26:32 server kernel: [  136.120754] ata3.00: failed command: WRITE FPDMA QUEUED
Jul  1 00:26:32 server kernel: [  136.120768] ata3.00: cmd 61/f8:08:08:25:d2/02:00:69:00:00/40 tag 1 ncq 389120 out
Jul  1 00:26:32 server kernel: [  136.120771]          res 40/00:0c:08:25:d2/00:00:69:00:00/40 Emask 0x10 (ATA bus error)
Jul  1 00:26:32 server kernel: [  136.120777] ata3.00: status: { DRDY }
Jul  1 00:26:32 server kernel: [  136.120789] ata3: hard resetting link
Jul  1 00:26:32 server kernel: [  136.616370] ata3: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
Jul  1 00:26:32 server kernel: [  136.626188] ata3.00: configured for UDMA/133
Jul  1 00:26:32 server kernel: [  136.626232] ata3: EH complete
Little time later...

Code: [Select]

Jul  1 00:27:40 server kernel: [  204.428986] ata3.00: exception Emask 0x10 SAct 0x3 SErr 0x400000 action 0x6 frozen
Jul  1 00:27:40 server kernel: [  204.429002] ata3.00: irq_stat 0x08000000, interface fatal error
Jul  1 00:27:40 server kernel: [  204.429014] ata3: SError: { Handshk }
Jul  1 00:27:40 server kernel: [  204.429026] ata3.00: failed command: WRITE FPDMA QUEUED
Jul  1 00:27:40 server kernel: [  204.429048] ata3.00: cmd 61/f0:00:08:35:12/03:00:6a:00:00/40 tag 0 ncq 516096 out
Jul  1 00:27:40 server kernel: [  204.429053]          res 40/00:04:08:35:12/00:00:6a:00:00/40 Emask 0x10 (ATA bus error)
Jul  1 00:27:40 server kernel: [  204.429063] ata3.00: status: { DRDY }
Jul  1 00:27:40 server kernel: [  204.429072] ata3.00: failed command: WRITE FPDMA QUEUED
Jul  1 00:27:40 server kernel: [  204.429092] ata3.00: cmd 61/08:08:f8:38:12/03:00:6a:00:00/40 tag 1 ncq 397312 out
Jul  1 00:27:40 server kernel: [  204.429097]          res 40/00:04:08:35:12/00:00:6a:00:00/40 Emask 0x10 (ATA bus error)
Jul  1 00:27:40 server kernel: [  204.429106] ata3.00: status: { DRDY }
Jul  1 00:27:40 server kernel: [  204.429124] ata3: hard resetting link
Jul  1 00:27:41 server kernel: [  204.924371] ata3: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
Jul  1 00:27:41 server kernel: [  204.933686] ata3.00: configured for UDMA/133
Jul  1 00:27:41 server kernel: [  204.933731] ata3: EH complete
Jul  1 00:28:22 server kernel: [  246.749768] ata3: limiting SATA link speed to 3.0 Gbps
Jul  1 00:28:22 server kernel: [  246.749786] ata3.00: exception Emask 0x10 SAct 0x7 SErr 0x400000 action 0x6 frozen
Jul  1 00:28:22 server kernel: [  246.749798] ata3.00: irq_stat 0x08000000, interface fatal error
Jul  1 00:28:22 server kernel: [  246.749809] ata3: SError: { Handshk }
Jul  1 00:28:22 server kernel: [  246.749820] ata3.00: failed command: WRITE FPDMA QUEUED
Jul  1 00:28:22 server kernel: [  246.749841] ata3.00: cmd 61/f0:00:08:e5:3c/03:00:6a:00:00/40 tag 0 ncq 516096 out
Jul  1 00:28:22 server kernel: [  246.749846]          res 40/00:04:08:e5:3c/00:00:6a:00:00/40 Emask 0x10 (ATA bus error)
Jul  1 00:28:22 server kernel: [  246.749856] ata3.00: status: { DRDY }
Jul  1 00:28:22 server kernel: [  246.749864] ata3.00: failed command: WRITE FPDMA QUEUED
Jul  1 00:28:22 server kernel: [  246.749883] ata3.00: cmd 61/08:08:f8:e8:3c/00:00:6a:00:00/40 tag 1 ncq 4096 out
Jul  1 00:28:22 server kernel: [  246.749888]          res 40/00:04:08:e5:3c/00:00:6a:00:00/40 Emask 0x10 (ATA bus error)
Jul  1 00:28:22 server kernel: [  246.749897] ata3.00: status: { DRDY }
Jul  1 00:28:22 server kernel: [  246.749905] ata3.00: failed command: WRITE FPDMA QUEUED
Jul  1 00:28:22 server kernel: [  246.749924] ata3.00: cmd 61/00:10:00:e9:3c/03:00:6a:00:00/40 tag 2 ncq 393216 out
Jul  1 00:28:22 server kernel: [  246.749928]          res 40/00:04:08:e5:3c/00:00:6a:00:00/40 Emask 0x10 (ATA bus error)
Jul  1 00:28:22 server kernel: [  246.749937] ata3.00: status: { DRDY }
Jul  1 00:28:22 server kernel: [  246.749953] ata3: hard resetting link
Jul  1 00:28:23 server kernel: [  247.240375] ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 320)
Jul  1 00:28:23 server kernel: [  247.249980] ata3.00: configured for UDMA/133
Jul  1 00:28:23 server kernel: [  247.250027] ata3: EH completeGoogle said, this could be a bug, cables and hdds are in good conditions - maybe because mdadm is working on resync the new array?

Can someone give further informations?

Thanks!

[Escorpiom]

Could you be more specific about the board or system you are using?
Asus CM601-I does not give me any direct hits on Google.
I'm interested in the sata controller you are using.

Cheers.

[christian]

It doesn't help here, sorry, but I'm just curious: what's the purpose of Raid 1 for /swap ?

[Commifreak]

Could you be more specific about the board or system you are using?
Asus CM601-I does not give me any direct hits on Google.
I'm interested in the sata controller you are using.

Cheers.

Sorry - its the C60M1-I - Typo in my post above...

--> http://www.asus.com/Motherboards/AMD_CPU_on_Board/C60M1I

The board is new, this is the first set-up with this hardware. The hard drives are used and ok. Maybe you're right: An issue with the sata-controller, AHCI and linux?
Can I list all drives connected to ata port 3? The log say its an device on ata3 - is it S-ATA port 3 on the board? Connected: 1-4.

Another hint was, to set libata.dma=0 as kernel option - but I want to know the reason! The previous AsRock-Board worked perfectly without any errors.

Or simply a Ubuntu-bug? I should try the latest zentyal-stable...

It doesn't help here, sorry, but I'm just curious: what's the purpose of Raid 1 for /swap ?

Linux puts different data to all hard drives. So, if one hard drive fails, it is still running if swap is on RAID, because all swap-data is still there. If its not on raid and something important was stored on a swap-partition of a disk which fails, linux crashes. Not directly tested, but it does not hurt to do it anyway

Thanks for all your answers!

[Commifreak]

So, I can leave this thread as solved 

I've checked all cable connections for errors. All seems fine. One cable was a little bit loose, so I've decided to change port 3 with 4.

The fault moves to port 4! So, I assume the sata-cable is defective But all hardware-tests (UBCD and WD-Tools) doesnt displayed any error.

Thanks for all your answers anyway!

[Escorpiom]

You're welcome!

That's a very nice alternative to the Atom platform, this APU will perform much better.
About the sata controller, I've read about issues with the latest Marvel controller and linux, but this does not apply here.

Cheers.

[christian]

That's a very nice alternative to the Atom platform, this APU will perform much better.

Indeed. I'm not sure difference is so big but, once you've added extra network card, it gives you a nice "green" platform for SMB server. C60 is slightly faster than N550, although only 2 cores compares to 4 due to hyper-threading.
Plenty of SATA and USB 

[Escorpiom]

You are right Christian, I looked it up and at processor level there is not much of a difference. Only the integrated graphics are much better on the AMD.
The Asus indeed has a lot of ports, I still choose this over the Atom N550 offering from (for example) JetWay.

Cheers.