Problem with hard disk - not arch related

[Dario-27]

Ciao!

This is unrelated to arch, but I wanted to tell this to the list because maybe
someone has a clue.

I have recently bought and installed a new sata hard disk. It has a couple of
partitions, one with XFS which I use as storage.
It generally works fine, but sometimes something happens and it cease to work.
I don't know if it's the disc's fault, because at the next restart the
problems are gone. It seems that the controller sometimes gets lost. Dmesg
says this:

ata4.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6                                                                                                    
ata4.00: BMDMA stat 0x25                                                                                                                                    
ata4.00: cmd 35/00:20:e7:a0:f1/00:01:04:00:00/e0 tag 0 dma 147456 out                                                                                        
         res 51/84:11:e7:a0:f1/84:00:00:00:00/e0 Emask 0x10 (ATA bus error)                                                                                  
ata4.00: status: { DRDY ERR }                                                                                                                                
ata4.00: error: { ICRC ABRT }                                                                                                                                
ata4: soft resetting link                                                                                                                                    
ata4.00: configured for UDMA/133                                                                                                                            
ata4: EH complete
...
(repeat some time)
....

ata4.00: qc timeout (cmd 0x27)                                                                                                                              
ata4.00: failed to read native max address (err_mask=0x4)                                                                                                    
ata4.00: HPA support seems broken, skipping HPA handling                                                                                                    
ata4.00: revalidation failed (errno=-5)                                                                                                                      
ata4: link is slow to respond, please be patient (ready=0)                                                                                                  
ata4: soft resetting link
ata4: SRST failed (errno=-16)                                                                                                                                
ata4: reset failed, giving up                                                                                                                                
ata4: EH complete
...
sd 3:0:0:0: [sdc] Unhandled error code                                                                                                                      
sd 3:0:0:0: [sdc] Result: hostbyte=0x04 driverbyte=0x00                                                                                                      
end_request: I/O error, dev sdc, sector 82953831
...
I/O error in filesystem ("sdc2") meta-data dev sdc2 block 0x104d6d83      
("xlog_iodone") error 5 buf count 6656                                            
xfs_force_shutdown(sdc2,0x2) called from line 1043 of file fs/xfs/xfs_log.c.  
Return address = 0xf9c8d923                                                    
Filesystem "sdc2": Log I/O Error Detected.  Shutting down filesystem: sdc2                                                                                  
Please umount the filesystem, and rectify the problem(s)                                                                                                    
...
Buffer I/O error on device sdc2, logical block 603726
lost page write due to I/O error on sdc2

So, what should I look for? Together with the disk, I installed an Nvidia 9800
GT, which could also be the culprit, because maybe I haven't a proper power
unit etc. - or not?

Thanks in advance to everybody!

Ciao a tutti

Dario
Chiacchiera con i tuoi amici in tempo reale!
 http://it.yahoo.com/mail_it/foot/*http://it.messenger.yahoo.com 

[Rogutes]

Dario (2009-11-06 21:26):

> Ciao!
>
> This is unrelated to arch, but I wanted to tell this to the list because maybe
> someone has a clue.
>
> I have recently bought and installed a new sata hard disk. It has a couple of
> partitions, one with XFS which I use as storage.
> It generally works fine, but sometimes something happens and it cease to work.
> I don't know if it's the disc's fault, because at the next restart the
> problems are gone. It seems that the controller sometimes gets lost. Dmesg
> says this:
>
> ata4.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6
> ata4.00: BMDMA stat 0x25
> ata4.00: cmd 35/00:20:e7:a0:f1/00:01:04:00:00/e0 tag 0 dma 147456 out
>          res 51/84:11:e7:a0:f1/84:00:00:00:00/e0 Emask 0x10 (ATA bus error)
> ata4.00: status: { DRDY ERR }
> ata4.00: error: { ICRC ABRT }
> ata4: soft resetting link

There's no clever advice I could give you, but I vaguely remember having
similar problems and similar logs some time ago (perhaps when kernel
2.6.29 became stable). The block device of my drive used to become
inaccessible, but the mounted filesystem continued to work, IIRC. But
I was rapidly switching SATA and IDE devices, playing with BIOS settings
at that time and eventually (perhaps after a kernel upgrade) the
symptoms disappeared...

Are you running Linux kernel 2.6.31?

--
--  Rogutės Sparnuotos

[Dario-27]

ciao!

In data sabato 07 novembre 2009 00:23:13, Rogutės Sparnuotos ha scritto:
> There's no clever advice I could give you, but I vaguely remember having
> similar problems and similar logs some time ago (perhaps when kernel
> 2.6.29 became stable). The block device of my drive used to become
> but the mounted filesystem continued to work, IIRC.

Indeed the filesystems is left intact, which at least is a good thing:)
 
> Are you running Linux kernel 2.6.31?

Yes, I am. Previously I pointed out the video card because the first time I
met this problem, I was doing dirty things with CUDA under Ubuntu 9.04, and
rebooting brought everything back in order. Could this be a sign of a
relationship between the closed source Nvidia drivers and the wild controller?

thanks for answering:) I'll give a look at the bios, and I'll try to find a
way to reproduce the bug, which up to now seems to appear randomly.

ciao:)

Dario
Chiacchiera con i tuoi amici in tempo reale!
 http://it.yahoo.com/mail_it/foot/*http://it.messenger.yahoo.com

[Rogutes]

Dario (2009-11-07 01:52):
> ciao!
>
> In data sabato 07 novembre 2009 00:23:13, Rogutės Sparnuotos ha scritto:
> > There's no clever advice I could give you, but I vaguely remember having
> > similar problems and similar logs some time ago (perhaps when kernel
> > 2.6.29 became stable). The block device of my drive used to become
> > but the mounted filesystem continued to work, IIRC.
>
> Indeed the filesystems is left intact, which at least is a good thing:)

Is KMail unable to quote mails or did you drop "inaccessible, " from the quote
above (after "used to become")?

> > Are you running Linux kernel 2.6.31?
>
> Yes, I am. Previously I pointed out the video card because the first time I
> met this problem, I was doing dirty things with CUDA under Ubuntu 9.04, and
> rebooting brought everything back in order. Could this be a sign of a
> relationship between the closed source Nvidia drivers and the wild controller?

I had a 13 year old Matrox PCI graphics card inside when the problem was
happening. And at that time I _think_ I had 2 SATA and 1 IDE drives
connected to the JMicron controller. Now the 2 SATA drives are connected
to Intel ICH9 Southbridge, and the IDE drive to JMicron (the motherboard
is Gigabyte GA-EP35-DS3).

--
--  Rogutės Sparnuotos

[Dario-27]

ciao:)

In data sabato 07 novembre 2009 12:24:29, Rogutės Sparnuotos ha scritto:
> Is KMail unable to quote mails or did you drop "inaccessible, " from the
>  quote above (after "used to become")?

KMail is able to quote, it was my fault:) What I meant is: the filesystem
becomes inaccessible after a series of errors. But the filesystem per se is
all right, simply cannot be reached in that particular moment due to something
at lower level.

I'm in need of further proofs of what's going on...

grazie:)

Dario
Chiacchiera con i tuoi amici in tempo reale!
 http://it.yahoo.com/mail_it/foot/*http://it.messenger.yahoo.com