FSCK causing ATA status error

[chupa]

9.1.0.1 - Sandstorm (revision 573)
x86-embedded on AMD Sempron(tm) Processor LE-1250
MSI MS-7309 Bios: V9.4 02/17/2009
With an add on 4-ch SATA card. (How can I find the specific chipset?) its a low end one. I know that. 3 assorted data disks on this card
Raid 1 array using SATA controller on motherboard. 2 disks in mirror software raid on these channels.


My NAS4Free box has been restarting at times when I try to access an SMB share from the network. It is one of the data disks (NOT the RAID 1 mount) on the add-on card SATA card controller but I am unable to determine exactly which one yet. It seems to cause the mount point of the disk (or sometimes a couple of the disks) to mess up and I have to reapply them when the server reboots. The issue seems to be intermittent and I am unsure how to duplicate it just yet. It has happened at least 4 times in the past week.

On boot up I am seeing the WARNING: /mnt/MyStuff was not properly dismounted error. I have tried to run FSCK on the suspect drive as described here: http://wiki.nas4free.org/doku.php?id=faq:0001

When I do so I don’t see any output in the WEBGUI however on the server console I see a repeating error:
COM status:
ATA Status Error
ATA status: 51 (DRDY SERV ERR), error: 40 (UNC )

And it repeats that error for about an hour. I see fsck is running from the processes page on the webGUI. But after an hour it just stops.

I have not tried Procedure #2 on FAQ as I don’t have the ability to make a liveCD right now.

Is this possibly my SATA card dyeing? Or what else would cause this?

[al562]

Hi Chupa,

How can I find the specific chipset?

Did you try using the report generator? Many times something that helps to identify it shows up there, if not then try pciconf -l and look up the results with Google, the next easyest way is to look at the chips on the card itself.

I think your particular error more likely indicates a communication problem between the disks and the controller or the controller and the Motherboard; BIOS setting/conflict or controller firmware misconfiguration, incorrect driver or unsupported controller, bad cable, bad port, bad controller, card not firmly seated in slot, bad disk? Test the disks on another computer with the manufacturers' diagnostic software (long test) and make sure they are good. If they pass then it is most likely the controller.

Regards,
Al

[chupa]

OK so I pulled 2 of the suspect drives out and the nas4free box seemed to run fine. I did a long test and one of them passed and the other did not due to bad sectors. I was unable to repair them so that one will be RMAed and remain out of the box. I put the other one back into the NAS system and got it configed back and tried to do a system restart and got the system hanging on shutdown deal as described here:
http://wiki.nas4free.org/lib/exe/fetch. ... uption.pdf


I did a Procedure #2 fsck as described here: http://wiki.nas4free.org/doku.php?id=faq:0001
The first drive comes up fine. The 2nd one gives me a Run Fsck Manualy message.

Where do I go from here?

Also on that page the process for fscking a raid 1 array does not seem to work. There is no /dev/gmirror so i was unable to run fsck on my raid 1 array which might also be a problem.
How do I fsck my raid 1 array?

Thanks!

[al562]

Hi Chupa,

The 2nd one gives me a Run Fsck Manualy message.

Try this instead:

Code: Select all

fsck_ufs -y /dev/YourDrivep1 &

and note down the results. If corrections are made, then try -p again.

There is no /dev/gmirror

Is there a /dev/mirror? Try that instead.

Let us know how it goes,
Al

[chupa]

Code: Select all

Is there a /dev/mirror? Try that instead.

That worked!

Code: Select all

fsck_ufs -y /dev/YourDrivep1 &

Worked! I have all file systems clean now.

Going to try and put it all back together tomorrow minus the bad sector drive.

[al562]

Happy to hear things are going better now .

All the best,
Al