Problem: UFS Random Reboots / Reboot loops --> missing volume.

[uncarvedblock]

Some background:
I converted my old computer to a FreeNAS box.
Motherboard: Gigabyte GA-MA790X-UD4
Processor: Phenom II 920 quad core 2.8 GHz
Memory: 4 GB DDR2
HDDs: 2x 3 TB WD Red
Powersupply: 500W Antec Earthwatts
Videocard: Radeon HD 4870
(not really a key part of the build, but it is still installed so I can hook up a monitor and in case the system would not work headless)

I am running FreeNAS-9.2.0-RELEASE-x64 (ab098f4).

I decided to go with UFS because I did not have the ram for ZFS and I decided I did not want a raid setup to start. I had a friend warn me that he had issues keeping his raid setups running and being a noob, I just created 2 UFS volumes and set up an rsync job to copy the contents of the main disk to the second disk daily.
I named the drives Master0 and Second0.

I started getting random reboots and the system would run through a lot of checks / fixes when it rebooted. The most common message I would get was something like this:

(ada0:ahcich1:0:0:0): WRITE_FPDMA_QUEUED. ACB: 61 40 80 9e 40 40 00 00 00 00 00 00
(ada0:ahcich1:0:0:0): CAM status: Uncorrectable parity/CRC error
(ada0:ahcich1:0:0:0): Retrying command

Then it would usually reboot again. I would power down and a couple times I tried booting again and then it would work after a few file repairs. Unfortunately I did not record the monitor and as far as I know FreeNAS does not by default save the logs after reboots because of being installed on an usb or something. Correct me if I am wrong...

It might have been foolish but I did the following:
I removed and reconnected the sata cables in case there was a connection issue.

I did not completely understand what was discussed in another post with a similar error, where they discussed possible issues with sata cables and I figured that was pretty low risk.
http://forums.freenas.org/index.php?threads/uncorrectable-parity-crc-error.14404/page-2

Then I rebooted and it worked fine but the disk Master0 showed and error.

I though it might have been due to switching sata ports and i tried to import volume. UFS, named Master0, ada1 but I got an error, "The selected disks were not verified for this import rules."

I thought maybe if I removed the old Master0 volume entry and then tried importing it again it might work so i tried, but now I still get the same error.
Hopefully that did not screw things up more than they already were.
If I screwed up the contents of Master0 I can always copy them over from the Second0 drive, but that still will not address the Uncorrectable parity/CRC error I was getting at the start of this.

Sorry for the long post. Thanks for any help.

 

[Yatti420]

Bad sata cable(s) [Try some new ones]... I'd start with those.. PSU? Capacitors? Bad ram? Everything tested? Is it limited to ada0 only?

 

[uncarvedblock]

As far as I know it was just the 1 drive, but like I said I do not have any logs to check. I did record one boot sequence to my cell phone camera before trying anything, but a lot of it flashes too fast to read.

Is there any way to get FreeNAS to recognize the disk that was Master0 (ada1) again?

From the webgui, I tried using the shell with the fsck command:

Code:

# fsck -t ufs /dev/ada1
** /dev/ada1
Cannot find file system superblock
ioctl (GCINFO): Inappropriate ioctcl for device
fsck_ufs: /dev/ada1: can't read disk label

I switched the cable on that drive, but since I removed the volume from the webgui, and it won't let me reimport it I dont know what else to do.
Will I have to recreate that volume or can I recreate the GPT data or something? How do I know if the cable helped?

 

[uncarvedblock]

Looking at gdisk for ada1 it seemed the gpt partition tables were messed up. Since both my disks were identical originally (both WD Red 3TB created as UFS volumes in FreeNAS), I decided to try copying the partition table from the working drive to the other drive.

I booted into a linuxmint live disk and used gdisk to backup the good partition table and load it into the other disk (ada1 in FreeNAS). Then the GUID was the same on both disks so I had gdisk create a new random GUID for the disk I was trying to fix (ada1).
Then I booted into FreeNAS and the disk named 'Second0' (ada0) still mount and I imported volume on the other disk on ada1, choosing the name Master0 again.
Now FreeNAS give the alert " WARNING: The volume Master0 (UFS) status is UNKNOWN"

I managed to mount the disk with the following commands:
mount -uw /
mkdir /mnt/Master0
mount -t ufs /dev/ada1p2 /mnt/Master0

It appears the contents are still intact, so I just need to figure out how to get FreeNAS to recognize the data again. Anyone have any suggestions? Again thank you for any help.

 

[cyberjock]

Hey.. look.. another AMD user with these dreaded errors. :(

Post the output of smarctl -a -q noserial /dev/ada0 and /dev/ada1 to pastebin. ;)

 

[cyberjock]

And, if you searched for the keywords for your errors, you'd realize that this question has been asked 10 times and you can probably solve the problem on your own. ;)

 

[uncarvedblock]

Sorry, for my poor search skills . Thank you for your help. :)

ada0
http://pastebin.com/zbQzGStZ
ada1
http://pastebin.com/pMmuFq7w

 

[cyberjock]

Hard drives look okay. But you need to read and do the wdidle tool on your drives. Check out the how-to guide in the forums for how to use the tool.