Out of the blue I noticed that I was unable to access files my tower through SMB and when I went to check on the tower through the web console, it started to act odd. In red text it said there was a "Parity Check in progress", so I went to the Syslog screen and of the 738 lines of text, the last few messages read the following;

 

Aug 25 02:51:48 Tower kernel: md: parity incorrect: 9400 (Errors)
Aug 25 02:51:48 Tower kernel: md: parity incorrect: 9408 (Errors)
Aug 25 02:51:48 Tower kernel: md: parity incorrect: 9416 (Errors)
Aug 25 02:51:48 Tower kernel: md: parity incorrect: 9424 (Errors)
Aug 25 02:51:48 Tower kernel: md: parity incorrect: 9432 (Errors)
Aug 25 02:51:48 Tower kernel: md: parity incorrect: 9440 (Errors)
Aug 25 02:51:48 Tower kernel: md: parity incorrect: 9448 (Errors)
Aug 25 02:51:48 Tower kernel: md: parity incorrect: 9456 (Errors)
Aug 25 02:51:48 Tower kernel: md: parity incorrect: 9464 (Errors)
Aug 25 02:51:48 Tower kernel: md: parity incorrect: 9472 (Errors)
Aug 25 02:51:56 Tower unmenu-status: Starting unmenu web-server
Aug 25 02:51:56 Tower unmenu[1602]: awk: ./unmenu.awk:306: fatal: can't open two way socket `/inet/tcp/8080/0/0' for input/output (No such file or directory) (Drive related)
Aug 25 02:51:56 Tower unmenu-status: Exiting unmenu web-server, exit status code = 2
Aug 25 02:51:56 Tower unmenu[1602]: unmenu.awk unable to open port.  It  may already be running
Aug 25 02:52:06 Tower unmenu-status: Starting unmenu web-server
Aug 25 02:52:06 Tower unmenu[1602]: awk: ./unmenu.awk:306: fatal: can't open two way socket `/inet/tcp/8080/0/0' for input/output (No such file or directory) (Drive related)
Aug 25 02:52:06 Tower unmenu-status: Exiting unmenu web-server, exit status code = 2
Aug 25 02:52:06 Tower unmenu-error: Fatal error:Exiting uu, unmenu may already be running, exit status=2 (Errors)

 

When I tried click refresh to see the status of the parity check, it was unable to connect to the tower. I tried several times with no luck. Because I am a rather newbie when it comes to unRAID I powered down the tower manually. After starting it back up, the BIOS screen indicated that the S.M.A.R.T. status is bad. I logged in to the tower and obtained a syslog and SMART log which are attached.

 

After I obtained those two logs for this thread, I booted it up one last time and the main BIOS screen alerted me of the bad S.M.A.R.T. status but it booted back up and suddenly everything is working again. All 3 balls were green and I was able to play a video file that is stored on the tower.

 

Seems like some quirky behavior though. Please advise what I should do next. I'm thinking a Parity Check?

smart.txt

syslog.txt

A NON-CORRECTING parity check  will not hurt anything that is not already broke.  I would NOT run a correcting parity check until you have found and corrected a a fault that required a correcting check. 

 

It almost sounds like you are getting a spontaneous reboot of your system.  (I understand that unRAID will start a correcting parity check on restart any time that the system does not have a normal shutdown!)  You could try rebooting your system and run a memory check rather then booting directly into unRAID.  If you do this, you should let the memory test run for at least twelve hours. 

 

You have not really told us much about your system  (the hardware) or which version of unRAID that you are running.  This information would be useful to those trying to help you. 

Here are my specs:

unRAID 4.7 (Basic)

 

Parity: 2TB Seagate

Two Data drives: both 2TB Hitachi

 

ASUS P8B75-M LE B75

Intel Celeron G1610 LGA 1155 (2.6 GHz)

GSkill GSK F3 2GBx2

 

The smart report is incomplete. Repost the entire text file.

You are right archedraft. Here is a smart test on the bad drive. See the attached for the full report but here is a snippet...

 

SMART overall-health self-assessment test result: FAILED!
Drive failure expected in less than 24 hours. SAVE ALL DATA.
See vendor-specific Attribute list for failed Attributes.

 

Is this something to be concerned about? Should I be ordering up a new drive replacement?

smart-disk1.txt

You are right archedraft. Here is a smart test on the bad drive. See the attached for the full report but here is a snippet...

 

SMART overall-health self-assessment test result: FAILED!
Drive failure expected in less than 24 hours. SAVE ALL DATA.
See vendor-specific Attribute list for failed Attributes.

 

Is this something to be concerned about? Should I be ordering up a new drive replacement?

Yes, you should be concerned, yes, you need to replace the disk ASAP.

 

Joe L.

Agreed. Replace that drive ASAP. If it were me I would buy a replacement drive from a local store and pay a little extra in order to get it then and there...

Yes, you should be concerned, yes, you need to replace the disk ASAP.

 

Joe L.

 

I just ordered the replacement drive (2-day shipping). I'll have the tower powered down until then. It has been about 2 years since I have had to touch my unRAID system, can anyone give me a quick rundown on the procedure on replacing the drive (including how to preclear again)?

Pre-clearing won't speed up the rebuild, but it IS a good way to confirm the new drive is good.

 

To pre-clear, you first need to copy Joe L's pre-clear script [preclear_disk.sh] to your flash drive.

 

Then you boot the system and log into UnRAID (password = root unless you've changed it).    Then change to the boot directory [ cd \boot ].

 

And finally run preclear ...  "preclear_disk.sh -l"  to list the drive available for preclear (it should only list the new drive you installed);  and  "preclear_disk.sh \dev\sdx"  to do the actual preclear (where x = the drive you want to preclear)

 

That will run for ~ 20 hours or so for a 2TB drive (possibly a good bit longer, depending on your controller).

 

Then you simply follow the instructions in the Wiki for replacing your bad disk:

http://lime-technology.com/wiki/index.php?title=UnRAID_Manual#Replace_a_failed_disk

... the link to download Joe's preclear script is at the end of the first (long) post here:

http://lime-technology.com/forum/index.php?topic=2817.msg23246#msg23246

 

That post also includes a good bit more detail on the specifics of using it.

 

Pre-clearing won't speed up the rebuild, but it IS a good way to confirm the new drive is good.

 

To pre-clear, you first need to copy Joe L's pre-clear script [preclear_disk.sh] to your flash drive.

 

Then you boot the system and log into UnRAID (password = root unless you've changed it).    Then change to the boot directory [ cd \boot ].

 

And finally run preclear ...  "preclear_disk.sh -l"  to list the drive available for preclear (it should only list the new drive you installed);  and  "preclear_disk.sh \dev\sdx"  to do the actual preclear (where x = the drive you want to preclear)

 

That will run for ~ 20 hours or so for a 2TB drive (possibly a good bit longer, depending on your controller).

 

Then you simply follow the instructions in the Wiki for replacing your bad disk:

http://lime-technology.com/wiki/index.php?title=UnRAID_Manual#Replace_a_failed_disk

 

Thanks! Should I disconnect my other two working drives while I am preclearing the new drive, just to be safe? Is that even possible to get that far in unRAID with only one drive that has not been formatted at all, when it is used to having 3 drives in the setup?

No need to disconnect the other drives -- just be SURE you are pre-clearing the correct drive (pay attention to the serial numbers).

 

As long as you Start the array, the preclear_disk.sh -l (that's an L ... not an I) will only list the disks not in the array (the one new drive) => so you'll know the proper drive designation to use for the preclear.

 

Then you simply start the preclear for that drive.

 

Note that if you don't have a console (monitor & keyboard) on the UnRAID machine, you should do this via Screen.    You could use Telnet, but if you do, the Telnet session has to remain open until the preclear is finished.

 

I put in my new drive, started up the tower, unRAID recognized it and has a blue ball next to it. The Commend Area has a red ball next to it saying "Stopped. Upgrading disk." I then ran the preclear_disk.sh -l but it said that there were "No un-assigned disks detected".

 

Do I have to Tick the "I'm sure" checkbox, and press "Start will bring the array on-line, start Data-Rebuild, and then expand the file system" BEFORE I run this preclear script?

 

Sorry I'm such a newbie and need such precise instructions, I just do not want to mess anything up.

Since you've already added the disk to the array, you can't pre-clear it.

 

The idea was to pre-clear the disk BEFORE you added it to the array to thoroughly test it.

 

I'm not sure what would happen now if you removed it from the array -- but just to be safe I wouldn't chance it.    Just check the box and Start the array ... and let it do the rebuild.    When that's done (tomorrow), run a parity check to ensure all was okay.    I'd also then look at the SMART data for the disk to confirm the parameters all look okay.    Doing all this will be a fairly good test anyway ... the entire disk will be written to;  it will then be read (for the parity check);  and any bad sectors will show up in the reallocated sectors parameter on SMART.

 

I have rebuilt the array with the new drive. I then did a parity check and received 0 sync errors. Everything seems to be in order. Thanks!