I would love some advice on my situation.  Running Unraid 5.05  and I had a disk get "red balled".  I caputured a syslog here is the link:  https://www.dropbox.com/s/txw62m9nebu3pvy/syslog-2014-07-21.txt

I also ran a smart report, which is attached to this post.  It looks like the smart test passed....am I looking at this wrong?

 

I have had random failures of disk 8 in the past.  For some strange reason it is always disk 8???  When it fails the drive disappears from unraid settings...here is what I mean...unraid notify sends me an email that the drive is disabled...I go to unraid web gui and the drive is red balled.  When I stop the array disk 8 is missing...no serial number, not available for running smart report, etc.  If I reboot the computer hangs on the boot saying there is an error on this drive.  If I power cycle the computer versus a reboot, the drive is visible and I am able to boot into unraid, open unmenu and grab a smart report...so the drive needs to be power cycled to be visible again by unraid.  Is this normal behavior?  I remember that there used to be power cycle issues with some of the WD Green drives a long time ago...this drive is a WD Red 2TB drive that is less than a year old.  Any suggestions would be greatly appreciated.  I am currently running a long smart report...it will be done in 277 minutes  

 

I checked and reseated all cables (power and sata).  I plan on ordering new locking sata cables tonight.  However, I do not think it is a cabling issue since a power  cycle is required to reset the drive.

 

Thanks for any input you can provide!

 

Dan

SmartReportDisk8.txt

Your smart report shows 55 ATA errors. ATA errors indicate a connection problem from controller too disk. Usually a bad or lose SATA cable.

 

(ATA errors basically mean that the commands from the controller to the disk are getting garbled and not understood by the disk.)

The long smart test just completed and it also passed. 

 

Disk 8 is connected to one of the motherboard sata ports (Super Micro C2SEE).  I have 4 open ports on my supermicro AOC-saslp 8mv, so I am thinking of moving the drive connection to this controller. 

 

Any suggestions on which procedure to follow?  I shut the array down as soon as the drive failed.  My wife may have been copying photos onto the array at the time (unsure if it finished as she kicked the process off and walked away)...however these pictures are still on the SD card so we have a backup.

 

Where I am at right not is that I have powered down and swithed the cabling to the AOC board.  I see the drive it is still assigned to slot 8 with a red ball and the array is stopped.

 

Thanks,

 

Dan

You basically have two choices.

 

- Rebuild disk8

- Redefine your array and rebuild parity

 

To rebuilding disk8, you need to absolutely know that the disk is properly connected. I would suggest using a different physical disk to do the rebuild connected to a different physical port with a fresh SATA cable. That way if something goes wrong you have the original disk8 as a backup.

 

If you rebuild parity, you would lose any files copied to disk8 after it red balled (you didn't mention if your wife copied the files to disk8).

 

I'd go for rebuilding disk8 if I had a spare disk, or go with rebuilding parity if I didn't.

Unfortunately, I do not have a spare disk.  I now know that I need to make this investment. 

 

The share she was writing to spans over multiple disks...so some files were writing to disk 8.  The reason I know this is that I unassigned disk 8 and then used unmenu to mount the drive as read only and made a samba share....I then  backed up the precious data (photos)...I have opened multiple photos including the ones she took yesterday and the files open correctly.  I know the process did not complete as there are some that are still on the SD card that are not on the photo share. 

 

Please forgive my ignorance...would it make sense to run a reiser fs check on disk 8 and then assuming it comes back fine then rebuild parity?

 

Any syntax or  link to a procedure would be greatly appreciated!

 

Dan

You don't want/need to run reiserfsck.

 

If you start the array as-is (with disk8 red-balled), the array will be simulating disk8. You should see all of the files as though nothing had failed. If you do that and disk8 looks good, you can rebuild disk8 over top of the physical disk8. I personally don't like doing that, but many many people have done it successfully.

BJP999,

 

Thanks for your help.  I did a backup of all critical data and rebuilt over the top of disk 8.  I agree this is not ideal, but since critical data was backed up, I moved forward with this option.  All went well, the data rebuilt and I have verified it.  I also ran a parity check and no errors were found.  I am going to pick up a spare drive for future issues.  I am going to mark this thread as solved, however if disk8 fails again, I will reopen this thread. 

 

 

Thanks again,

 

Dan

Glad to hear it worked out.

 

Red balls due to bad or loose cables are the bane of the unRaid user's existence. I strongly recommend removable drive cages because in swapping out a bad disk it is so incredibly easy to knock another cable loose and mess up a drive rebuild. Many a user have done this, tried to follow various procedures recommended to other users to recover from similar issues, and at the end of the day find themselves in deep doo doo running reiserfsck in a last ditch effort to recover data. Luckily it usually works (at least partially) but the stress is very high. But you did everything right. Asked the right question at the right time, understood your options, took a reasonable approach, and avoided the drama.

 

Good job, and enjoy your array!

BJP999,

 

Great suggestion on the drive cages.  You are correct it is always stressful.  I have had multiple failures and having someone like yourself who is willing to take their time and provide suggestions is priceless.

 

Thanks Again!

 

Dan

So - I removed the solved as my issue from July repeated last night.  For some reason disk8 red balls and I have to power cycle the drive for it to show back up in unraid.  I unassigned the disk and mounted it via unraid to backup critical data.  I completed this to know that all sensitive data is safe. 

 

Smart report passed with no issues again...so I reassigned in an attempt to rebuild the data on the same disk.  However I ran into a problem...immediately I started getting read errors on the parity drive (I assume this was a cabling error as I took the opportunity to replace my sata cables with new locking cables (this has been on my to do list).  I then stopped the data rebuild.  My problem is I now have an orange ball next to disk 8 and now my parity shows as a blue ball.

 

Any suggestions on how to proceed?  I feel like the parity disk is good, as the errors were read errors.  Is there any way to force this back to green so it can rebuild data on disk 8.

 

Syslog attached...error on disk 8 failing are Oct 29th at 21:23.

 

Thanks,

 

Dan

Syslog_2014-10-30.zip