So this happened.

 

Recently upgraded from v5.??? to 6.3. During the upgrade i slip the drives out so that if i have to do an ungraceful shutdown i dont have to go through the parity check.

The upgrade went well and all seemed fine for about a week.

 

Just did the first parity check (Write corrections to parity) and once it had been running for a while, i logged in to see 20k+ parity errors. I stopped the check to investigate.

One of the drives was inaccessible, showed IO errors in "Capabilities" but was still showing as green.

 

This is a remote server at work and i was accessing it via a pc using teamviewer, so i decided to reboot system. The drive appeared and i could access it no worries.

 

As the parity corrections were written to disk and i could access the data on the drive, i re-ran a partity check to correct what i assumed to be errors.

Check completed with another 20k errors which i hoped corrected the problem.

 

To confirm that the errors were fixed i ran a 3rd check and was expecting it to come back with zero errors as per normal.

 

Now though, the drive is red balled (or x now) and using parity. I can only hope that the drive was not re-inserted correctly, but i have reinserted it.

 

The problem i now have is that i have a red balled drive and a parity that i cannot trust.

 

 

This server was used for long term storage of videos and photos and the drive in question has data from 2008-2010. Due to cost, i only have a backup for the last 18 months, so cannot verify the data.

 

My gut feeling after typing this all out is that the drive is ok and that i just need to rebuild the parity, but would love to hear others thoughts.

 

 

BTW typing it out like this does allow you to think about it logically.

First thing I would do is to look at the Smart reports.  You can easily do this by going to 'Main'   >>>   'Array Devices'.   Then click on the drive in question and look at the "Attributes".  You can use the arrow buttons at the upper right of the screen to move between Drives.  

 

You should probably also post up your diagnostics file.    'Tools'   >>>   'Diagnostics'        with your next post.  As you probably realize, don't do anything that might result in a write to that drive.  Remember that the disk can be easily read in another Linux computer (and most likely, even in a Windows computer with the proper driver loaded).  If that is the case, file recovery would not be an issue...

I did check the smart reports and have done a few short test and they complete without errors. This is one of the reasons i thought the drive was ok and it was just a connection issue.

 

I may just pull the hard drive and do a comparison between the files on the drive and the computed files from the parity. This will then show the files with differences and from there i can probably figure out which data is correct.

torres-diagnostics-20170706-1153.zip

Diags are right after a reboot, so not much help, SMART reports are mostly fine, to have an idea if parity is valid or not we'd need the logs from when the errors happened.