I recently bought a number of 4tb drives to replace some of my 2tb drives in my Unraid setup.

 

I precleared 1 of the 4tb drives and added it successfully into the array, parity-sync completed and everything working fine (sdg).

 

Onto the next 4tb drive, replaced the 2tb, installed the 4tb - showing as device emulated as normal - once I began the parity-sync on the newest 4tb drive (sdh), the previous 4tb drive (sdg) started showing read errors and SMART shows "197 Current pending sector" as an increasing count.

 

I let the sync continue until about 68% it dramatically slowed down and continued to register read errors on the previous 4tb drive (sdg).

 

What are my options here? Is the system protected while the newest 4tb drive (sdh) is showing "device emulated" that I can remove the previous 4tb drive (sdg) and replace it with another 4tb drive I have yet to install? Would this cause data loss, etc?

 

I'm currently backing up some crucial files from the array while considering my next option.

 

Diagnostic files attached.

 

Thanks in advance.

 

tower-diagnostics-20181224-1104.zip

7 hours ago, rysith said:

What are my options here?

With only single parity Unraid can't rebuild a disk correctly when there are read errors on another drive, if you continue you'll just get a corrupt disk.

 

Assuming you still have the old disks intact, best way forward would be to installed them again and re-sync parity.

I tried putting in the previous 2tb drive that sdg replaced so I could get it to a state where I can replace the sdh error drive with another 4tb and I got the message "The replacement disk must be as big or bigger than the original".

 

Found this article: 

 

 

Trying this technique now with 

Tools -> New Config -> Retain: All -> Apply

 

Then go back to main and assign/unassign the drives as intended, data on data disks woun't be touched, parity will need to be re-synced though on first array start.

 

So it is currently rebuilding the parity drive - hope it works then I can rip out sdh and replace it...

Yes, using a new config is the correct procedure for this.

Damn, it gets to about 68% then slows down from about 100MB/s to 200Kb/s (when it seems to be reading from the error filled drive).

 

Not sure if it will complete (says 100+ days) though the speed is fluctuating. 

 

At this stage would it be worth cancelling the sync, starting the array unprotected, copying/deleting data from the error drive and trying sync again to see if it proceeds?

 

Or copying/deleting data from the error drive, then doing a new config and taking that drive out of the equation and rebuilding the sync?

6 hours ago, rysith said:

Damn, it gets to about 68% then slows down from about 100MB/s to 200Kb/s (when it seems to be reading from the error filled drive).

You should be doing the parity sync using the old good disks, without both 4TB disks installed last, assuming you still have them untouched.

I'm currently moving days off the problem drive just in case.

 

Was going to do the shrink array method to remove that drive.

 

But it should be ok to just take it out and use the old 2tb in its place even though the parity drive needs to be rebuilt now?

You need to use both old 2TB disks and re-sync parity, the one that was replaced by the now failing 4TB, and the other that wans't successfully rebuilt because of that failing 4TB, only then continue with the upgrades.

Done, parity is syncing now with both of the old 2tb drives.

 

Fingers crossed

Parity sync completed perfectly

Adding the good 4tb drive back in now and rebuilding.

 

Before RMA'ing the error 4tb drive, I think a full format in Windows and SMART scan might be in order

 

Thanks for your help, much appreciated

48 minutes ago, rysith said:

Before RMA'ing the error 4tb drive, I think a full format in Windows and SMART scan might be in order

Always a good idea to hold onto anything that has your data on it until you are confident in the results of the rebuild.