Best way to avoid data loss with one dead drive and one failing drive

[applejack]

I've been out of town for 6 weeks for work and left my Unraid array running for the family to use, and came home to find one drive dead (disk12) and one drive (disk5) showing tens of thousands of "Reported uncorrect" errors in SMART.  Right now I can still browse those drives and the few media files I tried seem to be okay, so parity is reconstructing on disk12.  Any advice on how to minimize potential data loss?  My first thought was that attempting a parity rebuild after replacing disk12 would be too taxing on disk5 (and might lead to errors anyways), so I was going to copy as many files as I can from those 2 drives off the array first, then replace the drives and redo parity, then copy them back. 

syslog.txt ST4000DM000-1F2168_W300859D-20220211-0904.txt

Edited February 11, 2022 by applejack
Typo

[JonathanM]

51 minutes ago, applejack said:

I was going to copy as many files as I can from those 2 drives off the array first, then replace the drives and redo parity, then copy them back. 

That would be my approach. Don't touch a running system if you can still read the files you need, it may be the last shot at a backup that you get.

 

After you are satisfied you have all the files you want to save, then start messing with reconfiguring things. I mean ALL important files, not just the ones in obvious jeopardy. Once you start playing around, you never know what's going to shake out, and a proper backup turns a VERY bad day into just a pain in the rear.

 

Only you can decide what percentage of files are worth the money and time to keep backed up. For many people, it's a small percent of what is on the server, for others, it's 100%. For those folks, I recommend setting up a second server to back up to.

[Michael_P]

1 hour ago, JonathanM said:

a proper backup turns a VERY bad day into just a pain in the rear.

+1

 

And turn on notifications