Two drives went disabled and loads of errors all of a sudden. What do do from here?

[detz]

I should have a spare, I'll run out today and grab one and do what you suggested as that seems the safer of the approaches. Thanks, I will report back!

[trurl]

15 hours ago, trurl said:

Unless you mounted it read-only in UD it will be slightly out-of-sync though.

5 hours ago, detz said:

I didn't mount it read-only but it doesn't appear to be re-built.

My point was that simply mounting it read/write alters the contents slightly so it is now slightly out-of-sync with parity. If you rebuild it, the rebuild would be in sync with parity of course, but if you New Config it back into the array, then you would have to correct parity.

 

Any way, just trying to educate.

 

Looks like itimpi has a good plan. Rebuilding to a spare is always safer than rebuilding to the same disk, but in simpler situations than you have, that extra safety maybe isn't worth the cost of the disk.

[detz]

new drive is rebuilding and I mounted the old disk and it's ready for copying once the rebuild is done. I did have to change the UUID in the settings for the mount to work. I'll report back tomorrow when the rebuild is done. Is there a way to rsync easily to just copy everything over or is this going to be a slow single dir move process?

[trurl]

I always use the builtin Midnight Commander (mc at the command line) to work with files directly on the server.

[detz]

The rebuild went fine and just like the emulated drive it was missing a lot of files. I used an external usb HD base to connect up the old drive and I've been copying files over for days now (so slow...). All the files appear to be fine on the old drive which makes this even more confusing but this approach does seem to be working.

 

Is there anything we can do to protect against this? Is this common, seems odd it's this easy to lose data in a protected array