How to remove a failed disk and recover data, without replacing the disk.

[R.M.H]

I do not have the issue but trying to get my head around the scenario as no doubt it will happen as I have 3 dead drives sat on the side already.

 

Setting up new unRAID (I am still new to it). This is a replacement of my old setup and the disks are moving from old to new so not all new disks. Parity is a new WD Red 20TB remaining disks are older WD Red 2TB and 6TB. My plan is to stick them all in the array and as one dies then get rid to the point I need new disks.

 

Looking at recovery options my plan seems flawed as all the ones I have seen require a replacement disk of same or greater size to be able to rebuild data from parity. Or you move the data off the disk then remove the disk from the system prior to disk failure.

 

Is there a way, in the event of a failed disk to rebuild the lost data to another drive already in the array. Or would I essentially need to keep a spare larger drive. Use that for rebuild, move the data off it after rebuild then remove it from the array to be a spare again. 

 

On that, are there any plans to add the ability for a hot spare in unRAID so that in the event of a disk failure the automatic replace \ rebuild can start.

[JorgeB]

Though rebuilding is what recommended you can move the data after a disk gets disable, by moving it from the emulated disk to other array disk(s), it will be slower than normal though.

[JonathanM]

5 hours ago, R.M.H said:

My plan is to stick them all in the array and as one dies then get rid to the point I need new disks.

Don't use disks in Unraid that are ready to fail. The way Unraid parity works it uses the entire capacity of ALL remaining drives to reconstruct the failed drive, so if a second drive dies you will lose all the data on both drives. The moment a drive fails, ALL rest of the data drives plus the parity drive are read to emulate the failed drive, and writes to the missing drive's slot update the parity drive.

 

As an example, say you have replaced all your older disks except one, and it doesn't have much content, so you aren't worried about replacing it. Now, out of the blue without warning, one of your new disks, full of data, dies on you. Now, that old drive that was on its last legs is called into constant use to emulate the failed drive, and it must survive reading the entire capacity to rebuild the failed new drive. You hope it survives long enough to get through the rebuild process, but chances are, it's not going to.

 

You must always be able to trust all your drives to perform perfectly, so when one of them inevitably fails the rest are in good shape. Any drive that shows signs of failure must be replaced ASAP.

[R.M.H]

1 hour ago, JonathanM said:

Don't use disks in Unraid that are ready to fail. The way Unraid parity works it uses the entire capacity of ALL remaining drives to reconstruct the failed drive, so if a second drive dies you will lose all the data on both drives. The moment a drive fails, ALL rest of the data drives plus the parity drive are read to emulate the failed drive, and writes to the missing drive's slot update the parity drive.

 

As an example, say you have replaced all your older disks except one, and it doesn't have much content, so you aren't worried about replacing it. Now, out of the blue without warning, one of your new disks, full of data, dies on you. Now, that old drive that was on its last legs is called into constant use to emulate the failed drive, and it must survive reading the entire capacity to rebuild the failed new drive. You hope it survives long enough to get through the rebuild process, but chances are, it's not going to.

 

You must always be able to trust all your drives to perform perfectly, so when one of them inevitably fails the rest are in good shape. Any drive that shows signs of failure must be replaced ASAP.

A very valid comment and makes me need to think twice about my approach. I have 2TB drives that I purchased several years before the 6TB ones. All the 2TB drives are still working fine and the newer (out of warranty) 6TB ones 3 have failed, leaving me with 4 left. From a reliability standpoint I probably have more faith in the 2TB drives then the 6TB ones.

 

I keep an eye on the SMART status of the drives and if any show signs of failure they would be removed after the data moved to other drives.

[JonathanM]

3 minutes ago, R.M.H said:

I keep an eye on the SMART status of the drives and if any show signs of failure they would be removed after the data moved to other drives.

Be sure you set up and verify that you get notifications so you have immediate feedback of any issues.

 

Since Unraid works with the entire surface of all the drives, it's a good idea to do an extended smart test on all your drives before you trust them.

 

The parity build and check will also help you gain confidence in the drives.

[R.M.H]

6 hours ago, JonathanM said:

Be sure you set up and verify that you get notifications so you have immediate feedback of any issues.

 

Since Unraid works with the entire surface of all the drives, it's a good idea to do an extended smart test on all your drives before you trust them.

 

The parity build and check will also help you gain confidence in the drives.

I did not have notifications enabled but I do now, thank you for mentioning it