Jump to content

Unable to replace failing data drive


southpaw5271

Recommended Posts

Hello, 

 

I'm pulling my hair out trying to recover my data. I had a drive fail (disk 2), and I was trying to follow this wiki to replace the drive, but I can't complete it the way it is described. https://lime-technology.com/wiki/Replacing_a_Data_Drive

 

The first screenshot shows what my array currently looks like. Keep in mind, that the Disk 2 is actually a failing disk, it refuses to mount due to bad sectors.

 

When I reboot the server and connect the new 1 TB drive to disk 2, I get a warning about my configuration being invalid due to a missing disk 2. See second screenshot for missing disk 2. See third screenshot for disk 2 replacement. I cannot start the array with the new disk.

 

 

What am I doing wrong?  I am unable to attach both of the disks to the system at the same time due to sata port limitations. Do I need to buy a raid card to attach both disks to the system at the same time? I attached my diagnostics for the old drive (grabbed during first screenshot).

1.PNG

2.PNG

3.PNG

sol-diagnostics-20180826-1817.zip

Link to comment

Tell me about Disk 3. You don't mention it in your OP.

 

In the first screenshot you have only one missing disk (Disk 3) and therefore unRAID is offering you the possibility of starting with Disk 3 emulated.

 

In the second and third screenshots you have two missing disks (Disks 2 and 3) and since you have only one parity disk that isn't enough to start the array.

Link to comment
45 minutes ago, John_M said:

Tell me about Disk 3. You don't mention it in your OP.

 

In the first screenshot you have only one missing disk (Disk 3) and therefore unRAID is offering you the possibility of starting with Disk 3 emulated.

 

In the second and third screenshots you have two missing disks (Disks 2 and 3) and since you have only one parity disk that isn't enough to start the array.

 

Disk 3 does not exist. That disk slot was just skipped over in the assignment period a few years ago. Should have elaborated earlier. Disk 2 is the one that is definitely failing.

Link to comment
5 minutes ago, southpaw5271 said:

Disk 3 does not exist.

 

The thing is, it does exist! Not as a physical drive, but it exists as an emulated disk in the array. Your logs show a /dev/md3 as proof. This is indicated by the red cross. If you had simply skipped over the slot it would appear as a grey square, like Disk 7, for example.

 

5 minutes ago, southpaw5271 said:

Disk 2 is the one that is definitely failing.

 

Yes, it is.

Link to comment

It used to exist but the drive failed years ago. Is that why I am having this issue? Should I delete disk 3 permanently? This is why it's not allowing me to start things properly right? It wants two parity drives as well as both failed drives? Or maybe I should get another disk to replace disk 3 and disk 2 at the same time?

Link to comment
20 minutes ago, southpaw5271 said:

It used to exist but the drive failed years ago.

 

And it's still being emulated. You should have replaced the disk then. Now you have a second disk that's causing problems.

 

Do you care about the data on Disk 3? If you revert to the situation you had in the top screen grab, can you start the array and does Disk 3 mount? How about Disk 2?

 

Do you have notifications enabled?

 

 

Link to comment

I'm pretty sure disk 3 is useless. It was an SSD for VMs that was moved to disk 6. Is there a way I can look at the contents of what was there to be sure it's useless?

 

If I go back to the state with the top screen, after clicking start, the array will be in a constant state of "starting array" but will say that disk 2 has an unmountable file system. 

 

There is currently no way I can actually start the array successfully. I am okay with double checking disk 3 and ensuring that data was useless and then removing it permanently.

Link to comment

Do you have a file backup of disk 2? Since disk3 is missing and you only have one parity drive, there is nothing unraid can do to reconstruct the data that was on disk2. You will need to recover disk2 from backups.

 

It's possible if you send the physical disk2 to a drive recovery shop they might be able to get your data, but it's going to be expensive.

 

If you are ok with giving up on disk2 and disk3 and just want the rest of the drives to show up, you will need to do a "new config" and include only the healthy remaining drives.

Link to comment
31 minutes ago, John_M said:

Do you have notifications enabled?

THIS^

 

Notifications would have been nagging you all this time to get your array healthy again, instead of ignoring the fact that you had no protection after you didn't do anything about the failed disk3.

 

Parity doesn't have any of your data. The parity calculation requires ALL of the other disks to allow the data for a failed or missing disk to be recovered.

 

Obviously you should have asked for help a long time ago.

 

Please setup Notifications.

Link to comment

Well it looks like I really screwed myself. I found some articles and some information about recovering the data from a btrfs drive. I will have to pull that drive and recover the data separately.

 

Thanks to everyone who replied. I just set up notifications for the future to avoid this.

Link to comment

UPDATE: The failing disk uses btrfs file system. There is a native recovery tool built into btrfs-prog for Linux called "restore" that was able to recover all of the files on the disk. Fortunately for me, this disk was non-critical data.

 

My final question is a simple one, can I just install the new drive and click the "new config" button in the tools menu to rebuild the array? I know that the contents of disk 2 will not be in the new build but everything else should be fine correct?

Link to comment

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...