So I had a drive failure in my MBLD 6TB (2x3TB) device. It was drive A. I found that out by removing one drive and letting the device reboot, and then swapping out the drives and rebooting again. Whenever drive A was in the system, it became extremely unreliable - it reboots over and over. Sometimes it would go a day or two without a reboot, sometimes it would reboot a dozen times in a day.
So I decided to replace BOTH drives, since I’ve had this thing for several years now. I bought two 6TB WD Green drives, identical to the 3TB drives that were in it except they were 6TB. Even the part numbers were the same, except for the 6TB designation. I followed Guide 2 with the new drives. I used the following downloads:
The debrick script from above
The SystemRescueCd from SystemRescue - Browse Files at SourceForge.net
The latest MyBookLive DUO firmware from https://download.wdc.com/nas/ap2nc-024310-048-20150507.zip
There’s a section in the debrick script that appears to write to the partition tables on the drive prior to creating the new partitions so it generates a bunch of errors. If you reboot your Linux box and run the script again, it runs without the errors, and appears to succeed. I did this with both of the new drives. Someone without any IT experience might think something is wrong there. They just need to run the script twice. And you can’t just run it twice, you have to reboot the Linux box between runs. If you just run it a second time without rebooting, it complains about /dev/md0 still being present and mdadm –S /dev/md0 gives you an error. Just reboot the stupid thing and run it again.
Now here’s where it gets interesting. Everywhere I’ve read I hear that you cannot upgrade the drives. I refuse to believe that. Since this thing is capable of creating a 6TB volume, I thought it would recognize and use the 6TB drives. Well, it kind of did. While waiting for the debrick script to build my second drive, I put the first drive into the device and booted it. It booted just fine and I was able to go through the initial agreement acceptance and get to the Dashboard. The LED was red, but it still let me log on. When I went to the Storage page, it said that the drive was a good 6TB drive, but that it FAILED. There was also a message on the dashboard saying that it couldn’t mount the data volume.
Then I logged into the shell. Parted had some really interesting things to say:
MyBookLiveDuo:~# parted
GNU Parted 2.2
Using /dev/sda
Welcome to GNU Parted! Type ‘help’ to view a list of commands.
(parted) p
Error: The backup GPT table is not at the end of the disk, as it should be.
This might mean that another operating system believes the disk is smaller.
Fix, by moving the backup to the end (and removing the old backup)?
Fix/Ignore/Cancel? i
Warning: Not all of the space available to /dev/sda appears to be used, you can
fix the GPT to use all of the space (an extra 1 blocks) or continue with the
current setting?
Fix/Ignore? i
Model: ATA WDC WD60EZRX-00M (scsi)
Disk /dev/sda: 6001GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt
Number Start End Size File system Name Flags
3 15.7MB 528MB 513MB primary raid
1 528MB 2576MB 2048MB ext3 primary raid
2 2576MB 4624MB 2048MB ext3 primary raid
4 4624MB 6001GB 5997GB ext4 primary raid
Sure doesn’t look like it can’t read the drive to me. I reran parted and told it to Fix all the errors that it found, and it fixed them. But the MBLD was still not happy. The drive was still a failure.
When debrick got done with drive B, I put it in the MBLD (by itself). Initially, I put it into slot B and got an eternal yellow LED and the Dashboard was stuck forever at “Initializing device. Please wait…” When I put it into the A slot, it had a yellow LED for a long time, but eventually turned red. On the web interface I got the same error about not being able to mount the data volume.
I had to do the same things to this drive with parted that I did to the first drive but when I did, it did not show partition 4 as being an ext4 partition. It was blank. I didn’t like that so I re-debricked it (three times). The first time, it didn’t fix partition 4. So I used GPARTED to delete all of the partitions and debricked again. Got the usual slew of errors and debricked again. Now partition 4 was an ext4 partition. Then I put it into the MBLD and re-fixed it with parted.
After doing that, I wanted to make sure that the device would boot up with either drive in either slot. It did. Then I put both drives in and booted one more time. It came up, but was still showing the two failed drives and messages about not being able to mount the data volume. When I did the quick factory reset – JACKPOT! I now have a 12TB MBLD. The default RAID configuration with a new MBLD build is RAID-0, which is why I have a 12TB device. And if I remember correctly, my device was 6TB out of the box. I had to change it to RAID-1 (my preference).
I wanted to check the swap partition. It bitched about there not being a /dev/md2 and swapon –s was not showing anything, but the silly thing was working so I left it alone – for now. The command to create /dev/md2 looks like it’s specific to RAID-1 and I didn’t want to muck with it since my device was sitting at RAID-0. I went ahead and used the web interface to switch it back to RAID-1 like it was before. After it gets done building the RAID-1, I’ll check it again and MAYBE I’ll run the commands to set up /DEV/MD2 and the swap. But I suspect that the reason the GUIDE tells you to do these steps manually is because all new debricks default to RAID-0. Just a hunch. To be honest, I really don’t know jack about Linux RAIDs. After about two hours, it’s about 19% complete with rebuilding the RAID-1. When it’s done, I’ll check to see if /DEV/MD2 and swapon –s show up differently.
