I've got an old QNAP TS-412 that's been in service far too long, and I'm facing some serious issues. Currently, I have two 3TB WD RedPlus drives set up in a RAID1 mirror (drives sda and sdd). Recently, I added two more identical drives and formatted them, successfully migrating to RAID5 with disk 2 (sdb). However, when I tried to migrate to RAID6 with disk 3 (sdc), I encountered an I/O error, and now my system is stuck in a recovery loop. Access is only possible through a very slow SSH connection since the web app is unresponsive due to high CPU usage.
Here's what I find perplexing: when I look at the mdstat output, it shows RAID6 attempting to involve both sdc and sda with an unusual status, while at the same time, RAID5 includes sdb and sdd. It seems like the initial RAID1 successfully moved sda over to the new array, but things fell apart during the transition to RAID6.
My main question is: Is it possible to remove these drives from the QNAP and mount them on my Debian machine to manually rebuild the RAID5? Any insights or resources would be greatly appreciated!
4 Answers
Given that your setup is based on `mdadm`, you're in a decent spot for recovery, especially compared to hardware RAID options, which could lock you out altogether. Your main goal should be to clarify whether you want to stick with the old QNAP or move everything over to Debian for better management. Don't rush into changing the OS until you fully wrap your head around the current state and how you got here; careful documentation will help avoid confusion as you troubleshoot this. And if necessary, using external resources or even posting clearer `mdstat` outputs could help others assist you better.
I've also got a TS412, and yes, you can take those disks and use them in a Debian machine. Just make sure you install `mdadm` first and then run `mdadm --assemble --scan`. This should help you get access to your RAID setup. If you're comfortable, I'd also be happy to help you get the latest Debian installed directly on your QNAP; it performs way better than the stock OS!
That `mdstat` output can seem confusing, but one way to approach this is by keeping a detailed log of all actions and drives' statuses. Start by creating a spreadsheet to chart the conditions of each hard drive. Once you have a clearer picture, it should help you figure out the next steps for recovery. Also, check out this recovery story I shared about dealing with a failed RAID controller; it has some useful insights linked inside!
The `mdstat` output you're seeing is quite unusual. It suggests some strange issues during the RAID level changes. My recommendation is to connect the drives to your Debian machine, try assembling the RAID setups, and see if you can access the data on either md0 or md3. If you can, back it up to another drive so you can rebuild the array safely. It sounds like the transitions are causing confusion; understanding the migration process will be key to recovery.
I suspect the transition from RAID5 to RAID6 might have locked things up. It could have mishandled the arrays, creating this dual setup. Clearing that up is crucial!
I'd love to know more about getting Debian on there! I haven't dealt with RAID in Debian before, just started looking into `mdadm`, but I'm a bit hesitant since I don't want to risk losing data.