Hello Unraid users,

 

I recently had a BTRFS pool that I was trying to remove a 1tb drive from and add two 12tb drives. I was also adding an SSD to a cache pool.

 

During the shutdown process after doing this to add the 12tb drives (I removed the 1tb drive first from the pool, but it wasn't using much storage, so I assumed that it didn't need much time to do it) I waited for a while and then did an ungraceful shutdown as it wasn't seeming to do anything anymore.

 

After rebooting the server, it wasn't bringing back up the server interface. I tried a safe boot and a GUI boot and found that localhost didn't bring anything up. I understand that I messed up by being too impatient.

 

I'm adding the log file that I got from a console login shutdown that I did afterwards.

 

I'm able to boot the device from the USB. I'm currently using Unraid 7.0.1.

tower-diagnostics-20250310-1643.zip

Do you have a flash backup? Doesn't look like much of your configuration is in those diagnostics.

 

5 minutes ago, muddles17 said:

recently had a BTRFS pool that I was trying to remove a 1tb drive from and add two 12tb drives.

Do you really mean the array? or are you not using the array?

20 minutes ago, trurl said:

Do you have a flash backup? Doesn't look like much of your configuration is in those diagnostics.

 

Do you really mean the array? or are you not using the array?

The flash drive itself still seems to be ok. It's booting up, and I can read it.

 

I didn't quite realize how it's supposed to be setup for a standard unraid setup. I ended up doing a btrfs pool with one SSD and a second btrfs pool with 4 hard drives.

 

I was doing the following:

cache pool:

2tb SSD -> 2x 2tb SSD with BTRFS RAID1

storage pool:

2tb HDD + 4tb HDD + 4tb HDD + 1tb HDD

to

2tb HDD + 4tb HDD + 4tb HDD + 12tb HDD + 12tb HDD

 

I have a backup of my flash drive, but it is from before I started doing these operations.

I managed to get everything to start up by backing up my flash as-is as an image and then running UNRAID USB creator on it and dragging and dropping the contents of the backup over the contents of the USB. I had to then add the second SSD again to get the BTRFS to start.

 

It's missing all of the docker configs though. Is there a way to get that back?

I tried copying my /config/plugins/dockerMan folder from the backup that I made to the USB drive, and my docker folder is still empty. I am able to get to the docker folder from my file browser. I'll search around a bit to see if I can add it to the UI. If that doesn't work, I'll copy the directory, rebuild it, and then copy everything back in.

42 minutes ago, muddles17 said:

storage pool:

2tb HDD + 4tb HDD + 4tb HDD + 1tb HDD

to

2tb HDD + 4tb HDD + 4tb HDD + 12tb HDD + 12tb HDD

Still not entirely clear. Is this the array, or is it another pool?

 

You typically don't want different sized disks in a pool. That is one of the great benefits of the Unraid array.

Ah - what worked was disabling docker in the docker tab, re-selecting the location in the interface, applying and then selecting enabled again and applying again.

 

Everything is back!

 

Thank you for the moral support on getting started.

New diagnostics would probably go a long way to understanding what you have and what you need to have.

 

 

6 minutes ago, trurl said:

Still not entirely clear. Is this the array, or is it another pool?

 

You typically don't want different sized disks in a pool. That is one of the great benefits of the Unraid array.

I'm using a BTRFS RAID1 array right now with this https://carfax.org.uk/btrfs-usage/ type of calculation for all of the drives. As-is, it doesn't have any blank space and is just space / 2.

 

I didn't realize that the pools and arrays were completely different when I started. everything. I have the 2x12tb that I still need to add.

 

Should I make a new array and copy everything over? As far as I can tell, it seems like BTRFS has some data integrity that it can run to make sure that everything is running properly / help with bit rot. Does the unraid array have something like that? When I ran scrub on my data, it picked up a significant amount of issues that could be helpful in the future. From what I could read on the documentation, if unraid info is incorrect, it primarily just grabs the data and writes it to parity without trying to correct data.

Edited March 10, 20251 yr by muddles17
adding detail

Just now, muddles17 said:

I'm using a BTRFS RAID1 array right now

This is not possible for the Unraid array. Do you mean it is actually another pool?

 

3 minutes ago, trurl said:

New diagnostics would probably go a long way to understanding what you have and what you need to have.

 

 

 

Here is a new diagnostics file.

 

You are correct. It is another pool.

holonet-diagnostics-20250310-1837.zip

btrfs raid1 can't use all the capacity when you only have 2 different sized disks, but it distributes things when you have more than 2 so there is always a mirror on other device(s). So I think that looks OK for the capacity of storage pool as originally configured but not clear whether anything else about it is broken.

 

I don't have any experience with more than 2 in btrfs raid1, and have since converted both of my 2 disk pools to zfs, so if you need help changing/adding disks someone else will have to get involved. I think the main thing is you have to wait for things to balance, which might take some time especially with HDDs, and perhaps that is where you got into trouble.

 

Not exactly related, but your Docker share is configured to only use cache pool, but it has files/folders on storage pool.

 

6 minutes ago, trurl said:

Not exactly related, but your Docker share is configured to only use cache pool, but it has files/folders on storage pool.

oops. I'll have to get that moved.

 

Could I have two zfs arrays in a pool or array that would contain sets of two that would provide the ability to do scrubs? Would those go into pools or do zfs mirrors go into the array?

 

zfs is a bit more confusing from that aspect. The main benefit that I considered with btrfs is that it allowed for slapping everything into a raid and then it would take care of the complicated parts in the background.

The Unraid array are all single data disks, with separate one or two parity disk. The array data disks can be xfs, btrfs, zfs, but they are separate independent filesystems that can be read all by themselves on any linux. Any file in the array is completely on a single disk. Folders can span disks and pools, that is what User Shares are all about.

https://docs.unraid.net/unraid-os/overview/nas/#parity-protected-array

 

This allows different sized disks in the array, and allows disks that aren't being accessed to spin down. And problems with one disk can't affect data on other disks.

 

I'm may not be as familiar with the finer points of btrfs/zfs as you are. I just wanted mirrors on SSDs for those use cases where speed is more important than capacity.