In my server I have 2 512gb SSDs, they were running as a cache pool. I was sure it was in RAID1 mode, but TBH I don't know how to check this.
I installed 2 new 1tb NVMe drives, intending to use them as a new cache pool.
(all 4 drives are installed at the same time).
 

After booting Unraid, cache running as before on the SSD drives, 2 new NVMe drives showing unassigned. So far so good.

I proceeded to follow the instructions found here https://forums.unraid.net/topic/46802-faq-for-unraid-v6/#comment-480419

 

 - I stopped the array
 - I changed Cache 2 drive to point to one of the NVMe drives.
 - I started the array

 

No sign of any brtfs device replace beginning or any pool activity.
The cache is offline, it shows as Unmountable: Unsupported or no file system.
The SSD that was removed from the cache shows as unassigned.

Hmm. I tried putting things back to how they were.

 

 - I stopped the array
 - I changed Cache 2 device to the now unassigned SSD.
 - I started the array

 

The cache is still offline, showing as Unmountable: Unsupported or no file system.

Following the instruction here https://forums.unraid.net/topic/141562-cache-unmountable-unsupported-or-no-file-system-after-update/?do=findComment&comment=1280258

 

 - I stopped the array
 - I changed Cache device to no-device
 - I changed Cache 2 device to no-device
 - I started the array (checking the no cache warning checkbox)

 

Array starts without any cache, as expected.

I then restored it:

 

 - I stopped the array
 - I changed Cache device to the SSD it was set to originally
 - I changed Cache 2 device to the SSD it was set to originally
 - I started the array

 

Array starts, cache is still offline, showing as Unmountable: Unsupported or no file system.

 

Please help!

 

At this point I would settle just to have things back to how they were, all my dockers are gone etc. 😢

server-diagnostics-20240925-2007.zip

I have found in my unraid flash drive (a backup taken a while ago) that the pool was indeed raid1, at least according to
config/pools/cache.cfg line 3:

diskFsProfile="raid1"

 

Also, in case it helps, i ran btrfs fi show

Quote

 

Label: none  uuid: 8ec33423-8912-41b5-8b37-e1c5d2886455
        Total devices 1 FS bytes used 388.00KiB
        devid    1 size 20.00GiB used 2.02GiB path /dev/loop2

Label: none  uuid: 960caaa5-b39f-4b75-9e59-7a823e8dbd9d
        Total devices 1 FS bytes used 408.00KiB
        devid    1 size 1.00GiB used 126.38MiB path /dev/loop3

warning, device 2 is missing
ERROR: cannot read chunk root
Label: none  uuid: 6ff47540-cf8f-4f92-9574-c652f94a843a
        Total devices 2 FS bytes used 98.39GiB
        devid    3 size 476.94GiB used 111.00GiB path /dev/sdc1
        *** Some devices missing

 

I don't know why it says some devices missing. They're both there as far as I can tell. Argh

 

Reading through this thread, it seems very similar.

I am wary of running the commands listed, they seem quite "deep", and i have only a shallow understanding of the underlying systems here.

It seems I'm not the only one who has fallen foul of following the instructions ☹️

Fortunately, unlike those people I have not rebooted, so the diagnostics posted maybe can be more helpful?

Edited September 25, 20241 yr by Proffles

type

sfdisk /dev/sdb

then type

2048

and hit enter, post the output of that

Ok thanks, here:

Hit CTRL + C to abort and repeat but with 64:

 

type

sfdisk /dev/sdb

then type

64

and hit enter, post the output of that

okiedokie:

Type N and Enter to keep the signature, then 

write

and enter, now post the output from 

btrfs fi show

 

Cheers, this is the output:

 

Restart the array and post new diags.

Great! thank you yes now looks like things are back to how they were to begin with. Phew!

But now, I am back to how I was to begin with, with the new cache drives waiting to be used. Do you know why following the instructions on this FAQ page did not work?

Can you suggest a revised procedure to use, as it seems the one listed is somewhat risky for some reason. Thank you for your help! 🙂

server-diagnostics-20240926-1316.zip

29 minutes ago, Proffles said:

But now, I am back to how I was to begin with, with the new cache drives waiting to be used. Do you know why following the instructions on this FAQ page did not work?

Can you suggest a revised procedure to use, as it seems the one listed is somewhat risky for some reason. Thank you for your help!

That procedure should work fine as long as you are careful that the pre-conditions are met.    For instance at the moment your pool is NOT fully redundant as it is using the Single profile for metadata and System (although it IS using RAID1 for data).   You need these to all be a redundant profile such as RAID1.

Ok, thank you. Evidently I'm not familiar enough with this topic.

Reading your reply and trying to understand it in terms of more familiar filesystems, am i correct in thinking that effectively "partitions" are RAID1 instead of whole drives.
BTRFS uses some partitions to work properly, and thus all partitions need to be set to RAID1 mode.

If this is correct, do you mind helping me understand how to check what partitions there are, what they are currently set to, and how to set them to RAID1 mode?

Thank you!

Looking at the details page of one of the cache drives, i can see this balance option



If i set the dropdown to "Convert to raid1 mode" and click Balance, will this handle it? for all partitions? What about the other cache drive, how does it know what to do with it?

Thanks!

4 minutes ago, Proffles said:

Ok, thank you. Evidently I'm not familiar enough with this topic.

Reading your reply and trying to understand it in terms of more familiar filesystems, am i correct in thinking that effectively "partitions" are RAID1 instead of whole drives.
BTRFS uses some partitions to work properly, and thus all partitions need to be set to RAID1 mode.

If this is correct, do you mind helping me understand how to check what partitions there are, what they are currently set to, and how to set them to RAID1 mode?

Thank you!

it is not the partitions - just the internal machinations of btrfs.

 

If you click on the pool name on the Main tab you can see what profile the pool is set to use (and potentially change it).    You probably at this point judt need to run a balance from their to get everything to RAID1 profile.

 

If there are any queries then @JorgeB is the expert on resolving such issues (ad you have found).

Ok thanks. I clicked the button! It now looks like this:

Can I safely attempt the upgrade again now? What is the GlobalReserve "not partition" bit, does that not need to also be RAID1?

1 hour ago, Proffles said:

Can I safely attempt the upgrade again now?

Should be fine now, global reserve profile can be ignored.

Thank you both for your help.
I followed the instructions again, and this time it appears to now be working, i can see a lot of writes on the pool and a BTRFS operation is running.

Hopefully it goes smoothly from this point on!

Woohoo. All done. thank you