Hi *,

first of all - I'm an absolute newbie to unraid.

I'ver recently purchased a small SSD only NAS (SyncStation N2). Now I'm in the process of setting up Unraid, which has been a bit of a mixed bag.

The system is set up with four 4TB nvme and two SATA 4TB SSDs (all new WD devices).

After registering for an initial 30 day trial period and updating to 7.1.2, I tried to set up the six disks as a ZFS pool. (I will not extend the system with further disks and see no advantage in using an array.) I'm roughly following the steps outlined in this Uncast Show: https://www.youtube.com/watch?v=z3oLhXEts-4&t=152s

1. I set the number of slots for the array to 0.

2. I created a pool with six slots and assigned all disks to the slots (beginning with the NVMEs /dev/nvme0n1 through to /dev/nvme3n1 followed by the two SATA drives /dev/sdb and /dev/sdc).

3. Then I set up the first device to utilize ZFS with raidz1 and configured one vdev with six devices.

4. After that I started the (non-existent) array and tried to format the pool devices. This fails reproducably.

The error message in the GUI is that the disk is unmountable with wrong or no file system.

The relevant portion (as far as I can see) from the syslog is as follows - I find the attemp to use wipefs on /dev/ strange:

May 23 15:35:19 nas emhttpd: creating volume: storage (zfs)
May 23 15:35:19 nas emhttpd: shcmd (2179): /sbin/wipefs -af --lock /dev/nvme0n1
May 23 15:35:19 nas root: /dev/nvme0n1: 8 bytes were erased at offset 0x00000200 (gpt): 45 46 49 20 50 41 52 54
May 23 15:35:19 nas root: /dev/nvme0n1: 8 bytes were erased at offset 0x3a3817d5e00 (gpt): 45 46 49 20 50 41 52 54
May 23 15:35:19 nas root: /dev/nvme0n1: 2 bytes were erased at offset 0x000001fe (PMBR): 55 aa
May 23 15:35:19 nas emhttpd: writing GPT on disk (nvme0n1), with partition 1 byte offset 1MiB, erased: 0
May 23 15:35:19 nas emhttpd: shcmd (2180): sgdisk -Z /dev/nvme0n1
May 23 15:35:20 nas root: Creating new GPT entries in memory.
May 23 15:35:20 nas root: GPT data structures destroyed! You may now partition the disk using fdisk or
May 23 15:35:20 nas root: other utilities.
May 23 15:35:20 nas emhttpd: shcmd (2181): sgdisk -o -a 8 -n 1:1M:0 /dev/nvme0n1
May 23 15:35:21 nas kernel: nvme0n1: p1
May 23 15:35:21 nas root: Creating new GPT entries in memory.
May 23 15:35:21 nas root: The operation has completed successfully.
May 23 15:35:21 nas emhttpd: shcmd (2182): udevadm settle
May 23 15:35:21 nas emhttpd: /bin/lsblk -lnbo TYPE,PARTN,NAME,START,SIZE,PTTYPE,PARTTYPE /dev/nvme0n1 2>&1
May 23 15:35:21 nas emhttpd: disk  nvme0n1         4000787030016 gpt
May 23 15:35:21 nas emhttpd: part  nvme0n1p1  2048 4000785964544 gpt
May 23 15:35:21 nas emhttpd: device nvme0n1: no partitions
May 23 15:35:21 nas emhttpd: shcmd (2183): /sbin/wipefs -af --lock /dev/
May 23 15:35:21 nas root: wipefs: error: /dev/: probing initialization failed: Is a directory
May 23 15:35:21 nas emhttpd: shcmd (2183): exit status: 1
May 23 15:35:21 nas emhttpd: shcmd (2184): /sbin/blkdiscard /dev/
May 23 15:35:21 nas root: blkdiscard: cannot open /dev/: Is a directory
May 23 15:35:21 nas emhttpd: shcmd (2184): exit status: 1
May 23 15:35:21 nas emhttpd: shcmd (2185): /sbin/wipefs -af --lock /dev/nvme1n1p1
May 23 15:35:21 nas emhttpd: shcmd (2186): /sbin/blkdiscard /dev/nvme1n1p1
May 23 15:35:22 nas emhttpd: shcmd (2187): /sbin/wipefs -af --lock /dev/nvme2n1p1
May 23 15:35:22 nas emhttpd: shcmd (2188): /sbin/blkdiscard /dev/nvme2n1p1
May 23 15:35:22 nas emhttpd: shcmd (2189): /sbin/wipefs -af --lock /dev/nvme3n1p1
May 23 15:35:22 nas emhttpd: shcmd (2190): /sbin/blkdiscard /dev/nvme3n1p1
May 23 15:35:23 nas emhttpd: shcmd (2191): /sbin/wipefs -af --lock /dev/sdb1
May 23 15:35:23 nas emhttpd: shcmd (2192): /sbin/blkdiscard /dev/sdb1

Now I'm stuck and would be happy to receive any hints about how to continue.

I'm also including the diagnostics.

Kind regards

Andreas

nas-diagnostics-20250523-1601.zip

So, I tried a few things, by more or less following the commands listed in the syslog, but changing them at the point that I found strange (see my post above):

root@nas:~# /sbin/wipefs -af --lock /dev/nvme0n1

/dev/nvme0n1: 8 bytes were erased at offset 0x00000200 (gpt): 45 46 49 20 50 41 52 54

/dev/nvme0n1: 8 bytes were erased at offset 0x3a3817d5e00 (gpt): 45 46 49 20 50 41 52 54

root@nas:~# sgdisk -Z /dev/nvme0n1

Creating new GPT entries in memory.

GPT data structures destroyed! You may now partition the disk using fdisk or

other utilities.

root@nas:~# sgdisk -o -a 8 -n 1:1M:0 /dev/nvme0n1

Creating new GPT entries in memory.

The operation has completed successfully.

root@nas:~# /bin/lsblk -lnbo TYPE,PARTN,NAME,START,SIZE,PTTYPE,PARTTYPE /dev/nvme0n1

disk nvme0n1 4000787030016 gpt

part 1 nvme0n1p1 2048 4000785964544 gpt 0fc63daf-8483-4772-8e79-3d69d8477de4

root@nas:~# udevadm settle

root@nas:~# /sbin/wipefs -af --lock /dev/nvme0n1p1

root@nas:~# /sbin/blkdiscard /dev/nvme0n1p1

Now follows the major change - I replaced the very strange /dev/ with the first partition on the first nvme: /dev/nvme0n1p1

root@nas:~# zpool create -f -o ashift=12 -O dnodesize=auto -O acltype=posixacl -O xattr=sa -O utf8only=on -m /mnt/storage storage raidz1 /dev/nvme0n1p1 /dev/nvme1n1p1 /dev/nvme2n1p1 /dev/nvme3n1p1 /dev/sdb1 /dev/sdc1

root@nas:~# zpool list

NAME SIZE ALLOC FREE CKPOINT EXPANDSZ FRAG CAP DEDUP HEALTH ALTROOT

storage 21.8T 864K 21.8T - - 0% 0% 1.00x ONLINE -

root@nas:~# ls /mnt/storage/

appdata/ domains/ isos/ system/

root@nas:~# df -h /mnt/storage/

Filesystem Size Used Avail Use% Mounted on

storage 18T 256K 18T 1% /mnt/storage

So now I seem to have a pool with about the expected size (6*4T using raidz1 should be something like 20T). The dashboard displayed the storage immedeately - but strange enough the web gui pages "Main" and "Shares" did not pick this up - only after rebooting the pool created by hand appeared. (This could be a simple cosmetic problem.)

Apparently the commands that the GUI generated where wrong, enforcing the very flaky first impression.

Other problems that had to be fixed by hand were

• a very uninformative Error message that the registration of the 30-day trial failed

• after copying the downloaded key file to the correct position an error message that internet access was not available, even though pings to google and other major sites worked as well as nslookups

So the first impression of Unraid leaves much to be desired.

3 hours ago, Zielke said:

May 23 15:35:21 nas emhttpd: disk  nvme0n1         4000787030016 gpt
May 23 15:35:21 nas emhttpd: part  nvme0n1p1  2048 4000785964544 gpt

This typically means there is a signature issue, post the output from:

wipefs /dev/nvme0n1p1

Hi Jorge,

thanks for taking the time to help! Much appreciated.

This is the output (after the manual steps I described in my own reply to my first post):

root@nas:~# wipefs /dev/nvme0n1

DEVICE OFFSET TYPE UUID LABEL

nvme0n1 0x3a381744000 zfs_member 2899005060788001305 storage

nvme0n1 0x3a381784000 zfs_member 2899005060788001305 storage

nvme0n1 0x200 gpt

nvme0n1 0x3a3817d5e00 gpt

nvme0n1 0x1fe PMBR

Kind regards

Andreas

Sorry, I edited my post, but not in time, it was missing the partition, it should be:

wipefs /dev/nvme0n1p1

Absolutely no problem, here you go:

root@nas:~# wipefs /dev/nvme0n1p1

DEVICE OFFSET TYPE UUID LABEL

nvme0n1p1 0x4000 zfs_member 2899005060788001305 storage

nvme0n1p1 0x44000 zfs_member 2899005060788001305 storage

nvme0n1p1 0x3a381644000 zfs_member 2899005060788001305 storage

nvme0n1p1 0x3a381684000 zfs_member 2899005060788001305 storage

I see that you manually created a pool after that, so it's going to be difficult to see what the problem was, but if there's still no data there, and you can try again, export the current pool first with zpool export storage, then do a full device trim on the problem device, and you should then be able to create the pool using the GUI

blkdiscard -f /dev/nvme0n1

If it still fails, please type LIBBLKID_DEBUG=all blkid -p -o udev /dev/nvme0n1p1 > /boot/debug.txt

This will have a few lines, so it will output to a text file on the flash drive, then please attach that here.

Ok, I tried to export the pool, but this fails when the (still non-existant) array is running. (Error message: cannot unmount '/mnt/storage/system': pool or dataset is busy).

When I stop the array, the pool cannot be exported, as it is not available.

The share system was auto-generated btw.

Jorge, sorry - I have to leave the computer until tomorrow.

Many thanks for your help up to this point!

Ok, some further observations and additional info:

I have not (yet) been able to figure out a way to get rid of the busy system share, which stops me from exporting the pool.

While examining the contents of the system share to get an idea of the potential damage I had a another thought:

I fought the GUI for some time, trying to set up a pool that would start up. One of the changes I made during these trials was to change the order of the disks.

On my first go I created a pool with six slots and added the two SATA drives as devices in the first two slots and continued with the NVMEs. When trying to format the disks, I ran into the same situation as now. (Judging only by the little information that the GUI provides; I did not consult syslog at this point.)

I then removed the pool, started fresh and changed the order of the drives.

Is this observation meaningful in any way - or could the error still be due to a potentially problematic NVME?

Moving on:

• I found out, that the system share contained to .img files which where mounted. Judging by name they belong to Docker and virtual machines.

• After stopping both options (using the GUI via Settings --> Docker --> Enable Docker --> No and Settings --> VM Manager --> Enable VMs --> No) the IMG files were not mounted any more.

• I was then able to export the pool.

root@nas:~# zpool export storage
root@nas:~# ll
root@nas:~# zpool list
no pools available
root@nas:~# blkdiscard -f /dev/nvme0n1
blkdiscard: Operation forced, data will be lost!

I then rebootet, as the system did not reflect the changes.

After rebooting the pool still showed up in the GUI, but I was able to remove it.

I re-created a new pool with six devices and added the devices. This resulted in a pool with five devices; the nvme0 was "Not installed". As all other device were picked up seemingly without the need to format them, I guess that I'll have to re-format the nvme0 after the blkdiscard.

I'm currently running a smart test on the device, which takes some time. I'll continue with my trials after that test.

The extended smart test did not show any errors.

The commands specified yielded no result, as in the current state there seems to be no partition on the vnme0n1:

root@nas:~# LIBBLKID_DEBUG=all
root@nas:~# blkid -p -o udev /dev/nvme0n1p1 > /boot/debug.txt
root@nas:~# cat /boot/debug.txt
root@nas:~# ls -lsa /dev/nvme0
root@nas:~# ls -lsa /dev/nvme0*
0 crw------- 1 root root 242, 0 May 24 06:53 /dev/nvme0
0 brw-rw---- 1 root disk 259, 1 May 24 06:53 /dev/nvme0n1
root@nas:~#

I ended preparing nvme0 using the manual steps listed in my first reply.

I was then able to create the pool via the GUI.

Were these brand new devices, especially nvme0n1, or had it been used before with a different OS?

All bought as new.

OK, thanks for confirming.