Hello there,
unfortunately I have two different sized NVMe's in my cache pool. 2TB the old one and 4TB the new one. I only realized later what a stupid thing I had done. I actually wanted to pack the SSD into my array.
I now have put the second one in single mode.

Question:
1. how can I make sure that there is no more data on the second NVMe? (it is just 3 days in my Server)
2. how can I remove the second NVMe from the pool?

 

Thanks for your help.

 

Best Regards
AceBurn90

If you have 2 SSDs in a pool in Single mode then you cannot remove one from the pool and keep its data.  I would suggest you post your system's diagnostics zip file so we can see how you actually have things set up.

 

Why do you want to put the new SSD into the array?    SSDs in the main array cannot be trimmed which might lead to performance loss over time, and anyway any writing is limited by the speed of the parity drive.  Perhaps if we understand better what you are trying to achieve we might be able to suggest a better way forward?

Thank you for the important information. The anonymized diagnostic file is attached.

I always plan to access an SSD quickly from my computer. My HDDs are too slow for this as they are always in sleep mode when I'm not using them for 15 minutes. Therefore, the fast access time and allways on of SSDs is important to me in this case.
If I could remove this second SSD from my first cache pool and create a second one for it, this should solve my problem, right?
The only problem I see is that my data would not be backed up in this case.
Is there a simple solution for this? Always have the data automatically duplicated in the array? Is that possible?

alent-diagnostics-20250114-1146.zip

You have less than 2TB on the pool (the size of the smaller drive), so you could change the profile to be RAID1 and run a Balance.  After that it would be possible to remove the larger drive from the current pool.

 

If the 4TB drive is then used in a new single drive pool then it would be optimal for performance but not redundant.   You could have regular backups done to the array via a script run using the User Scripts plugin - not sure if that will suit your needs?   It is worth pointing out that even redundancy/parity is not a replacement for backing up critical data so you should still have a backup strategy in place for really important stuff.

That sounds good. I am currently switching to RAID1.
That should let me remove the 4TB afterwards.
I was denied this in single mode.

The data in the 4TB pool won't be that important. All important long-term data is stored in the array.

20 minutes ago, AceBurn90 said:

That sounds good. I am currently switching to RAID1.
That should let me remove the 4TB afterwards.
I was denied this in single mode.

The data in the 4TB pool won't be that important. All important long-term data is stored in the array.

Once you are in RAID1 mode then you can remove the drive leaving all data intact.

I have just done something very stupid.
I deselected the 4TB as Unassigned under “Main”.
Then I selected “Array” under “Tools>New Config” under “Preserve current assignments:” and clicked Apply.
Unfortunately, I clicked too quickly. Can I undo this?

I still have my exact assignments in an Excel file. I have reassigned these accordingly.

 

But what is the next step?

Can I just start the array and have to create the parity again?

Can I remove the 4TB SDD from the Cache pool at the same time?

 

Edited January 14, 20251 yr by AceBurn90

If the array assignments are the same as before, you can check "parity is already valid" before array start, but you must start the array once before trying to remove the pool device.

3 minutes ago, JorgeB said:

If the array assignments are the same as before, you can check "parity is already valid" before array start, but you must start the array once before trying to remove the pool device.

Thanks for your help.
Looks good. Everything seems to be working.
What are the steps to remove the cache pool device?

Post diags before just to confirm the profile and cfg are correct.

1 minute ago, JorgeB said:

Post diags before just to confirm the profile and cfg are correct.

Here is the new diags

alent-diagnostics-20250114-1352.zip

Looks OK, you just need to unassign the device you want to remove and start the array, then wait for the balance to finish.

4 hours ago, JorgeB said:

Looks OK, you just need to unassign the device you want to remove and start the array, then wait for the balance to finish.

I have to thank you. It all worked out!

 

Now that I have two pools, I was wondering if I can convert the old cache pool from btrfs to “ZFS Encrypted”?

 

If not directly, I could use the second pool to backup the data and then reformat the cache pool, right?
What would be the easiest way to do this?

13 hours ago, AceBurn90 said:

I was wondering if I can convert the old cache pool from btrfs to “ZFS Encrypted”?

You can, but requires reformatting, so you need to backup first.

On 1/15/2025 at 9:47 AM, JorgeB said:

You can, but requires reformatting, so you need to backup first.

The cache currently only contains the “docker” and “system” folders.
How should I proceed?

All private data is already backed up.
I don't have to change my shares and Docker container configurations, because the storage path will be the same later, right?

Only the Docker and system folder would have to be moved before I can format the cache and then move them back again.

But I don't know exactly how to do this.
I only know that I have to deactivate Docker and the VM Manager in the settings. But how do I move the data and change the path to the newly formatted cache later?

You can use this:

https://forums.unraid.net/topic/46802-faq-for-unraid-v6/#findComment-511923

 

That's easier than I thought.^^
Thank you!

I now have a fully functional array including all data.
My two pools are the way I want them to be. (cache + quick access)
The appdata and system folders are back where they belong (cache pool)
Now I have the problem that my shares are all gone.
What has happened again?

alent-diagnostics-20250116-1916.zip

I would think that is because of:

Jan 16 19:13:38 Alent root: mkdir: cannot create directory '/mnt/user': File exists

Not sure what you have done that is causing that to happen as that is where shares would normally appear.   Sounds like a script or something running early in the boot sequence but I could be wrong about that.

 

You might want to try booting in Safe Mode just in case some plugin is inadvertently causing this to eliminate that possibility.

After you get your user shares working again, some other things I noticed.

22 minutes ago, AceBurn90 said:

The appdata and system folders are back where they belong (cache pool)

But they are configured to be moved to the array.

 

Also

On 1/14/2025 at 6:12 AM, AceBurn90 said:

All important long-term data is stored in the array.

You must still have a backup of anything important and irreplaceable Parity is not a substitute.

3 minutes ago, trurl said:

After you get your user shares working again, some other things I noticed.

30 minutes ago, AceBurn90 said:

The appdata and system folders are back where they belong (cache pool)

But they are configured to be moved to the array.

I know that the data on the array is configured as “move to array”.
I only changed this to move my data from the pool. I was just about to reset everything when I noticed that the shares were gone.

 

3 minutes ago, trurl said:

Also

On 1/14/2025 at 12:12 PM, AceBurn90 said:

All important long-term data is stored in the array.

You must still have a backup of anything important and irreplaceable Parity is not a substitute.

I am aware that the data is not backed up.
I have backed up the important data externally and I am still configuring further backup steps when everything is up and running again.

Everything works again.

 

Reason for error:
I moved the “docker” and “system” folders myself instead of using the mover to then set the shares as desired again (cache to array // system remains in the cache).

 

Solution:
I moved the files back in safe mode, then moved the data back again with the mover to where they belong in the endfect and then started the array to check. I then carried out a clean restart and restarted everything as it should be.

Incidentally, I noticed that the mover still didn't move everything, whether from cache to backup SSD or vice versa. I had to move the remaining Docker files and folders manually (without overwriting). It worked as desired.

I rejoiced too soon.
Now that I have also formatted the last three hard disks in the array in encrypted XPS, I have restarted the Docker service.

 

The settings of the storage locations are as follows:
Docker directory: /mnt/user/system/docker/installed/
Default appdata storage location: /mnt/user/appdata/

 

This also corresponds to the data in the cache pool.
Unfortunately, no Docker container is displayed. Can you perhaps help again?

Attached I have the last diagnostic report from a moment ago.

alent-diagnostics-20250117-0011.zip

I suspect the system > docker > installed is inconsistent or corrupted or something else that is causing the problems.
Can I just delete it and have it recreated after disabling the docker service?
Do I then just have to restart the Docker service and the docker > installed folder will generate all its data again?

Not likely to be the cause of this problem, but I see this from FCP

Jan 17 00:01:04 Alent root: Fix Common Problems: Warning: Share SABnzbd Downloads is set for both included (disk1,disk2,disk3,disk4,disk5) and excluded (disk10) disks ** Ignored

There is never any good reason to set both Include and Exclude.

 

Include already excludes all disks not in its list.

Exclude already includes all disks not in its list.

 

Also, setting these will not affect any files which already exist on the included or excluded disks. All they determine is where new files get written. All disks are included for reading.

 

Now for the actual point of this thread.

 

 

appdata and system shares are on cache and set to use cache only, but it doesn't look like there is much if any data on cache.

 

docker.img is easily recreated from your templates on flash, so you can reinstall your containers with Previous Apps. But unless you have appdata backed up somewhere you would be starting over with the actual applications.

 

https://docs.unraid.net/unraid-os/manual/docker-management/#re-create-the-docker-image-file

 

https://docs.unraid.net/unraid-os/manual/docker-management/#re-installing-docker-applications