I don't really know where to start here.  when I first set up my unraid I remember getting confused with setting up the cache but got something that worked so moved on.

the time has come now to try to fix that mess as I need the sata ports.

I have 3 SSD's in my cache pool,  1x2tb and 2x 120Gb I plan to get rid of the 2x120Gb ones leaving the 2Tb one in the cache pool on its own to give me 2 extra ports for other things.

I'm not entirely sure how to do it though  from this it looks to me that I've somehow created a hybrid pool 

 

 

is there a way without destroying anything to remove the cache completely and re create it with the one drive only?

 

First you need to convert the pool to raid1, then you can remove one device at a time, see here:

https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=480418

 

 

 

so i just select convert to raid1 mode in the gui and let it do its thing then remove the 2 small drives 1 by 1?

Correct.

thanks, ill let you know how i go its doing a preclear of a disk at the moment 

ok i converted to raid 1, current status is 

 

stopping the array and selecting no device for one of the 120gb drives gives this when starting array

 

ive stopped anything i can think of that will use the cache for now while its so small

You may need to change the pool profile manually, post the diagnostics.

I saw this this morning and was thinking the same 

problem at the moment is it wont stop the array from the gui.  I've tried from 2 different pc's

sag-a-star-diagnostics-20240221-0915.zip

Make sure you don't have an SSH window opened to a mount point.

 

To fix the pool issue the easiest way is to re-import it, it should then use the correct profile, to re-import, stop array, unassign all pool devices, start array, stop array, re-assign all pool devices, start array.

no ssh windows  open that i'm aware of.  rebooting worked and allowed me to stop the array.

before I do this should I set all shares to array primary and run mover to move all data off the cache pool or will data survive this?

 

Data survives.

still the same situation.  I try to remove disk and it wont start the array

 

then 

 

 

Post new diags.

I have repeated the removal of all drives, start array, stop array.  i figured out that if I reassign all the drives in the same places but before i start the array i can click the "cache" drive and I can change the file system type from "Auto" to "btrfs" and then choose "raid 1"  should I apply this or will i lose my data?

 

Leave in auto and post new diags after array start.

@JorgeB

Diagnostics attached as requested

 

sag-a-star-diagnostics-20240221-1359.zip

 

             Data      Data     Metadata  System                              
Id Path      single    RAID1    RAID1     RAID1    Unallocated Total     Slack
-- --------- --------- -------- --------- -------- ----------- --------- -----
 1 /dev/sdf1   5.00GiB 33.00GiB   2.00GiB 32.00MiB    71.76GiB 111.79GiB     -
 2 /dev/sdq1 807.44MiB 38.00GiB   2.00GiB        -    71.00GiB 111.79GiB     -
 3 /dev/sde1 244.95GiB 71.00GiB   2.00GiB 32.00MiB     1.51TiB   1.82TiB     -
-- --------- --------- -------- --------- -------- ----------- --------- -----
   Total     250.74GiB 71.00GiB   3.00GiB 32.00MiB     1.65TiB   2.04TiB 0.00B
   Used       34.12GiB 68.45GiB 154.41MiB 64.00KiB     

 

Conversation to raid1 didn't finish, or it aborted, it's using single and raid1 profiles for data, run the conversion to raid1 again and post new diags once it's done. 

@JorgeB

 

the balance for converting runs for about a second then stops

hope the video works

 

 

 

 

sag-a-star-diagnostics-20240221-1407.zip

Balance is aborting because btrfs is detecting data corruption:

 

Feb 21 14:07:29 SAG-A-STAR kernel: BTRFS info (device sdf1): relocating block group 8046271987712 flags data
Feb 21 14:07:30 SAG-A-STAR kernel: BTRFS warning (device sdf1): csum failed root -9 ino 257 off 433868800 csum 0x9b7bca66 expected csum 0x9ad9809d mirror 1
Feb 21 14:07:30 SAG-A-STAR kernel: BTRFS error (device sdf1): bdev /dev/sde1 errs: wr 0, rd 0, flush 0, corrupt 772, gen 0
Feb 21 14:07:30 SAG-A-STAR kernel: BTRFS warning (device sdf1): csum failed root -9 ino 257 off 433868800 csum 0x9b7bca66 expected csum 0x9ad9809d mirror 1
Feb 21 14:07:30 SAG-A-STAR kernel: BTRFS error (device sdf1): bdev /dev/sde1 errs: wr 0, rd 0, flush 0, corrupt 773, gen 0
Feb 21 14:07:32 SAG-A-STAR kernel: BTRFS info (device sdf1): balance: ended with status: -5

 

Run a scrub, it should list the corrupt files in the syslog, delete/restore those files from a backup, and then try again, alternatively, backup what you can from the pool and reformat with the single device.

 

It would also be a good idea to run memtest, data corruption can be the result of bad RAM.

scrub finished 

 

Syslog

 

memtest is a problem, server is in the loft without a screen as i lent it to a mate. i will have to run that at a later date

how do I go about backing up the pool? do i just share everything and back it up to another PC?

 

2 minutes ago, ApriliaEdd said:

pool? do i just share everything and back it up to another PC?

That is one option.  
 

Alternatively you can move it to the array using the process documented here in the online documentation accessible via the ‘Manual’ link at the bottom of the GUI or the DOCS link at the top of each forum page.

Looks like there's only one corrupt file, you can also delete it and try the conversion again.

I think you've cracked it!

 

 

Thanks for your help today I'll try to remove one of the smaller drives tomorrow.

 

Should work now.

@JorgeB

Morning,

new day same problem.  Stopped the array, removed one of the 120Gb disks and hit start 🤞........same thing as yesterday "Wrong Pool State".

 

ran a scrub: no errors, ran a balance which took 20 mins but completed without error, and tried again same story.

 

but, I had the logs open when I deselected the 120Gb drive and noticed this

 

 

Does it think i have 4 drives in the cache?

 

(sdy) is in another pool and has never been part of the cache

 

So I guess next question is how do I correct this?  that pool is used for a specific share for a proxmox backup server VM that didn't like being on the array.