HI all,

I was using co-pilot to help me pass though a keychron keyboard to my VM,as unairds 'usb' was seeing it but passing it through as a stadnard keyboard by editing the VM XML file. when i got/noticed an error (from the fix common problems plugin) that my var/log was getting full, even though my system had only been on for a few days. looking atthe logs. One of the ssd in my cache pool had lots of write and read errors - i assumed the drive was faulty/faulty cable. i reseated the drive and restarted the server. it didnt take while for erros to pop up. so i thought i would replace the drive, i stopped the array removed the drive and restated the array. the array seems to hang for an age - (to total sizeof the pool was 1.6 TB but the free space was ~1.2 TB) i assumed this would correct itself and then i could add the new drive. rather than stop the array again i restarted the server instead. this time it boots in unraid but the web GUI no longer works. looking around there are posts about it not working when there are problems with Cache, so as i kenw i had a problem with cache, i progressed through ther terminal to remove the fault drive from the cache pool and add the new drive. i have done that - hopefully sucessfully, but on restart of ther server, i still have the issue i can get into unriad but the web gui is not loading.

As an aside, i am currently looking at moving the server to a new cpu/motherboard. if the time taken to fix this would be too long or it would be better to start afresh please let me know as that might be the better option.

Any help anyone can give me would be grratfully recieved.

thanks

tower-diagnostics-20250901-1152.zip

Edited September 1, 2025Sep 1 by Belcraig

Delete /config/pools/cache.cfg from the flash drive and reboot; you should get the GUI back, the pool devices will be unassigned, then post the output from btrfs fi show

thanks Jorge

Label: none uuid: 9294ac5d-3e72-4a7d-85f3-a05d9527a03d

Total devices 1 FS bytes used 384.00KiB

devid 1 size 20.00GiB used 536.00MiB path /dev/loop2

Label: none uuid: a2c1fa2c-f854-4540-8ada-acd019fd6f37

Total devices 1 FS bytes used 416.00KiB

devid 1 size 1.00GiB used 126.38MiB path /dev/loop3

Label: none uuid: 01cc169c-ab3e-4f2b-86e3-5382b85f07e7

Total devices 3 FS bytes used 391.84GiB

devid 2 size 931.51GiB used 419.36GiB path /dev/sdc1

devid 3 size 232.88GiB used 232.88GiB path /dev/sdf1

devid 4 size 238.47GiB used 186.47GiB path /dev/sdg1

regards

seems i've lost the cache pool...

3 minutes ago, Belcraig said:

seems i've lost the cache pool...

18 minutes ago, JorgeB said:

the pool devices will be unassigned

Create a new pool with 3 slots, add the three pool devices, leave the filesystem set to auto, start the array, post new diags.

thanks

server-diagnostics-20250901-1258.zip

Pool looks OK. Which device do you think is failing?

Also, recommend updating to 7.1.4, it fixes the bug that caused the GUI issue with the pool cfg.

I'm removed the drive that was failing, having re plugged it back in the server nolonger even recognised the device.

do i just recreat the dockers etc.. and it will pick them back up?

A new Docker image was created on disk3 due to starting the array without the pool, but you can just delete it and revert to the old one, or just re-add the apps from CA:

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

Also see below if you have any custom docker networks:

https://docs.unraid.net/unraid-os/manual/docker-management/#docker-custom-networks

Last thing jorge if possible, i left the array until a while ago and i ran an extended smart scan ( on another computer) on the failed drive. it passed no problem, so i thought it might be cables so i've looked for some fresh cable but i've noticed that the Cache is now reporting unmountable - the oly thing i have done is moved the tower from under my desk to ontop of the desk so i can see into it better.

server-diagnostics-20250901-2007.zip

The other device appears to still be detected as part of the pool. Post the output from:

btrfs fi show

If it shows up there, the solution would be to disconnect it or wipe it.

Label: none uuid: 9294ac5d-3e72-4a7d-85f3-a05d9527a03d

Total devices 1 FS bytes used 376.00KiB

devid 1 size 20.00GiB used 536.00MiB path /dev/loop2

Label: none uuid: a2c1fa2c-f854-4540-8ada-acd019fd6f37

Total devices 1 FS bytes used 612.00KiB

devid 1 size 1.00GiB used 126.38MiB path /dev/loop3

Label: none uuid: 01cc169c-ab3e-4f2b-86e3-5382b85f07e7

Total devices 3 FS bytes used 406.78GiB

devid 1 size 1.82TiB used 1.05TiB path /dev/sdh1

devid 2 size 931.51GiB used 419.36GiB path /dev/sdc1

devid 4 size 238.47GiB used 186.47GiB path /dev/sdg1

could i delete the config and then rebuild with all thre drives again to get back the original cache?

1 hour ago, Belcraig said:

devid 1 size 1.82TiB used 1.05TiB path /dev/sdh1

This device is not assigned to the pool but is still being detected as part of it, instead of the assigned device in its place. With the array stopped, type :

echo 1 > /sys/block/sdh/device/delete

Wait a couple of seconds, refresh the GUI and start the array, then post new diags.

New diags

server-diagnostics-20250903-1057.zip

The syslog is being spammed with Nvidia related errors. It didn't log the last start. Remove the Nvidia driver, reboot, and don't start the array. You will need to drop the other device again. Check to see if it's still sdh, if it changed, adjust the command as needed, then post new diags after array start.

server-diagnostics-20250903-1350.zip

was still sdh used the same command as above.

What happened to the other pool device from before?

Sep 1 20:04:53 Server emhttpd: import 30 pool device: (sdc) CT1000MX500SSD1_2103E4E942E6

Sep 1 20:04:53 Server emhttpd: import 31 pool device: (sdf) CT250MX500SSD1_1846E1D7D8AD

Sep 1 20:04:53 Server emhttpd: import 32 pool device: (sdg) PLEXTOR_PX-256M5Pro_P02346100170

sdf was not assigned this time; sdh was the 2TB device:

Sep 3 13:44:16 Server emhttpd: online: CT2000MX500SSD1_2244E67F7926 (sdh) 512 3907029168

its down as sdf, now sitting as an unassigned disk.

And you want that device to be part of the pool now?

i originally added it to the pool so that when removing (what was at the time) what i thought was a faulty drive i (hopefully didnt lose data), so data in the pool could be spread.

as the faultly drive has now passed an extended smart test, i think it must have been a seating issue/calbe issue.

So i probably want the original cache type back which constisted ofthe 3 cruical disks, the plexor was a spare in the system not being used for anyhting.

You should be able to now to assign a device to replace the missing one first, then replace any other ones if desired, always one at a time, and don't try to remove a pool device before updating to 7.1.4.

i thought the drive had failed that's why i tried to add the plexor drive to the pool, for the pool to rebuild before i lost data and then remove the failed drive but something went wrong. once i've got everything back ill update Unraid to the latest version.

OK, please ask before proceeding if you have any doubts or issues.

Hi Jorge,

Thanks yes the question i have is as the previous 2TB drive issue was probably cabling (having been stable today and passing the extended smart tests, is there a way to rebuild the original cache pool without loing data. adding the drive back now one at a time prompt a data will be erased scenario.