Hi everyone,

I'm running into a critical issue with my Unraid server. The array starts up fine, but after running normally for a while, the user share system seems to crash.

When this happens, the "Shares" tab in the WebGUI goes completely blank, and all of my network shares become inaccessible.

If I try to check the directory via the terminal, I get the following FUSE error:

root@Tower:/mnt# ls
ls: cannot access 'user': Transport endpoint is not connected

Symptoms:

• The array remains started, but user shares drop out unexpectedly after some uptime.

• Terminal confirms /mnt/user is inaccessible with the "Transport endpoint is not connected" error.

• The "Shares" page in the WebGUI is completely empty.

tower-diagnostics-20260316-2339.zip

Your have corruption on disk1, but you also have btrfs csum errors on /dev/loop2 (docker.img)

btrfs csum errors are usually bad RAM. Very important you take care of that before trying to repair disk1.

memtest is on the boot menu

3 hours ago, harrymay said:

"Transport endpoint is not connected" error.

From what I recall this mean FUSE has crashed. Likely caused by whatever else is going bad.

I would consider a 24 hour minimum memory test which you can access from the unraid boot menu screen (blue square screen)

You must not attempt to run any computer unless memory is working perfectly. Everything goes through RAM. The OS and other executable code, your DATA. EVERYTHING! The CPU can't do anything with anything until it is loaded into RAM.

Thank you so much for pointing me in the right direction

I ran Memtest86+ from the boot menu as suggested. It didn't even need 24 hours—within just 3 minutes, the screen lit up with red errors.

1 hour ago, harrymay said:

Thank you so much for pointing me in the right direction

I ran Memtest86+ from the boot menu as suggested. It didn't even need 24 hours—within just 3 minutes, the screen lit up with red errors.

Yup that's no good, time to replace the RAM.

you can test each stick individually if you really want to only replace the bad DIMMs but that's your call.

I went ahead and let Memtest86+ run for a full 24 hours with both sticks installed just to see the extent of the damage, and it threw over 1.2 million errors

After that, I tested the sticks individually to isolate the issue. It turns out both sticks have physically failed and threw massive errors when tested alone

1 hour ago, harrymay said:

see the extent of the damage

Even one error is too many

On 3/16/2026 at 4:23 PM, trurl said:

You must not attempt to run any computer unless memory is working perfectly.

The new 32GB RAM kit arrived and is installed.

I ran Memtest86+ and it successfully completed a full pass with absolutely zero errors. Seeing that big green "PASS" was a huge relief (screenshot attached).

Now that the hardware foundation is finally stable and trustworthy, I have booted up Unraid and grabbed a fresh diagnostics file (attached) before attempting any repairs.

tower-diagnostics-20260325-2144.zip

20 minutes ago, harrymay said:

Memtest86+ and it successfully completed a full pass

On 3/16/2026 at 3:53 PM, MowMdown said:

24 hour minimum memory test

Then...

All 3 disks are mounted, disk3 is empty which I assume is expected.

On 3/16/2026 at 12:27 PM, trurl said:

corruption on disk1

Might be a good idea to check filesystem on disk1 anyway.

On 3/16/2026 at 12:27 PM, trurl said:

btrfs csum errors on /dev/loop2 (docker.img)

docker.img still showing the corruption it already had. Delete, recreate, reinstall Previous Apps.

https://docs.unraid.net/unraid-os/troubleshooting/common-issues/docker-troubleshooting/

Might also be a good idea to do a parity check.

Do you intend to have cache or other pool outside the parity array? It can have some benefits with docker even if you don't use it for caching.