I've been having consistent problems with the UnRaid GUI crashing. When this happens, I can't ssh in, either; I can only log in directly on the local console. The only way I've been able to recover is to power cycle the machine, which I'm not a big fan of.

Many places in the forums say to run the diagnostics from the console, but typing diagnostics from the local console results in diagnostics not found

What version of Unraid do you have.

35 minutes ago, ericbr said:

typing diagnostics from the local console results in diagnostics not found

This should definitely work unless your version is many years old or there is something wrong with your installation.

Do you have a current flash backup?

root@unSERVER:~# which diagnostics
/usr/local/sbin/diagnostics
root@unSERVER:~# 

So diagnostics command is definitely there unless you somehow did something to the OS files.

Check if diagnostics works with the array stopped.

So - when the Unraid GUI is running, diagnostics works (and I've attached the output). I did note down the path, and when the GUI stops working again, I'll follow up with the results.

zimacube-diagnostics-20260511-1604.zip

So when the Unraid GUI is unresponsive, the entire /usr/local directory is "missing". /usr/local/sbin doesn't exist, so (obviously) I can't run diagnostics.

The reboot command "works", but doesn't shut down any VMs/docker containers, nor does it actually reboot anything.

Based on some other threads, I have decided to run memtest86+, and so far (21 hours in) the 10 test passes have all passed. I also set up remote syslogging, but nothing has shown up in the logs.

The only thing that seems to have changed is that I set up a VM to run Proxmox Backup Server around the time that the crashes started to happen. Before setting up the VM, RAM usage was in the low GBs (~2); with the VM, RAM usage is around 8 GB. The system has 16 GB, so that seems like plenty of headroom.

18 hours ago, ericbr said:

when the Unraid GUI is unresponsive, the entire /usr/local directory is "missing". /usr/local/sbin doesn't exist

That suggests something is writing into the OS space.

When that happens, what do you get from command line with this?

df -h /

Actually, I think it might be because of some problems with the USB boot drive. On the most recent "crash", Unraid couldn't even start the web UI; the logs were showing read failures on sda1 (the boot drive); I moved it to Windows and repaired the disk, and it seems to have fixed the initial issue. I turned off the VM as well, and maybe that will resolve the issue.

Interestingly enough, I was able to cause the problem again w/o using a VM. I was using rclone to back up another server, and after backing up ~320 GB, the web UI crashed, and, after logging in on the console, /usr/local/sbin has exactly1 file:

lwrxrwxrwx 1 root root 17 Jun 22 09:35 unraid-api -> ../bin/unraid-api*

df shows that /var/log is at 100%, and the array at 65%. (Sorry, can't copy & paste from the console.)

Any ideas what might be triggering this? A bad spot on the disk array?

See if you can get the syslog cp /var/log/syslog /boot/syslog.txt

I had remote system logging set up, so I was able to find the remote logs. I didn't attach all of them, but these were the ones that had been most recently updated.

Clearly, at least part of the problem is that /var/log had filled up with ssh logs, but it's weird that the file system filling up would cause problems with the boot flash drive.

This seems like rclone with the sftp backend is, in fact, not the backup solution I was looking for, and maybe I should try rsync, since I'm already using rsync to back up my synology NAS.

Alternatively, is there a way to greatly reduce the amount of ssh logging generated by rclone/sftp?

I still have this suspicion that the problem really triggers when some large amount of data is sent, though.

zimacube.zip

That log is over 1GB, I'm sorry, but not even trying to open that. Get just the syslog or start a new persistent log syslog server just for the last boot, also try to stop all the extra spam

I edited down the logs until just before the failure. It looks like something goes wrong with the USB connection to the flash drive, at which point everything goes to hell.

kernel.zip

2026-06-27T17:42:28-07:00 ZimaCube kernel: usb 2-1.3: USB disconnect, device number 3

2026-06-27T17:42:28-07:00 ZimaCube kernel: device offline error, dev sda, sector 11716424 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2

2026-06-27T17:42:28-07:00 ZimaCube kernel: I/O error, dev loop0, sector 115904 op 0x0:(READ) flags 0x800 phys_seg 4 prio class 2

2026-06-27T17:42:28-07:00 ZimaCube kernel: SQUASHFS error: Failed to read block 0x38936dd: -5

2026-06-27T17:42:28-07:00 ZimaCube kernel: device offline error, dev sda, sector 11697752 op 0x0:(READ) flags 0x80700 phys_seg 10 prio class 2

2026-06-27T17:42:28-07:00 ZimaCube kernel: device offline error, dev sda, sector 11697752 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2

2026-06-27T17:42:28-07:00 ZimaCube kernel: device offline error, dev sda, sector 11715032 op 0x0:(READ) flags 0x80700 phys_seg 12 prio class 2

2026-06-27T17:42:28-07:00 ZimaCube kernel: device offline error, dev sda, sector 11715032 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2

2026-06-27T17:42:28-07:00 ZimaCube kernel: device offline error, dev sda, sector 12924448 op 0x0:(READ) flags 0x80700 phys_seg 10 prio class 2

Flash drive is dropping offline, you can first try using a different USB port and/or recreating the flash drive, if the issue persists, replace it.