Here is the backstory. I have been running this server for a year+ + now and DID have similar issues with this unmounting/hanging of the "unioned" filesystems mounted below /mnt/user/. The "user" mount point dies after N minutes. So back then, I uninstalled the SAS sleep plugin and let the SAS drives never sleep. it fixed it until last week. I ran the 7.2 RC builds with NO problem, then when I installed the 7.2 release, it started happening again. Here is what I have done to troubleshoot

Run in Safe Mode. Still hung.

Rebuilt my USB boot drive with serial and NO plugins and NO Docker. Runs fine. Although the UI for drive parity checking, due to multiple reboots, does not update as it used to. I have waited until the check is complete and Docker is enabled, and it still hangs.

I ordered a NEW SAS card from the list of supported ones in the forums, but there's still no real change.

If I keep Docker disabled, it works, but I need Docker. I WAS using a Docker file with plenty of free space and have moved it between drives to check that.

I have checked the SMART data on the drives and see a few read and write errors, but they didn't look too bad.

Not sure where to go from here. I thought that a total rebuild of the OS work had been done. But it has not.

Ideas? Anything you can see in the diag zip? dmesg and syslog didnt show me anything that i can recognize.

unraid-diagnostics-20251103-0649.zip

Edited November 4, 2025Nov 4 by dpackham

10 hours ago, dpackham said:

The "user" mount point dies after N minutes.

Are the diags after this happens?

I cant get the diagnostics after it happens, as the disk process hangs trying to get "df" info, etc, due to the missing /mnt/user mounts. i can try too setup a remote syslog server if that will help?

Edited November 4, 2025Nov 4 by dpackham

39 minutes ago, dpackham said:

i can try too setup a remote syslog server if that will help?

Yep, do that.

OK, did that and rebooted. And attached are the syslogs that were sent to my Synology backup

Some more info is that I can use the unraid shell and cd /mnt/disk1 and disk2 and disk3 just fine and run something like du -s etc., all works.

When I try to cd to /mnt/user, it hangs.

All_2025-11-4-9_23_23.csv

Does this log cover that happening? And if yes, do you know a date and time?

it should. i downloaded it from the synology after i could not CD to /mnt/user anymore. but not sure time when it went away.

Edited November 4, 2025Nov 4 by dpackham

BTW to reboot it via SSH i need to rename the /usr/local/sbin/diagnostic link to that it wont run as it hangs indefinitly and never finishes and therefore needs to ne kill -9 'd

OK it ran for a bit this time and i was able to install some dockers. then it hung again. around 10:45 MST. ish as I dont really know when it failed

All_2025-11-4-10_55_48.csv

Seems to be a SHFS issue. is there a way to turn on logging?

reading this thread. sounds exactly like whats happening to me

There's no additional logging for SHFS AFAIK. If possible, I would recommend using the server with disk shares or exclusive shares, to bypass FUSE and see if that makes any difference.

The/mnt/disk1 2 and 3 shares all work fine after/mnt/user has failed. my appdata is on FUSE at /mnt/user/appdata

You can move appdata to a single pool or disk, if it's not already, then use the disk path e.g., /mnt/cache/appdata or /mnt/disk1/appdata

I can do that to get going again, but that's no fun :). Anything else I can look at?

Would I have to move all my shares over to single disks as well?

Edited November 4, 2025Nov 4 by dpackham

Without anything relevant logged, I'm afraid that I don't have more ideas/suggestions.

33 minutes ago, dpackham said:

reading this thread. sounds exactly like whats happening to me

If this is the problem, it would leave the shfs crash logged.

OK right.   I'll keep working on getting something logged

Just now, JorgeB said:

it would leave the shfs crash logged.

like:

May 14 14:06:42 Tower shfs: shfs: ../lib/fuse.c:1451: unlink_node: Assertion `node->nlookup > 1' failed

I don't see that in the log posted, post the persistent syslog, ideally right after the issue starts.

The persistent syslog?  Do I need to enable syslog mirror to flash for that?

No, that's optional; the important part is to enable it to save to a share. See the link for the instructions.