Updated to 6.12.8 about two weeks ago, came back from a week away and my shares have disappeared. (And most docker containers have either stopped due to errors, or aren't running correctly due to no disk.)

 

Browsing from the Main --> Disk shares, the files all appear to be present

Shares --> all are missing

 

When I run:

 

Quote

root@Tower:/mnt# ls
/bin/ls: cannot access 'user': Transport endpoint is not connected
cache/  disk1/  disk2/  disk3/  disk4/  user/  user0/

 

 

 

i3-12100, Gigabyte Technology Co., Ltd. B660M DS3H AX 16Gb DDR4, LSI SAS2008

 

Found a pair of similar threads, but they are earlier versions. Solving it with a reboot doesn't seem to be a permanent fix, though, and I should probably try to understand why it's happening first. Diagnostics attached, anything else to check/get before I reboot the box?

 

 

 

 

 

 

tower-diagnostics-20240320-0842.zip

Shfs crashed and btrfs is detecting data corruption, would recommend running memtest.

 

P.S. unrelated but change the docker network to ipvlan.

Thanks for the reply JorgeB, I've been running memtest this afternoon and looks like one of the sticks are bad. Swapped in a different pair from the desktop, and running memtest on them to rule out the board or something else.

 

The ipvlan change is on my do-to list, didn't want to change too many things at once with the upgrade - since things looked to be stable before this.

 

 

Ran 4 passes of memtest6.2 on a different pair of 2x16 using the same slots in the motherboard, they passed. Rebooted the server and parity check has completed with no errors found.

 

Anything else I should check/run to catch errors that the bad memory could have caused?

 

 

(As an aside running memtest86+7 with only the suspect dimm in a different PC, will leave it for a while more, but it's done 4 passes without an error. I'll let it go until I need that PC tomorrow - is it likely that it was just a connection issue that was solved by removing/reinstalling in a diff PC?)

 

 

 

12 hours ago, seestray said:

Anything else I should check/run to catch errors that the bad memory could have caused?

Run a correcting scrub on the pool.

Had the same concern randomly today on 7.0.1
 

shfs: error: main, 4168: Transport endpoint is not connected (107): #012can't chdir: /mnt/user Transport endpoint is not connected

The user shares were working and detected. Installed a docker container or two and the main user-root shares disappeared... attempted to reboot the full server to see if it resolves

 

EDIT:
rebooting the server instance resolved the issue.... very strange 

Edited April 22, 20251 yr by bombz

Seeing this today - I can reliably reproduce it by manually running a kometa scan against my media library. Running 6.12.14

Edited May 6, 20251 yr by sgtsquiggs

Please post the diagnostics