Failed to set up mount namespacing: /run/systemd/unit-root/home: No such device

[richard]

Since the system updates applied the morning of 2 february 2023, I’m noticing a waterfall of startup problems, seemingly caused by NetworkManager (which, itself, restarts) but flakes out postgresql and uwsgi … postgresql needing manual restart which is a PITA.

What I notice in journalctl log file:

févr. 02 07:11:09 systemd[1]: Starting Network Manager...
févr. 02 07:12:39 systemd[1]: NetworkManager.service: start operation timed out. Terminating.
févr. 02 07:12:39 systemd[1]: NetworkManager.service: Failed with result 'timeout'.
févr. 02 07:12:39 systemd[1]: Failed to start Network Manager.
févr. 02 07:12:40 systemd[1]: NetworkManager.service: Scheduled restart job, restart counter is at 1.
févr. 02 07:12:40 systemd[1]: Stopped Network Manager.
févr. 02 07:12:40 systemd[1]: Starting Network Manager...
févr. 02 07:12:50 systemd[622]: NetworkManager.service: Failed to set up mount namespacing: /run/systemd/unit-root/home: No such device
févr. 02 07:12:50 systemd[622]: NetworkManager.service: Failed at step NAMESPACE spawning /usr/bin/NetworkManager: No such device
févr. 02 07:12:50 systemd[1]: NetworkManager.service: Main process exited, code=exited, status=226/NAMESPACE
févr. 02 07:12:50 systemd[1]: NetworkManager.service: Failed with result 'exit-code'.
févr. 02 07:12:50 systemd[1]: Failed to start Network Manager.
févr. 02 07:12:50 systemd[1]: NetworkManager.service: Scheduled restart job, restart counter is at 2.

currently at:

$ uname -a
Linux ... 6.1.9-arch1-2 #1 SMP PREEMPT_DYNAMIC Fri, 03 Feb 2023 18:49:53 +0000 x86_64 GNU/Linux
$ paclog-pkglist |grep -E "systemd|networkm"
lib32-systemd 252.5-1
networkmanager 1.40.12-1
networkmanager-openvpn 1.10.2-1
networkmanager-qt 5.102.0-1
systemd 252.5-1
systemd-libs 252.5-1
systemd-sysvcompat 252.5-1

Any ideas? for info, I do not use systemd-homed, which is disabled.
BTW - I don’t notice the supposed directory /run/systemd/unit-root/home, only the parent (which is empty)

$ sudo ls -lar /run/systemd/unit-root
total 0
drwxr-xr-x 21 root root 540  7 févr. 09:20 ..
drwx------  2 root root  40  7 févr. 09:18 .

[joekamprad]

sudo /usr/bin/NetworkManager --no-daemon

could show better messages about what exactly is failing…

But never had such issue or saw this reported …

[richard]

This doesn’t show anything at all, but as I mentioned above, probably because NetworkManager restarts automatically a bit later.
The turd is that postgresql doesn’t restart by itself, it needs a manual restart!

févr. 07 09:20:02 systemd[611]: postgresql.service: Failed to set up mount namespacing: /run/systemd/unit-root/home: No such device
févr. 07 09:20:02 systemd[611]: postgresql.service: Failed at step NAMESPACE spawning /usr/bin/postgresql-check-db-dir: No such device
févr. 07 09:20:02 systemd[1]: postgresql.service: Control process exited, code=exited, status=226/NAMESPACE
févr. 07 09:20:02 systemd[1]: postgresql.service: Failed with result 'exit-code'.
févr. 07 09:20:02 systemd[1]: Failed to start PostgreSQL database server.

[richard]

Well, it seems it was related to the nfs mount for “/home” in /etc/fstab
I restructured to use a bind mount from the root nfs mountpoint and put
the following clause in nfsmounts.conf

# Would cause all mount to /export/home would be done in
# the background
[ MountPoint "/export/home" ]
background=True

Booting up to the graphical login screen takes a few moments longer, but it seems stable enough [at least for now].