W: Couldn't identify type of root file system '/dev/pve/rootu/dev/pve/root' for fsck hook

[jacobrale]

I get text in the subject as a warning in /var/log/term.log when I run my PVE updates. I gather it's not a problem, because everything works fine. Also, there is no '/dev/pve/rootu/dev/pve/root'. Thoughts? I don't want to ignore a warning that might actually be tried to tell me something significant!

 

[Onslow]

Welcome, @jacobrale .

I would search for this string. E.g.:

grep -r rootu /etc
and maybe in and below /boot as well:
grep -r rootu /boot

 

[jacobrale]

Thanks! fstab has: "/etc/fstab:/dev/pve/rootu/dev/pve/root / ext4 errors=remount-ro 0 1" that is it. What dev is "dev/pve/rootu/dev/pve/root"? There is nothing there? Is this some sort of zfs thing I don't get? Does not seem right. When I run "zfs list -o name,mounted,mountpoint" i get:

NAME MOUNTED MOUNTPOINT
big18Tzfs yes /big18Tzfs
big18Tzfs/vm-100-disk-0 - -
big18Tzfs/vm-100-disk-1 - -
big18Tzfs/vm-102-disk-0 - -
big18Tzfs/vm-106-disk-0 - -

 

[Impact]

It's a LVM volume path. It's not called rootu by default though.

 

[Onslow]

Thanks! fstab has: "/etc/fstab:/dev/pve/rootu/dev/pve/root / ext4 errors=remount-ro 0 1" that is it. What dev is "dev/pve/rootu/dev/pve/root"? There is nothing there?

Most likely this is some typo or an erroneous pasting during editing fstab.

As this path doesn't exist, just comment out this line in fstab.
Then run an update as usual and verify that the warnings are gone .

 

[jacobrale]

Thanks Onslow! You're are undoubtedly right! Interestingly, I hadn't noticed before, but in fstab there is a NOTE above this line (the "/etc/fstab:/dev/pve/rootu/dev/pve/root / ext4 errors=remount-ro 0 1" line) that says "# Below added in recovery system and can probably be removed.". Presumably I put it there when I moved to 9? I had some real difficulty going to 9. Anyway, I could take it out, but I wanted to figure out how and where root (/) was actually being mounted. mount shows it as: "/dev/mapper/pve-root on / type ext4 (rw,relatime,errors=remount-ro)". Frankly I started using linux a quarter century ago, and to me, a mount is a logical (/home) to a "physical" (/dev/sda2). I am barely able to keep up when we start talking about zfs subvolumes or LVM (even UUIDs were a jump for me.) Anyway, I finally puttered around enough with LVM (lvscan etc.) and lsblk to convince myself I can remove it. I'm cautious! You can fix pretty much anythhing, but my proxmox is headless and not physically very accessiable. Commenting the line out in fstab! Thanks again.

 

[Impact]

I stripped the results down a bit but this should help to make sense of it

# ls -l /dev/pve/root /dev/mapper/pve-root
lrwxrwxrwx 1 root root 7 Jul 20 13:27 /dev/mapper/pve-root -> ../dm-0
lrwxrwxrwx 1 root root 7 Jul 20 13:27 /dev/pve/root -> ../dm-0

# dmsetup ls -oblkdevname
pve-root        (dm-0)

# lsblk -o+PATH,PKNAME,KNAME
NAME                 MAJ:MIN RM   SIZE RO TYPE MOUNTPOINTS PATH                       PKNAME KNAME
sda                    8:0    0    64G  0 disk             /dev/sda                          sda
└─sda3                 8:3    0  63.5G  0 part             /dev/sda3                  sda    sda3
  ├─pve-root         252:0    0  29.9G  0 lvm  /           /dev/mapper/pve-root       sda3   dm-0

See here for more.

 

[jacobrale]

Thanks!

 

[Onslow]

I am barely able to keep up when we start talking about zfs subvolumes or LVM

@jacobrale
I must admit I'm not good at ZFS either. But LVM can be easily imagined in relation to ordinary disks and partitions.

In the simplest form: first you create a PV from a disk or from a disk partition.
Then you build a VG of one or more PVs.
In a VG you create one or more LVs.
In this imaginary analogy: LV is like a disk partition.
On a LV you create a file system (e.g. ext4) like on a disk partition.

Here you are an image from Wikipedia:
https://upload.wikimedia.org/wikipedia/commons/thumb/e/e6/Lvm.svg/960px-Lvm.svg.png
Various elements of the LVM

 

[jacobrale]

Very nice! Thanks. Maybe now I'll remember better!

 

[jacobrale]

Hey, thanks for all the advice! Update: Yesterday I commented out the line ("/dev/pve/rootu/dev/pve/root / ext4 errors=remount-ro 0 1") in fstab.

So:
------
# <file system> <mount point> <type> <options> <dump> <pass>
UUID=2E4E-A2B4 /boot/efi vfat defaults 0 1
# Below added in recovery system and can probably be removed.
#/dev/pve/rootu/dev/pve/root / ext4 errors=remount-ro 0 1
/dev/pve/swap none swap sw 0 0
proc /proc proc defaults 0 0
------

Today, there were updates, all went well. Well, almost! One of the updates was a new kernel, so there was a suggestion to reboot.

Rebooted and no webui. SSH'ed in and root (/) was read-only. Anyway, poked around to see what it might be, but finally just put the line back in fstab and rebooted and back in business. Clearly, with that "rootu" thing, that line must be something I did months ago when I had problems with the move to V 9.

I'd like to understand what is going on, but not the most urgent thing in the world, as ProxMox comes up when I leave the line ("/dev/pve/rootu/dev/pve/root / ext4 errors=remount-ro 0 1") in. But when I take the line out, because the filesystem gets mounted ro, there are no logs to review. (At least, I don't see new logging.)

The thing is, that line in fstab ("/dev/pve/rootu/dev/pve/root / ext4 errors=remount-ro 0 1") doesn't even make any sense to me. There is **NO** /dev/pve/rootu/, at least not when fully booted; maybe during some transitional state. I tried to change the fstab line to the ultimate mount point which seems to be "/dev/mapper/pve-root" so, I changed the line to "/dev/mapper/pve-root on / ext4 errors=remount-ro 0 1", but that too resulted in a ro (readonly) root (/)?

When booted (wheather ro or rw) and I list mouint units, I get:
-------
root@pve:/dev/pve# systemctl list-units "*.mount"
UNIT LOAD ACTIVE SUB DESCRIPTION
-.mount loaded active mounted Root Mount
big18Tzfs.mount loaded active mounted /big18Tzfs
boot-efi.mount loaded active mounted /boot/efi
dev-hugepages.mount loaded active mounted Huge Pages File System
dev-mqueue.mount loaded active mounted POSIX Message Queue File System
proc-sys-fs-binfmt_misc.mount loaded active mounted Arbitrary Executable File Formats File System
run-lock.mount loaded active mounted Legacy Locks Directory /run/lock
run-rpc_pipefs.mount loaded active mounted RPC Pipe File System
run-user-1000.mount loaded active mounted /run/user/1000
sys-fs-fuse-connections.mount loaded active mounted FUSE Control File System
sys-kernel-config.mount loaded active mounted Kernel Configuration File System
sys-kernel-debug.mount loaded active mounted Kernel Debug File System
sys-kernel-tracing.mount loaded active mounted Kernel Trace File System
tmp.mount loaded active mounted Temporary Directory /tmp
var-lib-lxcfs.mount loaded active mounted /var/lib/lxcfs
-------

Confusingly to me, "find . -name "*.mount*"" (as/at: root@pve:/etc/systemd/system# ) shows no mount config files. I think on ubuntu (as opposed to debian, on which proxmox is based) that I found the config files for systemd mounts (both user and system, I think) in "/etc/systemd/system". There is nothing there that looks like a root mount to me.

So I don't really understand how root is even being mounted.

After booting *with* the line ("/dev/pve/rootu/dev/pve/root / ext4 errors=remount-ro 0 1") I found this refernce to rootu in syslog. Nothing else with root or rootu.
2026-07-20T00:15:00.371462-04:00 pve fstrim[711728]: /: 19.7 GiB (21146124288 bytes) trimmed on /dev/pve/rootu/dev/pve/root

I just don't get it. How is root being mounted? Again, not urgent, but I don't like not understanding what is being mounted how.