"w" no longer shows on which tty users are logged in

On current Tumbleweed w no longer shows where users are logged in:

bruno@LT-B:~> w
 12:56:41 up 9 min,  1 user,  load average: 0,34, 0,37, 0,22
USER     TTY      FROM             LOGIN@   IDLE   JCPU   PCPU  WHAT
bruno             -                12:47    9:35   0.02s  0.02s /usr/libexec/gnome-session-init-worker gnome
bruno@LT-B:~>

This happens on both Gnome and KDE; on Leap 16.0 the relevant TTY is still shown (for instance tty3 for the first KDE user, tty2 for the first Gnome user.
Anybody pointing me to where the change occurred?

(BTW, trying to diagnose a problem where switching to VT1 makes it impossible to switch back to any other VT, but the system looks otherwise working correctly.)

And

knurpht@Lenovo-P16:~> w
 13:38:05 up  6:17,  1 user,  load average: 0,37, 0,43, 0,67
USER     TTY      FROM             LOGIN@   IDLE   JCPU   PCPU  WHAT
knurpht           -                07:21    6:16m  0.17s  0.17s /usr/bin/startplasma-wayland
knurpht@Lenovo-P16:~> who
knurpht  seat0        2026-07-21 07:21
knurpht  tty2         2026-07-21 07:21
knurpht@Lenovo-P16:~>

Use w -t then there is loginctl and last…

1 Like

Thanks, but the problem is not how to find the relevant tty (after all there are only 6 or 7 to check) but if there was any recent major change in the way VT1 (or VTx in general) is managed.
Wondering if anything related might be at the roots of unstable-tumbleweed-last-7-days.

OK, what’s in the title is the trivial result of a change in procps v. 4.0.6:

  • w: Add terminal mode to show all terminal sessions issue #375

apparently in v. 4.0.5 (Leap 16.0) “terminal mode” was implied.

So the main question remains, any relevant recent change in the way VT1 is managed?

Not really sure, I see no issues here with GNOME and Hyprland (login via tty1), but I don’t run multiple DE’s, as in one system one environment.

Maybe Wayland changes?

Wayland 1.26.0 in TW 20260717… definitely something to look at!

Something interesting in Sometimes the graphical session doesn’t come up, stays at tty1 since booting with plymouth.enable=0 seems to fix things (but I need confirming with more reboots…)

Some time ago, I ran into an issue with “uptime”, but “w” still worked. I reported it as a bug.

As I recall, this was only a problem if using GDM, but everything was fine with SDDM. The explanation given was that the data base “utmp” was being deprecated. I think this was because of concerns about the 32 bit times used, and the replacement was using 64-bit times.

I’m guessing that what you seeing is part of a transition to the new database.

No idea about recent, but IME, VT1 has always been quirky, latent or otherwise, not exhibited by other VTs. Thus, the only time I use VT1 for more than glancing at boot messages is when booted to single. 2-6, if not more, when not tied to a DM and/or GUI session, all behave without surprises.

I have no idea if this was already earlier, but at least since my upgrade Leap 15.6 > 16.0 (and the switch to Wayland). my main graphical screen is at VT1

henk@boven:~> w
 18:31:47 up  9:43,  1 user,  load average: 0.20, 0.38, 0.26
USER     TTY      FROM             LOGIN@   IDLE   JCPU   PCPU  WHAT
henk     tty1     -                08:49    9:43m  0.27s  0.27s /usr/bin/startp
henk@boven:~> who
henk     seat0        Jul 21 08:49
henk     tty1         Jul 21 08:49
henk@boven:~>

Ten reboots without plymouth and VTs switch correctly, including VT1.
Two reboots with plymouth and two stuck on VT1 once switched to it.
This is enough for me to wait for a fixed plymouth without further digging; given that plymouth usually “lives” on VT1 makes it the perfect culprit.

Thanks all for hints and insight anyway!

Maybe that I should have added that I do not have Plymouth in the system.

@OrsoBruno:

Bruno, the RPM Change Log is a good place to inspect → there was a change committed on Friday January 30th 2026 –

 > which w
/usr/bin/w
 > rpm --query --whatprovides /usr/bin/w
procps-4.0.6-1.3.x86_64
 >

> rpm --query --changelog procps | less

  • OK … switch to coreutils-systemd and coreutils-systemd-doc
  • Conflict not only with systemd-coreutils but also with systemd-coreutils-doc
  • Conflict procps-uptime also with busybox-coreutils
  • Add upstream commit as d9c96ec0.patch
    • Fix Issues#413: watch in procps-ng 4.0.6 no longer uses full
      pane height in tmux
  • Now we have a .asc file for procps-ng-4.0.6 as well
  • Update to procps-ng-4.0.6
    • library
      version: inc revision to 1 now 1:1:0
      internal: Use openat for files under /proc/
      internal: Don’t check for sd_booted Debian #1108549
      internal: Address potential race leading to segfault issues #380, #390
    • local: guard SIGPOLL for MacOS merge !246
      .
      .
      .
    • w: Add terminal mode to show all terminal sessions issue #375
    • w: Use process TTY as backup for user TTY Debian #1080335
    • w: Use correct return value for sd_get_sessions Debian #1068904
    • w: Don’t check for sd_booted Debian #1108549
    • w: Don’t crash with pids in terminal mode issue #407
      .
      .
  • No asc file of the source tar ball
  • Remove patch procps-ng-4.0.5-bsc1246330.patch no upstream
  • Add patch glibc-2.43.patch to support glibc 2.43
    .
    .
    .

 > LANG=C.UTF-8 w
 16:45:37 up  7:49,  1 user,  load average: 0.17, 0.11, 0.12
USER     TTY      FROM             LOGIN@   IDLE   JCPU   PCPU  WHAT
xxx               -                10:20    7:49m  0.42s  0.42s /usr/bin/startplasma-wayland
 >
 > LANG=C.UTF-8 w -o
 16:45:55 up  7:49,  1 user,  load average: 0.12, 0.10, 0.12
USER     TTY      FROM             LOGIN@   IDLE   JCPU   PCPU  WHAT
xxx               -                10:20    7:49                /usr/bin/startplasma-wayland
 > 
 > LANG=C.UTF-8 w -t
 16:46:02 up  7:49,  1 user,  load average: 0.11, 0.10, 0.12
USER     TTY      FROM             LOGIN@   IDLE   JCPU   PCPU  WHAT
xxx      tty1     -                10:20    7:49m  0.42s  0.42s /usr/bin/startplasma-wayland
root     tty2     -                08:56    7:49m  2.95s  2.95s /usr/bin/Xorg.bin -nolisten tcp -background none -seat
xxx      pts/0    -                16:30    1.00s  0.07s  0.06s w -t
xxx      pts/1    -                16:35    2:02   0.02s  0.02s less
 > 
 > LANG=C.UTF-8 who
xxx      seat0        2026-07-22 10:20
xxx      tty1         2026-07-22 10:20
 > 
1 Like