Found left-over process plymouth on boot
Hi *, because of the recent problems with plymouth I checked my journal for plymouth messages more thoroughly and found: Nov 30 14:29:51 client systemd[1]: plymouth-switch-root.service: Deactivated successfully. Nov 30 14:29:52 client systemd[1]: plymouth-start.service: Found left- over process 463 (plymouthd) in control group while starting unit. Ignoring. Nov 30 14:29:52 client systemd[1]: plymouth-start.service: This usually indicates unclean termination of a previous run, or service implementation deficiencies. Is that an expected behavior? It seems, that the plymouth process from the early boot phase remained running - correct? TIA. Bye. Michael.
30.11.2024 16:55, mh@mike.franken.de wrote:
Hi *,
because of the recent problems with plymouth I checked my journal for plymouth messages more thoroughly and found:
Nov 30 14:29:51 client systemd[1]: plymouth-switch-root.service: Deactivated successfully. Nov 30 14:29:52 client systemd[1]: plymouth-start.service: Found left- over process 463 (plymouthd) in control group while starting unit. Ignoring. Nov 30 14:29:52 client systemd[1]: plymouth-start.service: This usually indicates unclean termination of a previous run, or service implementation deficiencies.
Is that an expected behavior?
SUSE sets KillMode=none for plymouth-start.service which means any process will not be terminated by systemd when (if) unit stops. So yes, it is sort of expected. OTOH I do not see it on any of my VM. Nov 30 17:00:09 tw systemd[1]: Starting Show Plymouth Boot Screen... Nov 30 17:00:10 tw systemd[1]: Started Show Plymouth Boot Screen. Nov 30 17:00:42 tw systemd[1]: Starting Terminate Plymouth Boot Screen... Nov 30 17:00:47 tw systemd[1]: Finished Terminate Plymouth Boot Screen. It is started in initrd and actually never stopped by systemd. GDM terminates plymouthd on its own. The "Terminate Plymouth Boot Screen" is plymouth-quit.service which in this case runs because I booted into run level 3. In your case the message is likely due to "systemctl daemon-reload".
It seems, that the plymouth process from the early boot phase remained running - correct?
This is actually the very idea of plymouth - it is supposed to show splash screen from the earliest possible moment without interruption.
On Samstag, 30. November 2024 15:19:28 Mitteleuropäische Normalzeit Andrei Borzenkov wrote:
30.11.2024 16:55, mh@mike.franken.de wrote:
Hi *,
because of the recent problems with plymouth I checked my journal for plymouth messages more thoroughly and found:
Nov 30 14:29:51 client systemd[1]: plymouth-switch-root.service: Deactivated successfully. Nov 30 14:29:52 client systemd[1]: plymouth-start.service: Found left- over process 463 (plymouthd) in control group while starting unit. Ignoring. Nov 30 14:29:52 client systemd[1]: plymouth-start.service: This usually indicates unclean termination of a previous run, or service implementation deficiencies.
Is that an expected behavior?
SUSE sets KillMode=none for plymouth-start.service which means any process will not be terminated by systemd when (if) unit stops. So yes, it is sort of expected.
OK. There is also a journal entry telling me, that KillMode=none is deprecated, so this should be changed sometimes in the future, I guess.
OTOH I do not see it on any of my VM.
Nov 30 17:00:09 tw systemd[1]: Starting Show Plymouth Boot Screen... Nov 30 17:00:10 tw systemd[1]: Started Show Plymouth Boot Screen. Nov 30 17:00:42 tw systemd[1]: Starting Terminate Plymouth Boot Screen... Nov 30 17:00:47 tw systemd[1]: Finished Terminate Plymouth Boot Screen.
It is started in initrd and actually never stopped by systemd. GDM terminates plymouthd on its own. The "Terminate Plymouth Boot Screen" is plymouth-quit.service which in this case runs because I booted into run level 3.
In your case the message is likely due to "systemctl daemon-reload".
Jepp.
It seems, that the plymouth process from the early boot phase remained running - correct?
This is actually the very idea of plymouth - it is supposed to show splash screen from the earliest possible moment without interruption.
Yes, that was clear to me, but not why systemd complains about a still running process. But your explanation helped a lot, thx! Bye. Michael.
On 11/30/24 8:35 AM, mh@mike.franken.de wrote:
SUSE sets KillMode=none for plymouth-start.service which means any process will not be terminated by systemd when (if) unit stops. So yes, it is sort of expected. OK. There is also a journal entry telling me, that KillMode=none is deprecated, so this should be changed sometimes in the future, I guess.
I have had problems with plymouth remaining running on my laptop causing hand on shutdown. I have removed plymouth completely (beginning in either Leap 42.3 or 15.0) and I have never had another issue. If you run into plymouth problems, that is always an option. Between a few seconds of fancy boot or having no problems on shutdown, I chose the latter. -- David C. Rankin, J.D.,P.E.
On Samstag, 30. November 2024 18:52:18 Mitteleuropäische Normalzeit David C. Rankin wrote:
On 11/30/24 8:35 AM, mh@mike.franken.de wrote:
SUSE sets KillMode=none for plymouth-start.service which means any process will not be terminated by systemd when (if) unit stops. So yes, it is sort of expected.
OK. There is also a journal entry telling me, that KillMode=none is deprecated, so this should be changed sometimes in the future, I guess. I have had problems with plymouth remaining running on my laptop causing hand on shutdown. I have removed plymouth completely (beginning in either Leap 42.3 or 15.0) and I have never had another issue.
If you run into plymouth problems, that is always an option. Between a few seconds of fancy boot or having no problems on shutdown, I chose the latter.
I prefer to get a graphical passphrase prompt for my encrypted partitions instead of a text prompt, that vanishes as soon as the next text messages show up. Bye. Michael.
On 2024-11-30 22:15, mh@mike.franken.de wrote:
On Samstag, 30. November 2024 18:52:18 Mitteleuropäische Normalzeit David C. Rankin wrote:
On 11/30/24 8:35 AM, mh@mike.franken.de wrote:
SUSE sets KillMode=none for plymouth-start.service which means any process will not be terminated by systemd when (if) unit stops. So yes, it is sort of expected.
OK. There is also a journal entry telling me, that KillMode=none is deprecated, so this should be changed sometimes in the future, I guess. I have had problems with plymouth remaining running on my laptop causing hand on shutdown. I have removed plymouth completely (beginning in either Leap 42.3 or 15.0) and I have never had another issue.
If you run into plymouth problems, that is always an option. Between a few seconds of fancy boot or having no problems on shutdown, I chose the latter.
I prefer to get a graphical passphrase prompt for my encrypted partitions instead of a text prompt, that vanishes as soon as the next text messages show up.
The prompt is still there and the keyboard remains active, but it is true that the text of the prompt flows up with the texts messages that continue appearing; it can even vanish and one doesn't know why the boot process has stopped. One has to tentatively press a key on the keyboard, and if an asterisk appears on the screen, one knows it is waiting for the passphrase. AND, if you have two or more partitions with the same passphrase, plymouth applies it automatically to them, saving typing. Supposedly. -- Cheers / Saludos, Carlos E. R. (from 15.5 x86_64 at Telcontar)
On Saturday 30 November 2024 08:19:28 Andrei Borzenkov wrote:
It is started in initrd and actually never stopped by systemd. GDM terminates plymouthd on its own. The "Terminate Plymouth Boot Screen" is plymouth-quit.service which in this case runs because I booted into run level 3.
GDM is the Gnome Display Manager, is it not? What if one is not running Gnome? Leslie -- Platform: Linux Distribution: openSUSE Leap 15.5 - x86_64
participants (5)
-
Andrei Borzenkov -
Carlos E. R. -
David C. Rankin -
J Leslie Turriff -
mh＠mike.franken.de