(solved) Testing: can no longer login with kernel 5.10/4.19

[Member sybok]

Hi, I have antiX testing (at home) and experienced troubles after a recent update.

Suspected cause:
Update, which included ‘dbus’, see the file attached.
Let me note that ‘libelogind0’ is installed.

Issue:
I could not log in after reboot/shutdown, the login screen (‘slimski’) accepted no keyboard input (I could not switch to no-X TTY using Ctrl+Alt+F*)
This occurs with the most recent 5.10 kernel(s) (both 5.10.137 and 5.10.142, i.e. linux-image-5.10.142-antix.2-amd64-smp and 5.10.137-antix.1-amd64-smp-1).
I did not see any errors reported during the boot.
I powered the PC down (by pressing the button at the PC-box) and booted to older kernel, namely 4.19.
The problem does not occur with 4.19.0-256, hence I removed the 5.10 completely and I intend to stick to the older kernel.

Info:
‘inxi’ output is the same as in the following post https://www.antixforum.com/forums/topic/mouse-freeze-rox-icewm/ apart from memory upgrade (increase from 4GB to 8GB).

Comment(s):
I know that @anticapitalista mentioned some issues with kernel 5.10, see e.g. https://www.antixforum.com/forums/topic/testing-5-10-kernel/
It is possible that with recent switch to libelogind0-free approach, my testing system has become a franken-antiX and a reinstall would help.

Did anybody experience the same issue or any hints on how to avoid it (apart from installing the not nosystemd ‘dbus’ which I thought could cause troubles but the installation went fine)?
E.g. use additional apt-pinning (is such a thing employed in the upcoming antiX-22 to prevent libelogind0 from being installed?).

• This topic was modified 7 months, 1 week ago by sybok. Reason: Another kernel added to the title
• This topic was modified 6 months, 2 weeks ago by christophe. Reason: Solved

Attachments:

1. APT_HISTORY_LOG.txt

[Forum Admin anticapitalista]

Helpful

Up

0

::

Was everything working ok with the the latest 5.10 kernel before this upgrade?

What services are running? Make sure dbus is running.

My testing and sid installs (frugal and installed to hard drive) show no issues.

Philosophers have interpreted the world in many ways; the point is to change it.

antiX with runit - leaner and meaner.

[Member sybok]

Helpful

Up

0

::

Things worked well with 5.10 before the last update.
VLC had issues (e.g. did not always respond, I had to hover with mouse over the bottom panel so that space = pause/play and other keyboard shortcuts were properly registered).
Will provide the other information later (when at home).

[Member stevesr0]

Helpful

Up

0

::

UPDATE:
No problem after reboot. Of course, my system is different, since I am running Sid not testing and using startx and Openbox. I don’t know your hardware -mine is an oldish Fujitsu Lifebook from 2006.

Happy to compare any packages you think might be causing problem with 5.10 kernels.

———————————————————————————–

I am running Sid with 5.10.142 antiX kernel. Last full-upgrade today. No problems that I am aware of. I will report back after rebooting. N.B. I am using startx not slimski.

stevesr0

• This reply was modified 7 months, 1 week ago by stevesr0.

[Member sybok]

Helpful

Up

0

::

For some reason, the problem occurred with the 4.19 kernel as well today morning after a reboot from Windows 10 (which adjusted the time +2 hours as seen in logs in antiX).
Power down and clean-boot to 4.9 (I keep a lots of series of kernels) solved it.
Girlfriend reported no such issue yesterday evening (boot from power down using kernel 4.19).

I rebooted (via ‘sudo reboot’) from the 4.9 to 4.19, login does not work again.
I tried to type in login, Enter, password, Enter – nothing.
Power down and clean-boot to 4.19 and it works again.
For some reason, reboot seems to be related the issue.

When I disconnect and reconnect the keyboard, the diodes briefly light up, hence I do not believe power/contact issue is the cause (unlike the mouse issue I reported in the past…).
The problem is hard to debug since I cannot log in.

SERVICES:
Running services (and I did press white-space to have the read-out complete):
service --status-all 2>/dev/null | grep ' \[ + \] '

acpid avahi-daemon bluetooth connman cron cups dbus elogind gpm haveged nfs-common resolvconf rpcbind rsyslog slimski ssh tlp udev

‘dbus’ is listed as one of the running services.envlang received as: en_US.UTF-8

slimski: waiting for X server to begin accepting connections

LOGS in ‘/var/log/’:
‘slimski.log’ contains many lines such as:

envlang received as: en_US.UTF-8
slimski: waiting for X server to begin accepting connections
.

and occasionally
slimski: pam_authentication(): Authentication failure
which is probably due to typo in the password.

‘auth.log’ includes complaints about GNOME keyring but these were present in the past as well:

Oct  3 10:28:12 antixacek slimski: PAM unable to dlopen(pam_gnome_keyring.so): /lib/security/pam_gnome_keyring.so: cannot open shared object file: No such file or directory
Oct  3 10:28:12 antixacek slimski: PAM adding faulty module: pam_gnome_keyring.so
Oct  3 10:28:13 antixacek elogind-daemon[1382]: New seat seat0.
Oct  3 10:28:18 antixacek sshd[1461]: Server listening on 0.0.0.0 port 22.
Oct  3 10:28:18 antixacek sshd[1461]: Server listening on :: port 22.

‘kern.log’ contains

Oct  3 10:40:09 antixacek kernel: [    2.330971] ima: Allocated hash algorithm: sha1
Oct  3 10:40:09 antixacek kernel: [    2.334505] ima: Error Communicating to TPM chip
...
Oct  3 10:40:09 antixacek kernel: [    0.409293] acpi PNP0A08:00: _OSC failed (AE_ALREADY_EXISTS); disabling ASPM
...
Oct  3 10:40:09 antixacek kernel: [    2.496585] fjes: module verification failed: signature and/or required key missing - tainting kernel

The _OSC may not be the cause, see e.g. https://askubuntu.com/questions/1145738/osc-failed-ae-error-disabling-aspm

• This reply was modified 7 months, 1 week ago by sybok. Reason: Reboot info added

[Member ModdIt]

Helpful

Up

0

::

Sounds like where, in exasperation I switched to unstable, that had nothing to
do with antiX directly. Was result of some discussion on timely security fixes
and upstream policy. Testing only gets fixes when they trickle in from sid.
Plus Boot issues plus NVidia driver issues.

did you already try deleting  /home/yourusername/.dbus/session-bus
which fixed a similar weird problem 

OT: I dumped VLC, with a config from subluminal mpv gives way better performance.
VLC is a great option but did not perform well anymore on this 12 year old system.

• This reply was modified 7 months, 1 week ago by ModdIt.

[Member sybok]

Helpful

Up

0

::

Hi @ModdIt, you got my hopes up, I removed all files in ‘~/.dbus/session-bus/’, rebooted but sadly, it did not work.
It seems to me that a clean-boot (no reboot) is the only option that works reliably.

[Member ModdIt]

Helpful

Up

0

::

;-( is the fail same on all desktops ?.
If not done so already please try completely deleting cache in case of corruption,

next updating initramfs and then update grub.
Has fixed boot issues on several occasions.

[Member sybok]

Helpful

Up

0

::

OK, (reinstalled) kernel 5.10 is not to blame – clean-boot (no re-boot) works well.

[Member ModdIt]

Helpful

Up

0

::

Thanks sybok,
reminded me to do an immediate ISO snapshot as backup.
a few weeks ago I was in a similar situation, also, after a lot of
trying to fix reinstalled the root fs leaving home.

[Member sybok]

Helpful

Up

0

::

FYI: The issue got fixed after a series of updates (which included ‘dbus’ among others) and me running ‘update-grub[2]’.

Can someone please mark as solved? Thanks very much!

[Author]

Posts