Interrupt signal during update process

z0id · 11 December 2021 00:24

Dear Forum,

during the recent update via terminal it happened that I accidentally typed Strg+C which led to the message “Interrupt signal received” and interrupted the update.

The try to continue the update led to the following message (in German):

sudo: Fehler in /etc/sudo.conf, Zeile 0, während Plugin »sudoers_policy« geladen wurde
sudo: »/usr/lib/sudo/sudoers.so« konnte nicht geladen werden: libldap-2.4.so.2: Kann die Shared-Object-Datei nicht öffnen: Datei oder Verzeichnis nicht gefunden
sudo: Schwerwiegender Fehler, Plugins konnten nicht geladen werden

Loosely translated:

sudo: error in /etc/sudo.conf, line 0, during plugin »sudoers_policy« was loaded
sudo:  »/usr/lib/sudo/sudoers.so« could not be loaded: libldap-2.4.so.2: cannot open Shared-Object File: File or Directory not found
sudo: serious error, plugins could not be loaded

I am honestly a little bit afraid to reboot or shutdown the system because of that fearing that it does not boot anymore. The update interruption happened during the update of kernel 5.4.

(224/330) Aktualisiert wird linux510-headers                                                                                                   [########################################################################################] 100%
(225/330) Aktualisiert wird linux510-virtualbox-host-modules                                                                                   [########################################################################################] 100%
In order to use the new version, reload all virtualbox modules manually.
(226/330) Aktualisiert wird linux54                                                                                                            [########################################################################################] 100%
^C
Interrupt signal received
(227/330) Aktualisiert wird linux54-headers

So my question is if I have to fix something (and if yes: how) before rebooting.

Thanks in advance for any help.

Regards,
z0id

Sierrafox · 11 December 2021 01:55

Found the info below hopefully it helps.

https://askubuntu.com/questions/637432/sudo-comes-up-with-an-error-cannot-run-anything-as-root

z0id · 11 December 2021 08:01

Thanks for the answer.

I tried to follow these steps, but there it is still not working. Maybe now there is some other ownership/permission issue.

What I did so far (as described in the link):

I opened a terminal and did

pkexec su
chown -R username /usr/lib/
chown -R root /usr/lib/

Now, I get the message:
sudo: /usr/lib/sudo/sudoers.so must be only be writable by owner

Following this:
https://askubuntu.com/questions/234603/how-do-i-deal-with-sudoers-so-must-be-only-be-writable-by-owner

I did
ls -l /usr/lib/sudo/sudoers.so
what showed
-rw-r--r-- 1 root root ...

Then, I did

chmod go-w /usr/lib/sudo/sudoers.so

Now, I get the first error message again:

»/usr/lib/sudo/sudoers.so« could not be loaded: libldap-2.4.so.2: cannot open Shared-Object File: File or Directory not found

Any further ideas?

z0id · 11 December 2021 08:26

Trying to open the application anki via terminal resulted in:

Traceback (most recent call last):
  File "/usr/bin/anki", line 3, in <module>
    import aqt
  File "/usr/lib/python3.9/site-packages/aqt/__init__.py", line 22, in <module>
    from aqt.qt import *
  File "/usr/lib/python3.9/site-packages/aqt/qt.py", line 13, in <module>
    from PyQt5.QtCore import *
ImportError: libicui18n.so.69: cannot open shared object file: No such file or directory

So it seems to be an issue with accessing shared libraries. (I have not rebooted the system since the (interrupted) update process, because of the reasons mentioned above.)

Wollie · 11 December 2021 08:40

https://wiki.manjaro.org/index.php/Pacman_troubleshooting

Nachlese · 11 December 2021 09:27

I’d run the update again, and let it finish.

Since sudo now doesn’t appear to work
you have to use su to get permission to do that.
… and you did:

pkexec su

What you then did was … blindly trying to fix something that wasn’t broken
and break even more in the process

but
you reverted that.

ls -l /usr/lib/sudo/sudoers.so
should indeed show:
-rw-r--r-- 1 root root ...

If that is the current state
just run the update, using su to get to be root:

pkexec su
pacman -Syu

and let it finish

or perhaps report what error messages you get when doing that.

If you run the command like this:

LANG=C pacman -Syu

the output will be in english - no need to translate …

z0id · 11 December 2021 11:59

ls -l /usr/lib/sudo/sudoers.so

showed

-rw-r–r-- 1 root root

pkexec su

resulted now in

Error executing command as another user: Not authorized

This incident has been reported.

Nachlese · 11 December 2021 12:12

should (!) result in a GUI confirmation dialogue where you give your root password

This is the response you get when you put the wrong password three times.
…

In a terminal, I’d have done just:
su -

I just copy/pasted that

pkexec su

in my reply
because it was already here and seemed to work.
I have never really used that - only always just su -

You could also try to login to a TTY as root directly.

z0id · 11 December 2021 12:48

pkexec su

worked in the beginning, but not anymore. Anyway,

su -

worked. The installation showed for most of the files:

Warnung: Verzeichnis-Berechtigungen unterscheiden sich für /usr/lib/systemd/system/
Dateisystem: 757  Paket: 755

So saying directory permissions differeing. File system: 757 Package: 755

Nachlese · 11 December 2021 13:25

That’s just a warning, an information, if you will.
Not sure whether it needs intervention - doesn’t look like it.
permissions are now a bit more restrictive than they where
I have seen these kind of messages and so far always acknowledged and then ignored them.

z0id · 11 December 2021 13:52

Maybe because of updating as root (?).

I noticed that during the post-transaction-hooks there was no update-grub. Doing it as root resulted in the following message:

Generating grub configuration file …
…
Warning: os-prober will be executed to detect other bootable partitions.
Its output will be used to detect bootable binaries on them and create new boot entries.
Adding boot menu entry for UEFI Firmware Settings …
Found memtest86+ image: /boot/memtest86+/memtest.bin
/usr/bin/grub-probe: warning: unknown device type nvme0n1.
done

So, no kernel was mentioned (?).

Nachlese · 11 December 2021 14:09

No.
Updates have to be done as root.
(sudo gives you these rights temporarily for one command
su makes you root - and any command you run after that is run as root)

It’s because the permissions of the directories and files in the package
(which just got installed in place of the previous one)
are different now - namely (your example): the “write” permission for “other”
was present
but is not anymore now

As you can see in any output of
ls -l /some/directory
there will be three rwx sections, the first for “user”, the second for “group”, the third for “other” - the last one got changed

I’m not qualified to comment.

It would help others to have the full output that you saw.
It’s probably available from the tail end of
/var/log/pacman.log

I do not think there is anything wrong here though.

Does your sudo now work again?

z0id · 11 December 2021 14:26

Yes, it does.

Do you mean from the update process?

There still seems to be a permission/ownership issue concerning /usr/lib. I just tried to update the package Oracle VM VirtualBox Extension Pack from AUR using paru and got the following error message:

[2021-12-11T15:08:28+0100] [ALPM] running ‘30-systemd-update.hook’…
[2021-12-11T15:08:49+0100] [PACMAN] Running ‘pacman --upgrade --noconfirm – /home/username/.cache/paru/clone/virtualbox-ext-oracle/virtualbox-ext-oracle-6.1.30-1-any.pkg.tar.zst’
[2021-12-11T15:08:49+0100] [ALPM] transaction started
[2021-12-11T15:08:50+0100] [ALPM-SCRIPTLET] 0%…
[2021-12-11T15:08:50+0100] [ALPM-SCRIPTLET] Progress state: NS_ERROR_FAILURE
[2021-12-11T15:08:50+0100] [ALPM-SCRIPTLET] VBoxManage: error: Failed to uninstall “Oracle VM VirtualBox Extension Pack”
[2021-12-11T15:08:50+0100] [ALPM-SCRIPTLET] VBoxManage: error: The installer failed with exit code 1: VBoxExtPackHelperApp: error: World writable: '/usr/lib’
[2021-12-11T15:08:50+0100] [ALPM-SCRIPTLET] VBoxManage: error: Details: code NS_ERROR_FAILURE (0x80004005), component ExtPackManagerWrap, interface IExtPackManager
[2021-12-11T15:08:50+0100] [ALPM-SCRIPTLET] VBoxManage: error: Context: “RTEXITCODE handleExtPack(HandlerArg*)” at line 1465 of file VBoxManageMisc.cpp
[2021-12-11T15:08:50+0100] [ALPM] upgraded virtualbox-ext-oracle (6.1.28-1 → 6.1.30-1)
[2021-12-11T15:08:50+0100] [ALPM-SCRIPTLET] 0%…
[2021-12-11T15:08:50+0100] [ALPM-SCRIPTLET] Progress state: NS_ERROR_FAILURE
[2021-12-11T15:08:50+0100] [ALPM-SCRIPTLET] VBoxManage: error: Failed to install “/usr/share/virtualbox/extensions/Oracle_VM_VirtualBox_Extension_Pack-6.1.30.vbox-extpack”
[2021-12-11T15:08:50+0100] [ALPM-SCRIPTLET] VBoxManage: error: Extension pack ‘Oracle VM VirtualBox Extension Pack’ is already installed. In case of a reinstallation, please uninstall it first
[2021-12-11T15:08:50+0100] [ALPM-SCRIPTLET] VBoxManage: error: Details: code NS_ERROR_FAILURE (0x80004005), component ExtPackManagerWrap, interface IExtPackManager
[2021-12-11T15:08:50+0100] [ALPM-SCRIPTLET] VBoxManage: error: Context: “RTEXITCODE handleExtPack(HandlerArg*)” at line 1424 of file VBoxManageMisc.cpp
[2021-12-11T15:08:50+0100] [ALPM] transaction completed
[2021-12-11T15:08:50+0100] [ALPM] running ‘30-systemd-update.hook’…

Nachlese · 11 December 2021 14:33

Setting correct permissions for /var directory recursively - #2 by megavolt

This is a very recent pertinent post from user @megavolt

perhaps in the course of your attempts … you managed to set the wrong permissions on that directory?

anyway - here is the description of the cause for the error - and the link gives the tools and shows how you can fix it

You where concerned about something regarding the grub or kernel installation - I tried to point to the file where the relevant messages will still be.
If you still have the same terminal open, you could just scroll up …

edit:
or, you could simply just reinstall that kernel and review the output again …
(probably via mhwd-kernel - but don’t ask me for specifics
I only run Manjaro in a VM and am not very familiar with the Manjaro tools)

z0id · 11 December 2021 15:35

I managed to reinstall the kernels using pacman and then the post-transaction-hooks updated grub and detected the kernels.

Thank you very much for this link. I followed megavolt’s steps applied for /usr.

Thank you so much so far.

So do you think rebooting could be the next step?

Nachlese · 11 December 2021 15:40

When the whole update has finished successfully:
yes.

You could add a safeguard before you do, if you are concerned and haven’t done so until now:
add at least one other kernel to be able to choose from for the unlikely case that the updated one will not work.
… have at least two kernels installed …

z0id · 11 December 2021 19:39

Thanks for the answer.

That seems to be the case.

I do have.

So I will try the reboot.

Nachlese · 11 December 2021 19:44

… can’t postpone it forever
(Man kann es ja nicht ewig verschieben )

z0id · 11 December 2021 19:55

Finally, the reboot went fine. Everything seems to work properly, incl. sudo. After resetting the permissions in /usr/lib the update of virtualbox-ext-oracle finished without any error.

@Nachlese: Thank you so much. You solved my problem and saved my day.

Nachlese · 11 December 2021 19:56

You are very welcome!
anytime - if I can, I will …