Hi Franta.

On Sun, 5 Oct 2025 14:39:30 +0200
Franta Hanzlík <[email protected]> wrote:

> thank You for help and akmods-sop-shutdown !

Thanks.

> I set it, and now after two three new kernel updates everything seems 
> to be fine - xt_geoip.ko is available at iptables startup and iptables 
> starts correctly.
> If I issue the reboot command right after the update is complete, it 
> is ignored:
>
> ... (dnf completing updates)
> [115/115] Removing intel-vpl-gpu-rt-0:25.3.2-1.fc42.x86_64
> Complete!
> root@ns:/etc# reboot
> Operation inhibited by "akmods" (PID 1319773 "systemd-inhibit", user root), 
> reason is "Akmods Transaction running".

Yeah: I discovered recently that reboot, even as root, checks now for
inhibitors if called interactively :-(
See: man systemctl:--check-inhibitors=
     /usr/share/doc/systemd/NEWS: CHANGES WITH 257:
       systemd-logind now always obeys block inhibitor locks, [...]

It was not the case past year.
See: https://bugzilla.redhat.com/show_bug.cgi?id=2011120#c81

I bet that a "reboot < /dev/null &> /dev/null" would thus have
rebooted at once even with the akmods inhibitor.

[...]

> root@ns:/etc# reboot
> 
> Broadcast message from [email protected] on pts/2 (Sun 2025-10-05 13:17:34 
> CEST):
> 
> The system will reboot now!
> ...
> 
> 
> and /var/log/messages contains:
> ...
> Oct  5 13:16:37 ns audit[1319801]: USER_END pid=1319801 uid=0 auid=4294967295 
> ses=4294967295 subj=kernel msg='op=PAM:session_close 
> grantors=pam_keyinit,pam_limits,pam_unix acct="akmods" exe="/usr/bin/runuser" 
> hostname=? addr=? terminal=? res=success'
> Oct  5 13:16:37 ns audit[1319801]: CRED_DISP pid=1319801 uid=0 
> auid=4294967295 ses=4294967295 subj=kernel msg='op=PAM:setcred 
> grantors=pam_rootok acct="akmods" exe="/usr/bin/runuser" hostname=? addr=? 
> terminal=? res=success'
> Oct  5 13:16:48 ns audit[1321403]: SOFTWARE_UPDATE pid=1321403 uid=0 
> auid=4294967295 ses=4294967295 subj=kernel msg='op=install 
> sw="kmod-xtables-addons-6.16.9-200.fc42.x86_64-3.29-1.fc42.x86_64" 
> sw_type=rpm key_enforce=0 gpg_res=0 root_dir="/" comm="dnf" 
> exe="/usr/bin/dnf5" hostname=? addr=? terminal=? res=success'
> Oct  5 13:16:49 ns systemd-inhibit[1319778]: Building and installing 
> xtables-addons-kmod[  OK  ]
> Oct  5 13:16:49 ns systemd[1]: Finished [email protected]_64.service 
> - Builds and install new kmods from akmod for a given kernel.
> Oct  5 13:16:49 ns audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 
> ses=4294967295 subj=kernel msg='[email protected]_64 
> comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? 
> res=success'
> Oct  5 13:17:34 ns systemd-logind[1077]: The system will reboot now!
> Oct  5 13:17:34 ns systemd-logind[1077]: System is rebooting.
> ... (system rebooted)
> Oct  5 13:18:57 ns systemd[1]: Reached target akmods-keygen.target.
> Oct  5 13:18:57 ns systemd[1]: Starting akmods-sop-shutdown.service - Builds 
> and install new kmods from akmod packages...
> Oct  5 13:18:57 ns systemd[1]: Starting akmods.service - Builds and install 
> new kmods from akmod packages...
> ...
> Oct  5 13:18:57 ns systemd[1]: Finished akmods-sop-shutdown.service - Builds 
> and install new kmods from akmod packages.
> Oct  5 13:18:57 ns audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 
> ses=4294967295 subj=kernel msg='unit=akmods-sop-shutdown comm="systemd" 
> exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
> ...
> 
> There is no mention of akmods-sop-shutdown service either on the screen 
> or in the log, it is in the log only after the system starts.

Weird. Have you started it **before** the reboot ?
(like does "systemcl --now enable").

Note that akmods-sop-shutdown.service if of type oneshot +
RemainAfterExit=yes, meaning it will stay active during the whole
boot, then will call akmods only at shutdown.

"Finished akmods-sop-shutdown.service" means in fact waiting for
shutdown to check for kmods.

> Which looks like the kernel module was compiled at the end of the update, 
> in some other way than by akmods-sop-shutdown (which it probably should 
> be).

Yes, by the [email protected]_64.service that sets the
inhibitor.
Ref: /usr/lib/kernel/install.d/95-akmodsposttrans.install (from the akmods RPM).

> That some bug was fixed, so that the kernel modules are now compiled 
> correctly?

Yes, except if a reboot is done without checking for inhibitors.

But this introduced a new problem:

    Bug 2394774 - DNF5 offline update does not reboot a machine because akmods 
holds an inhibition lock
    https://bugzilla.redhat.com/show_bug.cgi?id=2394774

> Anyway, all right now...

Let's hope ... until another problem :-)

-- 
francis
-- 
_______________________________________________
users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
Fedora Code of Conduct: 
https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: 
https://lists.fedoraproject.org/archives/list/[email protected]
Do not reply to spam, report it: 
https://pagure.io/fedora-infrastructure/new_issue

Reply via email to