number of lines on console
by Greg Woods
I've spent about an hour searching without finding an answer. Is there any
way from the kernel command line (i.e. by editing the GRUB entry) to
control the number of lines on the console? I installed Fedora 39 on a USB
stick so that I can use it to boot other computers to diagnose problems
with the installed OS. The problem is that when it is booting, the latest
console messages are off the bottom of the screen so that I can't see what
it's doing (this is on a standard 1920x1080 HD monitor in this case). If
the boot hangs, then of course the last few messages are the crucial ones
to see.
Once the machine is booted, of course, then I can log in and use tools such
as "stty" to set the number of rows and cure this problem, but I need
something that will take effect during system startup before the login
prompt comes up.
Thank you,
--Greg
2 weeks, 6 days
boot in read only
by Patrick Dupre
Hello,
I am trying to understand why I boot in a read-only mode
(I have to do mount -o rw,remount /)
I found in/var/log/message
gsd-usb-protect[1905]: Failed to fetch USBGuard parameters: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name is not activatable
systemd[1240]: Starting xdg-document-portal.service - flatpak document portal service...
Jun 6 13:08:06 Teucidide gsd-media-keys[1907]: Failed to grab accelerator for keybinding settings:playback-random
Jun 6 13:08:06 Teucidide gsd-media-keys[1907]: Failed to grab accelerator for keybinding settings:screensaver
Jun 6 13:08:06 Teucidide gsd-media-keys[1907]: Failed to grab accelerator for keybinding settings:hibernate
Jun 6 13:08:06 Teucidide gsd-media-keys[1907]: Failed to grab accelerator for keybinding settings:playback-repeat
Jun 6 13:08:06 Teucidide gsd-media-keys[1907]: Failed to grab accelerator for keybinding settings:rotate-video-lock
Dependency failed for xdg-desktop-portal-gnome.service - Portal service (GNOME implementation).
Jun 6 13:08:06 Teucidide systemd[1240]: xdg-desktop-portal-gnome.service: Job xdg-desktop-portal-gnome.service/start failed with result 'dependency'.
Jun 6 13:08:06 Teucidide systemd[1240]: graphical-session.target - Current graphical user session is inactive.
gnome-terminal-[2264]: Failed to load module "pk-gtk-module"
wireplumber[2291]: spa.bluez5: BlueZ system service is not available
Jun 6 13:08:09 Teucidide gsd-media-keys[1907]: gvc_mixer_card_get_index: assertion 'GVC_IS_MIXER_CARD (card)' failed
Jun 6 13:08:09 Teucidide gsd-media-keys[1907]: gvc_mixer_card_get_index: assertion 'GVC_IS_MIXER_CARD (card)' failed
Jun 6 13:08:09 Teucidide gsd-media-keys[1907]: Unable to get default sink
system_r:systemd_logind_t:s0 tcontext=system_u:object_r:unlabeled_t:s0 tclass=dir permissive=0
Jun 6 13:09:42 Teucidide systemd-logind[994]: Failed to open parent directory of "/boot": Permission denied
Jun 6 13:09:42 Teucidide kernel: audit: type=1400 audit(1717672182.777:304): avc: denied { getattr } for pid=994 comm="systemd-logind" path="/boot" dev="sdc4" ino=2 scontext=system_u:system_r:systemd_logind_t:s0 tcontext=system_u:object_r:unlabeled_t:s0 tclass=dir permissive=0
===========================================================================
Patrick DUPRÉ | | email: pdupre(a)gmx.com
===========================================================================
2 weeks, 6 days
fsck
by Patrick Dupre
Hello,
file -Ls /dev/sdc5
/dev/sdc5: Linux rev 1.0 ext4 filesystem data, UUID=1ea2d462-244f-4959-9a62-99a7288aa38c, volume name "root_sec" (extents) (64bit) (large files) (huge files)
but
fsck /dev/sdc5
fsck from util-linux 2.38.1
If you wish to check the consistency of a BTRFS filesystem or
repair a damaged filesystem, see btrfs(8) subcommand 'check'.
Why that?
===========================================================================
Patrick DUPRÉ | | email: pdupre(a)gmx.com
===========================================================================
2 weeks, 6 days
Re: How to remove settings from a systemd unit file
by R. G. Newbury
On 2024-06-05 00:06, users-request(a)lists.fedoraproject.org wrote:
> From: Jeffrey Walton<noloader(a)gmail.com>
> On Tue, Jun 4, 2024 at 7:24 PM Sam Varshavchik<mrsam(a)courier-mta.com> wrote:
>> So I was tearing my hair out trying to figure out why attempts to push via
>> DAV to a git repo were failing.
....
>> However this apparently did not work. I threw in the towel and just edited
>> /lib/systemd/system/httpd.service and commented this setting out, entirely,
>> to finally fix this issue, and happy git pushing resumed.
>>
>> But how do I fix this so that the next apache update doesn't clobber this?
> I think a better choice is to leave the systemd unit files alone. Then
> you don't have to worry about your changes getting reverted on updates
> and system upgrades.
>
> I also think it is better to avoid serving files from your home
> directory. Instead, use /var. Install your Git-managed project in
> /var/git (and your Subversion projects in /var/svn). Add a git user,
> and make ownership of /var/git as root:git. Finally, change the
> server's document root to/var/git/<project>.
>
> This setup works well for me. The only problem I have encountered is
> Git's fix for CVE-2022-24765 a/k/a safe directories. Safe directories
> caused a big DoS at my site. Also see
> <https://github.com/git/git/commit/8959555cee7e>.
>
> Jeff
I agree with your first point about not *serving* files from /home. I
disagree with your 'fix' in using /var.
Notwithstanding the anaconda installer's indications and instructions
NOT TO, the installer always seems to WIPE the /var partition.
Accordingly, you should use a symlink from /var/git (or in my case(s)
/var/lib/mysql and /var/www/html) to somewhere else which will remain
stable across a re-install.
In my case I point these instances to an actual "physical" partition
which is /misc, which allows me to easily back-up all the active
'stuff'. (It's really meta to describe a 'partition' on an SSD as
'physical'... but there we are.)
Yes, it's a little more work on an actual re-install but I have some
'run-once' script lines in /etc/rc.d/rc.local which handles all of the
stuff I would not remember how to do, much less remember to do...
And Jeff, you should remember that it is spelt:
Geoff
3 weeks
start services
by Patrick Dupre
Hello,
Is there a way to start an installed fedora in rescue mode
and then start the services one by one as fedora would have done
in normal boot?
Before the rescue mode was obtained by setting single-user
Thanks.
===========================================================================
Patrick DUPRÉ | | email: pdupre(a)gmx.com
===========================================================================
3 weeks
grub2-mkconfig
by Patrick Dupre
Hello,
I am still stuck with my installation.
Question
Why
grub2-mkconfig
says
Found Fedora Linux 36 (Workstation Edition) on /dev/sdb11
Found Fedora Linux 40 (Forty) on /dev/sdc3
Found Fedora Linux 36 (Workstation Edition) on /dev/mapper/VolSys_2-root
Adding boot menu entry for UEFI Firmware Settings ...
done
but in the generated file (grub.cfg) there is no reference to Fedora 40
or sdc3
In addition there are a lot of
references to
Linux 36 (Workstation Edition) on /dev/sdb11
?
Thanks
===========================================================================
Patrick DUPRÉ | | email: pdupre(a)gmx.com
===========================================================================
3 weeks
How to remove settings from a systemd unit file
by Sam Varshavchik
So I was tearing my hair out trying to figure out why attempts to push via
DAV to a git repo were failing.
Eventually I succeeded in stracing the httpd process sto capture the
request. It was getting an EROFS when it tried to write to the git repo.
Amusing.
To make a long story short, the culprit was:
ProtectHome=read-only
in /lib/systemd/system/httpd.service,(the git repo was in a directory inside
a mounted /home partition).
I tried using
systemctl edit httpd
And putting this in there:
[Service]
ProtectHome=
However this apparently did not work. I threw in the towel and just edited
/lib/systemd/system/httpd.service and commented this setting out, entirely,
to finally fix this issue, and happy git pushing resumed.
But how do I fix this so that the next apache update doesn't clobber this?
3 weeks
[Rawhide Test Week] DNF 5.2 2024-06-05 through 2024-06-11
by Sumantro Mukherjee
Hey Folks,
Many of you by now know that in Fedora, we try to open up more
opportunities to test upcoming features which are currently being
developed. This time DNF team is working closely with Fedora QA to run
the DNF 5.2 test week[0].
As a part of this test week, we will be specifically testing the
system-upgrade feature. Previously, DNF would leverage DNF plug-ins to
do system upgrade. This is now baked into DNF 5.2 and this test week
will focus on testing that feature.
Needless to say, we will also be testing the regular DNF features. You
can always go ahead and submit the results here[1].
[0] http://fedoraproject.org/wiki/Test_Day:2024-06-05_Fedora_41_DNF_5.2
[1] https://testdays.fedoraproject.org/events/189
--
//sumantro
Fedora QE
TRIED AND PERSONALLY TESTED, ERGO TRUSTED
3 weeks
Re: strange results of history command list
by AV
On Tue, 2024-06-04 at 16:26 +0000, Joe Wulf wrote:
It truly depends on how the default management of bash history is set
up in the virgin environment.
Variables like HISTCONTROL, HISTIGNORE can play a part. The setting
for 'shopt -s histappend' also matters.
The BASH man page briefly explains them.
R,
-Joe
/etc/profile only contains the, as far as I know, usual:
if [ "$HISTCONTROL" = "ignorespace" ] ; then
export HISTCONTROL=ignoreboth
else
export HISTCONTROL=ignoredups
fi
That does not seem to explain anything
AV
Sorry for replying directly. I forgot to check send to list
3 weeks, 1 day
