However, I found the 5.10 kernel already installed in the latest manjaro iso.
Perhaps the iso should only contain a stable LTS kernel
The reason for this is that the kernel 5.4 (last LTS) unfortunately causes problems with AMD CPUs.
That’s why we opted for kernel 5.10 - despite the shortcomings.
After the update my Xorg crashes. I found the problem to be the libepoxy package (upgrading from version 1.5.5 to 1.5.7 to be exact). Anyone else experiencing such issue? (KDE, AMDGPU)
Recently, after that update checking for updates (including the AUR) stopped working properly.
It couldn’t query AUR anymore and as a result some of my databases were corrupted. Attempts to update the databases sudo pacman -Syu resulted in:
error: GPGME error: No data
error: failed to update core (invalid or corrupted database (PGP signature))
After several things I tried, the solution suggested in the archived manjaro forum helped to solve the problem:
I used to type in the password only once, but after the last update (the previous Stable one), I have to type it twice: (i) once pre-boot, likely for "/", and (ii) another after booting, likely for "/home". (Yes, you probably guessed right: both partitions have a certain little thing in common). Whether this is bad practice or not, is not the point, but:
Is it somehow possible to revert to the previous behavior of having to type password only once? If so, how? (Command-level explanation, please).
That says /crypto_keyfile.bin should be used to unlock both those cryptdevices.
Does that unlocking of the /home cryptdevice still work with that keyfile?
If this needs more investigations/fiddling, open a separate thread to not pollute this announcement any further.
Initially I ran the usual upgrade commands in the text console:
Ctrl-Alt-F2
pacman -Syyu
shutdown -r now
This got my system in a state where the display manager would crash and the following messages would be printed in the logs:
systemd-modules-load[531]: Failed to insert module 'nvidia': Exec format error
kernel: nvidia: disagrees about version of symbol module_layout
systemd-modules-load[531]: Failed to insert module 'nvidia_drm': Exec format error
kernel: nvidia: disagrees about version of symbol module_layout
systemd[1]: Failed to start Load Kernel Modules.
systemd[1]: lightdm.service: Main process exited, code=exited, status=1/FAILURE
systemd[1]: lightdm.service: Failed with result 'exit-code'.
systemd[1]: Failed to start Light Display Manager.
After a quick search through the forums, in the “[Testing Update] 2021-04-26”, I found that Nvidia driver got downgraded to v460 from v465, and to accommodate that downgrade I needed to run these commands:
For all those who keep asking about the kernel, LTS, etc. - please have a look at the following link.
It shows what is currently changing in the kernel.
following some forum post but result still the same.
If I select y y n it will do this instead:
sudo pacman -Syu 1.6m Min 16 Mei 2021 09:32:26
[sudo] password for ifrit:
:: Synchronizing package databases...
core is up to date
extra is up to date
community is up to date
multilib is up to date
:: Starting full system upgrade...
:: Replace python-dbus with extra/dbus-python? [Y/n]
:: Replace python-dbus-common with extra/dbus-python? [Y/n]
resolving dependencies...
looking for conflicting packages...
:: dbus-python and python2-dbus are in conflict. Remove python2-dbus? [y/N]
error: unresolvable package conflicts detected
error: failed to prepare transaction (conflicting dependencies)
:: dbus-python and python2-dbus are in conflict
My system did not notify me about this upgrade.
I noticed malfunctions when I tried to download software using pamac.
Then Discover warned me about having to install a new kernel and language packages.
I concluded I had to refresh mirrors and upgrade via pacman (using the command line)
That’s fine, but I’m not a mind reader…I need the system to tell me when there is an upgrade pending.
Am I experiencing this bug, which a little research led me to:
sudo mhwd-kernel -i linux510 1.5m Sen 17 Mei 2021 06:10:38
:: Synchronizing package databases...
core 165,5 KiB 101 KiB/s 00:02 [##############################] 100%
extra 1968,1 KiB 113 KiB/s 00:17 [##############################] 100%
community 6,6 MiB 156 KiB/s 00:43 [##############################] 100%
multilib 180,5 KiB 151 KiB/s 00:01 [##############################] 100%
The following packages are out of date, please update your system first: adwaita-icon-
... the list goes on
Do you want to continue anyway? [y/N] y
resolving dependencies...
looking for conflicting packages...
error: failed to prepare transaction (could not satisfy dependencies)
:: installing nvidia-utils (460.73.01-1) breaks dependency 'nvidia-utils=460.56' required by linux54-nvidia
:: installing nvidia-utils (460.73.01-1) breaks dependency 'nvidia-utils=460.56' required by linux59-nvidia
I also tried with manjaro settings manager.
It said “New kernel packages are ready to install. Would you like to continue?”
pick yes, then looks like it’s success (quick process)
but nothing changed.
EDIT:
Oh wait sorry, maybe ill use the other command. sudo mhwd-kernel -i linux510 rmc
hmm no cant use that too
resolving dependencies...
looking for conflicting packages...
error: failed to prepare transaction (could not satisfy dependencies)
:: installing nvidia-utils (460.73.01-1) breaks dependency 'nvidia-utils=460.56' required by linux54-nvidia
:: installing nvidia-utils (460.73.01-1) breaks dependency 'nvidia-utils=460.56' required by linux59-nvidia
Error: 'rmc' aborted because the kernel failed to install or canceled on removal.
Then I tried to remove the LTS kernel which says “Recomended” then try sudo mhwd-kernel -i linux510 rmc but still not working.
Should I also manually remove the one I’m currently using? it sounds dangerous. I want to avoid reinstalling my system.