While trying to diagnose problems as a part of another thread in this list, I discovered that wayland is not functional on my workstation. I'm currently logged in to gnome using the top "GNOME" entry in the menu of desktops that is displayed in the upper right corner of the monitor during login. Results of tests suggested in that other thread:
bash.12[~]: env | grep -i wayland bash.13[~]: ps -ef | grep -i wayland weilian+ 128352 2160 0 11:15 pts/0 00:00:00 grep --color=auto -i wayland bash.14[~]: ps -ef | grep -i earlyoom weilian+ 128358 2160 0 11:15 pts/0 00:00:00 grep --color=auto -i earlyoom bash.15[~]:
Also, I tried logging in to plasma with wayland. After entering my credentials, the screens went black. After 2 minutes, I hit the hard reset and rebooted.
Through 8 years of weekly patches and semi-annual upgrades, I've never done anything to block or disable wayland. Shouldn't it have been automatically installed somewhere in those 8 years?
How do I get wayland properly working on my workstation?
On 20/04/2021 01:23, home user wrote:
While trying to diagnose problems as a part of another thread in this list, I discovered that wayland is not functional on my workstation. I'm currently logged in to gnome using the top "GNOME" entry in the menu of desktops that is displayed in the upper right corner of the monitor during login. Results of tests suggested in that other thread:
bash.12[~]: env | grep -i wayland bash.13[~]: ps -ef | grep -i wayland weilian+ 128352 2160 0 11:15 pts/0 00:00:00 grep --color=auto -i wayland bash.14[~]: ps -ef | grep -i earlyoom weilian+ 128358 2160 0 11:15 pts/0 00:00:00 grep --color=auto -i earlyoom bash.15[~]:
Also, I tried logging in to plasma with wayland. After entering my credentials, the screens went black. After 2 minutes, I hit the hard reset and rebooted.
Through 8 years of weekly patches and semi-annual upgrades, I've never done anything to block or disable wayland. Shouldn't it have been automatically installed somewhere in those 8 years?
How do I get wayland properly working on my workstation?
You have nVidia HW and are using the nVidia drivers packaged by rpmfusion.
To give any opinion about you chances to get wayland working one would need to know your HW and what version of the driver you're running. If your HW is 8 years old, then my guess would be your chances of getting wayland running are slim.
On Thu, 2021-04-22 at 15:14 +0800, Ed Greshko wrote:
On 20/04/2021 01:23, home user wrote:
While trying to diagnose problems as a part of another thread in this list, I discovered that wayland is not functional on my workstation. I'm currently logged in to gnome using the top "GNOME" entry in the menu of desktops that is displayed in the upper right corner of the monitor during login. Results of tests suggested in that other thread:
bash.12[~]: env | grep -i wayland bash.13[~]: ps -ef | grep -i wayland weilian+ 128352 2160 0 11:15 pts/0 00:00:00 grep -- color=auto -i wayland bash.14[~]: ps -ef | grep -i earlyoom weilian+ 128358 2160 0 11:15 pts/0 00:00:00 grep -- color=auto -i earlyoom bash.15[~]:
Also, I tried logging in to plasma with wayland. After entering my credentials, the screens went black. After 2 minutes, I hit the hard reset and rebooted.
Through 8 years of weekly patches and semi-annual upgrades, I've never done anything to block or disable wayland. Shouldn't it have been automatically installed somewhere in those 8 years?
How do I get wayland properly working on my workstation?
You have nVidia HW and are using the nVidia drivers packaged by rpmfusion.
To give any opinion about you chances to get wayland working one would need to know your HW and what version of the driver you're running. If your HW is 8 years old, then my guess would be your chances of getting wayland running are slim.
I don't currently use Wayland as I'm on KDE, but I have briefly tested it in F33 under Gnome. My hardware is of a similar age, i.e. the mobo is at least 8 years old and the GPU is a GTX-1050 from 2017, using the RPMfusion driver. It all worked fine. I sincerely hope it will continue to work under F34 when Wayland is supposed to be supported for KDE.
F33 was a clean install of Fedora Workstation, so I presume that's where Wayland came from. If the OP installed the KDE spin it may not have been installed by default.
poc
On 22/04/2021 19:36, Patrick O'Callaghan wrote:
I don't currently use Wayland as I'm on KDE, but I have briefly tested it in F33 under Gnome. My hardware is of a similar age, i.e. the mobo is at least 8 years old and the GPU is a GTX-1050 from 2017, using the RPMfusion driver. It all worked fine. I sincerely hope it will continue to work under F34 when Wayland is supposed to be supported for KDE.
Well, today there was an update to the kernel. And since I've not done this in a while and since the OP is asking about Wayland and said he got a black screen with KDE I figured I'd give it a try what with being a KDE user.
I only ran it for about 10 minutes as my time was needed elsewhere. I have a GeForce GTX 660 card and I'm running the 465.24.02-2 packaged by rpmfusion which was released a couple of days ago.
So far I only noticed a few hiccups.
1. I have some shortcuts defined and not all window placements were correct. In particular konsole didn't start where I wanted it to and it didn't appear in all the workspaces as it should.
2. Even though the font size in konsole was the same as with Xorg they appear smaller under Wayland.
3. Redshift doesn't work, but isn't expected to. But, neither does KDE's NightLight which is suppose to work on Wayland.
I may try more tomorrow, but it wasn't a disaster for me as it has been in the past.
F33 was a clean install of Fedora Workstation, so I presume that's where Wayland came from. If the OP installed the KDE spin it may not have been installed by default.
If his HW requires the 390xx or 340xx series drivers then he is pretty much out of luck.
--
Remind me to ignore comments which aren't germane to the thread.
On Thu, 2021-04-22 at 20:55 +0800, Ed Greshko wrote:
On 22/04/2021 19:36, Patrick O'Callaghan wrote:
I don't currently use Wayland as I'm on KDE, but I have briefly tested it in F33 under Gnome. My hardware is of a similar age, i.e. the mobo is at least 8 years old and the GPU is a GTX-1050 from 2017, using the RPMfusion driver. It all worked fine. I sincerely hope it will continue to work under F34 when Wayland is supposed to be supported for KDE.
Well, today there was an update to the kernel. And since I've not done this in a while and since the OP is asking about Wayland and said he got a black screen with KDE I figured I'd give it a try what with being a KDE user.
I only ran it for about 10 minutes as my time was needed elsewhere. I have a GeForce GTX 660 card and I'm running the 465.24.02-2 packaged by rpmfusion which was released a couple of days ago.
So far I only noticed a few hiccups.
1. I have some shortcuts defined and not all window placements were correct. In particular konsole didn't start where I wanted it to and it didn't appear in all the workspaces as it should.
2. Even though the font size in konsole was the same as with Xorg they appear smaller under Wayland.
3. Redshift doesn't work, but isn't expected to. But, neither does KDE's NightLight which is suppose to work on Wayland.
I may try more tomorrow, but it wasn't a disaster for me as it has been in the past.
Might be time to test F34 RC-1.1, which has just been announced.
F33 was a clean install of Fedora Workstation, so I presume that's where Wayland came from. If the OP installed the KDE spin it may not have been installed by default.
If his HW requires the 390xx or 340xx series drivers then he is pretty much out of luck.
I guess so.
poc
On 4/22/21 1:14 AM, Ed Greshko wrote:
On 20/04/2021 01:23, home user wrote:
How do I get wayland properly working on my workstation?
You have nVidia HW and are using the nVidia drivers packaged by rpmfusion.
To give any opinion about you chances to get wayland working one would need to know your HW and what version of the driver you're running. If your HW is 8 years old, then my guess would be your chances of getting wayland running are slim.
I'm at Fedora-33. The graphics card is nVidia GeForce GTX 660/PCIe/SSE2.
The driver (from dnf info): Name : kmod-nvidia-5.11.13-200.fc33.x86_64 Epoch : 3 Version : 460.67 Release : 1.fc33 Architecture : x86_64 Size : 43 M Source : nvidia-kmod-460.67-1.fc33.src.rpm Repository : @System From repo : @commandline Summary : nvidia kernel module(s) for 5.11.13-200.fc33.x86_64 URL : http://www.nvidia.com/ License : Redistributable, no modification permitted Description : This package provides the nvidia kernel modules built for the : Linux kernel 5.11.13-200.fc33.x86_64 for the x86_64 family of : processors.
It is my understanding that wayland should have been installed by default when I did an upgrade some time ago, a few releases ago. A year ago, just after upgrading to F-31, I started having problems with xeyes. I submitted this bug: "https://bugzilla.redhat.com/show_bug.cgi?id=1825370". The xeyes problem went away for unknown reasons prior to last November. I also see that "earlyoom" should have been installed by default in some earlier release, but I don't have it. I'm wondering if wayland and earlyoom did get installed and then later removed or de-activated. I'm also wondering what else I should have but am missing. In other words, do I have a broader problem?
I will shortly go off line for weekly patches. I should be back on early this afternoon (US mountain time).
On 4/22/21 9:50 AM, home user wrote:
I'm wondering if wayland and earlyoom did get installed and then later removed or de-activated. I'm also wondering what else I should have but am missing. In other words, do I have a broader problem?
If you're concerned about that, you might consider using distro-sync to make sure you've got everything.
On 4/22/21 11:10 AM, Joe Zeff wrote:
On 4/22/21 9:50 AM, home user wrote:
I'm wondering if wayland and earlyoom did get installed and then later removed or de-activated. I'm also wondering what else I should have but am missing. In other words, do I have a broader problem?
If you're concerned about that, you might consider using distro-sync to make sure you've got everything.
I did that as a part of the upgrade earlier this month. But I try again:
bash.1[~]: su - Password: -bash.1[~]: dnf distro-sync Last metadata expiration check: 0:51:59 ago on Thu 22 Apr 2021 10:35:59 AM MDT. Dependencies resolved. Nothing to do. Complete! -bash.2[~]:
On Thu, 22 Apr 2021 at 04:14, Ed Greshko ed.greshko@greshko.com wrote:
To give any opinion about you chances to get wayland working one would need to know your HW and what version of the driver you're running. If your HW is 8 years old, then my guess would be your chances of getting wayland running are slim.
Wayland runs with nouveau on this system purchased in 2010 and running Fedora 33:
% uname -a Linux dormarth 5.11.7-200.fc33.x86_64 #1 SMP Wed Mar 17 18:55:20 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux % env | grep WAY WAYLAND_DISPLAY=wayland-0 % sudo lshw -C video *-display description: VGA compatible controller product: GT218 [GeForce 310] vendor: NVIDIA Corporation physical id: 0 bus info: pci@0000:01:00.0 version: a2 width: 64 bits clock: 33MHz capabilities: pm msi pciexpress vga_controller bus_master cap_list rom configuration: driver=nouveau latency=0 resources: irq:32 memory:fa000000-faffffff memory:d0000000-dfffffff memory:ce000000-cfffffff ioport:ec00(size=128) memory:c0000-dffff
I have had problems with distros that no longer include modules to support older hardware. With Fedora I was stuck on 5.8 kernels until a simple mistake that broke nouveau on old hardware was corrected upsteam.
On 4/22/21 11:29 AM, home user wrote:
I did that as a part of the upgrade earlier this month. But I try again:
bash.1[~]: su - Password: -bash.1[~]: dnf distro-sync Last metadata expiration check: 0:51:59 ago on Thu 22 Apr 2021 10:35:59 AM MDT. Dependencies resolved. Nothing to do. Complete! -bash.2[~]:
OK, if those packages had been part of the original installation, they'd either be there or be replaced. Since they aren't there, they must be part of something you installed later. Why don't you just install them again and see what happens?
On 22/04/2021 23:50, home user wrote:
On 4/22/21 1:14 AM, Ed Greshko wrote:
On 20/04/2021 01:23, home user wrote:
How do I get wayland properly working on my workstation?
You have nVidia HW and are using the nVidia drivers packaged by rpmfusion.
To give any opinion about you chances to get wayland working one would need to know your HW and what version of the driver you're running. If your HW is 8 years old, then my guess would be your chances of getting wayland running are slim.
I'm at Fedora-33. The graphics card is nVidia GeForce GTX 660/PCIe/SSE2.
That should support Wayland.
You can see the actual version of the nvidia driver with....
strings /lib/modules/`uname -r`/extra/nvidia/nvidia.ko | grep ^version=
What display manager are you using?
On 4/22/21 1:17 PM, Ed Greshko wrote:
That should support Wayland.
You can see the actual version of the nvidia driver with....
strings /lib/modules/`uname -r`/extra/nvidia/nvidia.ko | grep ^version=
-bash.4[~]: strings /lib/modules/`uname -r`/extra/nvidia/nvidia.ko | grep ^version= version=465.24.02 -bash.5[~]:
I patched my system this morning, which is why the version changed since my earlier message.
What display manager are you using?
Is that referring to gnome vs. plasma vs. mate vs. (etc)? If yes, I usually log into the first GNOME entry, which is apparently the same as GNOME on Xorg. The version is 3.38.5. The windowing system is X11. If no, how do I determine that?
On 23/04/2021 04:04, home user wrote:
Is that referring to gnome vs. plasma vs. mate vs. (etc)? If yes, I usually log into the first GNOME entry, which is apparently the same as GNOME on Xorg. The version is 3.38.5. The windowing system is X11.
That is the DE or Desktop Environment
If no, how do I determine that?
ll /etc/systemd/system/display-manager.service
or
systemctl status display-manager
On Thu, 2021-04-22 at 11:29 -0600, home user wrote:
On 4/22/21 11:10 AM, Joe Zeff wrote:
On 4/22/21 9:50 AM, home user wrote:
I'm wondering if wayland and earlyoom did get installed and then later removed or de-activated. I'm also wondering what else I should have but am missing. In other words, do I have a broader problem?
If you're concerned about that, you might consider using distro- sync to make sure you've got everything.
I did that as a part of the upgrade earlier this month. But I try again:
bash.1[~]: su - Password: -bash.1[~]: dnf distro-sync Last metadata expiration check: 0:51:59 ago on Thu 22 Apr 2021 10:35:59 AM MDT. Dependencies resolved. Nothing to do. Complete!
Was your original installation the default Workstation or KDE Spin?
poc
On 4/22/21 3:18 PM, Ed Greshko wrote:
On 23/04/2021 04:04, home user wrote:
Is that referring to gnome vs. plasma vs. mate vs. (etc)? If yes, I usually log into the first GNOME entry, which is apparently the same as GNOME on Xorg. The version is 3.38.5. The windowing system is X11.
That is the DE or Desktop Environment
If no, how do I determine that?
ll /etc/systemd/system/display-manager.service
or
systemctl status display-manager
bash.8[~]: systemctl status display-manager ● lightdm.service - Light Display Manager Loaded: loaded (/usr/lib/systemd/system/lightdm.service; enabled; vendor preset: enabled) Active: active (running) since Thu 2021-04-22 10:57:46 MDT; 4h 27min ago Docs: man:lightdm(1) Main PID: 979 (lightdm) Tasks: 5 (limit: 19096) Memory: 76.8M CPU: 6min 44.528s CGroup: /system.slice/lightdm.service ├─ 979 /usr/sbin/lightdm └─129997 /usr/libexec/Xorg -core -noreset :0 -seat seat0 -auth /run/lightdm/root/:0 -nolisten tcp vt1 -novtswitch
Warning: some journal files were not opened due to insufficient permissions. bash.9[~]: ll /etc/systemd/system/display-manager.service lrwxrwxrwx. 1 root root 39 Sep 26 2015 /etc/systemd/system/display-manager.service -> /usr/lib/systemd/system/lightdm.service bash.10[~]:
lightdm (= Light Display Manager)
On 4/22/21 3:23 PM, Patrick O'Callaghan wrote:
On Thu, 2021-04-22 at 11:29 -0600, home user wrote:
On 4/22/21 11:10 AM, Joe Zeff wrote:
On 4/22/21 9:50 AM, home user wrote:
I'm wondering if wayland and earlyoom did get installed and then later removed or de-activated. I'm also wondering what else I should have but am missing. In other words, do I have a broader problem?
If you're concerned about that, you might consider using distro- sync to make sure you've got everything.
I did that as a part of the upgrade earlier this month. But I try again:
bash.1[~]: su - Password: -bash.1[~]: dnf distro-sync Last metadata expiration check: 0:51:59 ago on Thu 22 Apr 2021 10:35:59 AM MDT. Dependencies resolved. Nothing to do. Complete!
Was your original installation the default Workstation or KDE Spin?
It was the default workstation (GNOME?). This would have been early spring 2013. Other workstations were added later, though I rarely use them.
On 23/04/2021 05:26, home user wrote:
bash.8[~]: systemctl status display-manager ● lightdm.service - Light Display Manager Loaded: loaded (/usr/lib/systemd/system/lightdm.service; enabled; vendor preset: enabled) Active: active (running) since Thu 2021-04-22 10:57:46 MDT; 4h 27min ago Docs: man:lightdm(1) Main PID: 979 (lightdm) Tasks: 5 (limit: 19096) Memory: 76.8M CPU: 6min 44.528s CGroup: /system.slice/lightdm.service ├─ 979 /usr/sbin/lightdm └─129997 /usr/libexec/Xorg -core -noreset :0 -seat seat0 -auth /run/lightdm/root/:0 -nolisten tcp vt1 -novtswitch
Warning: some journal files were not opened due to insufficient permissions. bash.9[~]: ll /etc/systemd/system/display-manager.service lrwxrwxrwx. 1 root root 39 Sep 26 2015 /etc/systemd/system/display-manager.service -> /usr/lib/systemd/system/lightdm.service bash.10[~]:
lightdm (= Light Display Manager)
This may be your problem.....
Wayland session not working with duplicate GNOME entries in greeter Some greeters (lightdm-webkit2-greeter for example) do not support two sessions with the same name [1]. To check for duplicate entries: ls -1 /usr/share/wayland-sessions /usr/share/xsessions Rename the duplicate entry in /usr/share/xsessions. For example: mv /usr/share/xsessions/gnome.desktop /usr/share/xsessions/gnome.desktop.disabled
Or, you can try switching to a different DM.
If you have gdm or sddm installed you can do....
systemctl --force enable sddm
Then reboot.
On 4/22/21 3:34 PM, Ed Greshko wrote:
On 23/04/2021 05:26, home user wrote:
Or, you can try switching to a different DM.
If you have gdm or sddm installed you can do....
systemctl --force enable sddm
Then reboot.
Thinking of both the present and the next few years, which of lightdm, gdm, and sddm would be best? Which do you recommend? wikipedia isn't saying much about those choices.
On 23/04/2021 06:01, home user wrote:
On 4/22/21 3:34 PM, Ed Greshko wrote:
On 23/04/2021 05:26, home user wrote:
Or, you can try switching to a different DM.
If you have gdm or sddm installed you can do....
systemctl --force enable sddm
Then reboot.
Thinking of both the present and the next few years, which of lightdm, gdm, and sddm would be best? Which do you recommend? wikipedia isn't saying much about those choices.
I have no particular opinion. As a KDE user I tend to stick with sddm as it is part of KDE. The only departure from that is when I'm using KDE in a qemu VM. Recently using sddm in the VM has issues with display sizing. So, I've been using gdm in that case.
On Thu, 2021-04-22 at 15:29 -0600, home user wrote:
On 4/22/21 3:23 PM, Patrick O'Callaghan wrote:
On Thu, 2021-04-22 at 11:29 -0600, home user wrote:
On 4/22/21 11:10 AM, Joe Zeff wrote:
On 4/22/21 9:50 AM, home user wrote:
I'm wondering if wayland and earlyoom did get installed and then later removed or de-activated. I'm also wondering what else I should have but am missing. In other words, do I have a broader problem?
If you're concerned about that, you might consider using distro- sync to make sure you've got everything.
I did that as a part of the upgrade earlier this month. But I try again:
bash.1[~]: su - Password: -bash.1[~]: dnf distro-sync Last metadata expiration check: 0:51:59 ago on Thu 22 Apr 2021 10:35:59 AM MDT. Dependencies resolved. Nothing to do. Complete!
Was your original installation the default Workstation or KDE Spin?
It was the default workstation (GNOME?). This would have been early spring 2013. Other workstations were added later, though I rarely use them.
I'm not sure Wayland would have been part of the default install in 2013, and if all your subsequent installs have been updates than that could explain why you don't have it.
poc
On 23/04/2021 06:09, Patrick O'Callaghan wrote:
I'm not sure Wayland would have been part of the default install in 2013, and if all your subsequent installs have been updates than that could explain why you don't have it.
I would have no idea as to the minimum packages needed to support wayland.
My system was installed back in 2017 using the KDE live image. Sometime along the line I decided to give wayland a try. I learned that to enable wayland on KDE you would need to install plasma-workspace-wayland. Looking at dnf history this happened on January 13, 2018. Additional packages were installed.
Install plasma-workspace-wayland-5.11.4-1.fc27.x86_64 @updates Install kwayland-integration-5.11.4-1.fc27.x86_64 @updates Install kwin-wayland-5.11.4-1.fc27.x86_64 @updates Install xorg-x11-server-Xwayland-1.19.6-1.fc27.x86_64 @updates
And after that I don't recall explicitly installing anything to get wayland working.
I do recall that back in 2018 the nVidia drivers were not ready to support wayland and I couldn't use nouveau due to rendering issues of Chinese characters in certain applications. So, nothing worked in 2018.
Also, FWIW, the 470 version of nVidia drivers will be more "wayland friendly". But that version is not yet in beta.
On 4/22/21 3:34 PM, Ed Greshko wrote:
This may be your problem.....
Wayland session not working with duplicate GNOME entries in greeter Some greeters (lightdm-webkit2-greeter for example) do not support two sessions with the same name [1]. To check for duplicate entries: ls -1 /usr/share/wayland-sessions /usr/share/xsessions Rename the duplicate entry in /usr/share/xsessions. For example: mv /usr/share/xsessions/gnome.desktop /usr/share/xsessions/gnome.desktop.disabled
I tried this first.
As root, I did the rename. Then I re-booted.
The menu had only one GNOME entry, as expected. I tried it. After about a second, I got the login screen again. I switched to the GNOME on Xorg and logged in. I got a system problem and a SELinux alert. I think the system problem is the same as the one I reported in a newer thread earlier today. The SELinux alert complains...
The source process: gdb Attempted this access: open On this file: /car/cache/fwupd/metainfo.xmlb
and suggested doing this...
# ausearch -c 'gdb' --raw | audit2allow -M my-gdb # semodule -X 300 -i my-gdb.pp
which I did as root.
I rebooted.
Same results logging in to GNOME. Same results logging in to GNOME on Xorg.
As root, I put the gnome.desktop.disabled back to gnome.desktop.
The menu is back to the way it was (2 GNOME entries). But I'm still getting the SELinux alerts and the system problem.
I thought it best to check back with you before trying gdm (which I already have) or sddm (which I don't yet have).
On 4/22/21 4:56 PM, Ed Greshko wrote:
If you are a majority Gnome user I'd go with gdm
bash.1[~]: su - Password: -bash.1[~]: systemctl --force enable gdm Removed /etc/systemd/system/display-manager.service. Created symlink /etc/systemd/system/display-manager.service → /usr/lib/systemd/system/gdm.service. -bash.2[~]:
I rebooted.
The login screen is different. When I click the little gear in the lower right, I see only 2 gnome entries. I chose the GNOME one, I don't remember what the other was labelled. After logging in, I again get the system and SELinux alerts.
bash.1[~]: env | grep -i wayland bash.2[~]:
I'm still not in wayland.
On 23/04/2021 07:19, home user wrote:
bash.1[~]: su - Password: -bash.1[~]: systemctl --force enable gdm Removed /etc/systemd/system/display-manager.service. Created symlink /etc/systemd/system/display-manager.service → /usr/lib/systemd/system/gdm.service. -bash.2[~]:
I rebooted.
The login screen is different. When I click the little gear in the lower right, I see only 2 gnome entries. I chose the GNOME one, I don't remember what the other was labelled. After logging in, I again get the system and SELinux alerts.
bash.1[~]: env | grep -i wayland bash.2[~]:
I'm still not in wayland.
OK....
Try this for a test.
When you get the boot menu to select the kernel hit "e" to edit. Add "selinux=0" to the end of the linux line. Then "ctrl-x" to continue booting. This will disable selinux from the very start.
On 4/22/21 4:56 PM, Ed Greshko wrote:
If you are a majority Gnome user I'd go with gdm
With both gdm and the renamed file, the gear menu lists the following gnome choices... GNOME Classic GNOME on Xorg I went with GNOME on Xorg. Same system and SELinux alerts.
I just saw your post to disable SELinux. I'll try that next.
On 23/04/2021 07:49, home user wrote:
With both gdm and the renamed file, the gear menu lists the following gnome choices... GNOME Classic GNOME on Xorg I went with GNOME on Xorg. Same system and SELinux alerts.
Well, with the renamed file you don't seem to have "GNOME" only. That would be the selection for Wayland.
Best to revert that name change.
I just saw your post to disable SELinux. I'll try that next.
OK.... But that makes things slightly more "complicated".
The disabling of selinux was to determine if that would allow you to bring up wayland. But if you don't have that option, then nothing will be learned. You won't get the sealerts since selinux was disabled.
Now, the disabling of selinux is only temporary. But, you should be made aware that the next time you boot with selinux enabled the system will automatically do a relable which may take a bit of time depending on the size of your filesystems.
On 4/22/21 5:55 PM, Ed Greshko wrote:
OK.... But that makes things slightly more "complicated".
The disabling of selinux was to determine if that would allow you to bring up wayland. But if you don't have that option, then nothing will be learned. You won't get the sealerts since selinux was disabled.
Now, the disabling of selinux is only temporary. But, you should be made aware that the next time you boot with selinux enabled the system will automatically do a relable which may take a bit of time depending on the size of your filesystems.
The only difference the selinux=0 made was to stop the SELinux alerts, as you predicted.
Rebooting to re-enable SELinux....
I'm having to use my old windows-7 partition and system for this message.
Rebooting to re-enable SELinux....
It failed. The last 4 lines of boot messages were: [OK] Started Command Scheduler. Started GNOME Display Manager... Started Hold until boot process finishs up... [OK] Started GNOME Display Manager.
After that, nothing for nearly 3 hours, at which time I did a hard shutdown. Tried a second time. Same results except this time I hit the hard reset after about 30 minutes. Tried a third time, but this time I used the third kernel in the grub boot menu. That's still F33. Same results as the preceding try. I used to have something in this windows that could access the Linux partitions. But I can't remember what it was called, and I can't now find any hint of it. I do remember it was installed about 4 years ago.
What now?
On 23/04/2021 12:34, home user wrote:
I'm having to use my old windows-7 partition and system for this message.
Rebooting to re-enable SELinux....
It failed. The last 4 lines of boot messages were: [OK] Started Command Scheduler. Started GNOME Display Manager... Started Hold until boot process finishs up... [OK] Started GNOME Display Manager.
After that, nothing for nearly 3 hours, at which time I did a hard shutdown. Tried a second time. Same results except this time I hit the hard reset after about 30 minutes. Tried a third time, but this time I used the third kernel in the grub boot menu. That's still F33. Same results as the preceding try. I used to have something in this windows that could access the Linux partitions. But I can't remember what it was called, and I can't now find any hint of it. I do remember it was installed about 4 years ago.
What now?
As you did before, when you have the kernel selection menu hit e. Then add a 3 to the end of the linux line. Ctrl-x to continue to single user mode.
Login as yourself and then
sudo systemctl start display-manger
and see what error results.
On 4/22/21 11:50 PM, Ed Greshko wrote:
As you did before, when you have the kernel selection menu hit e. Then add a 3 to the end of the linux line. Ctrl-x to continue to single user mode.
No, runlevel 3 boots you into a multi-user system with networking, but text based, not graphical. However, starting a display manager will then take you into a normal graphical environment.
On 04/23/2021 12:19 AM, Joe Zeff wrote:
On 4/22/21 11:50 PM, Ed Greshko wrote:
As you did before, when you have the kernel selection menu hit e. Then add a 3 to the end of the linux line. Ctrl-x to continue to single user mode.
No, runlevel 3 boots you into a multi-user system with networking, but text based, not graphical. However, starting a display manager will then take you into a normal graphical environment.
(currently in windows-7)
My gut tells me the SELinux re-label either never finishedor never started. The 4 lines that I put into my most recent post came up in under 2 minutes after the grub menu disappeared, and I never saw any hint of output from a SELinux relabelling. So when I try what Ed and Joe suggest, should I also add "selinux=0" to the linux line? And how long should I wait before assuming something is stuck in a loop, frozen, etc. and do a hard reset?
On Fri, 2021-04-23 at 07:08 -0600, home user wrote:
My gut tells me the SELinux re-label either never finished or never started. The 4 lines that I put into my most recent post came up in under 2 minutes after the grub menu disappeared, and I never saw any hint of output from a SELinux relabelling. So when I try what Ed and Joe suggest, should I also add "selinux=0" to the linux line? And how long should I wait before assuming something is stuck in a loop, frozen, etc. and do a hard reset?
Have you removed rhgb and quiet from the kernel line, so you can see what your computer is doing as it boots up?
If you can get the computer to boot up, perhaps doing another reboot with selinux=0 to see if that helps, you can manually schedule a relabel.
touch /.autorelabel
Then reboot. When your system sees that file, it does a relabel, then removes it when the job is done.
For what it's worth, I think it's a ridiculously silly idea to hide the boot processes, by default, on an experimental OS that's always going through changes. Leave that kind of dopeyness to fixed installations.
On 23/04/2021 21:08, home user wrote:
My gut tells me the SELinux re-label either never finishedor never started. The 4 lines that I put into my most recent post came up in under 2 minutes after the grub menu disappeared, and I never saw any hint of output from a SELinux relabelling. So when I try what Ed and Joe suggest, should I also add "selinux=0" to the linux line? And how long should I wait before assuming something is stuck in a loop, frozen, etc. and do a hard reset?
No, don't disable selinux in this case.
Just boot the system, edit the linux line to add 3 and continue. This will get you to a text mode login prompt and gdm would not have been started.
I suspect when login and type
sudo systemctl start display-manager
you will get an error and then can more info from
systemctl status display-manager or looking at journal entries.
On 04/23/2021 7:28 AM, Ed Greshko wrote:
On 23/04/2021 21:08, home user wrote:
My gut tells me the SELinux re-label either never finishedor never started. The 4 lines that I put into my most recent post came up in under 2 minutes after the grub menu disappeared, and I never saw any hint of output from a SELinux relabelling. So when I try what Ed and Joe suggest, should I also add "selinux=0" to the linux line? And how long should I wait before assuming something is stuck in a loop, frozen, etc. and do a hard reset?
No, don't disable selinux in this case.
Just boot the system, edit the linux line to add 3 and continue. This will get you to a text mode login prompt and gdm would not have been started.
I suspect when login and type
sudo systemctl start display-manager
you will get an error and then can more info from
systemctl status display-manager or looking at journal entries.
(sigh) I had to remember all those log-in names! :(
It took some time to get rid of distractions. I did get in. I had to edit the .bash_profile in the user, admin, and root accounts to comment out xeyes.
I did a few tests. In all cases, the workstation ends up "frozen" but it sounds like a fan in the tower is surging, and I must do a hard shutdown.
1. (login as a regular user) sudo systemctl [etc] (enter the regular user password) "bill is not in the sudoers file. This incident will be reported." (sigh) Now y'all got me in trouble with me! :) su - (enter root password) systemctl [etc] (blank; freeze; fan surge; hard shutdown)
2. (login as regular user su - (enter root password) systemctl [etc] [FAILED] Failed to start Accounts Service. [FAILED] Failed to start GNOME Display Manager. (freeze; fan surge; hard shutdown)
3. (login as admin) systemctl [etc] (I'm prompted for the admin password) (enter admin password) [FAILED] Failed to start Accounts Service. [FAILED] Failed to start GNOME Display Manager. (freeze; fan surge; hard shutdown)
4. (login as root) systemctl [etc] [FAILED] Failed to start Accounts Service. [FAILED] Failed to start GNOME Display Manager. (freeze; fan surge; hard shutdown)
On 04/22/2021 10:34 PM, home user wrote:
I'm having to use my old windows-7 partition and system for this message.
Rebooting to re-enable SELinux....
It failed. The last 4 lines of boot messages were: [OK] Started Command Scheduler. Started GNOME Display Manager... Started Hold until boot process finishs up... [OK] Started GNOME Display Manager.
After that, nothing for nearly 3 hours, at which time I did a hard shutdown. Tried a second time. Same results except this time I hit the hard reset after about 30 minutes. Tried a third time, but this time I used the third kernel in the grub boot menu. That's still F33. Same results as the preceding try. I used to have something in this windows that could access the Linux partitions. But I can't remember what it was called, and I can't now find any hint of it. I do remember it was installed about 4 years ago.
What now?
Found that old tool to access the Linux partitions from windows-7. It's called ext2explore. It seems to give read-only access, and must be run as the windows-7 admin. But with it, I can copy files from the Linux partitions to windows-7. From there, I can get them to the google drive.
So if you tell me quick, simple, specific things I can do in Fedora run-level 3, such as: * command > command.out * grep logfile > grep.out I can then get results to windows, and then to the google drive (big things) or a post in this list (short things).
On 24/04/2021 02:56, home user wrote:
On 04/22/2021 10:34 PM, home user wrote:
I'm having to use my old windows-7 partition and system for this message.
Rebooting to re-enable SELinux....
It failed. The last 4 lines of boot messages were: [OK] Started Command Scheduler. Started GNOME Display Manager... Started Hold until boot process finishs up... [OK] Started GNOME Display Manager.
After that, nothing for nearly 3 hours, at which time I did a hard shutdown. Tried a second time. Same results except this time I hit the hard reset after about 30 minutes. Tried a third time, but this time I used the third kernel in the grub boot menu. That's still F33. Same results as the preceding try. I used to have something in this windows that could access the Linux partitions. But I can't remember what it was called, and I can't now find any hint of it. I do remember it was installed about 4 years ago.
What now?
Found that old tool to access the Linux partitions from windows-7. It's called ext2explore. It seems to give read-only access, and must be run as the windows-7 admin. But with it, I can copy files from the Linux partitions to windows-7. From there, I can get them to the google drive.
So if you tell me quick, simple, specific things I can do in Fedora run-level 3, such as:
- command > command.out
- grep logfile > grep.out
I can then get results to windows, and then to the google drive (big things) or a post in this list (short things).
Before that.....
Can you boot again with selinux disabled? And, if you can, does gdm start?
If both of those are "yes", would you do as Tim requested and
touch /.autorelabel
And then reboot with selinux enabled.
On 4/23/2021 4:34 PM, Ed Greshko wrote:
Before that.....
Can you boot again with selinux disabled? And, if you can, does gdm start?
Is this with or without run-level 3?
If both of those are "yes", would you do as Tim requested and
touch /.autorelabel
And then reboot with selinux enabled.
With or without run-level 3? With or without rhgb? With or without quiet? How much time should I allow before I "panic"?
On 24/04/2021 07:42, home user wrote:
On 4/23/2021 4:34 PM, Ed Greshko wrote:
Before that.....
Can you boot again with selinux disabled? And, if you can, does gdm start?
Is this with or without run-level 3?
Either way. If with 3 you'll have to login from the text prompt and manually start it.
If both of those are "yes", would you do as Tim requested and
touch /.autorelabel
And then reboot with selinux enabled.
With or without run-level 3? With or without rhgb? With or without quiet? How much time should I allow before I "panic"?
Well, the point is to get the system to relabel so getting logged in is the goal.
It doesn't matter if rhgb (Red Hat Graphical Boot) is on or off. And quiet makes things less verbose.
If things work it should be as a normal boot.
On 4/23/21 5:53 PM, Ed Greshko wrote:
On 24/04/2021 07:42, home user wrote:
On 4/23/2021 4:34 PM, Ed Greshko wrote:
Before that.....
Can you boot again with selinux disabled? And, if you can, does gdm start?
Is this with or without run-level 3?
Either way. If with 3 you'll have to login from the text prompt and manually start it.
Both ways, it boots, and gdm starts. I'm in now and using Fedora to do this reply.
If both of those are "yes", would you do as Tim requested and
touch /.autorelabel
And then reboot with selinux enabled.
With or without run-level 3? With or without rhgb? With or without quiet? How much time should I allow before I "panic"?
Well, the point is to get the system to relabel so getting logged in is the goal.
It doesn't matter if rhgb (Red Hat Graphical Boot) is on or off. And quiet makes things less verbose.
If things work it should be as a normal boot.
If the SELinux re-labelling displays nothing, I'll allow 2 hours before "panicking".
Here goes...
On 4/23/21 6:14 PM, home user wrote:
On 4/23/21 5:53 PM, Ed Greshko wrote:
On 24/04/2021 07:42, home user wrote:
On 4/23/2021 4:34 PM, Ed Greshko wrote:
If both of those are "yes", would you do as Tim requested and
touch /.autorelabel
And then reboot with selinux enabled.
With or without run-level 3? With or without rhgb? With or without quiet? How much time should I allow before I "panic"?
Well, the point is to get the system to relabel so getting logged in is the goal.
It doesn't matter if rhgb (Red Hat Graphical Boot) is on or off. And quiet makes things less verbose.
If things work it should be as a normal boot.
If the SELinux re-labelling displays nothing, I'll allow 2 hours before "panicking".
Here goes...
I'm BAAAaaack!
I did not stay at the workstation so I don't know if there were any errors, warnings, or messages. But I found the gdm login screen displaying. I'm currently logged in to my Fedora user account in GNOME on Xorg and doing this message in Thunderbird. So far there have been at least 3 problem alerts (kernel-core, tracker3-miners, and fwupd) and 1 SELinux alerts (gdb wanting to open /var/cache/fwupd/metainfo.xmlb). The gdb login screen did not offer GNOME wayland (GNOME).
On 24/04/2021 09:46, home user wrote:
I'm BAAAaaack!
I did not stay at the workstation so I don't know if there were any errors, warnings, or messages. But I found the gdm login screen displaying. I'm currently logged in to my Fedora user account in GNOME on Xorg and doing this message in Thunderbird. So far there have been at least 3 problem alerts (kernel-core, tracker3-miners, and fwupd) and 1 SELinux alerts (gdb wanting to open /var/cache/fwupd/metainfo.xmlb). The gdb login screen did not offer GNOME wayland (GNOME).
Well, the choices available in GDM are somewhat dependent on how and when things were installed.
On a F33 system installed from Workstation Live the choices would show as
GNOME GNOME Classic GNOME on Xorg
With the top being walyand.
On 4/23/21 8:02 PM, Ed Greshko wrote: (my apologies: I replied to him when I meant to reply to the list)
On a F33 system installed from Workstation Live the choices would show as
GNOME GNOME Classic GNOME on Xorg
With the top being walyand.
I don't have that top entry.
Since my initial install was 8 years ago, I don't remember for certain how I did it: CD or USB stick. It was definitely the Workstation option (spin?). Since then, it's been all patching and upgrading, all using yum, then dnf. So we're back to the starting questions. How do I get wayland working in gnome on this workstation? Am I missing something? If yes, what? Otherwise, what do I need to do to make wayland available and properly functioning in gnome?
On 24/04/2021 10:24, home user wrote:
On 4/23/21 8:02 PM, Ed Greshko wrote: (my apologies: I replied to him when I meant to reply to the list)
On a F33 system installed from Workstation Live the choices would show as
GNOME GNOME Classic GNOME on Xorg
With the top being walyand.
I don't have that top entry.
Since my initial install was 8 years ago, I don't remember for certain how I did it: CD or USB stick. It was definitely the Workstation option (spin?). Since then, it's been all patching and upgrading, all using yum, then dnf. So we're back to the starting questions. How do I get wayland working in gnome on this workstation? Am I missing something? If yes, what? Otherwise, what do I need to do to make wayland available and properly functioning in gnome?
Not being a GNOME user, and not caring all that much about wayland, I can't claim to know much.
But, I would check to see if the following are installed.
[egreshko@f33g ~]$ rpm -qa | grep -i wayland libwayland-client-1.18.0-2.fc33.x86_64 libwayland-server-1.18.0-2.fc33.x86_64 libwayland-egl-1.18.0-2.fc33.x86_64 libwayland-cursor-1.18.0-2.fc33.x86_64 gnome-session-wayland-session-3.38.0-1.fc33.x86_64 qt5-qtwayland-5.15.2-3.fc33.x86_64 xorg-x11-server-Xwayland-1.20.10-1.fc33.x86_64
On 4/23/21 8:28 PM, Ed Greshko wrote:
On 24/04/2021 10:24, home user wrote:
... So we're back to the starting questions. How do I get wayland working in gnome on this workstation? Am I missing something? If yes, what? Otherwise, what do I need to do to make wayland available and properly functioning in gnome?
Not being a GNOME user, and not caring all that much about wayland, I can't claim to know much.
But, I would check to see if the following are installed.
[egreshko@f33g ~]$ rpm -qa | grep -i wayland libwayland-client-1.18.0-2.fc33.x86_64 libwayland-server-1.18.0-2.fc33.x86_64 libwayland-egl-1.18.0-2.fc33.x86_64 libwayland-cursor-1.18.0-2.fc33.x86_64 gnome-session-wayland-session-3.38.0-1.fc33.x86_64 qt5-qtwayland-5.15.2-3.fc33.x86_64 xorg-x11-server-Xwayland-1.20.10-1.fc33.x86_64
If I understand the output from "dnf history" correctly, I have all of those, all since 2018 or earlier, all patched earlier this month.
On 24/04/2021 10:42, home user wrote:
On 4/23/21 8:28 PM, Ed Greshko wrote:
On 24/04/2021 10:24, home user wrote:
... So we're back to the starting questions. How do I get wayland working in gnome on this workstation? Am I missing something? If yes, what? Otherwise, what do I need to do to make wayland available and properly functioning in gnome?
Not being a GNOME user, and not caring all that much about wayland, I can't claim to know much.
But, I would check to see if the following are installed.
[egreshko@f33g ~]$ rpm -qa | grep -i wayland libwayland-client-1.18.0-2.fc33.x86_64 libwayland-server-1.18.0-2.fc33.x86_64 libwayland-egl-1.18.0-2.fc33.x86_64 libwayland-cursor-1.18.0-2.fc33.x86_64 gnome-session-wayland-session-3.38.0-1.fc33.x86_64 qt5-qtwayland-5.15.2-3.fc33.x86_64 xorg-x11-server-Xwayland-1.20.10-1.fc33.x86_64
If I understand the output from "dnf history" correctly, I have all of those, all since 2018 or earlier, all patched earlier this month.
2 questions....
1. rpm -qa | grep -i wayland returns all of them on your system?
2. Do you have the file /usr/share/wayland-sessions/gnome.desktop (You've not disabled it)
On 4/23/21 8:48 PM, Ed Greshko wrote:
On 24/04/2021 10:42, home user wrote:
On 4/23/21 8:28 PM, Ed Greshko wrote:
On 24/04/2021 10:24, home user wrote:
... So we're back to the starting questions. How do I get wayland working in gnome on this workstation? Am I missing something? If yes, what? Otherwise, what do I need to do to make wayland available and properly functioning in gnome?
Not being a GNOME user, and not caring all that much about wayland, I can't claim to know much.
But, I would check to see if the following are installed.
2 questions....
1. rpm -qa | grep -i wayland returns all of them on your system?
yes, and more. see below.
2. Do you have the file /usr/share/wayland-sessions/gnome.desktop (You've not disabled it)
I have it; see below. I don't know if I've disabled it; how would I check?
-bash.16[~]: rpm -qa | grep -i wayland libwayland-client-1.18.0-2.fc33.x86_64 ibus-wayland-1.5.23-2.fc33.x86_64 libwayland-cursor-1.18.0-2.fc33.x86_64 plasma-workspace-wayland-5.20.5-3.fc33.x86_64 libwayland-server-1.18.0-2.fc33.i686 gnome-session-wayland-session-3.38.0-1.fc33.x86_64 wayland-protocols-devel-1.20-2.fc33.noarch egl-wayland-1.1.6-1.fc33.i686 libwayland-server-1.18.0-2.fc33.x86_64 kf5-kwayland-5.79.0-2.fc33.x86_64 libwayland-client-1.18.0-2.fc33.i686 wayland-devel-1.18.0-2.fc33.x86_64 kwayland-server-5.20.5-1.fc33.x86_64 kwin-wayland-5.20.5-1.fc33.x86_64 egl-wayland-1.1.6-1.fc33.x86_64 kwayland-integration-5.20.5-1.fc33.x86_64 libwayland-egl-1.18.0-2.fc33.x86_64 xorg-x11-server-Xwayland-1.20.11-1.fc33.x86_64 qt5-qtwayland-5.15.2-5.fc33.x86_64 -bash.17[~]: ls -l /usr/share/wayland-sessions/gnome.desktop -rw-r--r--. 1 root root 7829 Sep 12 2020 /usr/share/wayland-sessions/gnome.desktop -bash.18[~]: cat /usr/share/wayland-sessions/gnome.desktop [Desktop Entry] Name[af]=GNOME ... (88 like the above) Name[zh_TW]=GNOME Name=GNOME Comment[af]=Die sessie laat u by GNOME aanmeld ... (88 lines like the above) Comment[zh_TW]=這個工作階段讓您登入 GNOME Comment=This session logs you into GNOME Exec=/usr/bin/gnome-session TryExec=/usr/bin/gnome-session Type=Application DesktopNames=GNOME X-GDM-SessionRegisters=true -bash.19[~]:
On 4/23/21 9:09 PM, home user wrote:
On 4/23/21 8:48 PM, Ed Greshko wrote:
On 24/04/2021 10:42, home user wrote:
On 4/23/21 8:28 PM, Ed Greshko wrote:
On 24/04/2021 10:24, home user wrote:
... So we're back to the starting questions. How do I get wayland working in gnome on this workstation? Am I missing something? If yes, what? Otherwise, what do I need to do to make wayland available and properly functioning in gnome?
Not being a GNOME user, and not caring all that much about wayland, I can't claim to know much.
But, I would check to see if the following are installed.
2 questions....
1. rpm -qa | grep -i wayland returns all of them on your system?
yes, and more. see below.
2. Do you have the file /usr/share/wayland-sessions/gnome.desktop (You've not disabled it)
I have it; see below. I don't know if I've disabled it; how would I check?
-bash.16[~]: rpm -qa | grep -i wayland libwayland-client-1.18.0-2.fc33.x86_64 ibus-wayland-1.5.23-2.fc33.x86_64 libwayland-cursor-1.18.0-2.fc33.x86_64 plasma-workspace-wayland-5.20.5-3.fc33.x86_64 libwayland-server-1.18.0-2.fc33.i686 gnome-session-wayland-session-3.38.0-1.fc33.x86_64 wayland-protocols-devel-1.20-2.fc33.noarch egl-wayland-1.1.6-1.fc33.i686 libwayland-server-1.18.0-2.fc33.x86_64 kf5-kwayland-5.79.0-2.fc33.x86_64 libwayland-client-1.18.0-2.fc33.i686 wayland-devel-1.18.0-2.fc33.x86_64 kwayland-server-5.20.5-1.fc33.x86_64 kwin-wayland-5.20.5-1.fc33.x86_64 egl-wayland-1.1.6-1.fc33.x86_64 kwayland-integration-5.20.5-1.fc33.x86_64 libwayland-egl-1.18.0-2.fc33.x86_64 xorg-x11-server-Xwayland-1.20.11-1.fc33.x86_64 qt5-qtwayland-5.15.2-5.fc33.x86_64
Looking more slowly and closely, I see 2 of mine are more recent than Ed's:
qt5-qtwayland-5.15.2-5.fc33.x86_64 (mine) qt5-qtwayland-5.15.2-3.fc33.x86_64 (Ed's)
xorg-x11-server-Xwayland-1.20.11-1.fc33.x86_64 (mine) xorg-x11-server-Xwayland-1.20.10-1.fc33.x86_64 (Ed's)
Does that matter?
On 24/04/2021 11:15, home user wrote:
Looking more slowly and closely, I see 2 of mine are more recent than Ed's:
qt5-qtwayland-5.15.2-5.fc33.x86_64 (mine) qt5-qtwayland-5.15.2-3.fc33.x86_64 (Ed's)
xorg-x11-server-Xwayland-1.20.11-1.fc33.x86_64 (mine) xorg-x11-server-Xwayland-1.20.10-1.fc33.x86_64 (Ed's)
Does that matter?
No. That system I got my information from hasn't been updated in a while.
On 24/04/2021 11:09, home user wrote:
On 4/23/21 8:48 PM, Ed Greshko wrote:
On 24/04/2021 10:42, home user wrote:
On 4/23/21 8:28 PM, Ed Greshko wrote:
On 24/04/2021 10:24, home user wrote:
... So we're back to the starting questions. How do I get wayland working in gnome on this workstation? Am I missing something? If yes, what? Otherwise, what do I need to do to make wayland available and properly functioning in gnome?
Not being a GNOME user, and not caring all that much about wayland, I can't claim to know much.
But, I would check to see if the following are installed.
2 questions....
1. rpm -qa | grep -i wayland returns all of them on your system?
yes, and more. see below.
Sure. Some of those are related to also have KDE installed with wayland support
2. Do you have the file /usr/share/wayland-sessions/gnome.desktop (You've not disabled it)
I have it; see below. I don't know if I've disabled it; how would I check?
That question was related to earlier in the thread in talking about lightdb. If one adds .disabled (or anything) after desktop then the display manager would ignore it.
On 4/23/21 9:18 PM, Ed Greshko wrote:
2. Do you have the file /usr/share/wayland-sessions/gnome.desktop (You've not disabled it)
I have it; see below. I don't know if I've disabled it; how would I check?
That question was related to earlier in the thread in talking about lightdb. If one adds .disabled (or anything) after desktop then the display manager would ignore it.
-bash.21[~]: ls -1 /usr/share/wayland-sessions /usr/share/xsessions /usr/share/wayland-sessions: gnome.desktop plasmawayland.desktop
/usr/share/xsessions: cinnamon2d.desktop cinnamon.desktop gnome-classic.desktop gnome.desktop.disabled gnome-xorg.desktop LXDE.desktop mate.desktop openbox.desktop plasma.desktop wmx.desktop xfce.desktop -bash.22[~]:
That applies to gdm as well as lightdm?
On 24/04/2021 11:26, home user wrote:
On 4/23/21 9:18 PM, Ed Greshko wrote:
2. Do you have the file /usr/share/wayland-sessions/gnome.desktop (You've not disabled it)
I have it; see below. I don't know if I've disabled it; how would I check?
That question was related to earlier in the thread in talking about lightdb. If one adds .disabled (or anything) after desktop then the display manager would ignore it.
-bash.21[~]: ls -1 /usr/share/wayland-sessions /usr/share/xsessions /usr/share/wayland-sessions: gnome.desktop plasmawayland.desktop
/usr/share/xsessions: cinnamon2d.desktop cinnamon.desktop gnome-classic.desktop gnome.desktop.disabled gnome-xorg.desktop LXDE.desktop mate.desktop openbox.desktop plasma.desktop wmx.desktop xfce.desktop -bash.22[~]:
That applies to gdm as well as lightdm?
That caveat only applied to lightdm. I'd revert that change now that you'e using gdm.
I just realized one of my KDE VM's which was installed from KDE live also has GNOME. It works, OK with wayland.
[egreshko@f33k ~]$ rpm -qa | grep wayland libwayland-server-1.18.0-2.fc33.x86_64 libwayland-client-1.18.0-2.fc33.x86_64 libwayland-cursor-1.18.0-2.fc33.x86_64 libwayland-egl-1.18.0-2.fc33.x86_64 gnome-session-wayland-session-3.38.0-1.fc33.x86_64 qt5-qtwayland-5.15.2-3.fc33.x86_64 kwayland-server-5.20.5-1.fc33.x86_64 kwayland-integration-5.20.5-1.fc33.x86_64 kwin-wayland-5.20.5-1.fc33.x86_64 plasma-workspace-wayland-5.20.5-3.fc33.x86_64 kf5-kwayland-5.79.0-2.fc33.x86_64 xorg-x11-server-Xwayland-1.20.11-1.fc33.x86_64
One thing, if I use sddm as my display manager there are 2 "GNOME" entries on the login menu. But one starts an X session and the other Wayland.
On 4/23/21 9:48 PM, Ed Greshko wrote:
That caveat only applied to lightdm. I'd revert that change now that you'e using gdm.
reverted.
rebooted. got the gdm login. The GNOME entry re-appeared; the GNOME on Xorg entry disappeared. logged into GNOME.
"env | grep -i wayland" returns nothing, so I assume I'm in GNOME on Xorg.
I just realized one of my KDE VM's which was installed from KDE live also has GNOME. It works, OK with wayland.
One thing, if I use sddm as my display manager there are 2 "GNOME" entries on the login menu. But one starts an X session and the other Wayland.
I logged out, and then logged in to Plasma. It was slow, and I got SELinux alerts, but I can use it. It's not wayland:
bash.1[~]: env | grep -i wayland bash.2[~]:
I do not have sddm.
I'm shutting down for the night.
On Fri, 2021-04-23 at 20:24 -0600, home user wrote:
Since my initial install was 8 years ago, I don't remember for certain how I did it: CD or USB stick.
For later on, once you're back working again:
Next time you do an install, leave some notes in the /root homespace about what you've done. I think the installation routine leaves a couple of files about the install in there, itself. And I write some post installation notes about any extra packages I add.
Since I only use root for installing stuff, there's only three files in there, and they can't get messed with during normal logins (e.g. they can't get accidentally deleted).
On 4/22/21 12:41 PM, Joe Zeff wrote:
On 4/22/21 11:29 AM, home user wrote:
I did that as a part of the upgrade earlier this month. But I try again:
bash.1[~]: su - Password: -bash.1[~]: dnf distro-sync Last metadata expiration check: 0:51:59 ago on Thu 22 Apr 2021 10:35:59 AM MDT. Dependencies resolved. Nothing to do. Complete! -bash.2[~]:
OK, if those packages had been part of the original installation, they'd either be there or be replaced. Since they aren't there, they must be part of something you installed later. Why don't you just install them again and see what happens?
Before starting this thread, and more since starting this thread, I spent a lot of time in Fedora docs and Fedora magazine trying to figure out wayland functionality and earlyoom in hopes of solving these for myself. Neither wayland functionality and nor earlyoom could have been a part of my original install (spring of 2013); they were not yet a part of Fedora. They should have been installed as a part of a semi-annual upgrade (currently "dnf system-upgrade download --refresh --releasever=nn" and "dnf system-upgrade reboot"), not as a part of patching (currently "dnf --refresh upgrade"). These should not have needed manual installation by me. I really believe it needs to be determined why the semi-annual upgrades fell short, so it can be fixed.
In the case of earlyoom, once this thread is solved, I hopefully can merely "dnf install earlyoom" and that is solved. But for wayland functionality, after 53 posts, we haven't yet figured out what packages (if any) are missing, or what is causing the problem.
Finally, I did not know I was missing earlyoom, or that there was such a thing as earlyoom, until I saw it in a post by Ed in previous thread of mine. Likewise, I thought I already had and was already using wayland until that same post from Ed showed me otherwise. I had no idea I was missing anything.
On 4/22/21 4:09 PM, Patrick O'Callaghan wrote:
On Thu, 2021-04-22 at 15:29 -0600, home user wrote:
On 4/22/21 3:23 PM, Patrick O'Callaghan wrote:
Was your original installation the default Workstation or KDE Spin?
It was the default workstation (GNOME?). This would have been early spring 2013. Other workstations were added later, though I rarely use them.
I'm not sure Wayland would have been part of the default install in 2013, and if all your subsequent installs have been updates than that could explain why you don't have it.
It almost certainly was not. Patching ("dnf --refresh upgrade" or some earlier dnf/yum equivalent) indeed almost certainly would not install wayland. But it should have been a part of a semi-annual upgrade ("dnf system-upgrade download --refresh --releasever=nn" + "dnf system-upgrade reboot", or some earlier dnf/yum equivalent). I have done the semi-annual upgrade every spring and fall since the initial install in 2013. So I should have it, and it should be functioning.
On 4/24/21 12:34 AM, Tim via users wrote:
For later on, once you're back working again:
Next time you do an install, leave some notes in the /root homespace about what you've done. I think the installation routine leaves a couple of files about the install in there, itself. And I write some post installation notes about any extra packages I add.
Since I only use root for installing stuff, there's only three files in there, and they can't get messed with during normal logins (e.g. they can't get accidentally deleted).
Thank-you. There have been a few times I wish I had done that.
I also use root for installing, patching, and upgrading.
I could not find any files as you described left in root's home by dnf or yum. Am I looking in the correct place? What should the file name be?
On Sat, 24 Apr 2021 at 16:43, home user mattisonw@comcast.net wrote:
On 4/24/21 12:34 AM, Tim via users wrote:
For later on, once you're back working again:
Next time you do an install, leave some notes in the /root homespace about what you've done. I think the installation routine leaves a couple of files about the install in there, itself. And I write some post installation notes about any extra packages I add.
Since I only use root for installing stuff, there's only three files in there, and they can't get messed with during normal logins (e.g. they can't get accidentally deleted).
Thank-you. There have been a few times I wish I had done that.
I also use root for installing, patching, and upgrading.
I could not find any files as you described left in root's home by dnf or yum. Am I looking in the correct place? What should the file name be?
https://docs.fedoraproject.org/en-US/Fedora/16/html/Installation_Guide/sn-au... says:
Every Installation Produces a Kickstart File The Fedora installation process automatically writes a Kickstart file that contains the settings for the installed system. This file is always saved as /root/anaconda-ks.cfg. You may use this file to repeat the installation with identical settings, or modify copies to specify settings for other systems.
I don't know how far back this goes, as my current Fedora system was put into service in Nov. 2019.
Regarding the lack if visibility for Wayland: I think Nvidia was very slow to support Wayland (or was it that Nvidia ran Wayland slowly?), so if you have been using Nvidia drivers it doesn't surprise me that Wayland was not on offer. This might change for Fedora 34 when Wayland gets a divorce from Xorg.
On 25/04/2021 03:42, home user wrote:
On 4/24/21 12:34 AM, Tim via users wrote:
For later on, once you're back working again:
Next time you do an install, leave some notes in the /root homespace about what you've done. I think the installation routine leaves a couple of files about the install in there, itself. And I write some post installation notes about any extra packages I add.
Since I only use root for installing stuff, there's only three files in there, and they can't get messed with during normal logins (e.g. they can't get accidentally deleted).
Thank-you. There have been a few times I wish I had done that.
I also use root for installing, patching, and upgrading.
I could not find any files as you described left in root's home by dnf or yum. Am I looking in the correct place? What should the file name be?
Read Tim's comment again. "He" wrote the notes in a file. Not dnf or another process.
BTW, have you tried plasma on wayland again?
You may want to do something like "journalctl -b 0 | grep wayland" to see if you find anything of interest. It may be best to do this just after booting and trying to start a wayland session so as to make sure you've go the right time frame.
general FYI for everyone...
This Fedora docs web page says that wayland has been the default gnome display server, and has been the default since F-25: "https://docs.fedoraproject.org/en-US/quick-docs/configuring-xorg-as-default-...".
On 4/24/21 4:24 PM, Ed Greshko wrote:
On 25/04/2021 03:42, home user wrote:
On 4/24/21 12:34 AM, Tim via users wrote:
For later on, once you're back working again:
Next time you do an install, leave some notes in the /root homespace about what you've done. I think the installation routine leaves a couple of files about the install in there, itself. And I write some post installation notes about any extra packages I add.
Since I only use root for installing stuff, there's only three files in there, and they can't get messed with during normal logins (e.g. they can't get accidentally deleted).
Thank-you. There have been a few times I wish I had done that.
I also use root for installing, patching, and upgrading.
I could not find any files as you described left in root's home by dnf or yum. Am I looking in the correct place? What should the file name be?
Read Tim's comment again. "He" wrote the notes in a file. Not dnf or another process.
The basis for my question was the sentence "I think the installation routine leaves a couple of files about the install in there, itself.". But I might easily be misunderstanding what Tim means by "the installation routine".
BTW, have you tried plasma on wayland again?
Not since last night. Will try again now as per your next paragraph.
You may want to do something like "journalctl -b 0 | grep wayland" to see if you find anything of interest. It may be best to do this just after booting and trying to start a wayland session so as to make sure you've go the right time frame.
A few minutes ago, I rebooted. There is only one plasma option in the menu: "Plasma". I chose that and completed the login. bash.1[~]: su - Password: -bash.1[~]: journalctl -b 0 | grep wayland -bash.2[~]: env | grep -i wayland -bash.3[~]:
While my first priority is wayland in gnome, I definitely do want: * wayland in plasma (when it's available and the default for plasma); * to retain gnome on Xorg; and * to retain plasma on xorg as fall-backs. This week certainly taught me the value of fall-backs. I was glad I still had windows-7 (slow, cumbersome, and bad as it is) on this workstation. And I do occasionally use plasma.
On 25/04/2021 10:12, home user wrote:
A few minutes ago, I rebooted. There is only one plasma option in the menu: "Plasma". I chose that and completed the login. bash.1[~]: su - Password: -bash.1[~]: journalctl -b 0 | grep wayland -bash.2[~]: env | grep -i wayland -bash.3[~]:
I recall you once did have an option for plasma on wayland.
You still have a
/usr/share/wayland-sessions/plasmawayland.desktop
file?
While my first priority is wayland in gnome, I definitely do want: * wayland in plasma (when it's available and the default for plasma); * to retain gnome on Xorg; and * to retain plasma on xorg as fall-backs. This week certainly taught me the value of fall-backs. I was glad I still had windows-7 (slow, cumbersome, and bad as it is) on this workstation. And I do occasionally use plasma.
F34 makes Plasma (Wayland) default.
You may want to consider using sddm, at least to test.
With the list of "desktop" files you showed in an earlier post it would see you have installed quite a bit of "extras".
On my "work" system I keep it simple and have only
[egreshko@meimei ~]$ ll -1 /usr/share/wayland-sessions /usr/share/xsessions /usr/share/wayland-sessions: total 4 -rw-r--r--. 1 root root 2769 Jan 15 01:46 plasmawayland.desktop
/usr/share/xsessions: total 4 -rw-r--r--. 1 root root 2611 Jan 15 01:46 plasma.desktop
On 4/24/21 8:37 PM, Ed Greshko wrote:
On 25/04/2021 10:12, home user wrote:
I recall you once did have an option for plasma on wayland.
You still have a
/usr/share/wayland-sessions/plasmawayland.desktop
file?
yes:
bash.2[~]: su - Password: -bash.1[~]: ls -l /usr/share/wayland-sessions/plasmawayland.desktop -rw-r--r--. 1 root root 2769 Jan 14 10:46 /usr/share/wayland-sessions/plasmawayland.desktop -bash.2[~]:
While my first priority is wayland in gnome, I definitely do want: * wayland in plasma (when it's available and the default for plasma); * to retain gnome on Xorg; and * to retain plasma on xorg as fall-backs. This week certainly taught me the value of fall-backs. I was glad I still had windows-7 (slow, cumbersome, and bad as it is) on this workstation. And I do occasionally use plasma.
F34 makes Plasma (Wayland) default.
I won't be upgrading to F-34 until October. I can wait until then for plasma on wayland.
You may want to consider using sddm, at least to test.
Is that for testing gnome on wayland or plasma on wayland or both?
I'll have to hold on this until tomorrow evening. (I have no desire to watch the oscars anyway!)
With the list of "desktop" files you showed in an earlier post it would see you have installed quite a bit of "extras".
On my "work" system I keep it simple and have only
[egreshko@meimei ~]$ ll -1 /usr/share/wayland-sessions /usr/share/xsessions /usr/share/wayland-sessions: total 4 -rw-r--r--. 1 root root 2769 Jan 15 01:46 plasmawayland.desktop
/usr/share/xsessions: total 4 -rw-r--r--. 1 root root 2611 Jan 15 01:46 plasma.desktop
Is that causing problems? I'm not sure why I did that. But most likely, because I was not familiar with the various desktops, I gave myself the chance to try all (or most?) of them and choose. I routinely use ksysguard and k3b (from plasma), and caja (from mate).
One of the major reasons I'm pursuing wayland is that I'm concerned that Xorg will eventually be "sunset". I don't want to be caught off-guard. I'd like to be ready well in advance. I'm not dissatisfied with Xorg.
On 25/04/2021 10:12, home user wrote:
Read Tim's comment again. "He" wrote the notes in a file. Not dnf or another process.
The basis for my question was the sentence "I think the installation routine leaves a couple of files about the install in there, itself.". But I might easily be misunderstanding what Tim means by "the installation routine".
He meant, when you initially install Fedora on your system. That leaves 2 files behind.
anaconda-ks.cfg and initial-setup-ks.cfg
On 25/04/2021 11:06, home user wrote:
Is that causing problems? I'm not sure why I did that. But most likely, because I was not familiar with the various desktops, I gave myself the chance to try all (or most?) of them and choose. I routinely use ksysguard and k3b (from plasma), and caja (from mate).
OK. FWIW, I tend not to use GUI apps from other distributions. And, if I want to "experience" other desktops I install them in dedicated Virtual Machines.
One of the major reasons I'm pursuing wayland is that I'm concerned that Xorg will eventually be "sunset". I don't want to be caught off-guard. I'd like to be ready well in advance. I'm not dissatisfied with Xorg.
I suspect I will reach EOL before Xorg is retired. :-)
On Sun, 2021-04-25 at 06:24 +0800, Ed Greshko wrote:
Read Tim's comment again. "He" wrote the notes in a file. Not dnf or another process.
Not quite. I've written my post-installation-log, but the system has written anaconda-ks.cfg files on two installations, and left me a initial-setup-ks.cfg file on one of them.
As I've dnf installed things, I've copied and pasted my commands into my log file. And I'll write any other notes I think will help me.
On Sat, 2021-04-24 at 19:41 -0600, home user wrote:
general FYI for everyone...
This Fedora docs web page says that wayland has been the default gnome display server, and has been the default since F-25: "https://docs.fedoraproject.org/en-US/quick-docs/configuring-xorg-as-default-... ".
Yes, the default for Gnome, but not for KDE/Plasma. That's about to change.
poc
On 4/24/21 10:02 PM, Ed Greshko wrote:
On 25/04/2021 10:12, home user wrote:
Read Tim's comment again. "He" wrote the notes in a file. Not dnf or another process.
The basis for my question was the sentence "I think the installation routine leaves a couple of files about the install in there, itself.". But I might easily be misunderstanding what Tim means by "the installation routine".
He meant, when you initially install Fedora on your system. That leaves 2 files behind.
anaconda-ks.cfg and initial-setup-ks.cfg
ok. Thank-you, Ed.
On 4/25/21 1:21 AM, Tim via users wrote:
On Sun, 2021-04-25 at 06:24 +0800, Ed Greshko wrote:
Read Tim's comment again. "He" wrote the notes in a file. Not dnf or another process.
Not quite. I've written my post-installation-log, but the system has written anaconda-ks.cfg files on two installations, and left me a initial-setup-ks.cfg file on one of them.
As I've dnf installed things, I've copied and pasted my commands into my log file. And I'll write any other notes I think will help me.
ok. Thank-you, Tim.
On 4/25/21 6:21 AM, Patrick O'Callaghan wrote:
On Sat, 2021-04-24 at 19:41 -0600, home user wrote:
general FYI for everyone...
This Fedora docs web page says that wayland has been the default gnome display server, and has been the default since F-25: "https://docs.fedoraproject.org/en-US/quick-docs/configuring-xorg-as-default-... ".
Yes, the default for Gnome, but not for KDE/Plasma. That's about to change.
That's how I've been understanding that also.
On 4/25/21 6:21 AM, Patrick O'Callaghan wrote:
On Sat, 2021-04-24 at 19:41 -0600, home user wrote:
general FYI for everyone...
This Fedora docs web page says that wayland has been the default gnome display server, and has been the default since F-25: "https://docs.fedoraproject.org/en-US/quick-docs/configuring-xorg-as-default-... ".
Yes, the default for Gnome, but not for KDE/Plasma. That's about to change.
(everyone)
For all the groping around I did in the fedora docs and magazine, I could very easily have missed something.
Might Fedora possibly maybe quietly disabled or turned off wayland sometime between F-25 and F-33 inclusively?
maybe just for Nvidia graphics card users?
possibly just for certain categories of users?
On 4/25/21 7:57 PM, home user wrote:
Might Fedora possibly maybe quietly disabled or turned off wayland sometime between F-25 and F-33 inclusively?
maybe just for Nvidia graphics card users?
possibly just for certain categories of users?
No, but there's always a possibility of it not working with the NVidia drivers sometimes depending on the kernel.
Have you checked the journal to see if there's any indication of why it's not working?
On 4/25/21 9:18 PM, Samuel Sieb wrote:
On 4/25/21 7:57 PM, home user wrote:
Might Fedora possibly maybe quietly disabled or turned off wayland sometime between F-25 and F-33 inclusively?
maybe just for Nvidia graphics card users?
possibly just for certain categories of users?
No, but there's always a possibility of it not working with the NVidia drivers sometimes depending on the kernel.
Have you checked the journal to see if there's any indication of why it's not working?
That's huge.
Please give me something to look for, or a full command.
On 2021-04-25 8:21 p.m., home user wrote:
On 4/25/21 9:18 PM, Samuel Sieb wrote:
Have you checked the journal to see if there's any indication of why it's not working?
That's huge.
Please give me something to look for, or a full command.
It's hard to give anything specific and I'm not sure what state your system is in now. Which window manager is active? Are you booting to graphical or starting from text mode?
What does "journalctl -b | grep nvidia" give you?
Run "journalctl -b" and search for "wayland" (use the "/" key). Once you've started a search, you can find the next ("n") or previous ("N") match. You can type "1G" to get back to the top again. "spacebar" goes forward one page, "b" goes back on page, arrow keys for one line up or down. It's just like "less" if you know the keys for that. Search for "logind" to find a line like: "systemd-logind[1003]: New session 1 of user <username>" Look around a bit before and after that to see if anything stands out. Look for lines from "gnome-session" and "gnome-shell".
On 26/04/2021 11:52, Samuel Sieb wrote:
On 2021-04-25 8:21 p.m., home user wrote:
On 4/25/21 9:18 PM, Samuel Sieb wrote:
Have you checked the journal to see if there's any indication of why it's not working?
That's huge.
Please give me something to look for, or a full command.
It's hard to give anything specific and I'm not sure what state your system is in now. Which window manager is active? Are you booting to graphical or starting from text mode?
What does "journalctl -b | grep nvidia" give you?
I just rebooted my system. I'm running a KDE system with sddm as my DM. I then logged in picking "plasma on wayland". For me things are working.
grepping for nvidia doesn't show much. The last entry is just after booting is complete.
I'd paste the output here but I've got to figure the magic for copy/paste under wayland.
Run "journalctl -b" and search for "wayland" (use the "/" key). Once you've started a search, you can find the next ("n") or previous ("N") match. You can type "1G" to get back to the top again. "spacebar" goes forward one page, "b" goes back on page, arrow keys for one line up or down. It's just like "less" if you know the keys for that. Search for "logind" to find a line like: "systemd-logind[1003]: New session 1 of user <username>" Look around a bit before and after that to see if anything stands out. Look for lines from "gnome-session" and "gnome-shell".
Or if using KDE plasma-shell. FWIW, since the OP is getting a blank screen with plasma it may be useful to reboot the system and use "journalctl -b -1 | grep plasma-shell"
On 26/04/2021 12:29, Ed Greshko wrote:
Or if using KDE plasma-shell. FWIW, since the OP is getting a blank screen with plasma it may be useful to reboot the system and use "journalctl -b -1 | grep plasma-shell"
This may be an issue with plasma wayland in F33.
https://bugs.kde.org/show_bug.cgi?id=435673
On 4/24/21 3:30 PM, home user wrote:
On 4/22/21 4:09 PM, Patrick O'Callaghan wrote:
On Thu, 2021-04-22 at 15:29 -0600, home user wrote:
On 4/22/21 3:23 PM, Patrick O'Callaghan wrote:
Was your original installation the default Workstation or KDE Spin?
It was the default workstation (GNOME?). This would have been early spring 2013. Other workstations were added later, though I rarely use them.
I'm not sure Wayland would have been part of the default install in 2013, and if all your subsequent installs have been updates than that could explain why you don't have it.
It almost certainly was not. Patching ("dnf --refresh upgrade" or some earlier dnf/yum equivalent) indeed almost certainly would not install wayland. But it should have been a part of a semi-annual upgrade ("dnf system-upgrade download --refresh --releasever=nn" + "dnf system-upgrade reboot", or some earlier dnf/yum equivalent). I have done the semi-annual upgrade every spring and fall since the initial install in 2013. So I should have it, and it should be functioning. _______________________________________________
Your propagation from your 2013 install carried what was Fedora at that time. Since it was xorg, xorg you have. You would have had to install the packages for wayland to get them. There seems to be an assumption that everyone is doing a scratch install.
(not replying to any specific post)
I discovered last night (Sunday night US mountain time) that some of the posts to this thread went to individual members (myself and Ed) rather than to the whole list. Apparently, I clicked "Reply" rather than "Reply List" when replying to a Friday night (US mountain time) post from Ed. This propagated through several posts.
This certainly was not intentional. My apologies for my carelessness.
Brief catch-up... Ed suggested trying sddm. I installed it, and activated it. The login screen offered three gnome choices and two plasma choices. Logging in to GNOME gave me gnome on Xorg. Trying to log in to Plasma on Wayland gave me black screens; I had to do a hard shutdown. I reverted back to gdm. I'm currently using gdm, but I've not yet uninstalled sddm.
The sddm install did result in several error messages and instructions. I'll make a separate post for that, and then try the journal tests that Samuel and Ed suggested.
On 4/26/21 7:35 PM, home user wrote:
The sddm install did result in several error messages and instructions. I'll make a separate post for that, ...
On 4/24 at 8:37pm US mountain time, Ed suggested that I consider using sddm, at least to test. So I installed it. I reported back on 4/25 at 8:06pm. See the attached text file for the results of the install.
There were several error messages and two instructions. In my 4/25 at 8:06pm message, I asked:
What do I do about all those errors (4 pairs)?
Regarding the 2 instructions, I noted:
I think I saw the last 2 messages (about kdm.conf) before, probably when I did this past Thursday's patches. But I saw no hint of them in the dnf log file for those patches. What files am I being asked to update, and where are they?
I showed my attempts to deal with those instructions:
--------------- -bash.7[~]: cd /usr/lib/tmpfiles.d/ -bash.8[tmpfiles.d]: cat kdm.conf d /var/run/kdm/ 0711 root root - d /var/run/xdmctl 0711 root root - -bash.9[tmpfiles.d]: cd /var/run/kdm -bash.10[kdm]: ls -bash.11[kdm]: ls -la total 0 drwx--x--x. 2 root root 40 Apr 25 12:57 . drwxr-xr-x. 57 root root 1520 Apr 25 19:49 .. -bash.12[kdm]: cd /run/kdm -bash.13[kdm]: ls -la total 0 drwx--x--x. 2 root root 40 Apr 25 12:57 . drwxr-xr-x. 57 root root 1520 Apr 25 19:49 .. -bash.14[kdm]: ---------------
In a 4/25 at 8:23pm message, Ed responded regarding the instructions:
I wouldn't worry about them. /usr/lib/tmpfiles.d/kdm.conf is related to the old display manager of KDE, kdm.
I would just enable sddm and see how it goes.
Other than the 2 instructions is there anything in the sddm installation output that I should give attention to?
Now on the the journal tests. This will take a while.
On 4/25/21 10:29 PM, Ed Greshko wrote:
On 26/04/2021 11:52, Samuel Sieb wrote:
... What does "journalctl -b | grep nvidia" give you? ... Run "journalctl -b" and search for "wayland" ... Search for "logind" to find a line like: "systemd-logind[1003]: New session 1 of user <username>" ... Look for lines from "gnome-session" and "gnome-shell". ...
I switched to sddm and rebooted.
The menu had one GNOME entry at the top, a few other entries, GNOME Classic, GNOME on Xorg, GNOME, and finally other entries. I tried the top GNOME entry a few times on 3 different accounts (root, admin, and bill). Each time, I was returned to the login screen. I changed my choice to the GNOME entry below GNOME on Xorg. That login was successful.
I saved the text of the journal file by doing "journalctl -b > jgnome.txt". I then put it on the google drive. Here's the link to the journal file: "https://drive.google.com/file/d/1tbS1WIKzAWQeiPvVMFgevywRJpD0BMUL/view?usp=s...".
I did all the suggested searches. Overall, a lot of hits, especially for "gnome-session". What I saw in the lines surrounding the hits did not make adequate sense to me. I did the searches again by doing "journalctl -b | grep -i -n [string]", and then pasted the results into another text file. I then put it on the google drive. Here's the link to the search results: "https://drive.google.com/file/d/1dssC9qrVdizFUGVHHuYV7U4-O2xq_VNY/view?usp=s...".
Next, I'll again switch to sddm, and try Plasma on wayland.
On 4/25/21 10:29 PM, Ed Greshko wrote:
Or if using KDE plasma-shell. FWIW, since the OP is getting a blank screen with plasma it may be useful to reboot the system and use "journalctl -b -1 | grep plasma-shell"
Once again, I switched to sddm and rebooted. In the login screen, I chose Plasma on wayland and logged in. The screen went black. After a minute, I did a hard shutdown. I booted into run-level 3, switched back to gdm, shutdown, booted up, and logged in.
I saved the text of the journal file by doing "journalctl -b -1 > jplasma.txt". I then put it on the google drive. Here's the link to the journal file: "https://drive.google.com/file/d/1HvBJI4RxSqXoAzTwSYtE69iiSQwvXdRo/view?usp=s...".
I did some searches. What I saw in the lines surrounding the hits did not make adequate sense to me. I did the searches again by doing "journalctl -b -1 | grep -i -n [string]", and then pasted the results into another text file. I then put it on the google drive. Here's the link to the search results: "https://drive.google.com/file/d/1KYzKYyoRD36e5X6wbER-1-eIqsqz3EQI/view?usp=s...".
On 27/04/2021 12:16, home user wrote:
On 4/25/21 10:29 PM, Ed Greshko wrote:
Or if using KDE plasma-shell. FWIW, since the OP is getting a blank screen with plasma it may be useful to reboot the system and use "journalctl -b -1 | grep plasma-shell"
Once again, I switched to sddm and rebooted. In the login screen, I chose Plasma on wayland and logged in. The screen went black. After a minute, I did a hard shutdown. I booted into run-level 3, switched back to gdm, shutdown, booted up, and logged in.
I saved the text of the journal file by doing "journalctl -b -1 > jplasma.txt". I then put it on the google drive. Here's the link to the journal file: "https://drive.google.com/file/d/1HvBJI4RxSqXoAzTwSYtE69iiSQwvXdRo/view?usp=s...".
I did some searches. What I saw in the lines surrounding the hits did not make adequate sense to me. I did the searches again by doing "journalctl -b -1 | grep -i -n [string]", and then pasted the results into another text file. I then put it on the google drive. Here's the link to the search results: "https://drive.google.com/file/d/1KYzKYyoRD36e5X6wbER-1-eIqsqz3EQI/view?usp=s...".
None of your logs make any sense. First off, you should be grepping for plasmashell. But that doesn't matter as there is no plasmashell in jplasma.txt. Also, there is no sign of sddm or a user login.
Booting a VM system here. (Not using nvidia, but not relevant). I would see....
Apr 27 13:42:58 f33k.greshko.com systemd-logind[678]: New session c1 of user sddm. Apr 27 13:42:58 f33k.greshko.com systemd[1]: Created slice User Slice of UID 990.
when sddm is starting. Note....
[egreshko@f33k ~]$ grep sddm /etc/passwd sddm:x:990:983:Simple Desktop Display Manager:/var/lib/sddm:/sbin/nologin
And I see no sign of a user login other than root.
But now I just realized you said....
I did a hard shutdown. I booted into run-level 3, switched back to gdm, shutdown, booted up,
So..... You have posted the wrong journal times. You need to keep track of the wall clock and boot times. Since you booted a second time you'd want -b -2.
journalctl --list-boots
will show all the boots and their relative numbers and start times that are currently in the journal logs.
journalctl --list-boots
will show all the boots and their relative numbers and start times that are currently in the journal logs.
Also, when you do things such as login/logout you can note the time and then use the -S (or --since) and -U (or --until) to narrow the output of journalctl to keep the amount of captured info to a minimum and to the time relevant to events you're interested in. That would eliminate the guesswork of others who may decide to look at your logs. a.k.a. making life easier on others.
On 4/27/21 12:30 AM, Ed Greshko wrote:
So..... You have posted the wrong journal times. ... Since you booted a second time you'd want -b -2.
(sigh) You're right. You got me. I'm busted.
The old plasma files have been removed from the google drive.
I switched to sddm, re-booted, chose plasma on wayland, and tried to log in as bill. The screens went black. After a minute, I did the hard shutdown. After booting up to run-level 3, I logged in as root, put the display manager back to gdm, and shutdown. I booted up, logged in to root to do the journal and grep work.
I saved the text of the journal file by doing "journalctl -b -2 > jplasma.txt". Here's the link to the journal file on the google drive: "https://drive.google.com/file/d/1oGjdw4kzDbJ5Qw8mfRi_jWkB9TUb66SD/view?usp=s..."
I did some searches. What I saw in the lines surrounding the hits did not make adequate sense to me. I did the searches again by doing "journalctl -b -2 | grep -i -n [string]", and then pasted the results into another text file. Here's the link to the search results on the google drive: "https://drive.google.com/file/d/1_iqm6DoC4YW3HmyOtQBe1ihggExVmW7T/view?usp=s...".
Oh, I forgot to grep for "sddm". grep -i sddm jplasma.txt gives 245 matching lines. And grep -i sddm /etc/passwd done as root gives sddm:x:960:957:Simple Desktop Display Manager:/var/lib/sddm:/sbin/nologin but that was done after switching back to gdm (if that matters).
The times at the beginning of both journal files caught my attention too. After several seconds pondering that, it occurred to me that those times are exactly 6 hours off, I'm in mountain time, and this place is in daylight savings time. So something is not handling time zone correction properly.
I don't know if the problems start before or after login. My "wall clock" does not show seconds, and its minutes don't exactly match the workstation's time display. So it seems best to leave all lines in the journal file.
On 28/04/2021 02:12, home user wrote:
The times at the beginning of both journal files caught my attention too. After several seconds pondering that, it occurred to me that those times are exactly 6 hours off, I'm in mountain time, and this place is in daylight savings time. So something is not handling time zone correction properly.
I don't know if the problems start before or after login. My "wall clock" does not show seconds, and its minutes don't exactly match the workstation's time display. So it seems best to leave all lines in the journal file.
Sounds like things are working correctly.
For consistent times in logs your system's hardware clock needs to be set to GMT/UTC. Then log entry times will almost always be in your local time. There may be exceptions to that, can't think of it at the moment, but if that is the case the time will have a Z appended to it.
I don't know if this is still the case since I don't dual boot. But I seem to recall that MS windows would reset the HW clock to the local time. That is why, unless absolutely necessary, I've only run Windows in a Virtual Machine for the longest time.
When it comes to "wall clock" it is quite simple. The best "wall clock" is your mobile phone. Unless you've changed it, mobile phones default to syncing time with the carrier whose time is also synced with the world. And it isn't necessary to have "seconds" accuracy. Minute is sufficient. You hit "return" at login when the minute reading on the phone changes. That way you know within a second or two when to start scanning the logs.
On 4/26/21 11:51 AM, Robert McBroom via users wrote:
On 4/24/21 3:30 PM, home user wrote:
On 4/22/21 4:09 PM, Patrick O'Callaghan wrote:
On Thu, 2021-04-22 at 15:29 -0600, home user wrote:
On 4/22/21 3:23 PM, Patrick O'Callaghan wrote:
Was your original installation the default Workstation or KDE Spin?
It was the default workstation (GNOME?). This would have been early spring 2013. Other workstations were added later, though I rarely use them.
I'm not sure Wayland would have been part of the default install in 2013, and if all your subsequent installs have been updates than that could explain why you don't have it.
It almost certainly was not. Patching ("dnf --refresh upgrade" or some earlier dnf/yum equivalent) indeed almost certainly would not install wayland. But it should have been a part of a semi-annual upgrade ("dnf system-upgrade download --refresh --releasever=nn" + "dnf system-upgrade reboot", or some earlier dnf/yum equivalent). I have done the semi-annual upgrade every spring and fall since the initial install in 2013. So I should have it, and it should be functioning. _______________________________________________
Your propagation from your 2013 install carried what was Fedora at that time. Since it was xorg, xorg you have. You would have had to install the packages for wayland to get them. There seems to be an assumption that everyone is doing a scratch install.
In a previous post to this thread, I listed the wayland-related packages that I have. No one gave any hint that I'm missing anything. I'd like to see authoritative confirmation and/or correction of what you seem to be saying about new things added to Fedora and whether manual install is required to get them.
Our conversational jargon and dnf jargon seem to have been mixed up a few times in this thread: patch vs. update vs. upgrade. I wish we had a good solution.
On 4/27/21 1:49 PM, Ed Greshko wrote:
On 28/04/2021 02:12, home user wrote:
The times at the beginning of both journal files caught my attention too. After several seconds pondering that, it occurred to me that those times are exactly 6 hours off, I'm in mountain time, and this place is in daylight savings time. So something is not handling time zone correction properly.
I don't know if the problems start before or after login. My "wall clock" does not show seconds, and its minutes don't exactly match the workstation's time display. So it seems best to leave all lines in the journal file.
Sounds like things are working correctly.
For consistent times in logs your system's hardware clock needs to be set to GMT/UTC. Then log entry times will almost always be in your local time. There may be exceptions to that, can't think of it at the moment, but if that is the case the time will have a Z appended to it.
I don't know if this is still the case since I don't dual boot. But I seem to recall that MS windows would reset the HW clock to the local time. That is why, unless absolutely necessary, I've only run Windows in a Virtual Machine for the longest time.
Your reply is bringing up a vague memory that I probably had a thread on this list 7 - 8 years ago regarding the system clock. I haven't changed anything relating to the clock since then.
When it comes to "wall clock" it is quite simple. The best "wall clock" is your mobile phone. Unless you've changed it, mobile phones default to syncing time with the carrier whose time is also synced with the world. And it isn't necessary to have "seconds" accuracy. Minute is sufficient. You hit "return" at login when the minute reading on the phone changes. That way you know within a second or two when to start scanning the logs.
I do not have a cell phone. I can afford neither a mobile phone nor mobile phone service.
On 28/04/2021 09:24, home user wrote:
On 4/27/21 1:49 PM, Ed Greshko wrote:
On 28/04/2021 02:12, home user wrote:
Your reply is bringing up a vague memory that I probably had a thread on this list 7 - 8 years ago regarding the system clock. I haven't changed anything relating to the clock since then.
Then the next time you boot enter your hardware's setup mode and check the time. Make sure that it is set to GMT/UTC time (HW generally has no knowledge of time zones). Also, do that each time you've also run Windows to make sure it hasn't altered your settings.
I have no recollection of what you've done last year, let along 8 years ago.
When it comes to "wall clock" it is quite simple. The best "wall clock" is your mobile phone. Unless you've changed it, mobile phones default to syncing time with the carrier whose time is also synced with the world. And it isn't necessary to have "seconds" accuracy. Minute is sufficient. You hit "return" at login when the minute reading on the phone changes. That way you know within a second or two when to start scanning the logs.
I do not have a cell phone. I can afford neither a mobile phone nor mobile phone service.
Then set whatever clock you have close enough to your system's time such that when the time advances to the next minute on the system you can start whatever process you want to track and make note of it. That way you have a better chance to correlate what is in the logs to the actions taken.
That being said. I found nothing interesting in jplasma.txt. There are no blatant errors. There is no suggestion that some package or executable is missing or lacking.
The system does record....
Apr 27 11:05:05 coyote sddm-helper[1102]: Starting: "/etc/sddm/wayland-session /usr/libexec/plasma-dbus-run-session-if-needed /usr/bin/startplasma-wayland"
But not much else after that suggests the process is proceeding to full completion. e.g. I don't see any entry for ksmserver. It would seem something may be "hanging" or "waiting". But you'd need to switch to another virtual terminal to look into it more.
I don't know if it worth it. But you may want to create a new user to see if that "virgin" user exhibits the same.
Other than that, I've no more to add to this thread.
On 4/27/21 8:20 PM, Ed Greshko wrote:
Then the next time you boot enter your hardware's setup mode and check the time. Make sure that it is set to GMT/UTC time (HW generally has no knowledge of time zones). Also, do that each time you've also run Windows to make sure it hasn't altered your settings.
done.
I don't know if it worth it. But you may want to create a new user to see if that "virgin" user exhibits the same.
I said in another thread I would submit bugs regarding system problem reports and SELinux alerts. I should get those done. Then I'll come back to the new user test.
Other than that, I've no more to add to this thread.
The "Wayland and Nvidia" thread that Patrick started yesterday has shifted my thinking. I now think it best to wait until I upgrade to F-34 (this coming fall), and (depending on how soon current F-34 wayland issues are resolved) maybe even F-35 (next spring). (But I will try the new user test before that.)
On Wed, 2021-04-28 at 10:23 -0600, home user wrote:
On 4/27/21 8:20 PM, Ed Greshko wrote:
Then the next time you boot enter your hardware's setup mode and check the time. Make sure that it is set to GMT/UTC time (HW generally has no knowledge of time zones). Also, do that each time you've also run Windows to make sure it hasn't altered your settings.
done.
I don't know if it worth it. But you may want to create a new user to see if that "virgin" user exhibits the same.
back to the new user test.
Other than that, I've no more to add to this thread.
The "Wayland and Nvidia" thread that Patrick started yesterday has shifted my thinking. I now think it best to wait until I upgrade to F-34 (this coming fall), and (depending on how soon current F-34 wayland issues are resolved) maybe even F-35 (next spring). (But I will try the new user test before that.)
You might want to monitor:
https://community.kde.org/Plasma/Wayland_Showstoppers
which mentions some current issues. I'm back on X11 until these are dealt with.
poc
On 28 Apr 2021, at 22:40, Patrick O'Callaghan pocallaghan@gmail.com wrote:
On Wed, 2021-04-28 at 10:23 -0600, home user wrote:
On 4/27/21 8:20 PM, Ed Greshko wrote:
Then the next time you boot enter your hardware's setup mode and check the time. Make sure that it is set to GMT/UTC time (HW generally has no knowledge of time zones). Also, do that each time you've also run Windows to make sure it hasn't altered your settings.
done.
I don't know if it worth it. But you may want to create a new user to see if that "virgin" user exhibits the same.
back to the new user test.
Other than that, I've no more to add to this thread.
The "Wayland and Nvidia" thread that Patrick started yesterday has shifted my thinking. I now think it best to wait until I upgrade to F-34 (this coming fall), and (depending on how soon current F-34 wayland issues are resolved) maybe even F-35 (next spring). (But I will try the new user test before that.)
You might want to monitor:
https://community.kde.org/Plasma/Wayland_Showstoppers https://community.kde.org/Plasma/Wayland_Showstoppers
which mentions some current issues. I'm back on X11 until these are dealt with.
One more data point. I use a GTX 1060 6GiB and when I login with nouveau drivers the mouse and display freeze very soon (a few seconds) after I see the first desktop draw.
I emailed the fedora KDE list with more details.
Barry
-
Barry Scott -
Ed Greshko -
George N. White III -
home user -
Joe Zeff -
Patrick O'Callaghan -
Robert McBroom -
Samuel Sieb -
Tim