packages disappeared from start menu and no sessions in sddm

gabespound · 2024-10-26 01:59:12

arch 6.11.5
kde plasma 6.2.2-1

tldr: programs started disappearing from start menu during very long (4 hour) ffmpeg-full build and install. eventually firefox told me i needed to restart and when i tried to open it again it was also gone from the start menu. at the end of the ffmpeg install i had no programs according to start menu. i restarted and found i couldn’t log in because sddm has no sessions listed (i have wayland and x11, default wayland). when i use a new tty i can see all my files and programs are there and should be fine.

i was noticing strange behavior today regarding ffmpeg, no Machine and some lib nx library i don’t recognize.

first: i downloaded Elder scrolls online with lutris and went to play, but kept getting errors related to lib NX. chat gpt suggested uninstalling noMachine, a virtual desktop client/server, and that fixed my problem.

later that day i saw the very same lib NX errors return, but instead in my terminal, and my terminal would default to /usr/bin rather than the usual ~ directory. then, when i started to use ffmpeg for some software im working on, a script that was working a couple weeks ago and had been for months prior all of a sudden suggested the “preset” flag was not recognized. chatGPT suggests this can happen when ffmpeg is built without libx264, but again i had been using this for awhile.

i tried uninstalling and reinstalling ffmpeg to no avail. i figured maybe obs or another app that uses ffmpeg may have updated and changed its version/build. i then tried installing ffmpeg-full which took about 4 hours to download make and build. over the next 4 hours i kept getting more frequent and strange errors regarding missing files or programs, and i could verify that they were not missing in terminal but kde start menu couldn’t see them. dolphin file explorer couldn’t open many directories, lutris stopped being able to see wine or any other installed dependencies. everything was installed, but they didn’t know about eachother. by the end of the ffmpeg-full install, start menu said i had no installed programs. on restart, sddm said i had no sessions (previously i had default wayland and also x11 as a backup). i can access all my programs and files just fine in a new tty, but cant access any sort of desktop environment.

How can i figure out what went wrong, and how can i figure out how to fix it

Anax · 2024-10-26 09:17:54

Can you post more Information?
Like How did you compile it?
Did you use an Build Script?
What was the build command?

V1del · 2024-10-26 09:26:28

Sounds like the disk is dying, check/post the links you get from

sudo journalctl -b | curl -F 'file=@-' 0x0.st #-b for the current boot, if this actively happened for the boot prior also throw in -b-1 for example or -2 -3 for respective prior boots
#Install smartmontools
sudo smartctl -a /dev/sdX | curl -F 'file=@-' 0x0.st #Replace with actual device path of the disk
sudo pacman -Qkk 2> /tmp/howbadisit.txt
curl -F 'file=@-' 0x0.st < /tmp/howbadisit.txt

Generally speaking, ffmpeg can be considered a core system library, you should really think thrice before opting to replace it with something from the AUR, especially if your asking chatGPT for a solution you're not going to be aware of the consequences of doing something like this. Another thing is NoMachine has a track record of breaking GL lookup paths and not properly cleaning them up on uninstall, so check whether you didn't run https://bbs.archlinux.org/viewtopic.php … 6#p2153376 and thus have wrong permissions for accessing the GPU

Last edited by V1del (2024-10-26 09:32:10)

gabespound · 2024-10-26 20:27:24

Anax wrote:

Can you post more Information?
Like How did you compile it?
Did you use an Build Script?
What was the build command?

i ran yay ffmpeg-full and selected it from the list of packages. the rest was done automatically.

i dont think this has to do with ffmpeg after reading the post linked by @V1del. ffmpeg was just the first thing i was messing with after what caused the issue which i believe to be no machine, however.

V1del wrote:

Another thing is NoMachine has a track record of breaking GL lookup paths and not properly cleaning them up on uninstall, so check whether you didn't run https://bbs.archlinux.org/viewtopic.php … 6#p2153376 and thus have wrong permissions for accessing the GPU

i read through this post and followed any instructions relevant to my setup, and the issue wasn’t fixed.
first i had no machine uninstalled, so i reinstalled it and ran the unconfig script. then i uninstalled nomachine and restarted. this didn’t work so i tried just installing nomachine and leaving it installed. this didnt work either. i also was unable to follow any steps related to adding myself to vglusers or deleting that group because it doesnt exist on my system.

i just tried reinstalling all my packages with a force overwrite, this did not work either. what commands can i run to gain some insight into what’s going wrong on my machine

V1del · 2024-10-26 20:41:49

The list of commands I gave you and the links they generate are needed to look at what's going on.

gabespound · 2024-10-26 23:28:29

@V1del

I thought those were just relating to my disk health, which i had already verified wasn’t the issue. i have 2 nvme drives setup in raid 0 using mdadm.

http://0x0.st/XUwu.txt
http://0x0.st/XUwo.txt

Last edited by gabespound (2024-10-26 23:31:17)

seth · 2024-10-27 07:35:12

disk health, which i had already verified wasn’t the issue

How?

/dev/md127: Unable to detect device type

Cause not like that.

V1del wrote:

sudo pacman -Qkk 2> /tmp/howbadisit.txt
curl -F 'file=@-' 0x0.st < /tmp/howbadisit.txt

i have 2 nvme drives setup in raid 0 using mdadm

https://en.wikipedia.org/wiki/RAID_0

RAID 0 provides no fault tolerance or redundancy, the failure of one drive will cause the entire array to fail

Oct 26 16:09:29 gabesArch systemd-journald[496]: File /var/log/journal/8098363ba87c4b54a37f2599dc2121cb/system.journal corrupted or uncleanly shut down, renaming and replacing.
Oct 26 16:09:30 gabesArch kernel: exFAT-fs (sda1): Volume was not properly unmounted. Some data may be corrupt. Please run fsck.
Oct 26 16:09:30 gabesArch kernel: FAT-fs (nvme1n1p1): Volume was not properly unmounted. Some data may be corrupt. Please run fsck.

Do you reboot w/ the power button? Is there a parallel OS (windows)?

The symptoms imply a FS corruption, it's either the FS, the raid or the raid members (and it's hard to sync data when the power suddenly drops out) - there's nothing in that 17 minute journal to explain any file loss, so also please post

sudo journalctl -b -1 | curl -F 'file=@-' 0x0.st

the journal of the previous boot so we can see a shutdown process.

gabespound · 2024-10-27 22:46:52

seth wrote:

i have 2 nvme drives setup in raid 0 using mdadm
https://en.wikipedia.org/wiki/RAID_0
RAID 0 provides no fault tolerance or redundancy, the failure of one drive will cause the entire array to fail
Oct 26 16:09:29 gabesArch systemd-journald[496]: File /var/log/journal/8098363ba87c4b54a37f2599dc2121cb/system.journal corrupted or uncleanly shut down, renaming and replacing.
Oct 26 16:09:30 gabesArch kernel: exFAT-fs (sda1): Volume was not properly unmounted. Some data may be corrupt. Please run fsck.
Oct 26 16:09:30 gabesArch kernel: FAT-fs (nvme1n1p1): Volume was not properly unmounted. Some data may be corrupt. Please run fsck.
Do you reboot w/ the power button? Is there a parallel OS (windows)?

1) i do often reboot with the power button
2) the not properly mounted drive sda1 is a usb drive i have in my fstab. it is there so it auto mounts i use it for key storage.
3) nvmen1p1 is a spare boot drive, i have two of them i believe i may of just never properly set up thre spare. my drive config is as follows

nvme[0-1]n1 are 2tb nvme drives.
they each have 256gb partitions (nvme[0-1]n1p2) reserved for other OSs (currently windows and fedora)
they each have a 1.6TB partition (nvme0n1p4 and nvme1n1p3) and they make up md127 with mdadm in raid 0.

my lsblk looks like this.

https://imgur.com/a/P7Qs15F

interestingly i don’t recognize the nvme0n1p3 partition, not sure where that came from, fairly certain i didn’t do that.

seth wrote:

The symptoms imply a FS corruption, it's either the FS, the raid or the raid members (and it's hard to sync data when the power suddenly drops out) - there's nothing in that 17 minute journal to explain any file loss, so also please post
sudo journalctl -b -1 | curl -F 'file=@-' 0x0.st
the journal of the previous boot so we can see a shutdown process.

http://0x0.st/XU0R.txt

does the fact that my system works fine via terminal and all my files seem to be here as well as the “missing programs” not suggest my drives are healthy? they are about a year and a half old which i realize guarantees nothing but i haven’t seen any signs that it is the drives. it seems very related to the nomachine uninstall as that is precisely when i began having issues.

seth · 2024-10-28 06:50:40

1) i do often reboot with the power button

Which is gonna cause file corruption because the filesystem will not sync to disk.
Stop doing that for crying out loud, your symptoms a perfectly expectable w/ that kind of behavior.

Also don't post screenshots of text and c/fdisk wants to see the entire path, fdisk -l /dev/nvme0n1

seth wrote:

Is there a parallel OS (windows)?

reserved for other OSs (currently windows and fedora)

3rd link below. Mandatory.
Disable it (it's NOT the BIOS setting!) and reboot windows and linux twice for voodo reasons.

all my files seem to be here as well as the “missing programs”

That largely depends on what you consider "missing programs" - cause the entries in some startmenu are not the same files as the program binaries, but desktop services in /usr/share/applications and ~/.local/share/applications

not suggest my drives are healthy?

So your health assertion is blind guesswork.
W/ the parallel windows installation and your fingers gravitation towards the power button, there's enough other explanation for the symptoms, but really check the disk health.

it seems very related to the nomachine uninstall as that is precisely when i began having issues.

i read through this post and followed any instructions relevant to my setup, and the issue wasn’t fixed.
first i had no machine uninstalled, so i reinstalled it and ran the unconfig script. then i uninstalled nomachine and restarted. this didn’t work so i tried just installing nomachine and leaving it installed. this didnt work either. i also was unable to follow any steps related to adding myself to vglusers or deleting that group because it doesnt exist on my system.

Stop flailing around, you're creating a moving target and probably some havoc on the way.

glxinfo -B
ls -l /dev/dri/card* /dev/dri/renderD*

and post your Xorg log, https://wiki.archlinux.org/title/Xorg#General

Arch Linux

#1 2024-10-26 01:59:12

packages disappeared from start menu and no sessions in sddm

#2 2024-10-26 09:17:54

Re: packages disappeared from start menu and no sessions in sddm

#3 2024-10-26 09:26:28

Re: packages disappeared from start menu and no sessions in sddm

#4 2024-10-26 20:27:24

Re: packages disappeared from start menu and no sessions in sddm

#5 2024-10-26 20:41:49

Re: packages disappeared from start menu and no sessions in sddm

#6 2024-10-26 23:28:29

Re: packages disappeared from start menu and no sessions in sddm

#7 2024-10-27 07:35:12

Re: packages disappeared from start menu and no sessions in sddm

#8 2024-10-27 22:46:52

Re: packages disappeared from start menu and no sessions in sddm

#9 2024-10-28 06:50:40

Re: packages disappeared from start menu and no sessions in sddm

Board footer