You are not logged in.
Hey,
I've been struggling with system freezes when I update my system via paru or pacman now for a while. The system freezes fully and I cannot do anything, not even trigger SysRq.
The last two times this happened when updating systemd, but I'm fairly sure that I've had this issue with other updates. Annoyingly, I cannot replicate the freeze if I downgrade systemd, reboot, and the update it again.
I'm running:
- The linux kernel.
- A LUKS encrypted EXT4 drive.
- zram swap.
I've ran memtest86+ and it passed without any errors. SMART also passes without any errors. Gist with some system information: https://gist.github.com/sQVe/504795a305 … b0d4b606f0. Tell me if I need to add something else.
I am now at my wits end, and don't know how to fix this - other than starting a fresh install.
Any input would be greatly appreciated
Last edited by sQVe (2024-04-30 11:54:30)
Offline
I just now got another freeze, that isn't related to pacman. I disabled TLP and it froze. Super weird.
Offline
I have now disabled TLP, removed setting swappiness to 1, disabled swap, and enabled the systemd-oomd.service. Unsure if this will help but we'll see...
Is the OOMD service supposed to be disabled by default?
Last edited by sQVe (2024-04-30 13:51:32)
Offline
Do you have an nvidia GPU (and use the nvidia 550xx proprietary driver)?
On a formal note, please don't bump your thread but edit your previous post if nobody has yet replied.
Offline
Do you have an nvidia GPU (and use the nvidia 550xx proprietary driver)?
Yes, I'm running exclusively on my NVIDIA GPU at the moment. I'm running version 550.67-1, since I had xrandr problems with some monitor setups with 550.76-1. Can this somehow be related to the freezes?
On a formal note, please don't bump your thread but edit your previous post if nobody has yet replied.
Ah, I didn't think that it would be an issue. Sorry, it won't happen again.
Offline
It's the clusterfuck of the last months, https://bbs.archlinux.org/viewtopic.php … 0#p2167660
Try https://aur.archlinux.org/packages/nvidia-535xx-dkms
Offline
It's the clusterfuck of the last months, https://bbs.archlinux.org/viewtopic.php … 0#p2167660
Try https://aur.archlinux.org/packages/nvidia-535xx-dkms
I will try that, thank you!
Do you still think it's worth disabling the things I mentioned earlier? It makes the issue harder to narrow down, I guess.
EDIT: Do I need to run the LTS kernel to run nvidia-535xx?
Last edited by sQVe (2024-04-30 16:11:06)
Offline
No, the version in the AUR should have a patch to build against the 6.8 kernels.
If you want to use the older version from the ALA you'lll however have to use the LTS kernel since it uses a symbol that got GPL protected and the module will fail to build.
Generally zswap XOR zram (don't use both at the same time) should augment a regular swap (in doubt a swapfile), as preferred swap, not replace it - you're not getting that much more space out of them.
Offline
No, the version in the AUR should have a patch to build against the 6.8 kernels.
If you want to use the older version from the ALA you'lll however have to use the LTS kernel since it uses a symbol that got GPL protected and the module will fail to build.Generally zswap XOR zram (don't use both at the same time) should augment a regular swap (in doubt a swapfile), as preferred swap, not replace it - you're not getting that much more space out of them.
Awesome, thank you for the input.
I have enabled zswap, and I'm going to run with that for a while. I'm mainly questioning if I should disable systemd-oom and enable TLP again.
Offline
No and yes.
OOM kills will be logged (and you'll kinda notice if your browser suddenly disappears
and while I don't think that TLP is the cause, if you pair it w/ the 535xx drivers an experience the halts again, it's becoming more likely the culprit.
Otherwise you just know (at best) that the system is now stable, but you don't know whether it's because of nvidia or tlp and you also don't know whether you can take the benefits from tlp.
Offline