| 10 Nov 2024 |
K900 | In reply to@eliasp:kde.org
I'm on current unstable and it seems, no Plasma user units at all are linked into the current environment... They're all present in the store, but none of them are present in /etc/systemd/user, none of them are shown in "systemctl --user list-units 'plasma'" ... This also means, logging in via sddm results in nothing except of a hanging sddm. Didn't do any changes to my config, just updated my flake inputs (and unfortunately, due to lack of space, deleted old generations...).
Anyone else noticed something similar? Oh wait, that's fine | 14:36:12 |
K900 | They're in /run/current-system/sw/share/systemd/user | 14:36:20 |
eliasp | Oh, you're right. They're there - so I'll have to figure out what else causes my user-session being unable to "see" those units. | 14:51:42 |
eliasp | But it looks like I've broken something locally/within my user config... Reverted back to an old flake.lock from my backup and the result is still the same 😜 | 14:53:21 |
eliasp | Hmm, but other local users (that haven't been active for a few weeks) on the same device experience the same issue... I feel like hunting ghosts | 14:54:50 |
mjm | maybe you have something in your home directory that’s overriding those global ones? | 14:58:08 |
mjm | you might if you messed around with systemctl enable/disable | 14:58:31 |
eliasp | That's why I just also checked with other users... | 14:58:40 |
mjm | Oh I misread that part sorry | 14:59:46 |
eliasp | must be something stupid on my part... Will have to re-check my flake/flake.lock etc | 15:00:29 |
K900 | Something unsetting/not setting XDGDATADIRS? | 15:02:50 |
eliasp | The last change I did to my config before I saw this breakage, was to customize a user slice, but since I reverted that change I can't imagine where this would've caused a persistent change affecting even users that didn't have a session while this change was active... | 15:10:07 |
eliasp | XDG_DATA_DIRS looks good in an affected session | 15:11:47 |
eliasp | Oh: displays manager's journal tells me, it starts xserver-wrapper, even though I've selected the Wayland session... Maybe something wrong within sddm...? Does it persist anything?
Will have to check later... Have to run now | 15:14:53 |
eliasp | sddm's state.conf looks good... | 15:15:34 |
K900 | SDDM runs X by default | 15:17:29 |
K900 | For the greeter itself | 15:17:47 |
eliasp | Ah, right... That explains it | 15:20:30 |
K900 | Today on things I did not expect to be touching: samba | 18:14:33 |
K900 | https://invent.kde.org/utilities/konsole/-/merge_requests/1042 | 18:26:40 |
K900 | YESSSSSSSSSSS | 18:26:42 |
K900 | Finally cgroups per tab in Konsole landed properly | 18:26:48 |
K900 | Finally I can oomd | 18:26:51 |
mjm | oooo | 18:27:41 |
K900 | Also forgot to link my samba touching: https://github.com/NixOS/nixpkgs/pull/355031 | 18:27:53 |
aloisw | In reply to @k900:0upti.me Finally cgroups per tab in Konsole landed properly Weren't they already introduced some months ago? | 19:13:34 |
aloisw | I might be confusing things with tmux, but I'm pretty sure Konsole added some systemd integration too. | 19:14:18 |
K900 | It used to only create cgroups if you had a memory limit set in the profile | 19:19:53 |
K900 | Now it just does it all the time | 19:19:58 |
eliasp | In reply to @k900:0upti.me https://invent.kde.org/utilities/konsole/-/merge_requests/1042 Oh, I was having some thoughts on a quite related topic today... https://mastodon.social/@eliasp/113458175173585295 will have to check out per-tab cgroups once everything is back up & running again here | 19:42:44 |