| Summary: | Unknown memory usage after screen lock | ||
|---|---|---|---|
| Product: | [Unmaintained] kscreenlocker | Reporter: | bwqr <ruzgardeniz.08> |
| Component: | general | Assignee: | Plasma Bugs List <plasma-bugs-null> |
| Status: | RESOLVED NOT A BUG | ||
| Severity: | normal | CC: | bshah, nate |
| Priority: | NOR | ||
| Version First Reported In: | 5.26.1 | ||
| Target Milestone: | --- | ||
| Platform: | Other | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed/Implemented In: | ||
| Sentry Crash Report: | |||
| Attachments: |
Memory usage after logging out plasma session
Memory usage before logging out plasma shell More detailed memory usages |
||
|
Description
bwqr
2022-10-21 15:38:06 UTC
Forgot to mention how I start the plasma, I run `dbus-run-session startplasma-wayland` in a tty console to start plasma. What process is using all that memory? System Monitor or `top` will tell you. That is the interesting part, nothing shows up in the top, htop or System Monitor. Neither running these tools as root or normal user changes anything. I also looked into slabtop, /proc/meminfo and dmesg and could not find a relevant information. Even though all tools display high memory usage, no process seems to consume that much memory. What else can I do to debug this? So what's telling you that the total usage is high? Can you attach a screenshot of it? When I encounter the same behavior, I will take screenshots and upload them. But I can say that `used` tab of the `free -h` command, `used mem` section of the `top` command and `mem` bar in the htop shows similar memory usages which is really high. All this observations are made in the tty console after exitting the plasma session. There are so few processes in the top or htop, and none of them uses that much memory. With high memory usage, I mean 10GB memory out of 32GB is used by something. Created attachment 153116 [details]
Memory usage after logging out plasma session
Here is a screenshot of htop after logging out plasma session when this behavior occurs.
Created attachment 153117 [details]
Memory usage before logging out plasma shell
And memory usage in htop before logging out plasma shell when this behavior occurs.
Created attachment 153124 [details]
More detailed memory usages
Here is a more detailed capture of memory usages. I have combined multiple htop screens into one image. Image at the top is captured first and the image at the bottom is captured last. Here are the things I have done before taking the specific image, the numbers indicate the image.
1. Computer resumed from suspend. Instead of unlocking the screen locker, I switched to tty2 and started the htop.
2. I switched back to tty1 and unlocked the screen locker. Workspace 1 is visible now.
3. Switched to workspace 3.
4. Killed all the applications.
5. Logged out from the session.
I also want to emphasize which application was running on which workspace, numbers indicates the workspace:
1. Firefox, chromium, evince
3. Thunderbird, rhythmbox
In the latest screenshot, your memory is mostly being used by by Thunderbird, Firefox, Evince, and Rhythmbox. Doesn't seem like there's a KDE bug there. In the one that shows multiple plasmashell processes open, each one is using a normal-ish amount of memory; the bug is that you have more than one running? How did that happen? Are you running `dbus-run-session startplasma-wayland` multiple times without killing the old sessions? When you do things manually, you become responsible for this sort of thing. (In reply to Nate Graham from comment #9) > In the latest screenshot, your memory is mostly being used by by > Thunderbird, Firefox, Evince, and Rhythmbox. Doesn't seem like there's a KDE > bug there. I also started to think like this. The only thing related to KDE is that this behavior is triggered when screen gets locked in plasma shell. I am normally using sway and this behavior did not happen for once, the reason may be that I do not use any screen locker while using sway. > In the one that shows multiple plasmashell processes open, each one is using > a normal-ish amount of memory; the bug is that you have more than one > running? How did that happen? Are you running `dbus-run-session > startplasma-wayland` multiple times without killing the old sessions? When > you do things manually, you become responsible for this sort of thing. Actually it is not running multiple times. Each image in the combined one, which contains 5 different images, just shows the htop taken at different times, not different scrolls of same htop. I will try another distro to check if this behavior occur in order to identify that issue is related to gentoo or not. It may be a bug in gtk wayland. I am going to close this bug. If you can guide me, I will appreciate it. Thanks for your time anyway. https://bugsfiles.kde.org/attachment.cgi?id=153117 does show multiple plasmashell instances. (In reply to Nate Graham from comment #11) > https://bugsfiles.kde.org/attachment.cgi?id=153117 does show multiple > plasmashell instances. Ooh, you were referencing this attachment. Well, they are just threads of plasmashell. Purplish rows indicate the threads. White ones indicate the process itself. Oops, thanks! |