Announcement

**DinoNerd** · Jun 16, 2026, 02:06 AM

Thinking a bit more: if the system has a plague of processes (ab)using mlock, or really huge in-kernel-memory data structures representing the virtual memory associated with each bloated process, the problem really could be physical memory. (In the limiting case, the data structures don't need to be huge if we're looking at a swap space a sufficiently large multiple of system memory. With 32 GB of main memory, and typical per-page data structures (64 bits per page), and the use of 4K pages, I'd expect trouble of that sort at a swap space size around 256 GB. Not what we have here.)

**die.boer** · Jun 16, 2026, 03:07 PM

Finally, is it just that linux is no longer good at running on "low end" hardware? I "only" have 32 GB of RAM,

not at all , i have a dual core machine from 2010 with 4Gig ram and it runs 26.04 like a champ
i have zram and a swapfile but the swapfile is basically never even being used

**die.boer** · Jun 16, 2026, 03:19 PM

output of these ?

Code:

free -h; swapon --show

Code:

journalctl -b -1 -g "oom\|Out of memory\|Killed process"

Code:

journalctl -u systemd-oomd -b

Code:

inxi -Fzxx

**DinoNerd** · Jun 16, 2026, 08:36 PM

Originally posted by die.boer View Post

not at all , i have a dual core machine from 2010 with 4Gig ram and it runs 26.04 like a champ
i have zram and a swapfile but the swapfile is basically never even being used

That's a lot more the kind of thing I was expecting. Clearly something specific is wrong with my system, or how I use it.

**DinoNerd** · Jun 16, 2026, 08:58 PM

Originally posted by die.boer View Post

output of these ?

Code:

free -h; swapon --show

Code:

journalctl -b -1 -g "oom\|Out of memory\|Killed process"

Code:

journalctl -u systemd-oomd -b

Code:

inxi -Fzxx

I spent much of today vigorously applying OneTab to reduce the suspected source of load on my system. If that was the right workaround, it's in better shape as I report these results.

Code:

$ free -h
total used free shared buff/cache available
Mem: 30Gi 13Gi 2.4Gi 5.9Gi 21Fi 17Gi
Swap: 511Mi 503Mi 8.9Mi

Code:

$ swapon --show
MAME TYPE SIZE USED PRIO
/swapfile file 512M 503.1M -1

Code:

$ journalctl -b -1 -g "oom\|Out of memory\|Killed process"
- - No entries - -

Code:

journalctl -u systemd-oomd -b
- - No entries - -

Code:

inxi -Fzxx
Command 'inxi' not found, but can be installed with:
sudo apt install inxi

By the looks of things, and guessing a bit what the commands did, it looks like (a) the workaround behaved as I expected and (b) logs wrap around fast, and were gone by the time I tried this.

This also doesn't look like an absurdly low amount of swap space for a system with 32 GB of RAM.

I OneTab'd 109 separate tabs from Safari on the 16 GB MacOS system with similar symptoms. The version of OneTab available for Firefox doesn't give me a count, but the list is of a similar length. I also outright closed a bunch of tabs on both systems, but more on the Kubuntu system.

At a guess, my habit of leaving lots of tabs around "to get back to later", is creating an awful lot of load, far more than I'd have expected. There may also be an issue with Protonmail (memory leak? enthusiastic use of mlock? keeping too much in memory when confronted with huge mailboxes? - I'm as prone to not deleting emails as I am to opening 100s of tabs).

**DinoNerd** · Jun 16, 2026, 09:10 PM

Originally posted by DinoNerd View Post

At a guess, my habit of leaving lots of tabs around "to get back to later", is creating an awful lot of load, far more than I'd have expected. There may also be an issue with Protonmail (memory leak? enthusiastic use of mlock? keeping too much in memory when confronted with huge mailboxes? - I'm as prone to not deleting emails as I am to opening 100s of tabs).

My choices right now seem to be:

- wait and see whether the problem has gone away, in which case I'll conclude that my tab habit is/was pathological, at least on modern browsers/modern web pages.
- re-open all the tabs now saved in OneTab, wait for the problem to recur. If this doesn't make it recur within a week, something even stranger is happening.

Either way, if/when the problem recurs I try to gather the above statistics when the system is as close to the recurrence as possible. (This time I couldn't even get my commands echoed in an ssh client window, so there will be some delay.)

I'm inclined to go with the lazy choice - presume use of OneTab has solved my problem, and don't try to replicate it.

FWIW, I was already looking for a solution to my tab proliferation. It has other unwanted side effects. Some of them were less problematic on MacOS, so my switch to linux had put "do something about this" on my agenda. It just hadn't previously got to the top of my list.

Also, if the problem recurs, and the OOM system picks ProtonMail as its target, that suggests some combination of reducing surplus emails and consulting its provider, as well as gathering data on it specifically.

**DinoNerd** · Jun 16, 2026, 09:18 PM

$ inxi -Fzxx
System:
Kernel: 7.0.0-22-generic arch: x86_64 bits: 64 compiler: gcc v: 15.2.0
Desktop: KDE Plasma v: 6.6.4 tk: Qt v: N/A wm: kwin_wayland dm: SDDM
Distro: Kubuntu 26.04 LTS (Resolute Raccoon) base: Ubuntu
Machine:
Type: Mini-pc System: System76 product: Meerkat v: meer6
serial: <superuser required> Chassis:type: 35 v: meer6
serial: <superuser required>
Mobo: Intel model: NUC11TNBi5 v: M11904-404 serial: <superuser required>
part-nu: BNUC11TNHi50000/1/2/5 Firmware: UEFI vendor: Intel
v: TNTGL357.0054.2021.0316.1417 date: 03/16/2021
CPU:
Info: quad core model: 11th Gen Intel Core i5-1135G7 bits: 64 type: MT MCP
arch: Tiger Lake rev: 1 cache:L1: 320 KiB L2: 5 MiB L3: 8 MiB
Speed (MHz):avg: 400 min/max: 400/4200 cores:1: 400 2: 400 3: 400 4: 400
5: 400 6: 400 7: 400 8: 400 bogomips: 38707
Flags-basic: avx avx2 ht lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx
Graphics:
Device-1: Intel TigerLake-LP GT2 [Iris Xe Graphics] driver: i915 v: kernel
arch: Xe ports:active: DP-1,HDMI-A-1 empty: DP-2,DP-3,DP-4 bus-ID: 00:02.0
chip-ID: 8086:9a49
Display: wayland server: Xwayland v: 24.1.10 compositor: kwin_wayland
driver:gpu: i915 d-rect: 7680x2160 display-ID: 0
Monitor-1: DP-1 pos: right model: ASUS VG289 res: 3840x2160 hz: 60
dpi: 157 diag: 708mm (27.9")
Monitor-2: HDMI-A-1 pos: primary,left model: Dell S2722QC res: 3840x2160
hz: 60 dpi: 163 diag: 685mm (27")
API: EGL v: 1.5 platforms:device: 0 drv: iris device: 1 drv: swrast gbm:
drv: iris surfaceless:drv: iris wayland:drv: iris x11:drv: iris
API: OpenGL v: 4.6 compat-v: 4.5 vendor: intel mesa v: 26.0.3-1ubuntu1
glx-v: 1.4 direct-render: yes renderer: Mesa Intel Iris Xe Graphics (TGL
GT2) device-ID: 8086:9a49 display-ID: :0.0
API: Vulkan v: 1.4.341 surfaces: N/A device: 0 type: integrated-gpu
driver: mesa intel device-ID: 8086:9a49 device: 1 type: cpu
driver: mesa llvmpipe device-ID: 10005:0000
Info:Tools:api: clinfo, eglinfo, glxinfo, vulkaninfo
de: kscreen-console,kscreen-doctor wl: wayland-info x11: xdriinfo,
xdpyinfo, xprop, xrandr
Audio:
Device-1: Intel Tiger Lake-LP Smart Sound Audio driver: snd_hda_intel
v: kernel bus-ID: 00:1f.3 chip-ID: 8086:a0c8
API: ALSA v: k7.0.0-22-generic status: kernel-api
Server-1: PipeWire v: 1.6.2 status: active with:1: pipewire-pulse
status: active 2: wireplumber status: active 3: pipewire-alsa type: plugin
Network:
Device-1: Intel Wi-Fi 6 AX201 driver: iwlwifi v: kernel bus-ID: 00:14.3
chip-ID: 8086:a0f0
IF: wlo1 state: down mac: <filter>
Device-2: Intel Ethernet I225-LM driver: igc v: kernel pcie:speed: 5 GT/s
lanes: 1 port: N/A bus-ID: 58:00.0 chip-ID: 8086:15f2
IF: enp88s0 state: up speed: 1000 Mbps duplex: full mac: <filter>
Bluetooth:
Device-1: Intel AX201 Bluetooth driver: btusb v: 0.8 type: USB rev: 2.0
speed: 12 Mb/s lanes: 1 bus-ID: 3-10:3 chip-ID: 8087:0026
Report: hciconfig ID: hci0 rfk-id: 0 state: down
bt-service: enabled,running rfk-block:hardware: no software: no
address: <filter>
Drives:
Local Storage:total: 7.28 TiB used: 1.6 TiB (21.9%)
ID-1: /dev/nvme0n1 vendor: Samsung model: SSD 980 PRO 2TB size: 1.82 TiB
speed: 63.2 Gb/s lanes: 4 serial: <filter> temp: 57.9 C
ID-2: /dev/sda vendor: Samsung model: SSD 870 QVO 2TB size: 1.82 TiB
speed: 6.0 Gb/s serial: <filter>
ID-3: /dev/sdb vendor: Seagate model: Expansion SW size: 3.64 TiB
type: USB rev: 3.2 spd: 5 Gb/s lanes: 1 serial: <filter>
Partition:
ID-1: / size: 1.79 TiB used: 756.46 GiB (41.3%) fs: ext4 dev: /dev/nvme0n1p2
ID-2: /boot/efi size: 299.4 MiB used: 10.7 MiB (3.6%) fs: vfat
dev: /dev/nvme0n1p1
Swap:
ID-1: swap-1 type: file size: 512 MiB used: 500.2 MiB (97.7%) priority: -1
file: /swapfile
Sensors:
System Temperatures:cpu: 82.0 C mobo: N/A
Fan Speeds (rpm): N/A
Info:
Memory:total: 32 GiB available: 30.97 GiB used: 14.41 GiB (46.5%)
Processes: 339 Power:uptime: 19d 1h 54m wakeups: 0 Init: systemd v: 259
default: graphical
Packages: 2846 pm: dpkg pkgs: 2834 pm: snap pkgs: 12 Compilers:
gcc: 15.2.0 Shell: Bash v: 5.3.9 running-in: konsole inxi: 3.3.40

**DinoNerd** · Jun 16, 2026, 09:46 PM

Originally posted by DinoNerd View Post

Code:

$ swapon --show
MAME TYPE SIZE USED PRIO
/swapfile file 512M 503.1M -1

This also doesn't look like an absurdly low amount of swap space for a system with 32 GB of RAM.

I totally misread the above, the first time around. All I noticed was that 512 is a lot more than twice 32.

When I read the units as well as the numbers it does seem extremely small, at 1/64th of system memory.

...

Obviously I don't have a feel for current linux swap space requirements, but I remember a rule-of-thumb of twice system memory for most/any *nix system. I think we were still using that with FreeBSD the last time I worked with it. Linux' memory management has always been a bit different from other *nixes, presumably due to some very basic decisions made long ago, so my heuristic may never have fit linux. (My ignorance is showing here.) But if it didn't, I don't see why.

Many *nix processes use the design pattern "read the whole thing into memory, however large, and let the kernel's paging system deal with it." In modern times, they may simply mmap the whole file. No problem if the file is mapped in such a way that changes to the process' mmap area are written back to the file. (Swap space won't be used, as the file is its own backing store.) More trouble if the app doesn't want changes saved unless/until the user tells them to do so. (It needs swap space in the meantime.) And loads of trouble if the app decides to improve its performance by mlocking as well as mmaping the whole thing, and someone feeds it a really large file.

Emacs would be an example that traditionally read the whole file into (virtual) memory. On an ancient system with small memory and large files, that would trigger lots of paging activity.

I've no idea how many other *nix/linux tools behave the same way, but I strongly suspect ProtonMail of being one of them, given what little I understand of its design, based mostly on its known features.

FWIW, I also use emacs as my text editor of choice, but I haven't had it read in anything huge - just the kind of .txt or .html file that's actually typed by a human, or could be.

---
Also, on my third reading I see that most of this swap file is in use, even with most of my tabs gone, and the system behaving properly.

Leaving stuff in swap once it's been paged out is normal. There may also be a swap space reservation process in the linux kernel. So having 503 Mb of 512 MB of swap in use is not problematic with 13 G of main memory available.

But it does suggest it has been pretty much totally full up.

Of course for all I know to the contrary, the swap file might grow and shrink dynamically. My knowledge is so very much out of date, and more general *nix than specifically linux.

**die.boer** · 2026-06-17T14:37:48

the swap is extremely small for a 32gb system , and firefox with LOTS of tabs open + two 4K monitors will eat up ram .
the swap size for you should be 8bg or 16gb...
let me know if you want to increase the swap

**DinoNerd** · 2026-06-17T15:22:11

Originally posted by die.boer View Post

the swap is extremely small for a 32gb system , and firefox with LOTS of tabs open + two 4K monitors will eat up ram .
the swap size for you should be 8bg or 16gb...
let me know if you want to increase the swap

I think that would be a good idea.

The 4K monitors won't be going away, and even if I use OneTab religiously, I'm sure to find some other activity that will use lots and lots of memory. (As one possibility, I suspect dosbox of having been involved in previous system slowdowns, perhaps as the last straw that was too much for an overloaded system.)

**oshunluvr** · 2026-06-17T16:24:20

I can't imagine FF even as a snap would eat up 32GB of RAM unless it has a huge memory leak. I'm using Vivalidi with 50ish tabs open using only 3.7gb. I would do some more testing before making changes.

I would start by opening Konsole (after a reboot) and enter "free". Then, leave Konsole open, launch FF, and enter "free" again. Note the amount of available RAM. Then open System Monitor > Applications tab and see what FF is using. Then use FF as normal but keep System Monitor so you can see it and watch the RAM usage.

I suspect you either have a buggy FF, or bad swap file (or partition?). Also, with 32GB RAM you should reduce swappiness to minimum. IME, no need to swap at all with that much RAM.

**oshunluvr** · 2026-06-17T16:45:35

You can check current swappiness by entering sysctl vm.swappiness in Konsole. If it's not one, you should reduce it to that. Edit /eetc/sysctl.d/99-sysswappiness.conf and edit "vm.swappiness=" to equals 1

**die.boer** · 2026-06-17T18:43:42

Originally posted by oshunluvr View Post

You can check current swappiness by entering sysctl vm.swappiness in Konsole. If it's not one, you should reduce it to that. Edit /eetc/sysctl.d/99-sysswappiness.conf and edit "vm.swappiness=" to equals 1

yes i agree ,the swappiness should be 1 or even 10 is ok, but i meant the swapfile is way to small to be of any use ,when needed

Announcement

My Kubuntu 26.04 LTS System Keeps Hanging; When it Recovers, I Get a Memory Shortage Avoided message from the kernel

My Kubuntu 26.04 LTS System Keeps Hanging; When it Recovers, I Get a Memory Shortage Avoided message from the kernel

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Users Viewing This Topic