I'm working through the stressors to see which ones are possibly causing issues. I did notice that the apparmor stressor eats up memory until the system runs out of memory. This stressor loads illegal apparmor profiles and then removes them. Perhaps there is a memory leak in the loading of profiles that don't pass the verification phase:
To show this issue, run the following, one can see that memory gets low over time before the user gets kicked off due to low memory: sudo ./stress-ng --apparmor 1 --vmstat 5 stress-ng: info: [1339] defaulting to a 86400 second (1 day, 0.00 secs) run per stressor stress-ng: info: [1339] dispatching hogs: 1 apparmor stress-ng: info: [1340] vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa st stress-ng: info: [1340] vmstat: 2 1 0 313824 32776 364352 0 0 16 18 4858 9752 4 25 70 0 0 stress-ng: info: [1340] vmstat: 5 0 0 257848 32776 366528 0 0 0 1091 4573 8435 4 23 72 0 0 stress-ng: info: [1340] vmstat: 5 0 0 198916 32784 368288 0 0 0 20 4642 8681 4 23 71 1 0 stress-ng: info: [1340] vmstat: 2 0 0 139496 32792 370600 0 0 0 16 4612 8500 4 23 71 1 0 stress-ng: info: [1340] vmstat: 2 0 0 85032 32740 363916 0 0 0 1751 4774 8710 4 23 71 1 0 stress-ng: info: [1340] vmstat: 5 0 0 92224 32748 310548 0 0 0 2020 5919 10123 4 24 70 1 0 stress-ng: info: [1340] vmstat: 2 0 0 93380 30068 268484 0 0 0 14 5590 10275 4 26 69 1 0 stress-ng: info: [1340] vmstat: 2 0 0 102152 23648 207872 0 0 0 3346 5277 9303 4 24 70 1 0 stress-ng: info: [1340] vmstat: 5 0 0 99184 18488 169084 0 0 48 2180 5614 9901 4 25 71 0 0 stress-ng: info: [1340] vmstat: 2 0 0 88068 7080 140392 0 0 359 2090 6146 11013 4 27 68 0 0 stress-ng: info: [1340] vmstat: 2 0 0 92368 564 82108 0 0 3568 2534 5899 10308 4 26 67 1 0 stress-ng: info: [1340] vmstat: 7 0 0 83784 100 47356 0 0 99834 4212 8540 14574 4 28 65 2 0 stress-ng: info: [1340] vmstat: 2 0 0 76784 188 44916 0 0 363427 7621 16647 28448 4 37 45 12 0 -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/2024599 Title: linux-image-5.15.0-1032-realtime locks up under scheduler test load Status in linux package in Ubuntu: Incomplete Status in linux source package in Jammy: Incomplete Bug description: lsb_release -a No LSB modules are available. Distributor ID: Ubuntu Description: Ubuntu 22.04.2 LTS Release: 22.04 Codename: jammy uname -a Linux jammie-amd64-efi 5.15.0-1032-realtime #35-Ubuntu SMP PREEMPT_RT Tue Jan 24 11:45:03 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux free total used free shared buff/cache available Mem: 4013888 200984 3439012 1204 373892 3744628 Swap: 4014076 0 4014076 Running in a kvm-qemu, 8 cpus, cpu Intel Core Processor (Skylake, IBRS): how to reproduce issue: git clone https://github.com/ColinIanKing/stress-ng sudo apt-get update sudo apt-get build-dep stress-ng sudo apt-get install libeigen3-dev libmpfr-dev libkmod-dev libxxhash-dev libglvnd-dev libgbm-dev cd stress-ng make clean make -j 8 sudo ./stress-ng --class scheduler --all 1 -v --vmstat 1 -t 30m ..wait for all the stressors to get invoked, system becomes unresponsive, can't ^C stress-ng, can't swap consoles on the VM, appears to be hard locked. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2024599/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
