You are not logged in.
For months now my system has been randomly slowing down every 10 minutes or so, for multiple minutes at a time. The symptoms are for example that opening new Firefox tabs is slow, opening new programs is very slow, even running a console command sometimes takes about 10 seconds or more to start, loading web pages is slow, etc. One thing that never seems to lag/hang is my mouse cursor though, I can basically always still move it around smoothly.
I think the issues started happening when I upgraded my system. It could have been before or after that though, but I don't think so.
My system before:
i5 6600K, RX480, Asus motherboard (forgot the model, can look it up if it's important), 32 GB RAM, 500 GB SSD and 2 TB hard drive
New system:
5900X, RX 6800, B550 AORUS ELITE V2, same 32 GB RAM, 1 TB SSD (cloned from the old one and stretched partition) and same 2 TB hard drive
Things I've checked:
dmesg, no suspicious messages as far as I can tell
iotop, no excessive reading/writing
htop, CPU cores are almost completely unused (when I'm not doing something intensive like compiling), no process that is using a lot of CPU, RAM usage is low, swap is basically unused because of that
journalctl, some things in there that I should probably clean up but nothing that would indicate a serious issue
Things I've tried:
Update BIOS
Use linux-zen instead of linux kernel
Unplug USB devices except keyboard, mouse, and headset
Disable everything related to virtualization, in BIOS, mkinitcpio, GRUB command line
Switch from KDE (X11) to Sway (Wayland), still had the same issues
Disabling KWin compositor
Probably some other things I'm forgetting right now
Apologies if this is the wrong subforum, I wasn't sure exactly where to put it.
Thanks to everyone willing to help, I really have no idea what else to try or where else to look. Feel free to ask for extra information.
Last edited by Visne (2023-06-14 18:58:23)
Offline
You're at least running NM and systemd-networkd concurrently, pick one, disable the other and post the output of
find /etc/systemd -type l -exec test -f {} \; -print | awk -F'/' '{ printf ("%-40s | %s\n", $(NF-0), $(NF-1)) }' | sort -fOther than that, the posted journal only covers 5 seconds, so there's not much else to see there.
Edit: except also
Jun 14 20:02:11 arch kernel: bridge: filtering via arp/ip/ip6tables is no longer available by default. Update your scripts to load br_netfilter if you need this.Last edited by seth (2023-06-14 20:14:17)
Online
Oops, sorry about the 5 seconds, not sure why it cut off.
I've disabled systemd-networkd, and rebooted. Thanks for pointing that out. Here's a longer journalctl output: https://pastebin.com/kbbyrnMz
I've also made a file /etc/modules-load.d/br_netfilter.conf, containing br_netfilter (I assume that is how I should fix it). Note that I haven't rebooted since adding that and that it's not part of the pastebin. I will reboot now to see if anything changes.
Edit: I'm aware that Steam spams a lot, I still have the issues while Steam isn't started. Would be nice to know how to fix them though.
Last edited by Visne (2023-06-14 20:53:28)
Offline
* You're also running resolved and dnsmasq (and dnsmasq-dhcp) - latter probably due to libvirtd
* You apparently saw the ipfs problem.
Do the hangs remain w/o mangohud and/or only one output attached?
Then there're a bunch of syncthing errors.
Ideally post a cleaner journal (sway, no libvirtd, syncthing, steam, discord) that covers an incident.
Online
Haven't had issues yet apart from some tiny freezes, this gets spammed in journalctl quite a bit:
Jun 15 00:42:17 arch NetworkManager[607]: <info> [1686782537.1079] device (wlan0): set-hw-addr: set MAC address to 8E:72:52:39:6A:EF (scanning)
Jun 15 00:42:17 arch NetworkManager[607]: <info> [1686782537.1253] device (wlan0): supplicant interface state: inactive -> interface_disabled
Jun 15 00:42:17 arch NetworkManager[607]: <info> [1686782537.1253] device (p2p-dev-wlan0): supplicant management interface state: inactive -> interface_disabled
Jun 15 00:42:17 arch NetworkManager[607]: <info> [1686782537.1255] device (wlan0): supplicant interface state: interface_disabled -> inactive
Jun 15 00:42:17 arch NetworkManager[607]: <info> [1686782537.1255] device (p2p-dev-wlan0): supplicant management interface state: interface_disabled -> inactiveIs that normal?
Edit: Forgot to mention that the MAC address changes every time.
Last edited by Visne (2023-06-14 23:04:52)
Offline
Firefox just crashed and then I got this: https://pastebin.com/pP1qjhvu
I'm in Sway, not sure what KDE is doing.
I've never seen this before though, so don't think it's actually the cause of the issues.
Last edited by Visne (2023-06-15 00:09:21)
Offline
https://wiki.archlinux.org/title/Networ … domization
Don't copypaste out of the pager, see the tip in the 1st link below on how to feed commands/text into a pastebin service.
There's no FF crash recorded in that journal segment, https://wiki.archlinux.org/title/Core_d … _core_dump
What there is, is some pipewire mess:
Jun 15 02:04:59 arch pipewire-pulse[682]: spa.loop: 0x556cb4d4f6a8: queue full 64, need 152https://gitlab.freedesktop.org/pipewire … ssues/1960
=> https://gitlab.freedesktop.org/pipewire … ssues/2747 & https://bbs.archlinux.org/viewtopic.php?id=280196
You can lock the AP by connecting to the BSSID (APs MAC address) instead of its SSID (the name it yells into the world) - might also justbe the randomization, but the pipewire error shows up before that.
https://discussion.fedoraproject.org/t/ … -use/69447 links the error to an OOM condition - do you have a SWAP partition/file?
Some KDE services are running btw.
Online
I've just disabled wifi scanning for now, since I don't need it. Definitely not an OOM, I have 32 GB RAM and a 16 GB swap file and I've never seen it nearly full.
Not sure why the KDE services are running, I guess they are just enabled systemd units?
I also noticed when running ulimit -a that my memlock was 64 kB, is that a normal value?
Offline
It's 8192kB here …
pacman -Qikk pam
cat /etc/security/limits.confOnline
Even after all these changes I still get the issues when doing anything remotely taxing. While updating packages a YouTube video starts buffering every few seconds.
Offline
Do you have an updated journal?
https://bbs.archlinux.org/viewtopic.php?id=57855
While updating packages a YouTube video starts buffering every few seconds.
What sounds very much IO related, either disk operations are eg. CPU intense or your network throughput is limited.
You could try to play a video and stress the disk a bit, https://wiki.archlinux.org/title/Stress_testing#stress
Online
Yes, I believe you are right. When stress testing my SSD my systems slows to a crawl. It's a Crucial BX500, if you search around you can see that many people have issues with it, as it doesn't have a DRAM cache. I guess my best option is to just buy a better SSD and use this for mass storage or something. Thanks for all the help.
Last edited by Visne (2023-06-16 11:10:37)
Offline