After having used Arch (i3) for years, accepting that I wasn’t able to configure some stuff (because I didn’t care and didn’t want to invest all the time investigating it), I came to a point where I started noticing that there were even important things I didn’t configure correctly. Battery life on my laptop was always very short and I never bothered configuring the dual gpu system correctly, just always using the dgpu.
So several days ago, I made the switch to Manjaro (i3), thinking that all basic configuration I had troubles with before were installed and configured for me. And it seems like that is the case. I thank all the contributors.
But I am having issues I didn’t have with my Arch installation, and I guess that that also has to do with configuration that Manjaro made for me. I used NetworkManager on my Arch and it comes installed with the i3 version of Manjaro. This time, however, the NetworkManager applet keeps hanging for up to 3 minutes (while wifi is already working fine), blocking itself (can’t select the VPN of my work for example).
Sometimes, my whole system starts slowing down. My mouse stops moving for 3 seconds, then it works for a split second, then it stops moving for 10 seconds, then moves again for a split second and then my computer seems to hang forever and only resetting is the solution. It usually comes when I am starting up network traffic (after playing a game on Steam → save file sync, and just now when I started the download of a game on Steam). Maybe it has nothing to do with NetworkManager/wifi, but I guess it does. I didn’t have these things with my Arch installations and I must have been using the driver that I have now on it as Manjaro’s version will most likely be one I have used on Arch before.
Any clues? What logs can I investigate?
The laptop is a BTO X-BOOK, intel i9 5Ghz, 32Gb memory, Intel GPU + Nvidia RTX2070, 4K OLED screen. It’s a beast
I’m sure you heard the saying Achilles’ heel … well, everything has a blind spot or something that can drag it down.
You described a couple of situations that might be good as correlation of issues, but not exactly pointing to the cause of such correlated or not issues. inxi -Fxza --no-host
Usually gives us a better idea of the system we talk about.
You can straight forward try to disable mac address randomization of wifi, disable IPv6 … and you can replace NetworkManager with connman (this is recommended on some special cases even tho there is no disadvantage doing it anyway).
I’ve disable IPv6 on the wifi connection, that is of course only on this specific connection, so can I disable it for every connection?
I don’t know how to disable mac randomization. The only thing I can see in the settings is ‘Cloned MAC address’ and it is empty.
FYI: The kernel parameters acpi_osi=! acpi_osi=Linux are for the backlight controls for my keyboard, otherwise the room is fully lit in blue and I can’t disable it. I found these somewhere on the internet and it ‘just worked’.
Done… done and done… Thanks, I’m learning already.
I rebooted but sadly, the problem is still there.
Edit: I don’t know if it’s related but I’m saying it just in case: sometimes Firefox may hang for 20 seconds and also xrandr can take up to 10 seconds to react. Both mostly work fine. I use xrandr to change the brightness & gamma on the OLED screen from a self-made python script. It worked fine on my Arch installation, but on this installation, when I start it by using the Fn-keys for brightness, it sometimes hangs for several seconds. After that it works fine for minutes and then it has this again.
NetworkManager
NetworkManager has been bugging me lately but that is the same whether on Arch or Manjaro. Sometimes network doesn’t come up at boot and nm-applet tray is telling me networkmanager is not running.
It usuallly appears after updates but otherwise without clues. Starting networkmanager manually usually solves the issue.
Hi, no, the main reason I switched to Manjaro is exactly for the availability stable branch. I use Linux on my work laptop too and every week there is at least one (small or big) problem with the unstable software on Arch that causes problem for my work. Last two weeks it was Virtualbox not being compatible with the latest kernel and I really needed to build, test and certify an executable of my software on the Windows VM.
I might try this connman then, but one of the reasons I like NetworkManager is that it has a plugin for Fortinet VPN which I definitely need for my work. I will have to investigate how that works in connman.
Ok, I just tried connman (disabled/stopped NetworkManager, rebooted to be sure) and I had the same issues. As soon as I started cmst to configure the wifi, its window popped up and then it all hung, never reaching the point where I could actually connect to something.
The problem seems to be rooted deeper than the network manager.
Just tested it, when I click on the NetworkManager tray icon, it takes 45 seconds for something to show up. I think that it hangs every time it refreshes the list of AP’s, and the same is probably happening with connman.
I just enabled pulseaudio and installed pa-applet for the tray icon. Loading the tray icon also seems to take forever and pulseaudio won’t work after boot. Pavucontrol will say it is searching for the pulseaudio server. But ps tells me it is running. Only if I kill it and restart it, pulseaudio will work.
When I then use the volume controls (I disabled them in my i3 configuration because pa-applet takes them over), it may again take up to 1 minute!!! before the volume changes (so it actually works).
It looks like something on my system is blocking events until some timeout happens. I guess this may no longer be a network issue but something deeper. Is there something I can check? Can I see if things are being deadlocked, stalled, whatever on a very low level?
Maybe try and install manjaro via architect, which gives you more ‘freedom of choice’, and trying to stay as close to your arch installations as possible
It sounds like some sort of I/O write and writeback issue, either to RAM or Disk. If it’s RAM then your swappiness might be set too high. If it’s to Disk then it could indicate a hardware problem.
swappiness is in /proc/sys/vm/swappiness. If it’s 60 lower it to 10. The I/O disk thing I’m unsure how to check or rectify that but htop or glances might show something. htop should be installed, glances probably not but it’s available.
I’m not sure if the above info shows the right place to look (I’m no expert user/guru!) but it may give some insights. Hope it’s helpful, good luck
I don’t use a swap disk, doesn’t seem necessary with 32GB memory and it has always worked fine without swap disk for me. Still though, I will try and set is lower. I hope the system isn’t trying to swap out memory just to discover that it doesn’t work… Setting it to 10 doesn’t seem to help.
The disks are around 6 months old and worked fine with other osses on it. I’ll check the SMART status to be sure, but I really doubt that would be the case. I have 10 year old Samsung SSD’s that still work fantastic. Smart checks say the disks are fine.
