You are not logged in.
Pages: 1
Hey, I've been experiencing "laggy UI behaviour" on both chrome and chromium, what I've tried:
- Deleted configuration files and reinstalled both
- Disabled all extensions
- Logged out of the account and disabled sync
- Tried on an anonymous tab, which is where it gets weird
- This arch wiki
Sometimes it works fine in anonymous mode, and sometimes it doesn't. Sometimes it works fine for a few seconds, right after opening it up, and then the "laggyness" strikes back. I can't seem to understand on which conditions it works fine or it doesn't.
Relevant information
- I know it started around 14:00 - 14:30, Nov 15, which was when I performed the upgrade.
- Firefox works fine
cat /var/log/pacman.log | grep -E '\-15.*upgraded'
: pastebin
journalctl -x | grep -E 'Nov 15.*chrome'
: pastebin
I noticed the nvidia driver was upgraded, could it be the culprit?
Last edited by VanillaLandlord (2018-11-16 22:53:41)
Offline
You put "laggy UI behaviour" in quotes, acknowledging it's unsharp meanig, but did not elaborate on how this "laggy UI behaviour" manifests.
Generic hint is to check CPU, RAM and I/O load.
Offline
By "laggy UI behaviour" I mean, every animation (opening/closing tab, loading icons, etc) is laggy. Scrolling and typing also have some drag.
Edit: CPU, RAM and I/O load are completely normal.
Last edited by VanillaLandlord (2018-11-16 13:13:16)
Offline
When you are having problems, what is the output of free ?
Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way
Offline
❯ free
total used free shared buff/cache available
Mem: 8074640 2442140 3032600 573900 2599900 4777248
Swap: 8388604 0 8388604
The problem is pretty much constant. I noticed when I open up the browser it runs fine for a few seconds and then the lag hits.
Last edited by VanillaLandlord (2018-11-16 16:31:36)
Offline
Not a problem with thrashing to swap. That was my first thought.
Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way
Offline
I've been looking at the dependency packages of chromium, and the only one that was upgraded was NSS, which seems completely unrelated. Any idea on how I might get some more hints?
Thanks!
Last edited by VanillaLandlord (2018-11-16 18:43:31)
Offline
I'm having the same problem. It started since 4th November, when I changed motherboard and cpu(intel->amd), but might be a coincidence. I've disabled firejail for it, changed DE(xfce to Cinnamon). Tried opening via terminal to see messages. Nothing in there while freezing is goind. No CPU spikes either. Disabled hardware acceleration. Still freezes. My GPU is amd.
Linux ArchPC 4.19.1-arch1-1-ARCH #1 SMP PREEMPT Sun Nov 4 16:49:26 UTC 2018 x86_64 GNU/Linux
[del@ArchPC ~]$ chromium
[3077:3084:1116/143606.354104:ERROR:browser_gpu_channel_host_factory.cc(137)] Failed to launch GPU process.
[3077:3141:1116/143607.347634:ERROR:database.cc(1828)] Web sqlite error 5, errno 0: database is locked, sql: PRAGMA auto_vacuum
[3077:3141:1116/143607.347708:ERROR:database.cc(1828)] Web sqlite error 5, errno 0: database is locked, sql: PRAGMA journal_mode=TRUNCATE
[3077:3084:1116/143608.946646:ERROR:browser_gpu_channel_host_factory.cc(137)] Failed to launch GPU process.
[3077:3077:1116/143608.996700:ERROR:gpu_process_transport_factory.cc(980)] Lost UI shared context.
[3077:3077:1116/143613.255658:ERROR:input_method_base.cc(146)] Not implemented reached in virtual ui::InputMethodKeyboardController *ui::InputMethodBase::GetInputMethodKeyboardController()Using InputMethodKeyboardControllerStub
[3077:3077:1116/143618.601300:ERROR:input_method_base.cc(146)] Not implemented reached in virtual ui::InputMethodKeyboardController *ui::InputMethodBase::GetInputMethodKeyboardController()Using InputMethodKeyboardControllerStub
The previous version that was working okay, based on pacman history, seems to be 70.0.3538.67-1. I could only test 67-2, which still lagged. 67-1 couldn't start because of icu library, which was also updated, so I didn't bother after that.
Is there something else I can test or troubleshoot?
Offline
Any idea on how I might get some more hints?
Use the chrome/ium task manager, shift+esc, to see resources for specific extensions, the browser, tabs etc. The dev-tools, F12/ctrl+shift+i, have a performance monitor for single tabs, a performance profiler you can use to record sessions and a bunch of other tools to debug an instance. Chrome/ium are pretty good debuggers for and by themselves.
Offline
VanillaLandlord wrote:Any idea on how I might get some more hints?
Use the chrome/ium task manager, shift+esc, to see resources for specific extensions, the browser, tabs etc. The dev-tools, F12/ctrl+shift+i, have a performance monitor for single tabs, a performance profiler you can use to record sessions and a bunch of other tools to debug an instance. Chrome/ium are pretty good debuggers for and by themselves.
I see no abnormal resource usage, everything looks fine.
For reference, here are the potentially relevant logs of both chrome and chromium.
Chrome
[5952:5952:1116/194321.328414:ERROR:sandbox_linux.cc(379)] InitializeSandbox() called with multiple threads in process gpu-process.
[5909:6116:1116/194325.250842:ERROR:get_updates_processor.cc(242)] PostClientToServerMessage() failed during GetUpdates
[5909:5909:1116/194325.251583:ERROR:http_bridge.cc(126)] Not implemented reached in virtual void syncer::HttpBridgeFactory::OnSignalReceived()
[5909:5909:1116/194348.092296:ERROR:input_method_base.cc(146)] Not implemented reached in virtual ui::InputMethodKeyboardController *ui::InputMethodBase::GetInputMethodKeyboardController()Using InputMethodKeyboardControllerStub
Chromium
[3382:3382:1116/194124.829933:ERROR:sync_control_vsync_provider.cc(141)] Calculated bogus refresh interval=0.146974 s, last_timebase_=372833209 bogo-microseconds, timebase=372980183 bogo-microseconds, last_media_stream_counter_=954, media_stream_counter=955
[3382:3382:1116/194140.344072:ERROR:gl_surface_presentation_helper.cc(237)] GetVSyncParametersIfAvailable() failed!
Offline
sync_control_vsync_provider.cc(141)] Calculated bogus refresh interval
Do you run a compositor or compositing WM? Wayland session maybe?
Offline
sync_control_vsync_provider.cc(141)] Calculated bogus refresh interval
Do you run a compositor or compositing WM? Wayland session maybe?
Yes, I run Wayland. It wasn't one of the upgraded packages though.
Edit: Found this, running chromium with
--disable-gpu-driver-bug-workarounds
seems to have fixed it! Hopefully, it won't magically strike back.
Thank you all!
Last edited by VanillaLandlord (2018-11-16 22:58:14)
Offline
It appeared at first fixed, but I still have the same issue even when running chromium --disable-gpu-driver-bug-workarounds.
Offline
Hey I was linked here by Afterglow.
I'm experiencing similar issues but mine seems to only be when I open a new tab in chrome / chromium and when I am playing League of Legends via wine. My setup has an amd cpu and nvidia gpu. The 0.5 - 1.5s freeze is slightly different for League of Legends vs chrome. For League, the entire game freezes but I can still hear the audio. No other application is frozen during that time. For chrome, I am still able to type and move my mouse. The key-presses aren't visually shown but appear in the URL bar after the freeze has finished. Both of these issues started sometime this month within the same time frame.
Do you run a compositor or compositing WM? Wayland session maybe?
I use x11 and gnome3.
--disable-gpu-driver-bug-workarounds
I tried this workaround but it does not work for me.
I am going to try downgrading the nvidia drivers to 396.54-6 from 410.73-3. Downgrading also requires me to downgrade the kernel so I will post back my test results if I am successful with the downgrade process.
Offline
I am going to try downgrading the nvidia drivers to 396.54-6 from 410.73-3. Downgrading also requires me to downgrade the kernel so I will post back my test results if I am successful with the downgrade process.
You might wanna just go with the snapshot route. It's way easier and you're not gonna mix old and up-to-date packages.
https://wiki.archlinux.org/index.php/Ar … cific_date
Offline
You do not have to downgrade the kernel in order to move from the nvidia to the nvidia-39xx package.
@Afterglow, the kernel is pretty much self-contained, so unlike w/ other packages an isolated downgrade will usually be no problem (w/ the prominent exception of the nvidia kernel modules)
A safe approach could be to ensure a proper and working linux-lts installation and then do to the main kernel version whatever you want - if you screw up, you can still just boot the lts kernel.
Offline
I couldn't figure out how to downgrade the nvidia drivers and have gdm load correctly but so far it seems that my issue has been fixed with the 415.18 nvidia drivers.
Offline
Still having this issue with nvidia-dkms 415.27-5 and Google Chrome Version 72.0.3626.81. I have a nvidia PRIME and also experience it when using Intel.
The best solution on the chromium project bug tracker is to uninstall noto-fonts, which I had installed via the big ttf-google-fonts-git package.
e.g.,
pacman -R noto-fonts ttf-google-fonts-git noto-fonts-extra noto-fonts
It's been running without the 2-5 seconds lag when changing tabs and re-focusing the window now.
Offline
I'm adding this here, in case somebody searches for my issues:
I was experiencing CPU loads over 10 and utilization of 99%, when loading some sites in Chrome, which was locking up i3wm for minutes at a time. And I'd see 30-50% utilization at idle.
I removed these packages and reloaded Chrome, now I see idle utilization ~5% and haven't seen a spike above 20%.
Offline
Still having this issue with nvidia-dkms 415.27-5 and Google Chrome Version 72.0.3626.81. I have a nvidia PRIME and also experience it when using Intel.
The best solution on the chromium project bug tracker is to uninstall noto-fonts, which I had installed via the big ttf-google-fonts-git package.
e.g.,
pacman -R noto-fonts ttf-google-fonts-git noto-fonts-extra noto-fonts
It's been running without the 2-5 seconds lag when changing tabs and re-focusing the window now.
I have experienced this bug for months and each time was worse. You have saved my life. Thanks!!!
Offline
Still having this issue with nvidia-dkms 415.27-5 and Google Chrome Version 72.0.3626.81. I have a nvidia PRIME and also experience it when using Intel.
The best solution on the chromium project bug tracker is to uninstall noto-fonts, which I had installed via the big ttf-google-fonts-git package.
e.g.,
pacman -R noto-fonts ttf-google-fonts-git noto-fonts-extra noto-fonts
It's been running without the 2-5 seconds lag when changing tabs and re-focusing the window now.
That proved very useful, thanks a lot!
I uninstalled the following packages: noto-fonts-extra, noto-fonts-cjk, noto-fonts-emoji. Uninstalling the noto-fonts package was not necessary in my case. And chromium started working again!
Last edited by ileska (2024-02-03 13:41:04)
Offline
In my case, this problem never really went away. Until I migrated from EncFS encrypted home folder to fscrypt. Then it was gone completely.
So it probably wasn't Chromium's problem at all despite similar symptoms or it was just exacerbated by encfs.
Last edited by Afterglow (2024-02-03 14:13:40)
Offline
Pages: 1