SYSTEM FREEZES - Pop!_OS causing daily, irrecoverable system lock-ups

[Adam Kadmon]

Hi Everyone,

If you're also experiencing irrecoverable freezes where only a hard restart will work, then please head on over to this GitHub ticket or let me know if you have any other avenues to get this fixed. Many others with this problem have already gone to other distros, I'm doing all I can not to join them but it's not easy.

Please also be warned that these freezes have already caused me to have to reformat and reinstall once. They get more and more frequent until they end your file system.

 

[derpOmattic]

I've seen Brave cause a multitude of problems in Pop. Maybe stop using it for while.

If excessive logging seems to initiate this, maybe turn it off or tone it down. Every new installation of Pop gets the following treatment for me. In Settings > File History & Trash, I enable the "Automatically Delete Trash Content" and "Automatically Delete Temporary Files". I change the period to a couple of days - a week at the most. Change the "File History Duration" to anything but forever.
I then go through Settings > Notifications and turn them off for almost everything. Even the ones I leave on, I turn off sound alerts and pop-ups. Only one or two apps ever have the whole lot enabled. Then go through the Settings > Applications individually and ensure things are ok.
Sounds tedious, but who needs so much logging, tracking and notifying. Worth a try at this stage.

 

[Adam Kadmon]

First of all, thank you for all these suggestions!

Brave is a recent addition to my system, the majority of the freezes happened before I installed it. The freezes have occured with Firefox (79, 80, & 81), Chromium, Brave, and now Qutebrowser. I pretty much always have a browser open (traditionally Firefox, recently Qutebrowser) and each browser above has been the sole browser open during a freeze at some point or other (with the exception of Brave, ironically, but I only use it for very specific work tasks and always in tandem with FF/QB).

I've gone through and done everything you described (it's not nearly as tedious as recovering from a freeze). I also cleared temporary files and file history (I keep trash empty anyway).

Is there any other way to tone down excessive logging? I didn't realise it was something we could affect, and so didn't dig deeper into @mmstick's comment on GitHub. I'll look into it myself as well, of course.

Thanks again! Any suggestions that aren't terribly risky and stand any non-zero chance of working I'm willing to try at this point. If the frequency of these freezes doesn't drop back down below one every couple of days then I'm going to have to try kernel 5.8. I've been holding off until Pop!_OS 20.10, but I might have to jump the gun to potentially save my system from being pulverised again.

 

[Adam Kadmon]

Although good advice as far as keeping your system light and lean, the above suggestions, unfortunately, did nothing to prevent another freeze. The devs don't seem to be doing anything at all to address this issue of frequent system lock-ups. As dedicated as I've been to staying with Pop!_OS, my work is being seriously affected by this problem. With no attention from the devs, a fix seems unlikely. I'm going to double-check I have everything backed-up, try a couple of long-shot solutions and call it a day. If I find a distro that supports a similar workflow but without all the bugs, I'll let everyone else with the same problem know.

I've invested an inordinate amount of time into making this work, but I'm not a programmer, developer or engineer. I fear this will go unresolved until the Pop team takes a look at it.

 

[derpOmattic]

This Arch wiki page may be helpful to learn more about systemd logging, and how to limit it. I've shared that before, but in the context of logs filling up 200GB of disk space.

 

[Adam Kadmon]

Thank you for this! I'd come across it before but didn't see any immediately useful information - now that you've drawn my attention to it I can see that I was wrong. At first glance everything seems OK with my systemd logging, but I need to sit down and work through it more systematically tonight.

In the meantime, I did something a litte drastic and updated my kernel to 5.8.14 (with the frequency of these freezes I really do have very little to lose). So far my laptop has booted up alright and is notably snappier, especially in Qutebrowser (which had been slow to load open tabs upon launch). This and the systemd logging are the last two things I have that I can try for now. I'll keep scrambling till the very end, but I'm running out of ideas and out of time - I have two jobs to hold down, both of which are 100% computer-based.

 

[gumballdead]

Any insights on fixing this? I'm a new Pop!_OS user, I recently bought a System76 Oryx Pro laptop and I love everything about it and the OS but I must say the random freezes are unbearably frustrating.

 

[derpOmattic]

I recently bought a System76 Oryx Pro laptop

You should definitely open a support ticket with System76 to get personalized professional support.

 

[Adam Kadmon]

Any insights on fixing this? I'm a new Pop!_OS user, I recently bought a System76 Oryx Pro laptop and I love everything about it and the OS but I must say the random freezes are unbearably frustrating.

I second what derpOmattic wrote above - as a System76 customer you won't have to hack away at your own troubleshooting. They might be able to reproduce the error on one of their Oryx Pros and this might lead to a solution for the growing number of us that are experiencing what looks to be the same thing.

To answer your question, though: I switched to kernel 5.8.14 and haven't had a freeze in just over a week now. After daily freezes that's awesome, but I'll hold off on celebrating for another couple of weeks. About two weeks is the longest I've gone without freezes thus far, so this fix needs to beat that record by a significant margin.

I'm finding a minority of reports that involve other distros, so it might not be a Pop!_OS thing exclusively. Also, I'm mainly reading about Pop!_OS, so that's biasing my perception.

 

[ozspeed]

do you see any NVRM Xid:61 errors in syslog? I too am getting random freezes, but with this error that seems to be releated to the nvidia driver - see thread here https://forums.developer.nvidia.com/t/random-xid-61-and-xorg-lock-up/79731

I have a ticket open with system 76.

Additionally do you see any wpa_supplicant errors in syslog, try setting net.ifnames=0 in your kernel paramters (use kernelstub -a "net.ifnames=0")

this has reduced the number of random freezes for me but I still have had one in the last 24 hours