Fedora 40 on the Framework Laptop 13

Framework Laptop 13 Linux
85

You mentioned running presentations, do you have any suggestions for the contents of said presentation… or the software used?
I need to reproduce this bug with something other than Firefox and I didn’t have much time to experiment the last week. : (

Edit: 67 unsafe shutdowns :smiley:

86

Hi everyone, sorry for being late with this issue. I went through the whole process with the framework support concerning this issue. At the end of the process, the conclusion was (for my case) a mainboard issue and lead to a replacement.

To arrive to this conclusion, many testing steps were required (because it could be more or less anything about hardware). Therefore, here is a summary of steps given by the support team I had to follow during the process. The aim of this list is to try to shorter the whole process by giving more details directly to the support team. Please note that I did not touch windows for a while so it was not question of it through this process, but some part may stay relevant. Also, I most of diagnostic on Fedora so some filenames could change over different distributions.

Also, do not hesitate correct me / give further details / provide links / give alternative tests or orders. This is a try to summarize > 2 months of exchanges. Maybe a comment from a support member could help re-organize / order the steps for a more coherent report.

For my case as described in this feed above, I went to the random crashing issues. I hopped from distros to distros (only linux ones), configuration to configuration (GNOME, KDE, etc…) and still got these crashes.

  1. Use the given template here to redact your message for the support team .

  2. Potentials updates: if you do not have the latest version of one element in the list, please update to the latest version available:

    • BIOS (see the framework guide for AMD).
    • Firmware update (Western digital SSD without WD Dashborad link for example).
    • OS version (depending on your OS).
    • software packets (depending on your packet manager).

If the problem is not solved, while contacting the support team:

  1. Logging and initial information:

    • Previous Troubleshooting:
      • Document any steps you’ve already taken to fix the issue, such as:
      • Updating firmware and drivers.
      • Changing kernel parameters.
      • Testing different Linux distributions and desktop environments.
      • Disabling hardware acceleration in applications (some such crashes were caused by some hardware acceleration in web browsers).
    • When you want to get some logs, use the newer log script from Framework GitHub to get logs:
      link. Downloading / usage instructions are given there.
    • Note when the problem first started occurring.
    • Note the BIOS version (Press F2 repeatedly during startup to enter the BIOS menu & look for “InsydeH2O Version” on the main BIOS screen).
    • List all external devices connected to your laptop:
      • Hubs
      • Docks
      • Monitors
      • eGPUs
      • Third-party power adapters (provide make and model)
    • Take some pictures of the opened laptop to see if there is no defaults without plastic shrouds (the small plastic part glued on top of more than one component with QR codes).
    • Run the following block of commands:
      github_link (the forum does not allow scripts t seems)
      You should get a tar archive to send to the support.

  2. Connections & expansions cards: try to see if the issue occurs when:

    • You are using a third party specific device connected (dock, hub, monitor, etc…)
    • If the issue occurs when charging / on battery
    • Using a specific expansion card

  3. Reset RAM & mainboard, Memtest86

    • Run Memtest86+ on the RAM sticks, make few rounds on it
    • Try to reset RAM (just remove and replace RAM stick(s) inside slot(s) and the mainboard tutorial

  4. Check kernel parameters:

    • Try to set the nomodeset parameter to 1 and see. To my understanding, this parameters put graphic processing in “default-and-no-error-mode”. The laptop become barely usable (since no “complex” graphics are used) but it can help determine the cause of the issue.

Now the point will be to test the following points on both your own system and another live USB using one of framework-officially-supported distros (Ubuntu 24.04 & Fedora 40 right now I think). In my case, crashes where easy to trigger so it was “fine”. But in the case yours are less frequent, it could become longer. The best is to try with both Ubuntu 24.04 and Fedora 40 (as I had to do) but I am not sure that the support team will ask for two distros.

  1. RAM swapping, CPU stress test on your installation and Fresh USB live:
    • Try to swap RAM slot
    • If you have more than one RAM stick, try to reproduce the issue with only one RAM stick. Swap the unique stick slot.
    • If you have two sticks, also try swapping slots between sticks
    • Run a filesystem check with the tool of your choice (for example use GNOME disk utility). You will may need to boot on an USB stick to test your FS without mount it.
    • Try to unplug every connected expension card. Replug one by one and see if you can discriminate either one expension or one mainboard’s internal USB-C connection (one of the fours slots).
    • Install s-tui & tmux with your package manager & run this

This command starts s-tui in a detached tmux session, runs a stress test for 2 minutes, logs data to s-tui_data.csv, and then stops.

Think to perform these steps on both your installation and a fresh live USB stick. If it does not matter, you could re-install the distro on your laptop. But in my case I had to not so I used USB sticks for testing fresh installation (plus it helps discriminating a faulty SSD). Here either you can try to do so with Fedora / Ubuntu installer, or install a full OS on a bigger USB (I did it with a 128GB USB that usually run Arch for instance).

In my case I only had to watch a YouTube video for a while on a fresh installation to trigger the crash. So during the process, the support team asked me to install lm-sensors and while running a video, launch a terminal (with the “keep on to” option) and run: ‘watch -n 1 sensors’ until the crash happen (while filming). I then shared videos samples on a drive solution.

Hoping this will help some of you with that issue.

1 Like
87

This is a very good summary of the same process I’ve been through with support in the past 4 weeks about my crashing/reset issues. I’m fairly confident I’ve exhausted all practical options to isolate the problem or cause; so I believe there’s an issue with mainboard. I’m documenting my adventure here: Framework 13 AMD Hard Crashing Issue - #9 by Joris_Kurst

I was so keen to be switching to AMD, but now I’m wondering if I make the jump back to Intel?