Bluefin stability issues (need to reboot to clear)

General
1

image
image1209×763 85.8 KB

I’m seeing several issues with my Bluefin install. Right now I can only give a general description as once I start seeing them just about nothing seems to work until I reboot.

I will try to capture them as they pop up again, but in general the issues are:

1 - Distrobox updates using ‘ujust update’ hang at “Setting up host’s nvidia integration…”

I need to ‘ctri-c’ out and select ‘retry’ (Y) after which that particular distrobox that hung updates successfully (or it seems to) but I have 11 distroboxes, and usually about 1/2 of them (or more) hang at this point forcing me to ‘ctrl-c’ and retry. (just tried now and 6 of them hung and I had to ctrl-c and retry.

2 - For no apparent reason Flatpaks stop working, even manually, I don’t have the exact error message as once that starts unless I have something opened already to take notes, then I can open anything other than a terminal. I just have to reboot before I can open anything of any significance (even the text editor won’t open.) I will try and take a screenshot next time, or a photo.)

Any ideas on what might be happening? I don’ t have any big changes made to my boot system, just btrfassistant so I can more easily take snapshots.

Even if someone can point me in the right direction as to where to look or logs to check, that might help.

3 - there is also an issue with Distrobox, in using a custome ‘home directory’ with the Docker Hub Container Image Library | App Containerization repo … it creates the container, but the custom home directory doesn’t get changed, I need to use the quay.io/toolbox/ubuntu-toolbox:22.04 repo instead. (docker used used to work, but suddenly the docker repo stopped working for some reason?)

I can of course open issues for these, but not sure to whom I should open them for? Bluefin, Silverblue, Fedora, Docker, Distrobox, Ubuntu, Google?

Thanks

image

image
image1340×546 46 KB

2

For #1 you’d need to see where it’s getting hung up on the boxes in the logs.

For #3 it’s always better to use an image from the toolboxes repo as those have the right modifications to work better in distroboxes. We also publish ghcr.io/ublue-os/ubuntu-toolbox:latest too if you want to try that one. You may also may get rate limited if you pull a bunch from docker hub.

Not sure on the flatpaks, try running them via the command line and see what error they spit out.

1 Like
3

Thanks.

1 - Which logs do you mean? Can you point me to them? Or do you mean the output from the ujust update command? I can indeed look into the logs, but I’m totally new to ‘just’.

2 - Awesome, I would prefer the toolbox versions then. Problem solved (for me, anyway!)

3 - I have tried that, but I need to make sure I always have something opened to take notes with, since when this starts happened, I cannot open anything new.

Note: I use my machine mostly for running AI (inference/llm apps) which often causes me to get OOM errors on the GPU … not sure if it is relevant but it might be a memory issue (although I don’t use my nvidia for my monitor, for that I use the integrated video, but I haven’t ruled that out as a possible cause?,

4

Use the distrobox command for this, do a distrobox upgrade blah, with blah being the name of the box you want to upgrade, and then see if it spits out an error.

5

That’s even worst lol … now it hangs at:

Starting container… [ OK ]
Installing basic packages… [ OK ]
Setting up devpts mounts… [ OK ]
Setting up read-only mounts… [ OK ]
Setting up read-write mounts… [ OK ]
Setting up host’s sockets integration…

hope you can read the attached screenshot.

image
image1148×854 73.7 KB

FYI: breaking out with ctrl-c displays something like:

^C/usr/bin/distrobox-enter: line 685: /var/home/david/.cache/distrobox/.DRv19Free.fifo: No such file or directory

Cheers.

Note 1: This seems to be an issue with ‘entering’ into a container on startup. If I start all the containers and then attempt to update (using “distrobox upgrade --all --verbose” or “ujust upgrade”) they all update with no issue.

Note 2: ujust update just gave me a new system update, so I’ll reboot and try against, see if that helped anything.

Note 3:
My system update is pulling from docker: “Pulling manifest: ostree-image-signed:docker://ghcr.io/ublue-os/bluefin-dx-nvidia:40”

Is this an issue? Should I be pulling from quay?

Back in a few minutes after I reboot and test again.

6

Update: Nope, still the same issue … as long as the containers are running I can update them no problem, but if they are not running, then it’s hit or miss on whether they hang (the terminal session it seems, as the container itself seems to start okay, or at least ‘ok’ as far as I can tell since I can then enter it and so useful work) or not.

7

Not able to poke around myself at the moment, but it might be worth attempting to remove the 2>/dev/null redirect from the install script (around line 271). I’m a rank novice with Distrobox, at best, but it seems like that could be suppressing the very error message you need to be seeing.

1 Like
8

The flatpak issue may be this:

https://bugzilla.redhat.com/show_bug.cgi?id=2301725#c7

2 Likes