Thank you for isolating the issue. It is definitely the entire snapshot repo with all its packages that need to have their keys renewed since they all [expired: 2023-01-31]. At least we now know who to report the issue to.
We need to email
ftpmaster@ports-master.debian.org
and request them to renew their key since it expired 2 days ago and re-package all the packages with the new key.
The 2023 package is in the updated package list. If not able to update with the mentioned args, try wget/curl the package directly and install with dpkg -i debian-ports-archive-keyring_2023.02.01_all.deb
Good news: aptitude and emacs-nox work. The full unstable repo snapshot is available to install which is impressive. Java was already installed. I installed Rust nightly with no issues via the standard https://rustup.rs way. Rust likes to have a cc around so I installed build-essential,clang-15-tools, lld-15, llvm 15 as well with no issues.
There is a much simpler solution:
1- download the debian-ports-archive-keyring_2023.02.01_all.deb package from https://deb.debian.org/debian-ports/pool/main/d/debian-ports-archive-keyring/debian-ports-archive-keyring_2023.02.01_all.deb with wget or curl,
install with dpkg -i debian-ports-archive-keyring_2023.02.01_all.deb
2-Use the regular unstable ports directory in /etc/apt/sources.list: deb http://ftp.ports.debian.org/debian-ports/ unstable main
instead of (or comment out) #deb https://snapshot.debian.org/archive/debian-ports/20220616T194833Z unstable main
Thanks @johanhenselmans; 1. by itself didn’t fix anything for me (maybe it wasn’t supposed to), but 1. and 2. seems to work (and boy, it’s working itself through 895 updates now).
I also like a CLI only image version. Many I suspect use as 24/7/365. Therefore only need a CLI version that ssh to when need access.
A method to easily enable SSH from base image would be most helpful so do not need to connect HDMI and keyboard to download/enable SSH. For example with RaspberryOS all one needs to do to enable ssh on image is to touch /boot/ssh. When RaspberyPiOS boots it enables ssh and deletes the /boot/ssh file.
(“Can’t boot” is a particularly useless bug report - you need to share how fair it gets at the very least)
It worked for me, but I saw one scary potential failure: the apt upgrade proposed a change to /etc/defaults/u-boot which would probably break boot if accepted:
You’re right, I should’ve been much more specific.
After a bit of analysis, my machine actually boots up, but the HDMI isn’t working. And when I try the SSH connection(I had to connect my sd card to another machine to change ssh conf), this works fine. Is there any logs I should provide to troubleshoot this?
Plus, The apt suggested two configuration changes, one is mime.conf and another one is uboot. I used previous configuration for uboot, and I changed mime conf file to new one. Like you said this debian image has unique uboot configuration, so it shouldn’t be changed.
Well HDMI has never worked for me (see other threads). @Michael.Zhu has proposed workaround which should give 1920x1080 on my 2560x1440 monitor. I haven’t tested it as I mostly use it remotely anyway. I’ve seen plenty of other people with HDMI issues, so I assume it’s a know issue being worked.
I just read your comments from other threads. HDMI worked for me (mine is 1920x1080) before I update it. Even I tried to update firmware, but it doesn’t helped. I should find out more or just wait for fixes. Anyways, thank you for information.
Thanks for all the help on update && upgrade. Yesterday I installed more than 800 upgrades!
Now I have a question on Docker.
I have it installed:
Version: 20.10.23+dfsg1
API version: 1.41
Go version: go1.19.5
Git commit: 7155243
Built: Fri Jan 20 08:04:03 2023
OS/Arch: linux/riscv64
Context: default
Experimental: true
But I can’t get it running.
after executing:
sudo systemctl start docker
Job for docker.service failed because the control process exited with error code.
See “systemctl status docker.service” and “journalctl -xeu docker.service” for details.
I tried it on the debian image and there was a few things missing in the kernel.
Is this all the logs you get ? iirc there was a lot more detailed things it complained on when I messed around with it.
all the stuff between these two lines, run with journalctl -u docker -n 50:
Jan 26 12:19:21 starfive dockerd[8267]: time="2023-01-26T12:19:21.271999471Z" level=info msg="Starting up"
...
Jan 26 12:19:22 starfive dockerd[8267]: (exit status 4))
The journalctl was truncated a bit, but not too important. If you want the full rows then add --no-pager to the command.
This is the main issue so far, and it is related to the kernel and modules available. “Running modprobe bridge br_netfilter failed with message: modprobe: WARNING: Module bridge not found in directory /l"
When I got docker running I used this build-script in the cross-compiler. Either try using the config from it (did not work for me back then) or just build that one and try it out. Remember to also bring the modules over to it, not only the kernel.
You can simply follow the guide here.