-
Notifications
You must be signed in to change notification settings - Fork 45
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Upgrade to Bookworm #231
base: devel
Are you sure you want to change the base?
Upgrade to Bookworm #231
Conversation
build: - add variables: cleanup (build.sh), debug_cache (update.sh) - /etc/mdev.conf: /dev/(null|zero|full|u?random) mode 666 (a=rw) fixes installer apt failure - output repo name for selected packages (update.sh) installer: - release: bookworm replaces bullseye - Debian Bookworm package changes: ntpsec replaces ntp, sysvinit-utils replaces lsb-base, systemd-resolved/systemd-timesyncd become presets - network time sync: remove ntpdate (python dep.), update rdate servers, http fallback: new http-date method replaces defunct method - add cdebootstrap_debug variable - resolve apt warnings, deprecated apt-key: install main repo gpg keys to /etc/apt/trusted.gpg.d/ doc: - preset packages reflect Debian Bookworm convention - add cdebootstrap_debug advanced install parameter
Hopefully this fixes #229 ! |
I just tried this and it tripped on
As far as I can tell, this is because |
Well, changing that path allows |
No, it was the 20 byte initramfs that did that. I was missing |
Would you prefer me to file separate issues on Bookworm-related issues, or keep them here in this PR, by the way? |
You can keep them here if you want. Thanks for your effort! |
Well, I built the installer, and it ran through almost the entire process on my ancient RPi 2B without issue 🎊 Unfortunately it fell over at installing packages into the new system:
Full log (~1000 lines!).
I am a bit baffled by this, because
...and
The first mirror definitely contains eg. aptitude. I don't see any errors relating to eg. signing keys or index downloads, so I don't understand why it isn't found. |
Can confirm this issue occurs today. It was not present when the PR was submitted. To make the build script (
Thanks for trying this out, hope you get it working. I actually have bookworm (built with this PR) running on a 2B without issue since early March.
The installer may have failed to update the package lists somehow. Your log indicates success: |
I can (I have a monitor and keyboard connected), but there was not much to look at. The install log contains only what I posted there, and I can't find any other incriminating details. However, it looks like this corresponds to L2267 in |
I'm getting a lot of "temporary failure in name resolution" errors, which is weird because a minute later
There's no
I suspect it's a DNS/systemd misconfiguration I've encountered before. Let me try again without |
Yes, setting My totally unverified hunch is that this is a variant of systemd's #4621/Ubuntu's issue #1624320, but I have not dug into it. I'd suggest that you look at how you configure DNS in the installer for |
I applied the patches from this conversation onto the "devel" branch and patched "build.sh". This was the commands I used to build the image:
Building went fine. I burned the image onto the sdcard. On the sdcard I created this config file "raspberrypi-ua-netinst/config/installer-config.txt":
Booting worked fine without "SOS"-code. But it doesn't continue from there. It seems like the installer script is not run. |
Connecting a display to your pi may help to see where it's stalling. Or if you comment out Another thing to try is a different mirror, or comment out the |
Have just been working through this PR trying to build on a Pi 4 and thought I'd add some feedback/new bugs...
I also had to update the However when it boots it fails almost immediately with I've had a look at Replacing Any hints on how to debug this further? |
Thanks for testing this PR and sharing a fix for the checksum issue.
Maybe try the installer on different storage media. |
Thanks for the suggestions...
I've tried this on different SD cards, and have rebuilt the image a couple of times and still get the same issue.
Nothing obviously 'wrong' with busybox - it has the same md5sum as the one installed from busybox_1.35.0-4_armhf.deb on a manually built working bookworm install, and is executable. When I strace busybox on the manual install it opens 3 libs - file:
readelf:
I've also tried changing config.txt to have |
Data corruption appears to be ruled out, at least with the busybox binary, assuming the SD cards you tried are otherwise fine. May help to check block devices with Were you able to |
Something else I just discovered (because I was using a RPi 4): commit b651d0d |
Good catch. Have not had a chance to test on a 64-bit host. Testing on a 32-bit host (RPi 2B) did not reveal any key issues. |
No description provided.