Lubuntu Hirsute daily install failed due to grub packaging errors

This bug has been reported on the Ubuntu ISO testing tracker.

A list of all reports related to this bug can be found here:
http://iso.qa.ubuntu.com/qatracker/reports/bugs/1917292

Lubuntu hirsute QA-test install on sony vaio thingy
testcase: uEFI, full disk, no internet, no encryption, no swap

Installation Failed
---
External command finished with errors.
Command apt install -y --no-upgrade -o Acquire::gpgv::Options::=--ignore-time-conflict grub-efi-$(if grep -q 64 /sys/firmware/efi/fw_platform_size; then echo amd64-signed; else echo ia32; fi) finished with exit code 100.
Output:
WARNING: apt does not have a stable CLI interface. Use with caution in scripts.

Reading package lists...
Building dependency tree...
Recommends: efibootmgr but it is not going to be installed
E: Unable to correct problems, you have held broken packages.

sorry I was intended to upload session.log

I didn't want to FAIL the QA-test, so I
- quit calamares
- restarted calamares
  - "manual partitioning" install..
but that too failed with the same error.

Not sure what's up with the Calamares logs, but the full error is:

Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.
The following information may help to resolve the situation:

The following packages have unmet dependencies:
 grub-efi-amd64-bin : Depends: grub-common (= 2.04-1ubuntu37) but 2.04-1ubuntu40 is to be installed
                      Recommends: efibootmgr but it is not going to be installed
E: Unable to correct problems, you have held broken packages.

This is the result of trying to install grub-efi-amd64-signed. src:grub2 and src:grub2-signed have been released (as 1ubuntu40) as of the 25th and efibootmgr hasn't changed in forever. Something is weird here with regards to packages.

It seems that 1ubuntu37 comes from the ship-live seed (i.e. on the ISO) while 1ubuntu40 is from the repos. Oddly, it seems to want to prefer the former? Still, I don't know why we have an outdated package.

Sorry for all the comments. Anyways, this is probably not a Calamares bug. Calamares may be contributing but a stale package on the ISO is a much bigger problem. Unfortunately, I'm not sure where the problem lies so I'm not entirely sure where to file it against. I've asked at #ubuntu-release and no response so far.

So our script for installing the EFI signed packages could use a refactor. If it was properly constructed, **assuming Internet was available**, one could have gotten around this issue. Right now, it forces the assumption that there is no Internet, so some small blame can be placed there.

One can resolve this by removing the following lines from /etc/calamares/modules/before_bootloader_context.conf:

        - command: sed -i '/deb http/d' /etc/apt/sources.list
              timeout: 10

There's a task to fix up the script here:
https://phab.lubuntu.me/T200

However, this is still not the cause of our problems.

We just checked with cjwatson and vorlon, it looks like this might be a stuck sync and mirror process. They're killing that and manually syncing it now, so the mismatch might go away in the next ISO build.

This issue was resolved by cjwatson and vorlon addressing the mismatching versions between what's in the repos and what was being put on the ISO (caused by a desync in the local ftp mirror the builder was using).

This is not a Cala bug nor a Grub bug, this was an ISO builder issue.

I can confirm 2 successful installs Lubuntu Hirsute daily http://cdimage.ubuntu.com/lubuntu/daily-live/20210302/hirsute-desktop-amd64.iso booting in UEFI+secure boot :

1. Lenovo X230i i3-3120M,8GB,Intel HD graphics,WiFi,Ethernet,Bluetooth,320GB hd
2. Lenovo V14 IIL i3-1005G1,20GB,Intel UHD Graphics,WiFi Realtek 802.11ac, bluetooth 4.2, 256GB M2NVMe SSD

http://iso.qa.ubuntu.com/qatracker/milestones/419/builds/227516/testcases/1701/results