I'm aware that the internet is packed with 'quiet splash' kernel config issues regarding the boot process on several hardware sets, which generally leads to graphics issues that can be prevented with 'nomodeset' or similar. This is not one of them.
On a fresh 20.04.1 Server installation (no additional packages installed, absolute installer-default minimum set), just adding quiet splash
to the GRUB_CMDLINE_LINUX_DEFAULT in /etc/default/grub (which is empty by default), breaks the boot process. Splash is shown and hangs there forever. No console login possible. Adding nomodeset
has no effect. Same thing on an Intel NUC w/ UEFI and within a Parallels VM.
When using the minimal set of of the corresponding Desktop image, quiet splash
is the default cmdline and splash works fine.
There are several questions I can't find an answer yet:
- What are the differences here? Shouldn't be a driver issue as only Kernel drivers are used, no proprietary sets. And AFAIK Ubuntu uses the same kernel config for desktop and sever since 12.04 or something.. Any hints which configurations to check for differences?
- Tips on how to debug this? As the issue only occurs with
quiet splash
set, I'm not able to see any logs from boot. I guess I could mount the partition with another system after a failed start and inspect log files, but is there a way that doesn't involve a second (or live) system?
Thanks!
Best Answer
It seem that remove the
$vt_handoff
option work, too. following is my boot option:I am using ubuntu 20.04 desktop,
vt_handoff
should be empty if$prefix/gfxblacklist.txt
exist, sotouch /boot/grub/gfxblacklist.txt
should fix this problem.