My Whonix-Workstation on QEMU/KVM (on Manjaro Linux) no longer boots.
It only booted the very first time, but after that it since hangs right after booting through the grub menu at the message sgx: There are zero EPC sections.
Anyway, after the encryption / timeout messages are spammed for another couple of minutes, I finally get to messages where the system appears to stop a number of services and unmount volumes.
Right after, a “system halted” message appears and the system stays completely hung there.
As someone running this exact setup with Whonix on KVM in the past, I have never experienced anything like this before.
Perhaps (if related) I do remember before reinstalling Whonix, my previous (very outdated) Whonix-Workstation would only boot into CLI, regardless of how much RAM was assigned.
Any ideas?
Note: Bear in mind this was a fresh install, and all I did at first boot was login as user and tried to run upgrade-nonroot, which did not work I’m guessing due to the new sysmaint user. My host machine is LUKS encrypted, but I did not encrypt anything in the VM.
Successfully reproduced the issue on Manjaro KDE. (Curiously, neither Fedora KDE nor Kubuntu 24.04 has this problem.) After a bit of inspection, it looks like something is corrupting the partition table on the workstation, specifically by corrupting the value that specifies the end location of the root partition. I’m not entirely sure why yet, but that seems to be what’s happening. (Maybe systemd-repart is misbehaving?)
edit: Gah, I was able to reproduce this one time and now subsequent attempts to reproduce are failing. Will need to do more research on this later. For now, my advice would be that if you encounter this, try deleting the Whonix-Workstation VM entirely, then define it from the XML file and copy the qcow2 image to /var/lib/libvirt/images as described in the documentation. On the first boot, boot into “PERSISTENT Mode | SYSMAINT Session”, then install updates from there and reboot into a user session. It’s possible this is a semi-random glitch.
I created a new Whonix-Workstation with a fresh qcow2 image and immediately booted into sysmaint. I ran all updates and rebooted but sadly I ended up at the same issue.
I’ll redo this once more with a new image and report back.
In the meantime, could you look a bit into what’s corrupting the partition table? That behaviour is quite strange.
Note: To whoever edited the title, I don’t think the issue is due to the sgx message. As mentioned above, I receive the exact message in Whonix-Gateway without issue.
