On Server netinst 20250408 nightly, Anaconda does not boot into GTK and ends before showing the following error: Service org.fedoraproject.Anaconda.Modules.Storage has failed to start: Process org.fedoraproject.Anaconda.Modules.Storage exited with status 1. In the bottom of the screen, it also says that the "Pane is dead". On a BIOS VM, Anaconda starts normally and the installation proceeds correctly. There are some errors seen during the boot process, however I am not sure whether they are related or not, see attachments. I am also attaching logs. Reproducible: Always Steps to Reproduce: 1. Download and boot the Server-netinst image, 20250408 nightly on an EFI VM. 2. Arrive into the situation. Actual Results: Anaconda dies and system cannot be installed. Expected Results: Anaconda works normally and installation is possible.
Proposed as a Blocker for 42-final by Fedora user lruzicka using the blocker tracking app because: https://fedoraproject.org/wiki/Basic_Release_Criteria#Installer_must_run
Created attachment 2084029 [details] Error message displayed in Anaconda
Created attachment 2084030 [details] Failures during boot
Created attachment 2084031 [details] anaconda.log
Created attachment 2084032 [details] dbus.log
Created attachment 2084033 [details] storage.log
The same happens on 20250405 nightly with Anaconda version 42.27.11-1.fc42 being present in both nightlies.
On F41 Final, with Anaconda 41.35-1.fc41, everything works as expected.
On Everything netinst, Anaconda works just fine in the same nightly.
With the same image and my UEFI VM, I don't see this problem. I saw plymouthd crash once (but not on the next try), but it didn't prevent anaconda from starting. Lukas, please attach full system journal after booting the netinst, and your VM configuration: $ virsh dumpxml VMNAME > vm.xml Does it happen always for you? Are you using the same VM for testing the other images?
I just tested Fedora-Server-netinst-x86_64-42-20250409.n.0.iso here at a VM and is started anaconda gtk correctly, and installed the system without issues Virt-manager KVM qemu-system-x86_64 Q35 UEFI cpu - host-passthrough MD Ryzen 7 5700U with Radeon Graphics Fedora-Server-netinst-x86_64-42-20250409.n.0.iso
I'm unable to reproduce this. Tested Fedora-Everything-netinst-aarch64-42-1.1 and Fedora-Server-netinst-aarch64-42-1.1 (both with anaconda 42.27.12-1.fc42).
Screencast booting Fedora-Server-netinst-x86_64-42-20250409.n.0.iso on virt-manager qemu/KVM Anaconda 42.27.11-1.fc42 started successfully https://geraldosimiao.fedorapeople.org/Bugzilla/Bug2358599-Fedora-Server-netinst-x86_64-42-20250409-Video_2025-04-09_09-31-15.mp4
I am able to reproduce it every single time and I have tried like 5 times in a row. Now, it breaks the 6th time, I am adding the journalctl from the affected machine. Also, the VM settings. Yes, it happens always. Yes, I am using the same VM for testing the other images. I plan to test the latest ISO, too, and will report right away.
Created attachment 2084055 [details] journalctl
Created attachment 2084056 [details] Virtual machine settings
The issue was caused by a passthrough device (USB flash drive) containing an older Fedora Server image possibly with identical partition UUIDs to the ISO being tested. This might have led the kernel to mistakenly load the wrong initrd from the flash drive instead of the intended one from the ISO. As a result, critical modules like zram and xfs failed to load during boot. Removing the flash drive (passthrough device) resolved the issue. The ISO now boots into Anaconda correctly, and no errors are seen in journalctl or zramctl.