Description of problem: Network installer ISO for custom "standard" components results in non-bootable system. Version-Release number of selected component (if applicable): How reproducible: We first encountered this with failed kickstart installs on 2018-01-24 where the same kickstart recipe worked for Fedora 29 the week before. The error can be reproduced with a manual GUI installation. Steps to Reproduce: 1. Prepare a qemu-kvm VM with blank disk via libvirt/virt-manager 2. Boot Fedora-Server-netinst-x86_64-29.1.2.iso 3. Do minimal custom install via GUI I chose custom partitioning but select the item to create /, /boot/, and swap automatically with the usual LVM partitioning scheme. I chose "nearest mirror" for software source. I chose custom install, minimal profile, and optional "standard" group. Actual results: Installer runs to completion, but resulting VM only bootstraps to the grub CLI. There are no bootable menu items. Expected results: New install should boot. Additional info: I can see that the /boot/ volume contains only an initramfs image for 4.20.3-200.fc29.x86_64 but no corresponding vmlinuz nor System.map. As a result, the grub.cfg doesn't have any boot items defined. On existing systems with Fedora 29, dnf automatic seems to have installed the same kernel normally. So there seems to be a difference between an update on a live system and via anaconda.
After further search, I guess this might be a duplicate of https://bugzilla.redhat.com/show_bug.cgi?id=1669256 which describes a systemd package dependency issue that causes the kernel-install step to fail if openssl-libs is not installed when the kernel-core package is installed from the updates repo.
*** This bug has been marked as a duplicate of bug 1669256 ***