Try to make the change as in the wiki for both boot@ and boot-user@ from within the system and not from chroot from live or other obarun installation.
It can be very confusing to explain, if I understood the process well myself which I feel I am not 100% and I have tried for months, when you make changes to a tree, a module, a service, you usually should be looking at the output from -v4 commands, 66-intree, and 66-inservice. Working from chroot some of the directories from where you might see output are mounted filesystems from the host, not the target you are modifying. So what you see may be confusing because the two systems may not be identical in names of trees, modules, and selection of services. So while in chroot trust yourself you followed the instructions correctly and don't look at the output of those commands because it may reflect the incompatibility of the two systems between them.
If you have made a mistake and your system didn't boot, obviously only through chroot can you modify and correct things. But don't let the output confuse you. This wasn't a problem in the past (although it was there) because eveyone's boot was identical, despite of things being enabled/disabled within that bundle of services that was called boot. Now each boot@ is nearly unique, from the name to what is part of the module and what has been left out.
Maybe using bare chroot and not arch-chroot might prevent some of this confusion, but I am still trying to figure it out myself as well. arch-chroot is a readable script in /usr/bin ... if you can study it and its options maybe you can deduce more than I have been able to associate with this issue... for example /run and /mnt/run ... and its contents.
It can be very confusing to explain, if I understood the process well myself which I feel I am not 100% and I have tried for months, when you make changes to a tree, a module, a service, you usually should be looking at the output from -v4 commands, 66-intree, and 66-inservice. Working from chroot some of the directories from where you might see output are mounted filesystems from the host, not the target you are modifying. So what you see may be confusing because the two systems may not be identical in names of trees, modules, and selection of services. So while in chroot trust yourself you followed the instructions correctly and don't look at the output of those commands because it may reflect the incompatibility of the two systems between them.
If you have made a mistake and your system didn't boot, obviously only through chroot can you modify and correct things. But don't let the output confuse you. This wasn't a problem in the past (although it was there) because eveyone's boot was identical, despite of things being enabled/disabled within that bundle of services that was called boot. Now each boot@ is nearly unique, from the name to what is part of the module and what has been left out.
Maybe using bare chroot and not arch-chroot might prevent some of this confusion, but I am still trying to figure it out myself as well. arch-chroot is a readable script in /usr/bin ... if you can study it and its options maybe you can deduce more than I have been able to associate with this issue... for example /run and /mnt/run ... and its contents.