fstab needs to have valid entries before running genkernel otherwise the system tries to mount the default /dev/boot. Update fstab entries should be before run genkernel in the handbook. This applies to the x86 manual and could apply to others as well. Reproducible: Didn't try Steps to Reproduce: 1. Follow install instruction from manual - system booted from live cd 2. 3.
genkernel maintainers, is this true? We've never had a problem with this in the past; that's why the handbook has remained the way it is, and we have lots of genkernel users. Is this a simple case of PEBKAC, or has genkernel's behavior changed?
Yes, genkernel tries to mount it. However, it should just display a warning if it can't mount it and continue. Jonathan, is this actually preventing you from using genkernel?
(In reply to comment #2) > Yes, genkernel tries to mount it. However, it should just display a warning if > it can't mount it and continue. > > Jonathan, is this actually preventing you from using genkernel? > genkernel displays error message but continues to function but attempt to mount /dev/boot which has not been correctly set means chroot environment does not function correctly & reboot with chroot is required to continue installation. Swapping order of fstab & genkernel in the manual would cure the problem - why not just fix it?
Why is a reboot with the chroot necessary? Unless /dev/BOOT is an actual device node, the mount will always fail and genkernel will plod along like normal. If the problem is that the kernel and initramfs are going in /boot on / instead of of a separate /boot partition, that's because you're not following the guide. You still haven't told us what the actual problem is.
Reading through the handbook it tells you to use genkernel and then after do other things like update fstab. Genkernel tries to mount partitions specified in fstab and does not manage to do so because it has invalid entries (this is my guess from the error message ... something like /dev/boot failed to mount). Genkernel completes and manages to write files correctly to /boot but because of the attempt to mount the chrooted environment gets confused ... I don't think it understood mount anymore. I don't think this is a problem with genkernel - its just a problem with the handbook. If fstab was updated before running genkernel it would have worked - after reboot I updated fstab before running genkernel & it worked.
Seems to be harmless for every single other genkernel user who follows the handbook. agaffney, I'm thinking of just closing this bug, unless you agree with the reporter that the handbook needs a rewrite (and it would be extensive), rather than any changes made to genkernel itself. I'm just not convinced that the handbook or genkernel itself need any changes.
The effects that he's describing sound to me like the "normal" behavior of a chroot where its /etc/mtab doesn't match the system one or /proc/mounts. I doubt it has anything to do with genkernel trying to mount a non-existant device. I'm all for leaving the handbook as it is.
(In reply to comment #7) > The effects that he's describing sound to me like the "normal" behavior of a > chroot where its /etc/mtab doesn't match the system one or /proc/mounts. I > doubt it has anything to do with genkernel trying to mount a non-existant > device. I'm all for leaving the handbook as it is. To quote Yul Brynner, "So let it be written, so let it be done."
*** Bug 206734 has been marked as a duplicate of this bug. ***
*** Bug 237443 has been marked as a duplicate of this bug. ***
