Description of problem: When installing silverblue on btrfs with boot directory (not subvolume), the installation fails. Version-Release number of selected component (if applicable): Rawhide 2020-07-31. How reproducible: Install silverblue with one of the following partition schemes (using advanced-custom with blivet, using UEFI and GPT partition scheme): sda1 vfat /boot/efi sda2 btrfs / subvolume home /var/home OR sda1 vfat /boot/efi sda2 btrfs (not-mounted) subvolume root / subvolume home /var/home In both cases, the /boot is a subdirectory. If it were subvolume, it would succeed (but hit a different bug, see bellow for additional info). Actual results: Installation fails with error message in acaconda GUI: The following error occurred while installing. This is a fatal error and installation will be aborted. mount ['--bind', '/mnt/sysimage/boot/efi', '/mnt/sysroot/boot/efi'] exited with code 32 and following entries in anaconda-log: Running... mount --bind /mnt/sysimage/boot/efi /mnt/sysroot/boot/efi mount: /mnt/sysroot/boot/efi: mount point does not exist. Return code: 32 Expected results: Successful installation. Additional info: These are scenarios #3 and #4 described in https://bugzilla.redhat.com/show_bug.cgi?id=1753485 This is only install time problem; after this is solved, these scenarios may still hig bug described in https://bugzilla.redhat.com/show_bug.cgi?id=1862783.
I think this happens because the /boot dir hasn't been created yet, so the /boot/efi mount fails. But then, I'd expect the same problem to happen on any other file system. At least in Custom partitioning, using a /boot directory used to work, so I wonder if it's only a problem in Advanced-Custom.
Problem happens on Fedora-Silverblue-ostree-x86_64-Rawhide-20200731.n.0.iso with ext4 in the same situation, /boot dir on sysroot (no separate /boot). Problem doesn't happen with Fedora-Workstation-Live-x86_64-Rawhide-20200731.n.0.iso, so it seems to be a uniquely silverblue/rpm-ostree problem.
Created attachment 1703208 [details] anaconda.log
Created attachment 1703209 [details] program.log
Created attachment 1703210 [details] storage.log
Created attachment 1703211 [details] syslog
Dup of bug 1395910 which says this won't be fixed soon. So arguably the installer should just disallow this on Silverblue.
This bug appears to have been reported against 'rawhide' during the Fedora 33 development cycle. Changing version to 33.
PR with "fix": https://github.com/rhinstaller/anaconda/pull/2848 Basically Anaconda won't allow /boot as a directory, same as suggested in https://bugzilla.redhat.com/show_bug.cgi?id=1395910#c11
Hold up, this might be fixed in ostree... cc: cwalters https://github.com/ostreedev/ostree/pull/2149
Switching to ostree, see discussion in https://github.com/rhinstaller/anaconda/pull/2848
*** Bug 1862783 has been marked as a duplicate of this bug. ***
*** Bug 1887116 has been marked as a duplicate of this bug. ***
This suggests the necessary work isn't done for GRUB yet https://github.com/ostreedev/ostree/pull/2149#issuecomment-692164435 Setting this to rawhide. Not sure if the component should be rpm-ostree but I'm gonna leave that alone.
I'd say if /boot directory on the same file system as / is supported by rpm-ostree, it can also be supported on Btrfs whether directory or subvolume so long as the subvolume name is "boot". But if the use case requires arbitrarily named subvolumes, then that's going to need separate work. And if so, that needs an upstream issue filed describing the use case, and the design, scope and coordination for that work.
OK so this bug means /boot has an actual path of <Btrfs top-level>/root/boot where "root" can be an arbitrarily named subvolume and "boot" is a directory. I expect this isn't going to work until rpm-ostree becomes more aware of Btrfs specifics. So I'd say it's (a) not a bug (b) works as designed (c) the installer should disallow it. The problem with (c) is that there are a number of things the installer allows for rpm-ostree installs that don't work so *shrug*. I'll leave it up to Colin if he just wants to close this as won't fix, and advise on how to track this upstream and where.
The error in the bug report comes from the following place: pyanaconda/payload/rpmostreepayload.py#L390 (_prepare_mount_targets): # Handle mounts like /boot (except avoid /boot/efi; we just need the # toplevel), and any admin-specified points like /home (really # /var/home). Note we already handled /var above. Avoid recursion since # sub-mounts will be in the list too. We sort by length as a crude # hack to try to simulate the tree relationship; it looks like this # is handled in blivet in a different way. for mount in sorted(mount_points, key=len): # my note: mount_points = payload_utils.get_mount_points() if mount in ('/', '/var') or mount in api_mounts: # my note: api_mounts = ["/dev", "/proc", "/run", "/sys"] continue self._setup_internal_bindmount(mount, recurse=False) it will try to mount /boot/efi: "mount --bind /mnt/sysimage/boot/efi /mnt/sysroot/boot/efi", while: 1. /mnt/sysroot/boot is empty dir, 2. thus /mnt/sysroot/boot/efi doesn't exist, 3. /mnt/sysimage/boot was populated by _copy_bootloader_data(), but since /boot is not a separate mountpoint (just a directory on /), it won't get bind-mounted just before mounting /boot/efi (which is always a separate partition). /, as the /boot parent, is not bind-mounted either. The comment claims that /boot/efi should be avoided; but the code doesn't nothing to handle that case. Suggested solution: if /boot is not in mount_points, hardcode it's bind-mount just like /usr and api-mounts? Or, if it's not necessary, add '/boot/efi' to the skip list?
This bug appears to have been reported against 'rawhide' during the Fedora 34 development cycle. Changing version to 34.
This message is a reminder that Fedora Linux 34 is nearing its end of life. Fedora will stop maintaining and issuing updates for Fedora Linux 34 on 2022-06-07. It is Fedora's policy to close all bug reports from releases that are no longer maintained. At that time this bug will be closed as EOL if it remains open with a 'version' of '34'. Package Maintainer: If you wish for this bug to remain open because you plan to fix it in a currently maintained version, change the 'version' to a later Fedora Linux version. Thank you for reporting this issue and we are sorry that we were not able to fix it before Fedora Linux 34 is end of life. If you would still like to see this bug fixed and are able to reproduce it against a later version of Fedora Linux, you are encouraged to change the 'version' to a later version prior to this bug being closed.
Fedora Linux 34 entered end-of-life (EOL) status on 2022-06-07. Fedora Linux 34 is no longer maintained, which means that it will not receive any further security or bug fix updates. As a result we are closing this bug. If you can reproduce this bug against a currently maintained version of Fedora please feel free to reopen this bug against that version. If you are unable to reopen this bug, please file a new report against the current release. Thank you for reporting this bug and we are sorry it could not be fixed.
Should we reopen? Probably still an issue on Rawhide.
Created attachment 1888173 [details] storage.log f37 Fedora-Silverblue-ostree-x86_64-Rawhide-20220608.n.0.iso when I set the /boot mountpoint to device type btrfs and ensure it's on the same btrfs as /, /var, /home - the installation succeeds but does not subsequently boot; the BLS snippet path to kernel+initrd lacks inclusion of /boot/ for the subvolume these files reside in. Same ISO, installed to XFS without a /boot mountpoint, thus it's a directory, fails installation with message "failed to write bootloader configuration", attaching storage.log.
This bug appears to have been reported against 'rawhide' during the Fedora Linux 37 development cycle. Changing version to 37.
This message is a reminder that Fedora Linux 37 is nearing its end of life. Fedora will stop maintaining and issuing updates for Fedora Linux 37 on 2023-12-05. It is Fedora's policy to close all bug reports from releases that are no longer maintained. At that time this bug will be closed as EOL if it remains open with a 'version' of '37'. Package Maintainer: If you wish for this bug to remain open because you plan to fix it in a currently maintained version, change the 'version' to a later Fedora Linux version. Note that the version field may be hidden. Click the "Show advanced fields" button if you do not see it. Thank you for reporting this issue and we are sorry that we were not able to fix it before Fedora Linux 37 is end of life. If you would still like to see this bug fixed and are able to reproduce it against a later version of Fedora Linux, you are encouraged to change the 'version' to a later version prior to this bug being closed.
Fedora Linux 37 entered end-of-life (EOL) status on None. Fedora Linux 37 is no longer maintained, which means that it will not receive any further security or bug fix updates. As a result we are closing this bug. If you can reproduce this bug against a currently maintained version of Fedora Linux please feel free to reopen this bug against that version. Note that the version field may be hidden. Click the "Show advanced fields" button if you do not see the version field. If you are unable to reopen this bug, please file a new report against an active release. Thank you for reporting this bug and we are sorry it could not be fixed.
This probably wasn't fixed, can we reopen?