system-reinstall-bootc: Do not warn on unmounted LVM volumes

[jeckersb]

If the system has a swap partition (or any other volume which is not
currently mounted) the findmnt command will (expectedly) fail to
find it. Don't early exit in this case, instead just ignore that
volume. If it wasn't mounted in the first place, we don't need to
warn about it being unmounted after the reinstall operation is
complete.

Signed-off-by: John Eckersberg [email protected]
Closes: #1659

[gemini-code-assist]

Code Review

This pull request correctly prevents failures when encountering unmounted LVM volumes by ignoring errors from findmnt. My review includes a suggestion to improve robustness by logging these errors at a debug level instead of silently swallowing them. This helps in diagnosing unexpected issues without altering the intended behavior for unmounted volumes.

[henrywang]

Thanks @jeckersb working no this issue! I made a rebase to include #1642 for testing this fix.

[henrywang]

Just ignore CS10, fedora 42 and 43 testing-farm: failures. PR #1664 fixed the failure.

Looks CS9 still has issue:

02:11:53             out: Image localhost/bootc-integration is already present locally, skipping pull.
02:11:53             out: 
02:11:53             out: Found only one user (root) with 1 SSH authorized keys.
02:11:53             out: Would you like to import its SSH authorized keys
02:11:53             out: into the root user on the new bootc system?
02:11:53             out:
                Then you can login as root@ using those keys. [Y/n] 
                Found only one user (root) with 1 SSH authorized keys.
02:11:53             out: Would you like to import its SSH authorized keys
02:11:53             out: into the root user on the new bootc system?
02:11:53             out: Then you can login as root@ using those keys. yes
02:11:53             out: error: No such file or directory (os error 2)

[jeckersb]

Just ignore CS10, fedora 42 and 43 testing-farm: failures. PR #1664 fixed the failure.

Looks CS9 still has issue:

02:11:53             out: Image localhost/bootc-integration is already present locally, skipping pull.
02:11:53             out: 
02:11:53             out: Found only one user (root) with 1 SSH authorized keys.
02:11:53             out: Would you like to import its SSH authorized keys
02:11:53             out: into the root user on the new bootc system?
02:11:53             out:
                Then you can login as root@ using those keys. [Y/n] 
                Found only one user (root) with 1 SSH authorized keys.
02:11:53             out: Would you like to import its SSH authorized keys
02:11:53             out: into the root user on the new bootc system?
02:11:53             out: Then you can login as root@ using those keys. yes
02:11:53             out: error: No such file or directory (os error 2)

I'll clean up my uncommitted local changes that add a bunch of #[context()] all over the place and push it, hopefully should help narrow down where it's breaking for you.

[jeckersb]

I'll clean up my uncommitted local changes that add a bunch of #[context()] all over the place and push it, hopefully should help narrow down where it's breaking for you.

Well I accidentally threw that away somewhere but easy enough to regenerate with claude, I just told it to add context annotations to every function in system-reinstall-bootc that returns a Result. @henrywang that should help narrow it down where it's breaking for you.

[henrywang]

We got error: run: mount_warning: check_root_siblings: parse_volumes: No such file or directory (os error 2) log.

Feel free to add findmnt or any debug command in

bootc/tmt/plans/integration.fmf

Line 7 in 24f2dd0

script:

just before the last provision script.

[jeckersb]

We got error: run: mount_warning: check_root_siblings: parse_volumes: No such file or directory (os error 2) log.

Ahh I suspect the lvs binary isn't available. It looks like there's a bug at:

bootc/crates/system-reinstall-bootc/src/lvm.rs

Lines 27 to 28 in 24f2dd0

	if which::which("podman").is_err() {
	tracing::debug!("lvs binary not found. Skipping logical volume check.");

It's checking for the podman binary but really it should be checking for lvs instead. I suspect inside of the integration environment on c9s for some reason (it works ok on my laptop using c9s... 🤷) that lvs is absent so it's failing trying to run it. I'll push another commit to fix that and see if things start working...

[cgwalters]

Note that lvs is just a symlink to lvm which is what we should probably be using.

[cgwalters]

These strings are intended to be somewhat human readable. Not a blocker but how about "Looking at root sibling mounts" or something?

[jeckersb]

Yeah I just had claude generate context for all of the Result-returning functions in the crate so we would at least have some idea where to look for the problem. Not great but at least more useful than No such file or directory with absolutely no context at all.