bootupctl: Clear failure status from previous runs

If for whatever reason a bootupd command fails, it will leave the
systemd service unit in a failed state and systemd will then refuse to
run a unit under the same name with `systemd-run` again until the
failure is cleared.

Thus systematically call `systemctl reset-failed` before calling
`systemd-run` to clear any potential failures from previous calls.

Do not check the return code of the systemctl command on purpose as it
may fail if the unit does not exists yet, i.e. if no bootupctl command
has been run yet.

Also ignore stdout/stderr to avoid showing unexpected errors messages to
users.

See: coreos#707
See: coreos#663

src/cli/bootupctl.rs

@@ -4,7 +4,7 @@ use clap::Parser;
 use log::LevelFilter;
 
 use std::os::unix::process::CommandExt;
-use std::process::Command;
+use std::process::{Command, Stdio};
 
 static SYSTEMD_ARGS_BOOTUPD: &[&str] = &[
     "--unit",
@@ -154,6 +154,14 @@ fn ensure_running_in_systemd() -> Result<()> {
     require_root_permission()?;
     let running_in_systemd = running_in_systemd();
     if !running_in_systemd {
+        // Clear any failure status that may have happened previously
+        let _r = Command::new("systemctl")
+            .arg("reset-failed")
+            .arg("bootupd.service")
+            .stdout(Stdio::null())
+            .stderr(Stdio::null())
+            .spawn()?
+            .wait()?;
         let r = Command::new("systemd-run")
             .args(SYSTEMD_ARGS_BOOTUPD)
             .args(std::env::args())