Commit ecd89093 authored by Thiago Kenji Okada's avatar Thiago Kenji Okada
Browse files

nixos-rebuild: run activation inside systemd-run 

Right now, running `nixos-rebuild switch` in a remote system via SSH may
put the system in an unusable state by restarting services (e.g.:
network ones like systemd-networkd.service) during the update. This will
cause the SSH to lose access to the TTY, stopping the process entirely.

This commit wraps up the call to `switch-to-configuration` inside a
systemd-run call, making the switch resiliant against TTY loss, since it
will allocate its own TTY. For the user this should be entirely
transparent though, with the only visible change is that the
`switch-to-configuration` messages will be logged through journalctl
(e.g.: `journalctl -u nixos-rebuild-switch-to-configuration`).
parent 3d007e8a

pkgs/os-specific/linux/nixos-rebuild/nixos-rebuild.8

+8 −0
Original line number Diff line number Diff line
@@ -421,6 +421,14 @@ nixpkgs=./my-nixpkgs 
Additional options to be passed to

.Ic ssh

on the command line.

.Ed

.

.It Ev NIXOS_SWITCH_USE_DIRTY_ENV

Expose the the current environment variables to post activation scripts. Will

skip usage of

.Ic systemd-run

during system activation. Possibly dangerous, specially in remote environments

(e.g.: via SSH). Will be removed in the future.

.El

.

.

pkgs/os-specific/linux/nixos-rebuild/nixos-rebuild.sh

+34 −4
Original line number Diff line number Diff line
@@ -653,18 +653,48 @@ fi 
# If we're not just building, then make the new configuration the boot

# default and/or activate it now.

if [[ "$action" = switch || "$action" = boot || "$action" = test || "$action" = dry-activate ]]; then

    # Using systemd-run here to protect against PTY failures/network

    # disconnections during rebuild.

    # See: https://github.com/NixOS/nixpkgs/issues/39118

    cmd=(

        "systemd-run"

        "-E" "LOCALE_ARCHIVE" # Will be set to new value early in switch-to-configuration script, but interpreter starts out with old value

        "--collect"

        "--no-ask-password"

        "--pty"

        "--quiet"

        "--same-dir"

        "--service-type=exec"

        "--unit=nixos-rebuild-switch-to-configuration"

        "--wait"

    )

    # Check if we have a working systemd-run. In chroot environments we may have

    # a non-working systemd, so we fallback to not using systemd-run.

    # You may also want to explicitly set NIXOS_SWITCH_USE_DIRTY_ENV environment

    # variable, since systemd-run runs inside an isolated environment and

    # this may break some post-switch scripts. However keep in mind that this

    # may be dangerous in remote access (e.g. SSH).

    if [[ -n "$NIXOS_SWITCH_USE_DIRTY_ENV" ]]; then

        log "warning: skipping systemd-run since NIXOS_SWITCH_USE_DIRTY_ENV is set. This environment variable will be ignored in the future"

        cmd=()

    elif ! targetHostCmd "${cmd[@]}" true &>/dev/null; then

        logVerbose "Skipping systemd-run to switch configuration since it is not working in target host."

        cmd=("env" "-i" "LOCALE_ARCHIVE=$LOCALE_ARCHIVE")

    else

        logVerbose "Using systemd-run to switch configuration."

    fi

    if [[ -z "$specialisation" ]]; then

        cmd="$pathToConfig/bin/switch-to-configuration"

        cmd+=("$pathToConfig/bin/switch-to-configuration")

    else

        cmd="$pathToConfig/specialisation/$specialisation/bin/switch-to-configuration"

        cmd+=("$pathToConfig/specialisation/$specialisation/bin/switch-to-configuration")



        if [[ ! -f "$cmd" ]]; then

        if [[ ! -f "${cmd[-1]}" ]]; then

            log "error: specialisation not found: $specialisation"

            exit 1

        fi

    fi



    if ! targetHostCmd "$cmd" "$action"; then

    if ! targetHostCmd "${cmd[@]}" "$action"; then

        log "warning: error(s) occurred while switching to the new configuration"

        exit 1

    fi