feat(install): add pre-flight environment and disk-target validation

archangel went straight from filesystem selection into a destructive install behind only a root check and a ZFS module load. A missing tool, a BIOS boot, a too-small or in-use disk, or a dead network surfaced as a confusing abort partway through, sometimes after partitioning had already run.

Two gates now fail fast. validate_environment runs after filesystem selection, before any disk is touched: it confirms UEFI boot mode and that every required command is present, with the list coming from a new required_commands helper built like pacstrap_packages. validate_install_targets runs after disk selection, before the first wipe: it refuses a target that's mounted, holds active swap, or belongs to an imported pool or md array, rejects disks under 20 GB, and confirms a mirror is reachable via DNS plus a TCP probe (no ICMP, since some networks drop it).

I folded the install_failure_cleanup hardening into the same change. It now falls back to lazy unmounts, so a pacstrap-interrupted target with busy bind mounts still releases the pool and unmounts the EFI partition. Without that, the disk-in-use guard would block the very retry the cleanup exists to enable. "Re-run to retry" only holds if the disk is genuinely freed first.

The 20 GB floor is decimal on purpose. It reads as the natural minimum and clears a 20 GiB disk image with headroom instead of sitting on the boundary.

3 files changed, 224 insertions, 0 deletions

diff --git a/tests/unit/test_archangel.bats b/tests/unit/test_archangel.bats
index c7bbc56..c38dcb8 100644
--- a/tests/unit/test_archangel.bats
+++ b/tests/unit/test_archangel.bats
@@ -239,3 +239,136 @@ setup() {
     [[ " ${CALLS[*]} " == *" zpool list zroot "* ]]
     [[ " ${CALLS[*]} " != *" zpool export"* ]]
 }
+
+@test "install_failure_cleanup ZFS path falls back to lazy unmount when a mount is busy" {
+    FILESYSTEM=zfs
+    POOL_NAME=zroot
+    CALLS=()
+
+    # A pacstrap-interrupted target can leave busy mounts that a plain
+    # umount can't release; cleanup must retry lazily so the retry sees a
+    # clean disk. Non-lazy umount fails here; the -l fallback succeeds.
+    umount() {
+        CALLS+=("umount $*")
+        [[ "$*" == *"-l"* ]] && return 0
+        return 1
+    }
+    zpool() { CALLS+=("zpool $*"); return 0; }
+    warn() { :; }
+    error() { return 1; }
+
+    install_failure_cleanup || true
+
+    [[ " ${CALLS[*]} " == *" umount -l /mnt/efi "* ]]
+    [[ " ${CALLS[*]} " == *" umount -R -l /mnt "* ]]
+    # The pool still gets exported after the lazy unmount.
+    [[ " ${CALLS[*]} " == *" zpool export zroot "* ]]
+}
+
+@test "install_failure_cleanup Btrfs path falls back to lazy unmount when EFI is busy" {
+    FILESYSTEM=btrfs
+    CALLS=()
+
+    umount() {
+        CALLS+=("umount $*")
+        [[ "$*" == *"-l"* ]] && return 0
+        return 1
+    }
+    btrfs_cleanup() { CALLS+=("btrfs_cleanup"); }
+    btrfs_close_encryption() { CALLS+=("btrfs_close_encryption"); }
+    warn() { :; }
+    error() { return 1; }
+
+    install_failure_cleanup || true
+
+    [[ " ${CALLS[*]} " == *" umount -l /mnt/efi "* ]]
+}
+
+#############################
+# validate_environment
+#############################
+# Boundary wrappers (is_uefi_boot, required_commands) are stubbed so the
+# composition's fail-fast wiring is exercised without depending on the
+# host's firmware mode or installed tools. The real command list lives in
+# test_common.bats; the real UEFI/network probes run in the VM harness.
+
+@test "validate_environment errors when not booted in UEFI mode" {
+    is_uefi_boot() { return 1; }
+    required_commands() { return 0; }
+    FILESYSTEM=zfs
+    run validate_environment
+    [ "$status" -eq 1 ]
+    [[ "$output" == *"UEFI"* ]]
+}
+
+@test "validate_environment errors when a required command is missing" {
+    is_uefi_boot() { return 0; }
+    required_commands() { echo "definitely-not-a-real-cmd-xyz"; }
+    FILESYSTEM=zfs
+    run validate_environment
+    [ "$status" -eq 1 ]
+    [[ "$output" == *"definitely-not-a-real-cmd-xyz"* ]]
+}
+
+@test "validate_environment passes when UEFI present and commands resolve" {
+    is_uefi_boot() { return 0; }
+    required_commands() { echo "bash"; }
+    FILESYSTEM=zfs
+    run validate_environment
+    [ "$status" -eq 0 ]
+}
+
+#############################
+# validate_install_targets
+#############################
+# disk_in_use / disk_size_bytes / network_available are the system-boundary
+# wrappers; stubbing them drives the real composition + real
+# disk_meets_min_size. Live probes run in the VM harness on the happy path.
+
+@test "validate_install_targets errors when a disk is in use" {
+    SELECTED_DISKS=(/dev/sda)
+    disk_in_use() { return 0; }
+    disk_size_bytes() { echo 500107862016; }
+    network_available() { return 0; }
+    run validate_install_targets
+    [ "$status" -eq 1 ]
+    [[ "$output" == *"in use"* ]]
+}
+
+@test "validate_install_targets errors when a disk is too small" {
+    SELECTED_DISKS=(/dev/sda)
+    disk_in_use() { return 1; }
+    disk_size_bytes() { echo 1000000; }
+    network_available() { return 0; }
+    run validate_install_targets
+    [ "$status" -eq 1 ]
+    [[ "$output" == *"too small"* ]]
+}
+
+@test "validate_install_targets errors when disk size is unreadable" {
+    SELECTED_DISKS=(/dev/sda)
+    disk_in_use() { return 1; }
+    disk_size_bytes() { echo ""; }
+    network_available() { return 0; }
+    run validate_install_targets
+    [ "$status" -eq 1 ]
+}
+
+@test "validate_install_targets errors when the network is unreachable" {
+    SELECTED_DISKS=(/dev/sda)
+    disk_in_use() { return 1; }
+    disk_size_bytes() { echo 500107862016; }
+    network_available() { return 1; }
+    run validate_install_targets
+    [ "$status" -eq 1 ]
+    [[ "$output" == *"network"* || "$output" == *"connectivity"* ]]
+}
+
+@test "validate_install_targets passes when disks idle, large enough, network up" {
+    SELECTED_DISKS=(/dev/sda /dev/sdb)
+    disk_in_use() { return 1; }
+    disk_size_bytes() { echo 500107862016; }
+    network_available() { return 0; }
+    run validate_install_targets
+    [ "$status" -eq 0 ]
+}
diff --git a/tests/unit/test_common.bats b/tests/unit/test_common.bats
index 8ce7280..a639a4e 100644
--- a/tests/unit/test_common.bats
+++ b/tests/unit/test_common.bats
@@ -520,3 +520,51 @@ Boot0001* ZFSBootMenu"
     grep -qF 'HOOKS=(base udev)' "$f"
     rm -f "$f"
 }
+
+#############################
+# required_commands
+#############################
+
+@test "required_commands zfs includes zpool and zfs" {
+    run required_commands zfs
+    [ "$status" -eq 0 ]
+    [[ "$output" == *"zpool"* ]]
+    [[ "$output" == *"zfs"* ]]
+}
+
+@test "required_commands btrfs includes mkfs.btrfs and grub-install" {
+    run required_commands btrfs
+    [ "$status" -eq 0 ]
+    [[ "$output" == *"mkfs.btrfs"* ]]
+    [[ "$output" == *"grub-install"* ]]
+}
+
+@test "required_commands zfs excludes Btrfs-specific commands" {
+    run required_commands zfs
+    [ "$status" -eq 0 ]
+    [[ "$output" != *"mkfs.btrfs"* ]]
+    [[ "$output" != *"grub-install"* ]]
+}
+
+@test "required_commands btrfs excludes the zpool command" {
+    run required_commands btrfs
+    [ "$status" -eq 0 ]
+    [[ "$output" != *"zpool"* ]]
+}
+
+@test "required_commands includes partitioning + pacstrap commands for both filesystems" {
+    for fs in zfs btrfs; do
+        run required_commands "$fs"
+        [ "$status" -eq 0 ]
+        [[ "$output" == *"sgdisk"* ]]
+        [[ "$output" == *"wipefs"* ]]
+        [[ "$output" == *"partprobe"* ]]
+        [[ "$output" == *"mkfs.fat"* ]]
+        [[ "$output" == *"pacstrap"* ]]
+    done
+}
+
+@test "required_commands unknown filesystem returns 1" {
+    run required_commands ext4
+    [ "$status" -eq 1 ]
+}
diff --git a/tests/unit/test_disk.bats b/tests/unit/test_disk.bats
index 1b3cfba..6bc6d8d 100644
--- a/tests/unit/test_disk.bats
+++ b/tests/unit/test_disk.bats
@@ -232,3 +232,46 @@ partition_disks_setup() {
     [[ " ${CALLS[*]} " != *"sgdisk"* ]]
     [[ " ${CALLS[*]} " != *"wipefs"* ]]
 }
+
+#############################
+# disk_meets_min_size / MIN_DISK_BYTES
+#############################
+
+@test "MIN_DISK_BYTES is 20 GB (decimal)" {
+    [ "$MIN_DISK_BYTES" -eq 20000000000 ]
+}
+
+@test "disk_meets_min_size: exactly the minimum passes" {
+    run disk_meets_min_size 20000000000
+    [ "$status" -eq 0 ]
+}
+
+@test "disk_meets_min_size: one byte under the minimum fails" {
+    run disk_meets_min_size 19999999999
+    [ "$status" -eq 1 ]
+}
+
+@test "disk_meets_min_size: a 20 GiB disk image clears the 20 GB floor" {
+    run disk_meets_min_size 21474836480
+    [ "$status" -eq 0 ]
+}
+
+@test "disk_meets_min_size: a large disk passes" {
+    run disk_meets_min_size 500107862016
+    [ "$status" -eq 0 ]
+}
+
+@test "disk_meets_min_size: zero fails" {
+    run disk_meets_min_size 0
+    [ "$status" -eq 1 ]
+}
+
+@test "disk_meets_min_size: non-numeric input fails" {
+    run disk_meets_min_size notanumber
+    [ "$status" -eq 1 ]
+}
+
+@test "disk_meets_min_size: empty input fails" {
+    run disk_meets_min_size ""
+    [ "$status" -eq 1 ]
+}