diff --git a/CHANGELOG.md b/CHANGELOG.md index 642a2d6..034b727 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -204,6 +204,24 @@ which records not just what changed but what each drill run proved. ### Fixed +- **A wedged `incus launch` fails loudly, not forever — the mint's launch + phase is narrated and time-boxed** (#93) — twice in the 2026-07-19 + release drill (Debian 13, Incus 6.x, /dev/kvm present, images cached), + the child `incus launch` under `box new` hung with *no server-side + operation*: `incus operation list` empty, the instance never created, the + daemon journal quiet — one wedge ran 56 minutes before being killed by + hand, and an immediate retry of the identical command succeeded in + minutes, both times. `box new` inherited that as an indefinite silent + hang, indistinguishable from a cold mint working. It now prints + `launching instance …` before the call, and the call rides + `timeout -k 5 $BOX_LAUNCH_TIMEOUT` (seconds, default 600 — generous: the + coldest measured mint is minutes, never an hour; the same scripting-knob + shape as `BOX_CPU`/`BOX_MEMORY`), with stdin pinned per the drill's own + trap list. On the budget firing it says exactly what was measured — the + client wedged with no server-side operation, an immediate retry has been + observed to succeed — and points at `box doctor` for the host. The + `--from` clone path is untouched: `incus copy` of a local instance is a + different operation and has never been observed to wedge this way. - **UFW's gateway carve-out converges with the bridge, and the doctor can see it** (the #86 review's blind spot) — `box-firewall` gated its whole UFW block behind "a `DENY on boxnet` rule exists", pinning every UFW host diff --git a/bin/box b/bin/box index a669144..7c03554 100755 --- a/bin/box +++ b/bin/box @@ -1017,14 +1017,43 @@ cmd_new() { # The template's identity is stamped ONTO the instance: which template, # which user. 'incus copy' preserves user.* keys (audit B2), so a clone # knows what it is without ever consulting the template again. - incus launch "$T_IMAGE" "$instance" --profile box-net \ + # + # The launch is narrated and TIME-BOXED (#93). Twice in the 2026-07-19 + # release drill the child 'incus launch' wedged before the create was + # even accepted — 'incus operation list' empty, the instance never + # existed, the daemon journal quiet — once for 56 minutes until killed + # by hand. Without a line here that wedge reads exactly like a cold mint + # working; without a budget it lasts forever. Ten minutes is generous — + # the coldest measured mint (first VM on a fresh pool, see wait_agent) + # is minutes, never an hour — and BOX_LAUNCH_TIMEOUT (seconds) overrides + # it, the same scripting knob shape as BOX_CPU / BOX_MEMORY. The drill's + # lore applies verbatim (RUNS.md trap 13): 'timeout -k' so a launch that + # shrugs off TERM still dies, and stdin pinned like every other + # non-interactive incus call — only shell/exec/tmux may own the terminal. + local budget="${BOX_LAUNCH_TIMEOUT:-600}" rc=0 + echo "box: launching instance $instance (incus launch, $m mode)..." + timeout -k 5 "$budget" incus launch "$T_IMAGE" "$instance" --profile box-net \ --config user.box=1 \ --config user.box.template="$t" \ --config user.box.user="$T_USER" \ --config limits.cpu="$T_CPU" \ --config limits.memory="$T_MEMORY" \ --config cloud-init.user-data="$(render_userdata "$root/templates/$t/user-data.yaml")" \ - "${extra[@]}" + "${extra[@]}" &2 + echo "box: 'incus launch' WEDGED — killed after ${budget}s, and the instance was never created." >&2 + echo "box: this is the #93 failure: the incus client hangs with NO server-side operation" >&2 + echo "box: ('incus operation list' is empty, the daemon journal is quiet). An immediate" >&2 + echo "box: retry of the exact same 'box new' has been observed to succeed, both times it" >&2 + echo "box: was measured. If it persists, diagnose the host: box doctor" >&2 + echo "box: (a genuinely slower mint can raise the budget: BOX_LAUNCH_TIMEOUT=)" >&2 + die "incus launch wedged with no server-side operation — retry the same command (#93)" + elif [ "$rc" -ne 0 ]; then + # Not a wedge: incus refused and said why on stderr, right above. + die "incus launch failed (exit $rc)" + fi wait_agent "$instance" echo "box: waiting for phase-1 (cloud-init)..." echo "box: (its full narration, live: incus exec $name -- tail -f /var/log/cloud-init-output.log)" diff --git a/test/cli.sh b/test/cli.sh index a287993..edc3896 100644 --- a/test/cli.sh +++ b/test/cli.sh @@ -373,6 +373,38 @@ check "new: the auto-run sits under the T_BOOTSTRAP_ROLE guard" 0 "" bash -c ' check "new: a failed tenant role names the re-run (the role converges)" 0 "" bash -c ' awk "/^cmd_new\(\) \{/,/^\}/" "'"$ROOT"'/bin/box" \ | grep -q "sudo rig bootstrap"' + +# The launch phase, narrated and time-boxed (#93) — grepped the way the other +# mint-path guards are (a daemon-free run cannot mint). Twice in the +# 2026-07-19 release drill the child 'incus launch' wedged silently before +# the create was even accepted, once for 56 minutes. The narration must order +# BEFORE the launch call (a wedge after the line is visible at a glance; a +# wedge before it is the old silent hang), the call itself must sit under +# 'timeout -k' with the BOX_LAUNCH_TIMEOUT override and pinned stdin (RUNS.md +# trap 13: bare 'timeout N' cannot kill an incus call that owns a TTY), and +# the budget's failure must be LOUD — no server-side operation, the measured +# retry-succeeds hint, and the doctor as the next move. +# shellcheck disable=SC2016 # the $-strings are literals in the target file +check "new: the launch narration orders before incus launch (#93)" 0 "" bash -c ' + fn="$(awk "/^cmd_new\(\) \{/,/^\}/" "'"$ROOT"'/bin/box")" + say="$(printf "%s\n" "$fn" | grep -n "launching instance" | head -1 | cut -d: -f1)" + run="$(printf "%s\n" "$fn" | grep -n "timeout -k.*incus launch" | head -1 | cut -d: -f1)" + [ -n "$say" ] && [ -n "$run" ] && [ "$say" -lt "$run" ]' +check "new: incus launch is time-boxed (timeout -k on the budget)" 0 "" bash -c ' + awk "/^cmd_new\(\) \{/,/^\}/" "'"$ROOT"'/bin/box" \ + | grep "timeout -k" | grep "budget" | grep -q "incus launch"' +check "new: the budget is BOX_LAUNCH_TIMEOUT, default 600s (the BOX_CPU knob shape)" 0 "" bash -c ' + awk "/^cmd_new\(\) \{/,/^\}/" "'"$ROOT"'/bin/box" \ + | grep "budget=" | grep -q "BOX_LAUNCH_TIMEOUT:-600"' +check "new: the launch pins stdin (RUNS.md trap 13)" 0 "" bash -c ' + awk "/^cmd_new\(\) \{/,/^\}/" "'"$ROOT"'/bin/box" \ + | grep -F "extra[@]" | grep -qF "