fix(gateway): detect legacy hermes.service + mark --replace SIGTERM as planned (#11909)
* fix(gateway): detect legacy hermes.service units from pre-rename installs Older Hermes installs used a different service name (hermes.service) before the rename to hermes-gateway.service. When both units remain installed, they fight over the same bot token — after PR #5646's signal-recovery change, this manifests as a 30-second SIGTERM flap loop between the two services. Detection is an explicit allowlist (no globbing) plus an ExecStart content check, so profile units (hermes-gateway-<profile>.service) and unrelated third-party services named 'hermes' are never matched. Wired into systemd_install, systemd_status, gateway_setup wizard, and the main hermes setup flow — anywhere we already warn about scope conflicts now also warns about legacy units. * feat(gateway): add migrate-legacy command + install-time removal prompt - New hermes_cli.gateway.remove_legacy_hermes_units() removes legacy unit files with stop → disable → unlink → daemon-reload. Handles user and system scopes separately; system scope returns path list when not running as root so the caller can tell the user to re-run with sudo. - New 'hermes gateway migrate-legacy' subcommand (with --dry-run and -y) routes to remove_legacy_hermes_units via gateway_command dispatch. - systemd_install now offers to remove legacy units BEFORE installing the new hermes-gateway.service, preventing the SIGTERM flap loop that hits users who still have pre-rename hermes.service around. Profile units (hermes-gateway-<profile>.service) remain untouched in all paths — the legacy allowlist is explicit (_LEGACY_SERVICE_NAMES) and the ExecStart content check further narrows matches. * fix(gateway): mark --replace SIGTERM as planned so target exits 0 PR #5646 made SIGTERM exit the gateway with code 1 so systemd's Restart=on-failure revives it after unexpected kills. But when a user has two gateway units fighting for the same bot token (e.g. legacy hermes.service + hermes-gateway.service from a pre-rename install), the --replace takeover itself becomes the 'unexpected' SIGTERM — the loser exits 1, systemd revives it 30s later, and the cycle flaps indefinitely. Before calling terminate_pid(), --replace now writes a short-lived marker file naming the target PID + start_time. The target's shutdown_signal_handler consumes the marker and, when it names this process, leaves _signal_initiated_shutdown=False so the final exit code stays 0. Staleness defences: - PID + start_time combo prevents PID reuse matching an old marker - Marker older than 60s is treated as stale and discarded - Marker is unlinked on first read even if it doesn't match this process - Replacer clears the marker post-loop + on permission-denied give-up
This commit is contained in:
Teknium
@@ -687,6 +687,195 @@ def has_conflicting_systemd_units() -> bool:
|
||||
return len(get_installed_systemd_scopes()) > 1
|
||||
|
||||
|
||||
# Legacy service names from older Hermes installs that predate the
|
||||
# hermes-gateway rename. Kept as an explicit allowlist (NOT a glob) so
|
||||
# profile units (hermes-gateway-*.service) and unrelated third-party
|
||||
# "hermes" units are never matched.
|
||||
_LEGACY_SERVICE_NAMES: tuple[str, ...] = ("hermes.service",)
|
||||
|
||||
# ExecStart content markers that identify a unit as running our gateway.
|
||||
# A legacy unit is only flagged when its file contains one of these.
|
||||
_LEGACY_UNIT_EXECSTART_MARKERS: tuple[str, ...] = (
|
||||
"hermes_cli.main gateway",
|
||||
"hermes_cli/main.py gateway",
|
||||
"gateway/run.py",
|
||||
" hermes gateway ",
|
||||
"/hermes gateway ",
|
||||
)
|
||||
|
||||
|
||||
def _legacy_unit_search_paths() -> list[tuple[bool, Path]]:
|
||||
"""Return ``[(is_system, base_dir), ...]`` — directories to scan for legacy units.
|
||||
|
||||
Factored out so tests can monkeypatch the search roots without touching
|
||||
real filesystem paths.
|
||||
"""
|
||||
return [
|
||||
(False, Path.home() / ".config" / "systemd" / "user"),
|
||||
(True, Path("/etc/systemd/system")),
|
||||
]
|
||||
|
||||
|
||||
def _find_legacy_hermes_units() -> list[tuple[str, Path, bool]]:
|
||||
"""Return ``[(unit_name, unit_path, is_system)]`` for legacy Hermes gateway units.
|
||||
|
||||
Detects unit files installed by older Hermes versions that used a
|
||||
different service name (e.g. ``hermes.service`` before the rename to
|
||||
``hermes-gateway.service``). When both a legacy unit and the current
|
||||
``hermes-gateway.service`` are active, they fight over the same bot
|
||||
token — the PR #5646 signal-recovery change turns this into a 30-second
|
||||
SIGTERM flap loop.
|
||||
|
||||
Safety guards:
|
||||
|
||||
* Explicit allowlist of legacy names (no globbing). Profile units such
|
||||
as ``hermes-gateway-coder.service`` and unrelated third-party
|
||||
``hermes-*`` services are never matched.
|
||||
* ExecStart content check — only flag units that invoke our gateway
|
||||
entrypoint. A user-created ``hermes.service`` running an unrelated
|
||||
binary is left untouched.
|
||||
* Results are returned purely for caller inspection; this function
|
||||
never mutates or removes anything.
|
||||
"""
|
||||
results: list[tuple[str, Path, bool]] = []
|
||||
for is_system, base in _legacy_unit_search_paths():
|
||||
for name in _LEGACY_SERVICE_NAMES:
|
||||
unit_path = base / name
|
||||
try:
|
||||
if not unit_path.exists():
|
||||
continue
|
||||
text = unit_path.read_text(encoding="utf-8", errors="ignore")
|
||||
except (OSError, PermissionError):
|
||||
continue
|
||||
if not any(marker in text for marker in _LEGACY_UNIT_EXECSTART_MARKERS):
|
||||
# Not our gateway — leave alone
|
||||
continue
|
||||
results.append((name, unit_path, is_system))
|
||||
return results
|
||||
|
||||
|
||||
def has_legacy_hermes_units() -> bool:
|
||||
"""Return True when any legacy Hermes gateway unit files exist."""
|
||||
return bool(_find_legacy_hermes_units())
|
||||
|
||||
|
||||
def print_legacy_unit_warning() -> None:
|
||||
"""Warn about legacy Hermes gateway unit files if any are installed.
|
||||
|
||||
Idempotent: prints nothing when no legacy units are detected. Safe to
|
||||
call from any status/install/setup path.
|
||||
"""
|
||||
legacy = _find_legacy_hermes_units()
|
||||
if not legacy:
|
||||
return
|
||||
print_warning("Legacy Hermes gateway unit(s) detected from an older install:")
|
||||
for name, path, is_system in legacy:
|
||||
scope = "system" if is_system else "user"
|
||||
print_info(f" {path} ({scope} scope)")
|
||||
print_info(" These run alongside the current hermes-gateway service and")
|
||||
print_info(" cause SIGTERM flap loops — both try to use the same bot token.")
|
||||
print_info(" Remove them with:")
|
||||
print_info(" hermes gateway migrate-legacy")
|
||||
|
||||
|
||||
def remove_legacy_hermes_units(
|
||||
interactive: bool = True,
|
||||
dry_run: bool = False,
|
||||
) -> tuple[int, list[Path]]:
|
||||
"""Stop, disable, and remove legacy Hermes gateway unit files.
|
||||
|
||||
Iterates over whatever ``_find_legacy_hermes_units()`` returns — which is
|
||||
an explicit allowlist of legacy names (not a glob). Profile units and
|
||||
unrelated third-party services are never touched.
|
||||
|
||||
Args:
|
||||
interactive: When True, prompt before removing. When False, remove
|
||||
without asking (used when another prompt has already confirmed,
|
||||
e.g. from the install flow).
|
||||
dry_run: When True, list what would be removed and return.
|
||||
|
||||
Returns:
|
||||
``(removed_count, remaining_paths)`` — remaining includes units we
|
||||
couldn't remove (typically system-scope when not running as root).
|
||||
"""
|
||||
legacy = _find_legacy_hermes_units()
|
||||
if not legacy:
|
||||
print("No legacy Hermes gateway units found.")
|
||||
return 0, []
|
||||
|
||||
user_units = [(n, p) for n, p, is_sys in legacy if not is_sys]
|
||||
system_units = [(n, p) for n, p, is_sys in legacy if is_sys]
|
||||
|
||||
print()
|
||||
print("Legacy Hermes gateway unit(s) found:")
|
||||
for name, path, is_system in legacy:
|
||||
scope = "system" if is_system else "user"
|
||||
print(f" {path} ({scope} scope)")
|
||||
print()
|
||||
|
||||
if dry_run:
|
||||
print("(dry-run — nothing removed)")
|
||||
return 0, [p for _, p, _ in legacy]
|
||||
|
||||
if interactive and not prompt_yes_no("Remove these legacy units?", True):
|
||||
print("Skipped. Run again with: hermes gateway migrate-legacy")
|
||||
return 0, [p for _, p, _ in legacy]
|
||||
|
||||
removed = 0
|
||||
remaining: list[Path] = []
|
||||
|
||||
# User-scope removal
|
||||
for name, path in user_units:
|
||||
try:
|
||||
_run_systemctl(["stop", name], system=False, check=False, timeout=90)
|
||||
_run_systemctl(["disable", name], system=False, check=False, timeout=30)
|
||||
path.unlink(missing_ok=True)
|
||||
print(f" ✓ Removed {path}")
|
||||
removed += 1
|
||||
except (OSError, RuntimeError) as e:
|
||||
print(f" ⚠ Could not remove {path}: {e}")
|
||||
remaining.append(path)
|
||||
|
||||
if user_units:
|
||||
try:
|
||||
_run_systemctl(["daemon-reload"], system=False, check=False, timeout=30)
|
||||
except RuntimeError:
|
||||
pass
|
||||
|
||||
# System-scope removal (needs root)
|
||||
if system_units:
|
||||
if os.geteuid() != 0:
|
||||
print()
|
||||
print_warning("System-scope legacy units require root to remove.")
|
||||
print_info(" Re-run with: sudo hermes gateway migrate-legacy")
|
||||
for _, path in system_units:
|
||||
remaining.append(path)
|
||||
else:
|
||||
for name, path in system_units:
|
||||
try:
|
||||
_run_systemctl(["stop", name], system=True, check=False, timeout=90)
|
||||
_run_systemctl(["disable", name], system=True, check=False, timeout=30)
|
||||
path.unlink(missing_ok=True)
|
||||
print(f" ✓ Removed {path}")
|
||||
removed += 1
|
||||
except (OSError, RuntimeError) as e:
|
||||
print(f" ⚠ Could not remove {path}: {e}")
|
||||
remaining.append(path)
|
||||
|
||||
try:
|
||||
_run_systemctl(["daemon-reload"], system=True, check=False, timeout=30)
|
||||
except RuntimeError:
|
||||
pass
|
||||
|
||||
print()
|
||||
if remaining:
|
||||
print_warning(f"{len(remaining)} legacy unit(s) still present — see messages above.")
|
||||
else:
|
||||
print_success(f"Removed {removed} legacy unit(s).")
|
||||
|
||||
return removed, remaining
|
||||
|
||||
|
||||
def print_systemd_scope_conflict_warning() -> None:
|
||||
scopes = get_installed_systemd_scopes()
|
||||
if len(scopes) < 2:
|
||||
@@ -1220,6 +1409,19 @@ def systemd_install(force: bool = False, system: bool = False, run_as_user: str
|
||||
if system:
|
||||
_require_root_for_system_service("install")
|
||||
|
||||
# Offer to remove legacy units (hermes.service from pre-rename installs)
|
||||
# before installing the new hermes-gateway.service. If both remain, they
|
||||
# flap-fight for the Telegram bot token on every gateway startup.
|
||||
# Only removes units matching _LEGACY_SERVICE_NAMES + our ExecStart
|
||||
# signature — profile units are never touched.
|
||||
if has_legacy_hermes_units():
|
||||
print()
|
||||
print_legacy_unit_warning()
|
||||
print()
|
||||
if prompt_yes_no("Remove the legacy unit(s) before installing?", True):
|
||||
remove_legacy_hermes_units(interactive=False)
|
||||
print()
|
||||
|
||||
unit_path = get_systemd_unit_path(system=system)
|
||||
scope_flag = " --system" if system else ""
|
||||
|
||||
@@ -1258,6 +1460,7 @@ def systemd_install(force: bool = False, system: bool = False, run_as_user: str
|
||||
_ensure_linger_enabled()
|
||||
|
||||
print_systemd_scope_conflict_warning()
|
||||
print_legacy_unit_warning()
|
||||
|
||||
|
||||
def systemd_uninstall(system: bool = False):
|
||||
@@ -1381,6 +1584,10 @@ def systemd_status(deep: bool = False, system: bool = False):
|
||||
print_systemd_scope_conflict_warning()
|
||||
print()
|
||||
|
||||
if has_legacy_hermes_units():
|
||||
print_legacy_unit_warning()
|
||||
print()
|
||||
|
||||
if not systemd_unit_is_current(system=system):
|
||||
print("⚠ Installed gateway service definition is outdated")
|
||||
print(f" Run: {'sudo ' if system else ''}hermes gateway restart{scope_flag} # auto-refreshes the unit")
|
||||
@@ -3137,6 +3344,10 @@ def gateway_setup():
|
||||
print_systemd_scope_conflict_warning()
|
||||
print()
|
||||
|
||||
if supports_systemd_services() and has_legacy_hermes_units():
|
||||
print_legacy_unit_warning()
|
||||
print()
|
||||
|
||||
if service_installed and service_running:
|
||||
print_success("Gateway service is installed and running.")
|
||||
elif service_installed:
|
||||
@@ -3594,3 +3805,14 @@ def gateway_command(args):
|
||||
else:
|
||||
print(" hermes gateway install # Install as user service")
|
||||
print(" sudo hermes gateway install --system # Install as boot-time system service")
|
||||
|
||||
elif subcmd == "migrate-legacy":
|
||||
# Stop, disable, and remove legacy Hermes gateway unit files from
|
||||
# pre-rename installs (e.g. hermes.service). Profile units and
|
||||
# unrelated third-party services are never touched.
|
||||
dry_run = getattr(args, 'dry_run', False)
|
||||
yes = getattr(args, 'yes', False)
|
||||
if not supports_systemd_services() and not is_macos():
|
||||
print("Legacy unit migration only applies to systemd-based Linux hosts.")
|
||||
return
|
||||
remove_legacy_hermes_units(interactive=not yes, dry_run=dry_run)
|
||||
|
||||
@@ -6442,6 +6442,31 @@ For more help on a command:
|
||||
# gateway setup
|
||||
gateway_subparsers.add_parser("setup", help="Configure messaging platforms")
|
||||
|
||||
# gateway migrate-legacy
|
||||
gateway_migrate_legacy = gateway_subparsers.add_parser(
|
||||
"migrate-legacy",
|
||||
help="Remove legacy hermes.service units from pre-rename installs",
|
||||
description=(
|
||||
"Stop, disable, and remove legacy Hermes gateway unit files "
|
||||
"(e.g. hermes.service) left over from older installs. Profile "
|
||||
"units (hermes-gateway-<profile>.service) and unrelated "
|
||||
"third-party services are never touched."
|
||||
),
|
||||
)
|
||||
gateway_migrate_legacy.add_argument(
|
||||
"--dry-run",
|
||||
dest="dry_run",
|
||||
action="store_true",
|
||||
help="List what would be removed without doing it",
|
||||
)
|
||||
gateway_migrate_legacy.add_argument(
|
||||
"-y",
|
||||
"--yes",
|
||||
dest="yes",
|
||||
action="store_true",
|
||||
help="Skip the confirmation prompt",
|
||||
)
|
||||
|
||||
gateway_parser.set_defaults(func=cmd_gateway)
|
||||
|
||||
# =========================================================================
|
||||
|
||||
@@ -2242,8 +2242,10 @@ def setup_gateway(config: dict):
|
||||
_is_service_running,
|
||||
supports_systemd_services,
|
||||
has_conflicting_systemd_units,
|
||||
has_legacy_hermes_units,
|
||||
install_linux_gateway_from_setup,
|
||||
print_systemd_scope_conflict_warning,
|
||||
print_legacy_unit_warning,
|
||||
systemd_start,
|
||||
systemd_restart,
|
||||
launchd_install,
|
||||
@@ -2261,6 +2263,10 @@ def setup_gateway(config: dict):
|
||||
print_systemd_scope_conflict_warning()
|
||||
print()
|
||||
|
||||
if supports_systemd and has_legacy_hermes_units():
|
||||
print_legacy_unit_warning()
|
||||
print()
|
||||
|
||||
if service_running:
|
||||
if prompt_yes_no(" Restart the gateway to pick up changes?", True):
|
||||
try:
|
||||
|
||||
Reference in New Issue
Block a user