fix(update): stop a timed-out git from leaving .git/index.lock behind

subprocess.run's timeout SIGKILLs only the direct git child, so the lock
is left behind and git's helpers are orphaned. On a slow SD card the same
pull then times out on every retry and recreates the lock, so removing it
by hand does not help.

- run_git starts git in its own session, stops the group with SIGTERM
  before SIGKILL, and clears the lock if nothing owns it. Update Code's
  pull, checkout, reset and stash go through it, with a 600s pull timeout.
- Read-only calls (describe --dirty, status, the update-check fetch) set
  GIT_OPTIONAL_LOCKS=0 so a timed-out poll cannot take the lock.
- A timed-out pull returns a clear error; a rebase left by an interrupted
  update is aborted before pulling.
- The manual-fix hint says to stop the services and kill running git.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
This commit is contained in:
Chuck
2026-10-29 11:27:45 -04:00
co-authored by Claude Sonnet 5.5
parent fa51ebab5f
commit e506c1b576
8 changed files with 176 additions and 37 deletions
+38 -20
View File
@@ -14,6 +14,7 @@ from web_interface.blueprints.api_v3 import (
)
import threading
from web_interface import git_index_lock
from web_interface.cache import get_cached, set_cached
from web_interface.system_metrics import collect_system_metrics, format_uptime
import web_interface.blueprints.api_v3 as _pkg
@@ -163,6 +164,7 @@ def check_for_update():
['git', 'fetch', '--quiet', '--tags', '--force', update_channel.REMOTE,
update_channel.BETA_BRANCH],
capture_output=True, timeout=10, cwd=cwd,
env=git_index_lock.read_only_env(),
)
if fetch_result.returncode != 0:
stderr = fetch_result.stderr.decode(errors='replace').strip()
@@ -332,7 +334,7 @@ def channel_status(project_dir=None, fetch=True):
def _perform_core_update_locked(stash_local_changes=True):
project_dir = str(PROJECT_ROOT)
from web_interface import update_channel, git_index_lock
from web_interface import update_channel
# A git killed mid-command (a timed-out pull, a power cut) leaves
# .git/index.lock behind, and every later update then fails with
@@ -343,6 +345,14 @@ def _perform_core_update_locked(stash_local_changes=True):
'message': 'Another git command is running in the LEDMatrix folder; '
'try again in a minute.'}
# A pull killed mid-rebase leaves .git/rebase-merge behind, and every
# later `pull --rebase` then fails. Nothing is running (checked above).
git_dir = Path(project_dir) / '.git'
if any((git_dir / name).exists() for name in ('rebase-merge', 'rebase-apply')):
logger.warning("aborting a rebase left by an interrupted update")
git_index_lock.run_git(['git', 'rebase', '--abort'], capture_output=True, text=True,
timeout=60, cwd=project_dir)
# Which code to move to: the newest release (stable) or main (beta).
# See web_interface/update_channel.py; it never picks an older commit.
channel, fetch_error = channel_status(project_dir)
@@ -434,25 +444,33 @@ def _perform_core_update_locked(stash_local_changes=True):
# Stable checks out the release tag, carrying edits across the way the
# pull's --autostash does.
channel_note = ''
if action == update_channel.ACTION_CHECKOUT_TAG:
result, autostash_note = update_channel.checkout_release(project_dir, channel.newest_release)
channel_note = f"Now on release {channel.newest_release} (stable channel). {autostash_note}".strip()
elif action == update_channel.ACTION_SWITCH_TO_BETA:
result, autostash_note = update_channel.checkout_beta_branch(project_dir)
if result.returncode == 0:
result = subprocess.run(['git', 'pull', '--rebase', '--autostash'],
capture_output=True, text=True, timeout=60, cwd=project_dir)
channel_note = f"Now following {update_channel.BETA_BRANCH} (beta channel). {autostash_note}".strip()
else:
result = subprocess.run(
pull_args,
capture_output=True,
text=True,
timeout=60,
cwd=project_dir
)
if channel.waiting and channel.newest_release:
channel_note = channel.message
try:
if action == update_channel.ACTION_CHECKOUT_TAG:
result, autostash_note = update_channel.checkout_release(project_dir, channel.newest_release)
channel_note = f"Now on release {channel.newest_release} (stable channel). {autostash_note}".strip()
elif action == update_channel.ACTION_SWITCH_TO_BETA:
result, autostash_note = update_channel.checkout_beta_branch(project_dir)
if result.returncode == 0:
result = git_index_lock.run_git(['git', 'pull', '--rebase', '--autostash'],
capture_output=True, text=True,
timeout=git_index_lock.MUTATING_TIMEOUT_SECONDS, cwd=project_dir)
channel_note = f"Now following {update_channel.BETA_BRANCH} (beta channel). {autostash_note}".strip()
else:
result = git_index_lock.run_git(
pull_args,
capture_output=True,
text=True,
timeout=git_index_lock.MUTATING_TIMEOUT_SECONDS,
cwd=project_dir
)
if channel.waiting and channel.newest_release:
channel_note = channel.message
except subprocess.TimeoutExpired:
# run_git has stopped git and removed the lock it left.
logger.warning("git timed out during Update Code")
return {'status': 'error', 'restart_required': False, 'dependency_failures': [],
'message': 'Update failed: git took too long (a slow network or SD card). '
'The partial update was cleaned up; try again.'}
# Give the branch tracking information so the next pull is a plain
# `git pull` — otherwise every update repeats the fallback.