ISSUE: Home Assistant Core crashes when reconfiguring or deleting Keymaster integration entry

[cjhesse]

Describe the bug

When I recently set up three locks (all Kwikset HC620 Z-Wave locks connected via zwave js), I didn't define the script entity for each one. Then, when I attempted to define that later (using the "reconfigure" option), it didn't take but eventually gave me an error at the top of the configuration window. I now believe Home Assistant Core crashed at this time, but I didn't make note of it then.

Whether or not Home Assistant Core crashed then, it's definitely crashing now anytime I try to modify or delete a Keymaster integration entry. Unfortunately, nothing related to the crash is being logged except exit code 256.

When Home Assistant Core reloads, the integration entity is unchanged. The reconfigure or delete has no lasting results.

Home Assistant version

2026.7.0 (running on HAOS 18.1 installed on Home Assistant Yellow with a CM5)

Keymaster version

v0.5.2

Home Assistant installation type

Home Assistant OS

Lock provider

zwave_js

Lock make and model

Kwikset HC620

Steps to reproduce

1. Go to the Keymaster integration page,
2. On any integration entry (not a device), click the three dots,
3. Click "Delete" and then confirm.

or

1. Go to the Keymaster integration page,
2. On any integration entry (not a device), click the three dots,
3. Select "Reconfigure",
4. Modify any configuration option and then submit.

At this point Home Assistant core crashes and has to restart. HAOS and all installed apps (add-ons) appear unaffected.

Expected behavior

The integration entry can be modified or deleted.

Logs

2026-07-03 11:21:22.014 ERROR (MainThread) [homeassistant.components.template.validators] Received invalid sensor state: unavailable for entity sensor.mal_phone_next_alarm_in_primary_bedroom, expected a valid timestamp
2026-07-03 11:21:24.762 ERROR (MainThread) [frontend.js.modern.202606243] Uncaught error from Edge 149.0.0.0 on Windows 10.0
Error: Failed to execute 'define' on 'CustomElementRegistry': the name "focus-trap" has already been used with this registry
A.define (node_modules/@webcomponents/scoped-custom-element-registry/src/scoped-custom-element-registry.ts:180:12)
/hacsfiles/advanced-camera-card/card-294f2ffb.js:70:3686

<This is when I attempted to delete a Keymaster integration entry.>

[16:23:40] INFO: Home Assistant Core finish process exit code 256
[16:23:40] INFO: Home Assistant Core finish process received signal 9
s6-rc: info: service legacy-services: stopping
s6-rc: info: service legacy-services successfully stopped
s6-rc: info: service legacy-cont-init: stopping
s6-rc: info: service legacy-cont-init successfully stopped
s6-rc: info: service fix-attrs: stopping
s6-rc: info: service fix-attrs successfully stopped
s6-rc: warning: service s6rc-oneshot-runner is marked as essential, not stopping it
s6-rc: info: service s6rc-oneshot-runner: starting
s6-rc: info: service s6rc-oneshot-runner successfully started
s6-rc: info: service fix-attrs: starting
s6-rc: info: service fix-attrs successfully started
s6-rc: info: service legacy-cont-init: starting
s6-rc: info: service legacy-cont-init successfully started
s6-rc: info: service legacy-services: starting
services-up: info: copying legacy longrun home-assistant (no readiness notification)
s6-rc: info: service legacy-services successfully started

Screenshots

Additional context

No response

[firstof9]

Is this the information from the home-assistant.log.1 file?

[secondof9]

Investigation Summary

I have performed a code audit of the keymaster repository in relation to issue #668. The reported crash (Exit 256/SIGKILL) suggests the Home Assistant event loop is being blocked, causing the OS/watchdog to terminate the process.

Potential Root Causes:

1. Blocking async_unload during Reconfiguration:
   When async_reload is called during reconfiguration, it triggers async_unload_entry. This function executes await kmlock.provider.async_unload(). If the provider's unload method performs synchronous/blocking I/O or hits a deadlock (especially when interacting with external controllers like Z-Wave JS), the main event loop will hang, leading to a SIGKILL.

2. Reconfiguration Reload Loop:
   config_flow.py triggers async_reload immediately after async_update_entry. If the entry state or the resulting async_unload triggers additional update events, it could initiate a recursive reload cycle that exhausts resources and crashes the core.

3. Synchronous Deletion/Storage I/O:
   The async_remove_entry and async_unload_entry functions perform significant cleanup, including interacting with hass.config_entries.async_unload and the Coordinator's storage. Any blocking filesystem operation or lock contention during this phase would cause the observed symptoms.

Suggested Next Steps:

• Audit all async_unload implementations in the providers directory to ensure they are fully non-blocking.
• Implement a guard in update_listener to prevent multiple overlapping async_reload calls.
• Review the interaction between async_update_entry and the unload logic to ensure no race conditions exist during the transition.

[cjhesse]

Is this the information from the home-assistant.log.1 file?

No, this was copied from the Home Assistant Core log in the UI (Settings > System > Logs > three dots > Show raw logs). As of 2025.11, /config/home-assistant.log is no longer updated to reduce writes to storage.

[firstof9]

@cjhesse please try this coordinator.py it goes in custom_components/keymaster/ and let me know if that helps with your issue at all.

Thanks.

[cjhesse]

I swapped out the original file for the variant you linked with the try and finally blocks, but it appears the result is the same. FWIW, it seemed like Home Assistant was taking longer to load that it had prior to installing and setting up Keymaster, but after changing out that file, startup time was much faster.

[firstof9]

Can you try adding this __init__.py as well

[cjhesse]

Same result, unfortunately. That's all for me tonight, so if you have other ideas, I won't be able to test them until sometime tomorrow.

[pshempel]

This issue for me is very much the same, but it actually will cause Home Assistant to restart. All I am doing is reconfiguring a door with new entity names. I see in the logs that HA says that the system was restarted abnormally and needs to clean up the db because of it. I will on ocassion, get the same error as the OP. In most cases it will force a restart of HA.

[firstof9]

If anyone can provide debug logs it would be very helpful.
Basically enable the debug logging, then cause the issue.

[cjhesse]

Here's the logs with custom_components.keymaster set to debug. I tried everything on debug in case that was helpful but that was... overwhelming.

2026-07-05 08:29:59.718 DEBUG (MainThread) [custom_components.keymaster.config_flow] [async_step_reconfigure] entry_id: 01KWK2Z75BR6YHF4KZQZ3ZJ11P, data: {'advanced_date_range': True, 'advanced_day_of_week': True, 'alarm_level_or_user_code_entity_id': 'sensor.back_door_lock_alarmlevel', 'alarm_type_or_access_control_entity_id': 'sensor.back_door_lock_alarmtype', 'hide_pins': True, 'lock_entity_id': 'lock.back_door_lock', 'lockname': 'Back Door', 'notify_script': 'keymaster_notification', 'parent': None, 'parent_entry_id': None, 'sensorname': 'binary_sensor.back_door_sensor_open', 'slots': 10, 'start_from': 1}
2026-07-05 08:29:59.718 DEBUG (MainThread) [custom_components.keymaster.config_flow] [async_step_reconfigure] notify script keymaster_notification not found, setting to NONE_TEXT
2026-07-05 08:29:59.718 DEBUG (MainThread) [custom_components.keymaster.config_flow] [start_config_flow] step_id: reconfigure, initial user_input: {'lockname': 'Back Door', 'lock_entity_id': 'lock.back_door_lock', 'parent': '(none)', 'slots': 10, 'start_from': 1, 'sensorname': 'binary_sensor.back_door_sensor_open', 'alarm_level_or_user_code_entity_id': 'sensor.back_door_lock_alarmlevel', 'alarm_type_or_access_control_entity_id': 'sensor.back_door_lock_alarmtype', 'notify_script': 'script.keymaster_notification', 'advanced_date_range': True, 'advanced_day_of_week': True, 'hide_pins': True}, errors: {}
2026-07-05 08:29:59.718 DEBUG (MainThread) [custom_components.keymaster.config_flow] [start_config_flow] step_id: reconfigure, final user_input: {'lockname': 'Back Door', 'lock_entity_id': 'lock.back_door_lock', 'parent': None, 'slots': 10, 'start_from': 1, 'sensorname': 'binary_sensor.back_door_sensor_open', 'alarm_level_or_user_code_entity_id': 'sensor.back_door_lock_alarmlevel', 'alarm_type_or_access_control_entity_id': 'sensor.back_door_lock_alarmtype', 'notify_script': 'script.keymaster_notification', 'advanced_date_range': True, 'advanced_day_of_week': True, 'hide_pins': True}
2026-07-05 08:29:59.718 INFO (MainThread) [custom_components.keymaster] Unloading Back Door
2026-07-05 08:29:59.753 DEBUG (MainThread) [custom_components.keymaster.coordinator] [unsubscribe_listeners] Back Door: Removing all listeners
[13:30:41] INFO: Home Assistant Core finish process exit code 256
[13:30:41] INFO: Home Assistant Core finish process received signal 9

One of the messages says that the selected script wasn't found, but it definitely exists. It's also being run by keymaster when a code is used despite the fact that the configuration entry still says "(none)", so the reconfigure must have updated something.

[firstof9]

Thanks @cjhesse this helps pinpoint where it's breaking.

[firstof9]

Ok 3 files for you guys to give a try to:

__init__.py
config_flow.py
coordinator.py

Swap those in and see if the issue persists.
Thanks again for testing.