Hi guys,
Just wondering how you guys organize your rule sets and rules. I wish I had an extra layer to have some more granularity, but my feature request was not popular.
Maybe there are better ways to organize my rule sets.
I use the rule sets now primarily for rooms. So a rule set per room. But maybe grouping by functionality works better. Any examples/ suggestions would be appreciated.
I have installed MSR on a RP5 bare metal and then copied the config and store files. Everything seems to be on the RP5 but I am missing all global expression and I don't see the controller time.
Screenshot 2025-04-06 201446.png
My browser is Microsoft Edge Version 135.0.3179.54 (Official build) (64-bit)
I have probably done something stupid or missed a step but I am stuck.
Thanks for any help.
Hi,
It seems that the widget deletion does not work. I tried to drag the widget to the left (as explained in here https://smarthome.community/topic/1071/deleting-widgets?_=1744037333660)
but it does not delete it. Anyone else experiencing the same problem?
Also the landing page after login is empty and seems be have some JS issues on the dev console:
Screenshot from 2025-04-07 18-06-19.png
fd7f0424-debb-49d3-86d8-9a3b09ad3868-image.png
Using dockerized version of Reactor (Multi-hub) latest-25082-3c348de6 on Chromium 135.0.7049.52 (Official Build) snap (64-bit)
br,
mgvra
Good day all,
I have an notification set up for my washing machine to let me know when it's complete. I have a templete sensor set up in HAAS to let me know if it's Washing, in Standby, or off, based upon the power consumption (Shelly 1PM on outlets)
The MSR code is relatively simple. I have a built in false positive attribute for if MSR gets rebooted, because I would suddenly get tons of notifications when I upgraded MSR.
6774565a-06cc-443f-99c1-e71301b33d83-image.png
What I'm trying to introduce, is a way to verify that I just didn't bump the knob on the washing machine when transferring loads from the washer to the dryer, which turns on the display and brings the power above the Standby threshold.
The power goes up to 3.7W for about 4 minutes if the selector knob is bumped/turned.
What would the best way to do this be? I have most of my MSR code set up for a couple of years now, and my coding logic is struggling a bit.
I think I need a power threshold to be substained for a minimum time period (say 2 or 3 minutes, above 10W), before the other triggers can act. What would the best way to do that be?
Running: latest-25082-3c348de6
Fedora 41 Server
HAAS:
Core
2025.3.4
Supervisor
2025.03.4
Operating System
15.1
Frontend
20250306.0
Build 21228 has been released. Docker images available from DockerHub as usual, and bare-metal packages here.
Home Assistant up to version 2021.8.6 supported; the online version of the manual will now state the current supported versions; Fix an error in OWMWeatherController that could cause it to stop updating; Unify the approach to entity filtering on all hub interface classes (controllers); this works for device entities only; it may be extended to other entities later; Improve error detail in messages for EzloController during auth phase; Add isRuleSet() and isRuleEnabled() functions to expressions extensions; Implement set action for lock and passage capabilities (makes them more easily scriptable in some cases); Fix a place in the UI where 24-hour time was not being displayed.
I have the following yaml configuration in local_mqtt_devices file
x_mqtt_device: set_speed: arguments: speed: type: str topic: "command/%friendly_name%" payload: type: json expr: '{ "fan": parameters.speed }'While this works fine, I'm wondering how this could be changed to "fixed" parameters, as in this case "fan" only accepts "A", "Q" or a numeric value of 1-5?
Hi!
I get this message when I'm on the status tab:
System Configuration Check
The time on this system and on the Reactor host are significantly different. This may be due to incorrect system configuration on either or both. Please check the configuration of both systems. The host reports 2025-04-01T15:29:29.252Z; browser reports 2025-04-01T15:29:40.528Z; difference 11.276 seconds.
I have MSR installed as a docker on my Home Assistant Blue / Hardkernel ODROID-N2/N2+. MSR version is latest-25082-3c348de6.
HA versions are:
Core 2025.3.4
Supervisor 2025.03.4
Operating System 15.1
I have restarted HA as well as MSR multiple times. This message didn´t show two weeks ago. Don´t know if it have anything to do with the latest MSR version.
Do anyone know what I can try?
Thanks in advance!
Let's Be Careful Out There (Hill Street reference...) 🙂
/Fanan
I have a very strange situation, where if InfluxDB restarts, other containers may fail when restarting at the same time (under not easy to understand circumstances), and InfluxDB remains unreachable (and these containers crashes). I need to reboot these containers in an exact order, after rebooting InfluxDB.
While I understand what's going on, I need a way to reliable determine that InfluxDB is not reachable and these containers are not reachable, in order to identify this situation and manually check what's going on - and, maybe, in the future, automatically restart them if needed.
So, I was looking at HTTP Request action, but I need to capture the HTTP response code, instead of the response (becase if ping is OK, InfluxDB will reply with a 204), and, potentially, a way to programmatically detect that it's failing to get the response.
While I could write a custom HTTP controller for this or a custom HTTP virtual device, I was wondering if this is somewhat on you roadmap @toggledbits
Thanks!
Hi ,
I'm on
-Reactor (Multi-hub) latest-25067-62e21a2d
-Docker on Synology NAS
-ZWaveJSUI 9.31.0.6c80945
Problem with ZwaveJSUI:
When I try to change color to a bulb RGBWW, it doesn't change to the RGB color and the bulb remains warm or cold white.
I tryed with Zipato RGBW Bulb V2 RGBWE2, Hank Bulb HKZW-RGB01, Aentec 6 A-ZWA002, so seems that it happens with all RGBWW bulb with reactor/zwavejsui.
I'm using from reator the entity action: "rgb_color.set" and "rgb_color.set_rgb".
After I send the reactor command, It changes in zwavejsui the rgb settings but doesn't put the white channel to "0", so the prevalent channel remains warm/cold White and the bulb doesn't change into the rgb color.
This is the status of the bulb in zwavejsui after "rgb_color.set" (235,33,33,) and the bulb is still warmWhite.
x_zwave_values.Color_Switch_currentColor={"warmWhite":204,"coldWhite":0,"red":235,"green":33,"blue":33}The "cold white" and "warm white" settings interfer with the rgb color settings.
Reactor can change bulb colors with rgb_color set — (value, ui8, 0x000000 to 0xffffff) or rgb_color set_rgb — (red, green, blue, all ui1, 0 to 255) but if warm or cold white
are not to "0", zwavejsui doesn't change them and I can't find a way to change into rgb or from rgb back to warm white.
So if I use from reactor: rgb_color set_rgb — (235,33,33) in zwavejsui I have
x_zwave_values.Color_Switch_targetColor={"red":235,"green":33,"blue":33} 14/03/2025, 16:43:57 - value updated Arg 0: └─commandClassName: Color Switch └─commandClass: 51 └─property: targetColor └─endpoint: 0 └─newValue └──red: 235 └──green: 33 └──blue: 33 └─prevValue └──red: 235 └──green: 33 └──blue: 33 └─propertyName: targetColor 14/03/2025, 16:43:57 - value updated Arg 0: └─commandClassName: Color Switch └─commandClass: 51 └─property: currentColor └─endpoint: 0 └─newValue └──warmWhite: 204 └──coldWhite: 0 └──red: 235 └──green: 33 └──blue: 33 └─prevValue └──warmWhite: 204 └──coldWhite: 0 └──red: 235 └──green: 33 └──blue: 33 └─propertyName: currentColorIn zwavejsui, the bulb changes rgb set but warm White remains to "204" and the bulb remais on warm White channel bacause is prevalent on rgb set.
x_zwave_values.Color_Switch_currentColor_0=204 x_zwave_values.Color_Switch_currentColor_1=0 x_zwave_values.Color_Switch_currentColor_2=235 x_zwave_values.Color_Switch_currentColor_3=33 x_zwave_values.Color_Switch_currentColor_4=33Is it possible to targetColor also for "warmWhite" and "coldWhite" and have something similar to this?
x_zwave_values.Color_Switch_targetColor={"warmWhite":0,"coldWhite":0,"red":235,"green":33,"blue":33}Thanks in advance.
Good day all,
I have a reaction set up, that I use for both troubleshooting and changing home modes when one of my family members either arrive or are leaving. I use the companion app for HAAS on our iPhones, and HAAS reports if the person associated with the iPhone enters or leaves the geofenced area around my home. I'm sure most MSR and HAAS users are familiar with this.
I use this rule set mainly as a condition for other rules, however, as part of troubleshooting, a notification is sent through HAAS to the companion app when the rule becomes true. The problem is that I'm getting notifications now for both arriving and departing simultaneously.
96b3f7db-ba09-499e-a78c-86903b603857-image.png
36903cdd-a87f-473b-82ef-af9ef96d3c44-image.png It used to work fine as intended. I'm not sure exactly when it changed, but now I'm getting two notifications when either of these conditions change.
Any idea what could be happening?
Edit:
Running: latest-25082-3c348de6, bare-metal Linux
ZWaveJSControllerr [0.1.25082]
MSR had been running fine, but I decided to follow the message to upgrade to 25067. Since the upgrade, I have received the message "Controller "<name>" (HubitatController hubitat2) could not be loaded at startup. Its ID is not unique." MSR throws the message on every restart. Has anyone else encountered this problem?
I am running MSR on a Raspberry Pi4 connecting to two Hubitat units over an OpenVPN tunnel. One C8 and a C8 Pro. Both are up-to-date. It appears that despite the error message that MSR may be operating properly.
Similarly as for local expressions, global expressions evaluate and update fine when getEntity(...) structure is used. However, at least when certain functions are in use, expressions do not update.
Consider the following test case:
Screenshot 2025-03-13 at 16.29.42.png
Even though auto-evaluation is active, value does not change (it changes only if that expression is manually run). MSR restarts do not help.
Screenshot 2025-03-13 at 16.31.43.png
Note: Tested using build 25067 on Docker. I have also a PR open (but couldn't now get details or PR number as my Mantis account was somehow expired?).
Trying to understand what cause a local expresssion to be evaluated. I have read the manual but I am still not clear about it. Using the test rule below, I can see in the log that the rule is being automatically evaluated every time the temperature entity is changing. That is great...
What I am trying to understand is why the expression is not evaluated based on time as well since the "case" statement has time dependencies.
Any help would be appreciated
I have the following test rule:
eba6a3ea-ff61-4610-88c9-9b9864f11ff8-Screenshot 2025-01-21 095244.png
2d9c1ff5-7b73-4005-b324-9029c2709db9-Screenshot 2025-01-21 095302.png
Here is the expressioncode:
vFrom1 = "09:25:00", vFrom2 = "09:30:00", vFrom3 = "09:41:00", vTo = "10:55:00", # Get current time (format HH:MM:SS) vToDay = strftime("%H:%M:%S"), #Get current house temperature CurrentHouseTemp = getEntity( "hass>Thermostat2 " ).attributes.temperature_sensor.value, case when CurrentHouseTemp <= 19 and vToDay >= vFrom1 && vToDay <= vTo: "true1" # From1 when CurrentHouseTemp <= 20 and vToDay >= vFrom2 && vToDay <= vTo: "true2" # From2 when CurrentHouseTemp < 26 and vToDay >= vFrom3 && vToDay <= vTo: "true3" # From3 else "false" endI am getting a Runtime error on different browsers when I click exit when editing an existing or creating a new global reaction containing a group. If the global reaction does not have a group I don't get an error. I see a similar post on the forum about a Runtime Error when creating reactions but started a new thread as that appears to be solved.
The Runtime Error is different in the two browsers
Safari v18.3
Google Chrome 133.0.6943.142
TypeError: self.editor.isModified is not a function at HTMLButtonElement.<anonymous> (http://192.168.10.21:8111/reactor/en-US/lib/js/reaction-list.js:171:34) You may report this error, but do not screen shot it. Copy-paste the complete text. Remember to include a description of the operation you were performing in as much detail as possible. Report using the Reactor Bug Tracker (in your left navigation) or at the SmartHome Community.Steps to reproduce:
Click the pencil to edit a global reaction with a group.
Click the Exit button.
Runtime error appears.
or
Click Create Reaction
Click Add Action
Select Group
Add Condition such as Entity Attribute.
Add an Action.
Click Save
Click Exit
Runtime error appears.
I don’t know how long the error has been there as I haven’t edited the global reaction in a long time.
Reactor (Multi-hub) latest-25060-f32eaa46
Docker
Mac OS: 15.3.1
Thanks
I am trying to delete a global expression (gLightDelay) but for some strange reason, it comes back despite clicking the Delete this expression and Save Changes buttons.
I have not created a global expression for some times and just noticed this while doing some clean-up.
I have upgraded Reactor to 25067 from 25060 and the behaviour is still there. I have restarted Reactor (as well as restarting its container) and cleared the browser's cache several times without success.
Here's what the log shows.
[latest-25067]2025-03-08T23:50:22.690Z <wsapi:INFO> [WSAPI]wsapi#1 rpc_echo [Object]{ "comment": "UI activity" } [latest-25067]2025-03-08T23:50:26.254Z <GlobalExpression:NOTICE> Deleting global expression gLightDelay [latest-25067]2025-03-08T23:50:27.887Z <wsapi:INFO> [WSAPI]wsapi#1 rpc_echo [Object]{ "comment": "UI activity" }Reactor latest-25067-62e21a2d
Docker on Synology NAS
Morning, experts. Hard on learning about the internet check script in MSR tools, I was wondering what suggestions anyone has about a local (i.e. non-internet dependent) notification method.
This was prompted by yesterday's fun and games with my ISP.
I've got the script Cronned and working properly but short of flashing a light on and off, I'm struggling to think of a way of alerting me (ideally to my phone)
I guess I could set up a Discord server at home, but that feels like overkill for a rare occasion. Any other suggestions?
TIA
C
Hi,
I'm trying to integrate the sonos-mqtt (https://sonos2mqtt.svrooij.io/) with the MSR and it's coming along nicely so far.
But cannot wrap my head around how to define custom capabilities in MQTT templates. I need this for the TTS announcements and similarly for the notification sounds where I would pass the sound file as parameter.
So this is what I have in the local_mqtt_devices.yaml
capabilities: x_sonos_announcement: attributes: actions: speak: arguments: text: type: string volume: type: int delay: type: intAnd this is the template:
templates: sonos-announcement: capabilities: - x_sonos_announcement actions: x_sonos_announcement: speak: topic: "sonos/cmd/speak" payload: expr: > { "text": parameters.text, "volume": parameters.volume, "delayMs": parameters.delay, "onlyWhenPlaying": false, "engine": "neural" } type: jsonSo the speak action should send something like this to topic sonos/cmd/speak
{ "text": "message goes here", "volume": 50, "delayMs": 100, "onlyWhenPlaying": false, "engine": "neural" }At startup the MSR seems to be quite unhappy with my configuration:
reactor | [latest-25016]2025-02-09T08:19:59.029Z <MQTTController:WARN> MQTTController#mqtt entity Entity#mqtt>sonos-announcement unable to configure capabilities [Array][ "x_sonos_announcement" ] reactor | i18n: missing fi-FI language string: Configuration for {0:q} is incomplete because the following requested capabilities are undefined: {1} reactor | i18n: missing fi-FI language string: Configuration for {0:q} has unrecognized capability {1:q} in actions reactor | Trace: Configuration for {0:q} is incomplete because the following requested capabilities are undefined: {1} reactor | at _T (/opt/reactor/server/lib/i18n.js:611:28) reactor | at AlertManager.addAlert (/opt/reactor/server/lib/AlertManager.js:125:25) reactor | at MQTTController.sendWarning (/opt/reactor/server/lib/Controller.js:627:30) reactor | at MQTTController.start (/var/reactor/ext/MQTTController/MQTTController.js:268:26) reactor | at async Promise.allSettled (index 0) Configuration for "sonos-announcement" has unrecognized capability "x_sonos_announcement" in actions Controller: MQTTController#mqtt Last 10:21:37 AM Configuration for "sonos-announcement" is incomplete because the following requested capabilities are undefined: x_sonos_announcement Controller: MQTTController#mqtt Last 10:21:37 AMThis is probably a pretty stupid question and the approach may not even work at all, but maybe someone or @toggledbits for sure, could point me to the right direction.
Basically the idea is to be able to send TTS messages from reactions using entity actions. I've previously used HTTP requests to Sonos HTTP API (https://hub.docker.com/r/chrisns/docker-node-sonos-http-api/) for the same functionality, but since moving to sonos-mqtt, I need a way to send the TTS notifications using MQTTController. Along with the actual message, volume and delay must also be parameterizable.
br,
mgvra
MSR latest-25016-d47fea38 / MQTTController [0.2.24293]
Hi, @toggledbits
I just noticed that following a reboot of my raspberry pi, some of the rules, that I was expecting to recover, are not catching up following a reboot. I have made a simple test rule (rule-m6rz6ol1) with only "after Date/time" as trigger and "turn on a lamp" as a set reaction. All my infrastructure is on the same board so Reactor, Hass, Zwavejs, ... are all rebooting.
Here is the sequence of the test case (All time converted to Zulu to match logs):
Rule "after Date/Time" set to 14:05:00z Shutdown on Raspberry Pi at 14:04:00z Power back up at 14:08:00z Rule overview shows true as of 14:08:14z waiting for 00:00:00 in GUIFrom the log I can see that MSR is picking up the rule and knows that the state of the rule has changed from false to true and tries to send the update to HASS but failed with websocket error.
Here is what I see from the log:
14:04:04z shutdown complete 14:08:08z Power up 14:08:13.111z websocket connection 14:08:15:323z Reaction to the light failed, Websocket not opened After there is a series of websocket connection attempt until 14:08:51z where it seemed to be really ready.Back in 2021 we had a discussion (https://smarthome.community/topic/700/solved-start-up?_=1738766986566) and you proposed to add a startup_delay:xxxx and startup_wait:xxxx parameter in the engine section of "reactor.yaml". When I try the startup_delay (this used to be a hard delay), the engine failed to start (I think). I then try the startup_wait:xxxx without any success. Since it wait for the connection status to be up to cancel the delay, it does not do anyting since Hass is reporting the socket up without really being up ( I think...).
Questions:
Did I figured it all wrong? should the startup_delay:xxxxx have worked? Any ideas?Here is the log:
OK now I am stuck. I did add the log but when I submit the editor complained saying that I am limited to 32767 characters. The log from the shutdown to the time the websocket is stable is about 300000 character long. What are my options?
Not a big issue simply a request if easily doable.
The MSR logs files inside the container are owned by root witch is fine however, the permissions are very restrictive. I do not know if there is something wrong with my installation but the logs permission are set to 222 (write only). Even if the docker volume is set for Read/Write the log files are retaining these values.
I go around the problem by doing a chmod 777 on all reactor logs but every time there is an MSR log rotation the permissions are set back to 222. So unless the permission are implemented in the container there is no permanent solution to this (that I know of).
I do not know much about Docker container so I do not know what is involved here.
Can the logfiles permission be simply chaged in the container to at least allow "other" read permission?
Could the MSR log rotation routine implement a chmod to set the permission?
Just a small anoyance
Thanks
[Solved] latest-22328 restart fails
-
Solution: An update to openLuups MQTT implementation on handling acknowledge packets for QoS > 0 solved this issue.
EDIT: Seems related to any restart without any configuration changes. If I revert to 22310 I can restart Reactor from UI and with systemd within seconds, but with 22328 it fails to restart., both from UI and systemd. I have to stop the service and then start again.
I tested to comment out my http (not https)?baseurlin the config on my bare metal Ubuntu install latest-22328 and triggered a restart from the UI but Reactor would not start after that.
If I uncomment the key and restart the service, Reactor comes back to life.
Is my setup an exceptional circumstance or is this only applicable on new installs -
It's working for me, and I've done some fresh installs of 22328 while testing alternatives to Raspberry Pi (a couple of promising boards so far). I hate to say it, but your post is right on the line of "I tried X and it didn't work for me," so without more detail, I can't really guide you.
Author of Multi-system Reactor and Reactor, DelayLight, Switchboard, and about a dozen other plugins that run on Vera and openLuup.
-
@toggledbits said in latest-22328 and baseurl [EDIT]: restarts fail:
It's working for me
This is at first all I wanted to know, if anyone else was having the same issue or it's just my setup.
Sorry for the lack of details, I did not know what else to provide as the log is dead silent after shutting down...
I will debug further.
-
Looks like it's related to openLuups MQTT server. The shutdown process hangs after sending LWT to openLuup (MQTTController#mqtt in log).
Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.775Z <app:NOTICE> Closing Structure... Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.776Z <Structure:INFO> Structure#1 Stopping controllers... Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.776Z <Controller:NOTICE> VeraController#vera stopping Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.778Z <Controller:ERR> Controller VeraController#vera is off-line! Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.799Z <EzloController:NOTICE> EzloController#ezlo stopping Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.800Z <wsapi:WARN> client close from unknown connection? "192.168.1.2#5" Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.800Z <wsapi:WARN> client close from unknown connection? "192.168.1.238#4" Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.800Z <wsapi:WARN> client close from unknown connection? "192.168.1.238#3" Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.800Z <wsapi:WARN> client close from unknown connection? "192.168.1.238#2" Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.800Z <wsapi:WARN> client close from unknown connection? "192.168.1.2#1" Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.802Z <EzloController:NOTICE> EzloController#ezlo connection closed: 1000 closing Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.804Z <Controller:ERR> Controller EzloController#ezlo is off-line! Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.804Z <Controller:NOTICE> EzloController#ezlo stopping Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.806Z <DynamicGroupController:null> DynamicGroupController#groups stopping Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.806Z <Controller:NOTICE> DynamicGroupController#groups stopping Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.808Z <Controller:ERR> Controller DynamicGroupController#groups is off-line! Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.809Z <HassController:NOTICE> HassController#hass stopping Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.810Z <HassController:NOTICE> HassController#hass websocket closing, 1000 Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.810Z <Controller:NOTICE> HassController#hass stopping Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.811Z <Controller:ERR> Controller HassController#hass is off-line! Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.823Z <Controller:NOTICE> OWMWeatherController#weather stopping Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.824Z <Controller:ERR> Controller OWMWeatherController#weather is off-line! Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.825Z <Controller:NOTICE> SystemController#reactor_system stopping Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.826Z <Controller:ERR> Controller SystemController#reactor_system is off-line! Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.827Z <MQTTController:NOTICE> MQTTController#mosquitto-mqtt stopping, sending LWT Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.870Z <MQTTController:NOTICE> LWT sent; closing broker connection Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.870Z <Controller:NOTICE> MQTTController#mosquitto-mqtt stopping Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.871Z <Controller:ERR> Controller MQTTController#mosquitto-mqtt is off-line! Nov 26 10:45:58 homebridge node[686832]: [latest-22328]2022-11-26T09:45:58.874Z <MQTTController:NOTICE> MQTTController#mqtt stopping, sending LWT Nov 26 10:46:03 homebridge node[686832]: [latest-22328]2022-11-26T09:46:03.452Z <httpapi:INFO> HTTP server closed.If I disable that controller Reactor restarts fine as usual.
@toggledbits Is this the moment openLuup MQTT support ends in Reactor or can I change something to make it work again?
@akbooer I've updated openLuup from 22.9.3 to 22.11.22 but no success to this issue.The thing is it's running great when it finally comes up again after a long wait of a forced restart with systemd but it doesn't feel right to force kill the process even though the shutdown process is almost finished.
-
@crille said in latest-22328 restart fails:
Looks like it's related to openLuups MQTT server. The shutdown process hangs after sending LWT to openLuup (MQTTController#mqtt in log).
I think we've seen an issue like that previously. I don't think I see that line in the log you posted?
There may be a problem with retained messages (ie. LWT) and wildcard subscriptions in the openLuup server... I'd have to check.
-
I use a popular package to handle the MQTT broker connection, so I can't see the innards of communications and confirm, but from the rhythm of the log output, it appears that the broker is not sending an ACK to the publish of the LWT; the
publish()call appears to be sitting there waiting for it.Author of Multi-system Reactor and Reactor, DelayLight, Switchboard, and about a dozen other plugins that run on Vera and openLuup.
-
Does this mean 22328 require an ACK but 22310 does not? or has something else changed?
-
22310 still required an ACK, in a sense. It didn't wait for it. But because the ACK never arrived, the task in the
mqttpackage also never cleared (they don't have a time-out mechanism), so it just stayed in the queue forever... as did every other topic sent with a non-zero QoS. This actually causes a memory leak that could lead to exhaustion and a crash, because the ACK never comes so the task is never removed from the queue, and those tasks remain and proliferate in the queue.IMO, I think it's fine if @akbooer doesn't truly support QoS levels 1 and 2 in his MQTT implementation just for basic use, but not sending an ACK regardless isn't the right choice, in my view, and it's going to cause problems for a lot of clients that may not be immediately evident (like memory leaks/exhaustion).
Author of Multi-system Reactor and Reactor, DelayLight, Switchboard, and about a dozen other plugins that run on Vera and openLuup.
-
@toggledbits said in latest-22328 restart fails:
I think it's fine if @akbooer doesn't truly support QoS levels 1 and 2 in his MQTT implementation just for basic use
It only supports QoS 0 .
Nevertheless, the protocol should be respected. AFAIK all PUBLISH requests receive an ACK unless the connection goes down in between times.
-
This is the end of
parse.PUBLISH():-- ACKNOWLEDGEMENT -- The receiver of a PUBLISH Packet MUST respond according to Table 3.4 - Expected Publish Packet -- response as determined by the QoS in the PUBLISH packet [MQTT-3.3.4-1] --[[ Table 3.4 - Expected Publish Packet response QoS Level Expected Response QoS 0 None QoS 1 PUBACK Packet QoS 2 PUBREC Packet --]] local ack -- None, because we only handle QoS 0 return ack, nil, TopicName, ApplicationMessage, RETAIN endComments to the contrary, it appears it returns
ack, which is declared butnil... so... no ACK? -
@toggledbits does Reactor publish it's LWT message with QoS > 0 even though the MQTTController config is at
qos: 0? otherwise the expected response would be none. -
Yes, it uses QoS 1 (and retain true) because it's a "vital" message. The
qosyou can set in config is for the echo/entity publish functionality; it does not affect other messages. Still, this only requires that the broker acknowledge its receipt (3.3.4), not any delivery, and does not even enforce that QoS on subscribers (3.8.3) -
"The receiver of a PUBLISH Packet MUST respond according to Table 3.4 - Expected Publish Packet response as determined by the QoS in the PUBLISH Packet."
So even though the server only supports QoS 0 it's obligated to send a PUBACK on a QoS 1 level packet as described in table 3.4, correct?
-
Yes, PUBACK for QoS 1, and PUBREC for QoS 2. I don't think that would be a big problem for @akbooer when he gets the time, because all of the information contained in the response can be sourced from the request (i.e. topic, packet identifier, etc.). And otherwise the treatment of the PUBLISH packet can be the same (no further changes beyond sending those ACKs). Not actually having guaranteed delivery behind that is, in my view, an acceptable variance. On the sending side (repeat to subscribers), even though a subscriber may request QoS 1 or 2 for packets from the broker, it still must accept QoS 0 packets (because the requested QoS is a maximum, not an absolute), so everything outbound at QoS 0 isn't likely going to cause problems, especially in this world.
To summarize: if he just provides the PUBACK and PUBREC responses to fix that layer of the protocol, that's good enough. No further actions required above that layer. Not fully compliant (no guaranteed delivery), but at that point, few if any would ever notice.
The other question I don't have an answer for (or I've forgotten; and haven't dug through code to figure out)... @akbooer, does it support retain? If so, what's the storage mechanism, and is it persistent?
-
Well, I must be reading the MQTT 3.1.1 spec all wrong.
When I read:
"The SUBACK Packet sent by the Server to the Client MUST contain a return code for each Topic Filter/QoS pair. This return code MUST either show the maximum QoS that was granted for that Subscription or indicate that the subscription failed [MQTT-3.8.4-5]. The Server might grant a lower maximum QoS than the subscriber requested. The QoS of Payload Messages sent in response to a Subscription MUST be the minimum of the QoS of the originally published message and the maximum QoS granted by the Server. The server is permitted to send duplicate copies of a message to a subscriber in the case where the original message was published with QoS 1 and the maximum QoS granted was QoS 0 [MQTT-3.8.4-6]."
...I understood it to mean that, since I only ever grant QoS 0, then no message would require a PUBACK or a PUBREC.
I realise that this is my bad for writing my own MQTT broker, but it made sense in the context of openLuup, especially in support of Shelly devices which was the reason I did it in the first place. I apologise if this has led to some difficulties, and I'm starting to look at an MQTT validation suite to check out my implementation further. I also realize that Mosquitto is the de-facto standard, but it turns out that having an internal server confers some significant benefits in terms of the internal openLuup architecture.
However, if anyone can clarify the about QoS response issue further, I'm very happy to comply. If a simple fix is to send PUBACK or PUBREC, then I'll do it, but I want to know the reason why. I do understand that the original CONNECT request contains a LWT QoS, per this paragraph:
3.1.2.6 Will QoS
"Position: bits 4 and 3 of the Connect Flags."
"These two bits specify the QoS level to be used when publishing the Will Message."
"If the Will Flag is set to 0, then the Will QoS MUST be set to 0 (0x00) [MQTT-3.1.2-13]."
"If the Will Flag is set to 1, the value of Will QoS can be 0 (0x00), 1 (0x01), or 2 (0x02). It MUST NOT be 3 (0x03) [MQTT-3.1.2-14]."
...but I had assumed that QoS to be overridden by the actual level established in SUBSCRIBE / SUBACK. However, now that I write that, it seems the LWT actually may have a separate life from standard messages?
-
@toggledbits said in latest-22328 restart fails:
The other question I don't have an answer for (or I've forgotten; and haven't dug through code to figure out)... @akbooer, does it support retain? If so, what's the storage mechanism, and is it persistent?
Yes, it supports retained messages.
Is it persistent? Not across openLuup restarts.
The openLuup console page: openLuupIP:3480/console?page=mqtt shows all current subscriptions and also (at the bottom) retained messages.
-
@akbooer said in latest-22328 restart fails:
...I understood it to mean that, since I only ever grant QoS 0, then no message would require a PUBACK or a PUBREC.
This section is about your response to a SUBSCRIBE (SUBACK)and what QoS you use to PUBLISH. If you only grant QoS 0 on subscribe, then you will never expect a PUBACK or PUBREC because you never publish anything to a subscriber other than QoS 0.
It in no way limits what a client may publish to the broker, which could include any QoS.
