-
Notifications
You must be signed in to change notification settings - Fork 12
V3.7 release #138
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
V3.7 release #138
Conversation
commit 90f1783 Author: IgorYbema <[email protected]> Date: Mon Aug 26 16:18:47 2024 +0200 remove table refresh functions commit afbca72 Author: IgorYbema <[email protected]> Date: Mon Aug 26 16:08:18 2024 +0200 better loading of table values
commit efec856 Author: CurlyMoo <[email protected]> Date: Sat May 18 20:17:19 2024 +0200 Resync NTP regurlarly
I am still seeing a crash reboot on the esp8266 when I open the webconsole after some hours/days of running. It immediatly crashreboots just after loading the first screen. Probably due to memory issue but hard to reproduce. |
Just installed this build here: https://github.com/IgorYbema/HeishaMon/actions/runs/10656111818 When do you plan to merge? Forget about it, this build is really unstable on the ESP8266... |
Reverted back to https://github.com/IgorYbema/HeishaMon/actions/runs/10637426833... |
@stumbaumr what are you expierencing? I am running this build fine on my ESP8266 |
I was experiencing reboot after reboot after reboot... Do you have rules active in your setup? I have 9 running... |
No i don't have rules. Do you want to share your rules? I hope I can replicate your issue so I can see what is going wrong |
Sorry, have seen is a newer Build Egyras#613. Console and Variables are ok ;-) |
I believe this was never available on the MQTT log output. The code used for the rules logging is seperate and is only printed to webconsole and serial port. But I'll modify that in a later version |
@McMagellan can you try https://github.com/IgorYbema/HeishaMon/actions/runs/10706629932 with your large ruleset and a reboot? I added some dog feeding rules during rules loading. |
Feedback Firmware Egyras#614 Signature Alpha e37c69b You wrote that it might be because of my surroundings. 1 ruleset deleted. Then I activated the 1Wire IF in the settings to which 7 Dallas sensors are connected. On this occasion I noticed something else: Then I repeated the test with the new version Egyras#614. |
Ok thanks. And did you test the new build also? |
@IgorYbema |
Ok thanks. I'll try to reproduce with 1wire enabled. Will be later this week |
Good news, I managed to reproduce this. With 5 temp sensor ok, with 6 sensor it fails to boot with the rules. Now time for debugging :) |
Quick question in between:
|
No I get: 423703: ERROR: Expected a parenthesis block, function, number or variable And I fixed the issue. Was indeed watchdog reset. You can test with the new build https://github.com/IgorYbema/HeishaMon/actions/runs/10724868618 |
I believe there is more wrong with the usage of the double-quotes. Following example will crash the system... so this is something for @CurlyMoo to look at
|
Feedback FW Igor Egyras#617 Signature Alpha ffa04b6 (Test: Big Rule crashes after reboot) The test was successful and the ruleset was started again even after a reboot. I also wanted to say that this ruleset adjusted the heating curve parameters with SetCurves. So check if used. I'm currently expanding the ruleset and using the "print" command several times. Is there already a guideline for how large the ruleset can be? So far I have never used more than 15 variables. How many can I use now without conflicts in level 3.7? |
I can't say for sure. The memory is the final limit but it is hard to tell what the max is. With the new larger board with ESP32 you can go much further if you reach the max on the older model as it has more memory. |
I'll declare this v3.7 as final now and push it to egyras repo |
Feedback V3.8 Spontaneous crash without reloading the rules. Let me say that version 3.8 runs very well and is largely stable. The recording begins one hour before the crash. I programmed a watchdog in iobroker Blockly that carries out a Heishamon reboot. |
Funny thing. At boot it says rst reason: 4 (which is a software reset) while the rules check says reason: 2 (which is a exception failure). If it was 4 it would not have disable the rules so I believe it wasn't a 4 or a watchdog reset but a exception. What is the reason you keep the webclient open? Can you check if it keeps running without webclients? |
The reason why I currently have the web server open with the console for hours is that I continue to work on my CoPilot ruleset. So is there a limit on the web server or is it a browser problem on the PC that logs in so often? Here is another recording of a reset from Heishamon with firmware 3.8 that is triggered by the message: The log file also includes the hour before when I first connected to the console and the 15 minutes after the restart. Is there a better place for feedback on 3.8 than here? Please let me know. |
You could use the 'log to mqtt' and monitor the console log through there. We know there is still a memory issue with the webserver so eventually we can not distuingish between maybe a rules error or a webserver error if you keep monitor it using the web console. |
@IgorYbema: log to mqtt sounds good but doesnt work with any Rulesset values. See the same timerange as following. Part1: mqtt- log in iobroker
Part2: Log from the console
|
Alright yes I understand. This printing of the variables is some left over debug code. Shouldn't be there anyway :) Maybe those variables should be visible as a new mqtt topic so you can monitor any change. I'll think about how to add that. But for now, could you just stop monitoring and let it run for a few hours/days to check if it crashes then also or not? |
OK, now I'm taking a 24 hour console break. However, if you set up an extra MQTT channel for rulesset values, keep in mind that it is important to include the "set" commands and the "received TOP xxx" messages. |
I have another suggestion in connection with an independent MQTT channel. When uploading a rule, you can observe the progress on the debug port. With older firmware versions it was possible to establish two web server connections to Heishamon. In V3.8 this is only possible with small rulesets. If I use my larger ruleset in the manner described, there is always a crash with two running web server tasks and the loading of the rules fails. I would therefore like to make a suggestion: |
Feedback test without Heishamon Console connection. Here are the key data from the debug log file. 9/23/24 16:11: Start test at uptime 5h 47min Rules always running.
9/25/24 17:44: Start working with Heishamon, permanent console connection.
|
Feedback Test 3.8, Memoryfight between rules and webserver. After I found a small error in the ruleset, I reestablished a web server connection to load the new ruleset. 4018: (1st time) New webserver client: 192.168.178.86:57095 (Uptime: 3 days 2 hours 39 minutes 46 seconds) |
release v3.7 adds to v3.62 the following: