-
-
Notifications
You must be signed in to change notification settings - Fork 96
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
[BUG] Plugin stops updateing once a day #341
Comments
This is potentially the same issue as #340 |
Thanks i will have a look, it stopped about 10am yesterday. I think it started happening after a recent HA update |
@J14111 you are using TCP, but how? is that through a data logger? If so, which one do you have? |
Im using the USR-W610 RS485/TCP adapter |
Interestingly turning that on and setting it to 5088 just crashed my HA instance, just restarting now. |
@J14111 I would try this for now #340 (comment) |
Thanks I will test that now. My HA is really not happy with that modbus polling setting turned on above for some reason. |
@wills106 Didnt fix it crashed again today around 7am |
Still happening, I have set an automation to restart HA when it stops. There is a HA update out today so will see if that fixes |
The following PR #346 may help? It might be that the Inverter is sending bad data every so often and the Integration gets stuck. |
yeab but odd it only started happening this week |
Were you on a really old HA installation before? Just wondering if you were on an old pyModbus v2.5.3 I don't know if pyModbus version would alter how it handles bad data? |
No I usually update within a day or so of new versions so dobut I was on an old version before, I will keep an eye on it for a day or so and report back |
Its been over 24h, two HA updates and no hang so hopefully the issue is resolved, will report back in a few days |
did you apply a lower block size? @J14111 |
Yeah it didnt make any difference |
I have no idea, my logic was a HA update started it so an HA update would eventually fix and it seems to have done |
pyModbus was updated in HA 2023.03.x and I matched the version in this Integration to it. That PR doesn't actually do anything yet. I need to implement it in the plugin_xyz |
Appears this is still an issue, it just stopped updating, i restarted ha manually about an hour ago which makes it even odder as it doesnt seem like its a buffer or something filling up, anything else i can try? |
Just released 2023.03.2b2 implementing #346
This is what @alienatedsec provided here #340 (comment) |
Thanks i will test and report back |
Just wondering if you have been able to confirm if 2023.03.2b2 has fixed the issue or not? |
I believe its still an issue, looking at the automation the last restart was 18 hours ago but I cant rememebr if I did that, I also havnt tested taking the block size down so I will keep an eye for another day or so |
Ok, well hopefully it won't need a restart over tonight / tomorrow? |
Just had a restart so its still happening, I have set the block size down to 40 so will keep an eye on it. There has also been a HA update to 2023.3.4 so I have installed that. Where can I get the the block config to send you? |
If you add:
To your configuration.yaml you will get more logging information. Could you provide me with the log file with it crashing as it is now (With the above added to your config). Then we can try and work out what it's actually failing on. |
Ok will do |
Here is the startup info: 2023-03-14 13:37:10.869 INFO (MainThread) [custom_components.solax_modbus.button] hub.wakeuButton: None |
Is that with the block_size at 48? |
Are you using the release 2023.03.b2 ? |
@J14111 you need to amend the line homeassistant-solax-modbus/custom_components/solax_modbus/plugin_solis.py Lines 1683 to 1692 in 9211bec
|
One warning about using block_size and ignore_readerror I may not have read this thread with sufficient attention so forgive me if I am off-topic |
@J14111 I don't want you to change anything in the Integration. I am trying to establish if #341 (comment) was the output of block_size = 40 or 48? I released 2023.03.b2 / 2023.03.b3 which I believe had |
Yes that output was with block_size = 40 |
@J14111 I have just released 2023.03.2b4 with the block_size set to 40 and I have reconfigured Could you upgrade to that version and confirm in your error log you still have:
It needs to match exactly. Unless something else has changed.
It will probably be worth automating this, assuming this fixes the issue. Saves having to reconfigure again if we add in more registers. |
Ok updated, log: 2023-03-14 18:40:20.711 INFO (MainThread) [custom_components.solax_modbus.config_flow] detected HA core version 2023 3 |
So your new block groups matches up with how So we just need to wait an see how it behaves over the next day or two. |
Of course will do, thanks for your help :) |
Still having the issue, I found this in the logs before the restart: 2023-03-17 09:37:14.032 ERROR (MainThread) [custom_components.solax_modbus] Something went wrong reading from modbus |
Oops, should be block.start ... |
Just pushed @infradom PR out in 2023.03.2b6 |
Thanks, installing now will let you know how it goes |
I had another restart yesterday, not as much in the logs that I can see this time: 2023-03-19 22:28:26.131 INFO (MainThread) [custom_components.solax_modbus] read_error : Solis cannot read input registers at device 1 position 0x81e3 and File "/config/custom_components/solax_modbus/number.py", line 117, in native_value |
Ok, I will try to catch that error in a new PR |
Just pushed the PR out in 2023.03.2b7 |
Thanks, installing now |
As mentioned to Will within a private conversation before, I too have the problem of the integration loosing the connection during the night. But I am not sure if the problem is the same. In my configuration one of my two Solis inverters will actually go offline during the night as it does not have a battery and will therefore lose energy once the sun is gone. Before the beta 2023.03.2b4 (maybe b3 as last running version?) the integration would not go to "unavailable" during that time, but just keep the last seen values. At least as long as I don't restart my HA instance. Since 2023.03.2b5 this behaviour has changed to all values going unavailable after some time. I think the behaviour is actually ok (the inverter actually is unavailable) but it will not regain connection by itself in the morning, when the inverter comes back online. And that is a problem - change from the behaviour before. I noticed that other integrations seam to have solved similar issues (e.G. the Denon integration). I power down my entertainment Center using a smart plug during the times I won't use it. So during these times the integration will go "unavailable" as well, but seems to have a loop function trying to reconnect every few minutes. This way it comes back quite quick once I power up the entertainment Center. Maybe this could be an approach to look into that could be adapted to get this integration back without the need to reload the integration by hand? Update: I just went though the different commits: |
Hey, I did try to set the block size back to 48 again. But that did not solve the issue. I run 23.3.2b8 and still have the problem that the integration will not resume working in the morning. Any ideas what I could try? Thanks! |
I was just thinking what else did change since the behavior of the integration has changed. So I realized that I also updated to Home Assistent 2023.03.05. So the change in behavior could actually not be caused by changes within the integration but it's surroundings. Just a thought I wanted to share with you! |
Im 99.99% sure this is fixed for me, not had any issues for 3 days, I will let you know if I see any. Thanks so much! |
Mandatory details
I have noticed the last 2 or 3 days the plugin seems to stop updating every day sometime between 1-6am it only seems to effect this plugin and a restart of HA resolves it, is anyone else seeing this?
I will check the logs when it happens tomorrow and update
You can see the gap here on the graphs:
The text was updated successfully, but these errors were encountered: