Grott 3.0.0 now available in Alpha branche!!!!! #613
Replies: 29 comments 93 replies
-
|
new dependency |
Beta Was this translation helpful? Give feedback.
-
|
Oh yes. I will document that! Thans. |
Beta Was this translation helpful? Give feedback.
-
|
Working fine with https://github.com/muppet3000/homeassistant-grott |
Beta Was this translation helpful? Give feedback.
-
|
Hi There, working fine here. Two SPF5000es inverters. No issues whatsoever. Confirm also the sending of commands is working fine. Awesome job as usual. PS - let me know if you need any tests with "auto" and the SPF inverters. |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for testing. The SPF server will be a challenge while I do not have found the unique "signature" yet in the data record. But I keep looking for it. |
Beta Was this translation helpful? Give feedback.
-
|
Many thanks Johan cheers, |
Beta Was this translation helpful? Give feedback.
-
|
I know it is in another thread, but just to make it clear here also, if you set |
Beta Was this translation helpful? Give feedback.
-
|
HI Johan, I Might have found a bug, but I am unsure. I have two SPF inverters and noticed one of them (the same always) sometimes the updates gets stalled for 5-6 minutes as the other is fine, reporting every 10 seconds. This never ever happened before, so I am suspecting that might be something with the new version. I have debug log, any clues of what i should search for to debug? This doesn't happens always. Since i changed versions i caught it happening twice as because i put the icon in my homeassistant to turn red in case something's wrong with the inverters. It might have happened more. Ah one more thing, it goes back and start to report without any intervention... |
Beta Was this translation helpful? Give feedback.
-
|
It happens also with the first inverter. I have restarted Grott service for the first time. Nothing strange in the log though. When it stops reporting, i have pinged the inverters and they are reachable. |
Beta Was this translation helpful? Give feedback.
-
|
If you have the log lets see if we can try to debug it. I will change my update frequentie ( now 2 minutes) to see if it also happens with my 2 inverters. Can you send the log to grott@ledidobe.nl? |
Beta Was this translation helpful? Give feedback.
-
|
Sure, but since yesterday it didn't happened. I'll set an automation to send me a push when it happens and I'll send one hour before the event until the inverter starts to speak again. PS. I have rebooted the server yesterday given i updated a bunch of packages and kernel. |
Beta Was this translation helpful? Give feedback.
-
|
It happened now, will send the logs as soon as it gets back reporting. Many thanks, Tabascoz |
Beta Was this translation helpful? Give feedback.
-
|
Hi @tabascoz I send some question via email but you can reply it here also. Is the inverter coming back after circa 5 minutes (300 seconds)? I have build in a mechanism that if a inverter is not sending information anymore then after (now a pre-defined) timeout, Grott will remove the connection. After that the inverter needs to reinitialize again. If this is working, then we can at least shorten the time of the stall by making this timeout configurable. At this moment I have not found the reason why your inverter stops sending or Grott stops receiving. (at least that is what I am seeing at this moment. For know I think it has to do with (network) timings, but I have to see if I can create a good test environment. Is it possible for you to change the update time (by example to 30 seconds) to see if become stable then? if we found what causes the problem we can fix it. The idea of grott 3.0.0 with adding more multithreading (a thread per inverter/datalogger) was to make things faster (make an even higher update frequency possible) and more stable. I am not convinced we already achieved this at this moment. But be aware it is the 0.0 version so there maybe some issues/challenges. |
Beta Was this translation helpful? Give feedback.
-
|
HI there
No, i don't need to do anything, it just reports back. I'm suspecting it might have something to do with the server's load average - will test it now.
I remember first time i saw it was off for 8 minutes but I'll check this, my feeling is more common with 2-3 minutes. My datalogger update is set to 0.1; I can change it to test, but would like to firstly test the timeout. Reason is because i am relying on this to do some PID with the short time i have to recharge batteries in winter and do not trip the AC outside breaker (due to crossing contracted power capacity) - but i can do it. Hard to say as it is completely random and not frequently at all. it might be something to do with threading. what i was thinking about trying also is some sort of "ping" to Grott just to check it's alive and ready to serve connections. How could i do it? If you could set your update interval to 0.1 ( 10 seconds) please see if this changes. PS. i purchased another inverter and planning to install it in this weekend - perhaps the third makes the issue more common and easy to identify. Thanks, |
Beta Was this translation helpful? Give feedback.
-
|
Hi, as a heads up, it hasn't happened again. This is the problem of intermittent issues.... will report back and I hope it stays working |
Beta Was this translation helpful? Give feedback.
-
|
Ok. I am running with 0.1 as well. Had also no issue today. But have to be fair that because of the bad wetter I have also periods during the day that one string did not generated any power at all. I will keep on monitoring. I will also make the timeout configurable. |
Beta Was this translation helpful? Give feedback.
-
|
Thanks Johan. Another day without any issue. Wondering if the reboot has resolved it (LoL). PS. not related to load. I compiled some updates (gentoo box) and kernel, leading to an 16 average load (4 processors) and Grott was running steady... |
Beta Was this translation helpful? Give feedback.
-
|
Hi there Johan, I had the issue a few times - didn't have opportunity to post before. Some facts/tests:
I noticed that i changed also grott.service, in the old one i had these lines to do periodically restarts: I've increased debugging to DEBUGV and will send you the file once it starts to happen. Thanks once more. Cheers |
Beta Was this translation helpful? Give feedback.
-
|
Ok lets see what happens. I am running now with the 0.1 interval for days and had now real issues with the inverter. I lost contact ones with the growatt server (running in server mode with passthrough enabled). |
Beta Was this translation helpful? Give feedback.
-
|
ok thanks, it might be something related to stale connections ? Does it helps to send you further logs? cheers, |
Beta Was this translation helpful? Give feedback.
-
|
Ok I have issue's as well at this moment. It seems that cleaning the threads does not work that well. |
Beta Was this translation helpful? Give feedback.
-
|
thank you Johan, happy to test here :) for now I'm restarting the service every three hours, for me it looks it does the job. |
Beta Was this translation helpful? Give feedback.
-
|
Hi there Johan, just passing to check if you found out the issue, cheers Tabascoz |
Beta Was this translation helpful? Give feedback.
-
|
Hi Fernando, I have not found anything yet. To be honest it was also hard to test for me while the days with sun and working inverter were very limited and I need a "live" environment to test with (not possible in the evening). The challenge now is that with a lot of sunny days I do not seem to have a real problem. My Grott v3 stopped working only once in the last 3 weeks. I will look again at your log data and will try to replay it. But it is not an easy problem to solve. Just to be sure do your run Grott as native application or in a docker container or as add-on to Home Assistent. The good news is I now have a Raspberry Pi running HA so I am able to perform additional test for HA. |
Beta Was this translation helpful? Give feedback.
-
|
Johan, totally understand and yep anything intermittent is a superior challenge. I cannot replicate myself. Sometimes it happens sometimes don't. So some bits of my setup:
I have also enabled in GROTT the influxdb output, next time i see it happening, will check if there are also gaps in influxdb data. Cheers, |
Beta Was this translation helpful? Give feedback.
-
|
quick update - indeed it happens also in Grafana and happened last hour. one of the inverters stopped reporting for a while, you can see here at the graphs - 1st from HA and second from Grafana (influxdb):
|
Beta Was this translation helpful? Give feedback.
-
|
I made a quick ( and dirty) automation in HA to restart Grott when time is > 15s. it should help to log it better when it happens. Perhaps there should be a pattern. |
Beta Was this translation helpful? Give feedback.
-
|
here's something... the automation above were not entirely effective. Some of the times when issuing "systemctl grott restart" it fixes the connection but sometimes it doesn't. However if i change to a stop, 5s delay and start, it always does. I am not sure why this behavior, but thought it may be relevant info. |
Beta Was this translation helpful? Give feedback.
-
|
Johan, some more (likely) useful information. I captured the traffic in Grott server and the gaps does start with the occurences below:
Spurious retransmissions happens every now and them, but they shouldn't affect the application right? think of anything? cheers, tabascoz |
Beta Was this translation helpful? Give feedback.
-
|
Ho Fernado, Yes that might be the problem (see also #637). I think that if there is a network instability and the connection is still there but not working properly/intermittent that can cause delays or errors in Grott. I am working on improving the connection cleanup routines (I detected that this behave different in my develop/test Windows environment then under Linux, my "production" environment). In this update cycle I will also look if I can improve the connection / thread health check that is already there to see if we can detect this delays and can cleanup/restart the connection if necessary. This is really advanced (network) programming, nice challenge! |
Beta Was this translation helpful? Give feedback.
-
|
@tabascoz are you able to change the python code? in grottserver.py there is a statement: This is the wait time before grott (grottserver) wait before it will cleanup a "dead" connection. 300 means it will wait for 5 minutes (the default update frequency). If you make this lower the connection will be closed earlier and a new connection will be created as soon the inverter will start communication. Maybe this can solve your problem. I will make this parameter configurable in the next version. |
Beta Was this translation helpful? Give feedback.
-
|
Back with some news... I think the parameter changed something, but still not able to resolve it fully, even with the service to restart. I also played with inverterrespwait and dataloggerrespwait and was hard to find any real conclusions. This is the very nature of intermittent issues... I came to the conclusion there must be something related to the "environment" as the restarts all group together in a 3-5 hour period. One theory I have it now is that it is related to influxdb slow writes. My influxdb got really big (25G) and i learned that if i restart or stop influxdb , i can force the issue to happen - BUT ( always a but right?) Grott logs the error.... when the issue happens alone there's no error logged whatsoever in Grott. Would it be possible that a slow or problematic insertion in either MQTT or in influxdb could be the real cause behind the issue? cheers, |
Beta Was this translation helpful? Give feedback.
-
|
So i tested now disabling Influxdb and the issue is now gone for two hours... @johanmeijer may ask you to test if it resolves in you ? I don't think the issue is any timeouts or a real performance issue, the system is fast and load average is < 1. cheers, |
Beta Was this translation helpful? Give feedback.
-
|
Yesterday I spend great time trying to figure this out. So I'm focusing now in Influxdb. The situation got to the point that connections were failing in a matter of minutes, so it is now easy to try fixes. I am convinced that removing influxdb writes does eliminate the problem, but it doesn't mean necessarily that the problem is with it. It may be that it is removing some server load and making things easier.... But not really being an option to disable Influxdb. I decided to play some cards:
So with the results above, i am convinced there's something with performance. Although nothing noticeable in day2day. Since I am using Influxdb V1.8 I am thinking now about finally moving to 2.0 and thus changing the library from influxdb-python to influxdb-client-python; @johanmeijer can you confirm if you're using Influxdb V1 still? Cheers, Tabascoz |
Beta Was this translation helpful? Give feedback.
-
|
Nice debugging work! Yes in my daily implementation I still using influxdb 1.8 natively on a rpi. For testing I use a v2 docker container. I noticed that if something goes wrong in the influxdb write grott stops the dataprocessing with an error (or wait for a answer for a long time?). Maybe that can be the issue. Do you see something strange in the log after th influxdb insert? |
Beta Was this translation helpful? Give feedback.
-
|
HI there, i think so... Was suspecting of old v1 (unmaintained) influxdb-python library and did a rough implementation of V2 to access Influx on 1.8 ( as per https://github.com/influxdata/influxdb-client-python?tab=readme-ov-file#influxdb-18-api-compatibility ) . It worked OK but, nothing changed on the issue... dead end Tried more tweaking and I enabled more logging (that you previously commented ) in Grott to give more logging to influxdb writing process. What i noticed is that it happens more when the sticks "flood" grott with historical + current data. But still i can't see any logs anywhere. So i changed some bits and looks way better now: reverted most of the changes I did in grott back to stock, except the log increase. grottconf.py (635) changed timeout to 1 grottserver.py: grottdata.py (717) commented SystemExit WIth these changes it has improved stability a lot, but still i have some zombie stalled connections every now and them. Here's an example as of now: `Every 0,1s: netstat -ant | grep -P '.51.34|.51.181' tcp 0 0 192.168.51.103:5279 192.168.51.34:17352 ESTABLISHED above when the issue happens, there is always a connection wtih recv-q being >0 |
Beta Was this translation helpful? Give feedback.
-
|
If I do netstat -ant | grep 8086 I see a lot of connections with influxdb. Most of them are in time_wait status and comming from node-red. I do not think this is a good status but it doesn't seem to be a problem. I do not have a lot connections comming from grott to influx. Is it possible that something else in your netwerk is blocking influx? I agree that the raise systemexit maybe is not the best solution if influxdb fails. I will do some additional testing on that and change it in the next release. |
Beta Was this translation helpful? Give feedback.
-
|
Hi Johan That netstat isn't from influx (sorry for lack of context). This is the traffic between the two inverters and grott server. 51.34 and .51.181 are the inverters and .103 the grott server. Influxdb is running in the same box, I changed from the box ip to 127.0.0.1 and it happens the same. What I'm suspecting is that something happens in Grott (I'm still learning your code and protocol) that prevents Grott from processing packages, then first the connection stays open (Estabilished) but queue starts increasing and data stays there and nevermore gets processed by Grott. Grott then recognizes the timeout and closes the connection (from server side), happens that the connection is still opened from client, therefore the closed_wait status. Client recognizes the closed_wait and opens another connection to grott and it all begins again. I don't think there's any problem at all with Grott - influxdb. The issue could even been a bad data / bad formatting or whatever bad that gets Influx to send ( or not send ) the proper "OK" and the it all begins. I don't know the cause nor what happens, but this is my two cents for now. THe reason why - the only things i know for "sure" :
May you investigate in the code what could a connection to stall and stop being processed by grott? I may do a video to show, will see if i can get it happening. |
Beta Was this translation helpful? Give feedback.
-
|
Oh sorry I still thought it was related to influxdb. I will upload 3.1 to the alpha branch today. I changed the connection cleanup routine totally. Maybe that fixes things. It is strange that I have no issues. But at this moment I run only with 1 inverter (but used 2 until 3 weeks ago). Just to be sure,how does your configuration looks like (#inverters, type, datalogger_type etc). What I can do if this is not working, is making the data handling part also a separate thread so it will block the normal grott (grottserver) processing flow. |
Beta Was this translation helpful? Give feedback.
-
|
yep that is a good idea... really thanks for looking into it. I'm still learning the code, hope i can contribute more. I'm new to python and there's more than 20 years that i don't do any coding myself(lol). But as it's like riding a bike, the challenge i have is to understand your code and how you manage queue connections etc. I'm getting there... I'm a little puzzled to the trigger / reasons whatever justifies the issue to happen with a crazy frequency and sometimes it stays stable for 6h+. May I ask you some crazy questions lol ? The way the code is now, how would it react if the shine sticks should send garbage data? Is there any error handling or would a debug message show it? I'm asking as i was dealing with some crazy (i don't know what it is) situation that i suspect started since I installed my second inverter about a year ago (in parallel). Some of my devices, when powered by batteries, behave strangely. For ex. two of 5 temp sensors running esp32 /bme680 started to produce garbage every now and them, just like what's happening in Grott. I was able to fix by changing the 5v source and i suspect some sort of harmonics or AC frequency noise/weirdness that the inverters are producing when in Batt may disturb some devices; Another example is that my car (Tesla) behaves weird when i charge it between 10-12A. It reduces power itself as a "protection" sense like if there was some bad contact in the AC socket - which is not the case. If i set to 16A+, it doesn't happen. so weird. This is the reason I'm asking. What if the sticks doesn't have a good data checksum routine and are passing some weird data to Grott? Would it explain? I'm happy to test the new branch and if you think could contribute more to the tests, would a downgrade to 2.8 just to test is worth it? Many thanks once more - really appreciate. |
Beta Was this translation helpful? Give feedback.
-
|
Running: Gentoo -systemd Grott.ini: |
Beta Was this translation helpful? Give feedback.
-
|
The setup now is: 2X Growatt SPF5000es in parallel, with two Shine-F sticks, both in firmware 3.0.0.3 Growatts are on most updated fw one 63/64 in release 4 and other 103/104 in release 3; Ps. I'm about to install the third inverter this weekend if the weather allows me. |
Beta Was this translation helpful? Give feedback.
-
|
THis is the grott.service file - note that as part of my tests i raised the nice # to -20 |
Beta Was this translation helpful? Give feedback.
-
|
Wow Fernando you pointed me to something I had forgotten. I disabled the CRC testing in Grottserver ( I do not remember when and why, probably for testing). I will enable it again. And it will give a debug message then. What I also realized that Grott/Grottserver is ignoring records that it does not recognize as Growatt records, but does not write a debug message at that moment. I changed some things and have it running it test now. If it works fine I will published it (probably tomorrow evening). It should work fine in 2.8.3 btw. It would be very strange and difficult to debug if your data is being crippled by some interfering signal. But I can tell you also some stories about my new MIN TL-XH inverter and some strange things happening there with suddenly appearing earth leaks that are hard to solve. |
Beta Was this translation helpful? Give feedback.
-
|
I'm having some earth issues with my mains RCD tripping when switching back and forth from batteries to grid. I think i kind of solve it, let me know your situation is somehow related and we can open a new thread and i'll tell you my year of research on the topic lol So I'm still wtih 3.0 alpha ( checked out the last week), i saw you released some changes to alpha in 3.1. I'm planning to stay in alpha and contribute to the test. Does it worth it to sync with #37bf53e or should i wait your CRC changes? The interfering signal theory is just a theory for now, as i don't have any supporting evidence whatsoever that it is happening in the inverters, but may be. No garbage, nothing weird in Grott log, i analyzed about 48hs of logs and find nothing strange. Cheers, |
Beta Was this translation helpful? Give feedback.
-
|
AH BTW i just enabled +1 inverter so my setup now has 3 SPFs in parallel. Unsure if it helps on anything cheers, Fernando |
Beta Was this translation helpful? Give feedback.
-
Yes lets start a other discussion on RCD Tripping and strange behavour when switching from batteries to grid (because I have some issue in that area as well). |
Beta Was this translation helpful? Give feedback.
-
|
I opened discussion #653 |
Beta Was this translation helpful? Give feedback.
-
|
If possible switch to 3.1. I made a redesign of the session cleanup routine that should make Grottserver more stable. With the changes I posted today CRC checking is switched on again and I add some addition record testing The debug messages you can expect:
|
Beta Was this translation helpful? Give feedback.
-
|
GREAT! so far, no issues, no invalid record and connections are all looking great :) I removed all the automation / periodic restarts and will report back tomorrow, but looks really promising !!! :) |
Beta Was this translation helpful? Give feedback.
-
|
So far, 1 hour running, no functional issue whatsoever. Connections are getting destroyed and recreated, all three inverters updating in .1 sec. Sending commands are functioning fine (think better than before) and all great. The only thing i noticed was only one socket with close_wait status. I'm unsure if it will get cleaned-up by the cleanup routine but I'll see tomorrow. 103 is Grott server, 181, 34 and 203 are the inverters. Cheers, Fernando |
Beta Was this translation helpful? Give feedback.
-
|
Hi there, morning reports it looks we're progressing but with disabling of hourly restarts, the close_wait connections started to get pilling. I send you the logfile through email. I think it started to behave a little bit like 2.8X, now there's no need for constant restarts, but there must be still some issue that lefts some of the connections in close_wait state. I saw no invalid* entries in the log. But will give a more close look now. |
Beta Was this translation helpful? Give feedback.
-
|
Some interesting logs:
|
Beta Was this translation helpful? Give feedback.
-
|
I increased the log to DEBUGV and will send you the new file. Here's another catch i think may be relevant |
Beta Was this translation helpful? Give feedback.
-
|
inverter: 5000MTL-S grott 3.1.0 proxy mode working fine (data to mqtt is ok and data in HA) |
Beta Was this translation helpful? Give feedback.
-
|
When did you downloaded grott? I published a new build (grottserver) that generates more debug info on the received data. Maybe that can help (or that is the problem). |
Beta Was this translation helpful? Give feedback.
-
|
I downloaded the alfa branche on 29/04/25 when I take "debugv" on server mode I get a bunch of lines counting up until 120seconds |
Beta Was this translation helpful? Give feedback.
-
|
Can you download the version I published yesterday |
Beta Was this translation helpful? Give feedback.
-
just done that. |
Beta Was this translation helpful? Give feedback.
-
|
this is debugV |
Beta Was this translation helpful? Give feedback.
-
|
Oke thanks. THAT make things more clearing but I think I have to log more
record information to debug.
I have to change grott for that. I will let you know when it is available
(and hope you want to test it)
Op vr 2 mei 2025, 13:05 schreef WimBra ***@***.***>:
… Can you download the version I published yesterday
just done that.
this is the debug out of the command line:
...
2025-05-02 12:59:35,095 - grottconf - INFO: Grott external record whitelist 'recwl.txt' not found
2025-05-02 12:59:35,095 - grottconf - DEBUG: - Grott records whitelisted :
{'5219', '0105', '501b', '0116', '5216', '5119', '0103', '5250', '5103',
'5129', '5229', '5150', '5016', '5050', '5005', '0120', '5116', '5004', '5104',
'5105', '5003', '0104', '0150', '5019', '0119'}
2025-05-02 12:59:35,096 - grottserver - INFO: Grottserver inititialisation started, grottserver version 3.1.0_250430
2025-05-02 12:59:35,096 - grottserver - INFO: Grott server started
2025-05-02 12:59:35,096 - grottserver - INFO: mode: server
2025-05-02 12:59:35,101 - grottserver - INFO: GrottHttpserver - Ready to listen at: 0.0.0.0:5782
2025-05-02 12:59:35,101 - grottserver - INFO: Grottserver - Ready to listen at: 0.0.0.0:5279
2025-05-02 12:59:35,102 - grottserver - INFO: GrottHttpserver - server listening
2025-05-02 12:59:35,102 - grottserver - INFO: Grottserver - server listening
2025-05-02 12:59:35,102 - grottserver - INFO: GrottHttpserver - Response interval wait time: 0.5
2025-05-02 12:59:35,103 - grottserver - INFO: GrottHttpserver - Datalogger ResponseWait: 5
2025-05-02 12:59:35,103 - grottserver - INFO: GrottHttpserver - Inverter ResponseWait: 10
2025-05-02 12:59:35,394 - grottserver - DEBUG: handle_readble_socket, input received on socket : <socket.socket fd=5, family=AddressFamily.AF_INET, type=SocketKind.SOCK_STREAM, proto=0, laddr=('0.0.0.0', 5279)>
2025-05-02 12:59:35,394 - grottserver - DEBUG: handle_readable_socket, no connection with peer yet, will be established
2025-05-02 12:59:35,394 - grottserver - DEBUG: handle_new_connection, new connection request received:
<socket.socket fd=5, family=AddressFamily.AF_INET, type=SocketKind.SOCK_STREAM, proto=0, laddr=('0.0.0.0', 5279)>
2025-05-02 12:59:35,395 - grottserver - DEBUG: handle_new_connection, send queue created for: 192.168.50.15_43083
2025-05-02 12:59:35,444 - grottserver - INFO: forward, connection with growatt server established: server.growatt.com:5279
2025-05-02 12:59:35,444 - grottserver - DEBUG: forward, socket: <socket.socket fd=7, family=AddressFamily.AF_INET, type=SocketKind.SOCK_STREAM, proto=0, laddr=('192.168.50.42', 46160), raddr=('47.254.130.145', 5279)>
2025-05-02 12:59:35,445 - grottserver - DEBUG: [thread: forward_192.168.50.42:46160] starting
2025-05-02 12:59:35,445 - grottserver - DEBUG: [thread: client_192.168.50.15:43083] starting
2025-05-02 12:59:35,445 - grottserver - DEBUG: handle_readble_socket, input received on socket : <socket.socket fd=6, family=AddressFamily.AF_INET, type=SocketKind.SOCK_STREAM, proto=0, laddr=('192.168.50.42', 5279), raddr=('192.168.50.15', 43083)>
2025-05-02 12:59:35,446 - grottserver - DEBUG: start message buffer processing
2025-05-02 12:59:35,446 - grottserver - DEBUG: Invalid Record (length) detected
2025-05-02 12:59:36,649 - grottserver - DEBUG: handle_readble_socket, input received on socket : <socket.socket fd=6, family=AddressFamily.AF_INET, type=SocketKind.SOCK_STREAM, proto=0, laddr=('192.168.50.42', 5279), raddr=('192.168.50.15', 43083)>
2025-05-02 12:59:36,650 - grottserver - DEBUG: start message buffer processing
2025-05-02 12:59:36,650 - grottserver - DEBUG: Invalid Record (length) detected
2025-05-02 13:00:16,589 - grottserver - DEBUG: handle_readble_socket, input received on socket : <socket.socket fd=6, family=AddressFamily.AF_INET, type=SocketKind.SOCK_STREAM, proto=0, laddr=('192.168.50.42', 5279), raddr=('192.168.50.15', 43083)>
2025-05-02 13:00:16,590 - grottserver - DEBUG: start message buffer processing
2025-05-02 13:00:16,590 - grottserver - DEBUG: Invalid Record (length) detected
2025-05-02 13:00:35,160 - grottserver - DEBUG: Main, available connections:
2025-05-02 13:00:35,167 - grottserver - DEBUG: pconn(fd=4, family=<AddressFamily.AF_INET: 2>, type=<SocketKind.SOCK_STREAM: 1>, laddr=addr(ip='0.0.0.0', port=5782), raddr=(), status='LISTEN')
2025-05-02 13:00:35,167 - grottserver - DEBUG: pconn(fd=5, family=<AddressFamily.AF_INET: 2>, type=<SocketKind.SOCK_STREAM: 1>, laddr=addr(ip='0.0.0.0', port=5279), raddr=(), status='LISTEN')
2025-05-02 13:00:35,167 - grottserver - DEBUG: pconn(fd=6, family=<AddressFamily.AF_INET: 2>, type=<SocketKind.SOCK_STREAM: 1>, laddr=addr(ip='192.168.50.42', port=5279), raddr=addr(ip='192.168.50.15', port=43083), status='ESTABLISHED')
2025-05-02 13:00:35,168 - grottserver - DEBUG: pconn(fd=7, family=<AddressFamily.AF_INET: 2>, type=<SocketKind.SOCK_STREAM: 1>, laddr=addr(ip='192.168.50.42', port=46160), raddr=addr(ip='47.254.130.145', port=5279), status='ESTABLISHED')
2025-05-02 13:00:35,168 - grottserver - DEBUG: Main, available threads:
2025-05-02 13:00:35,168 - grottserver - DEBUG: - MainThread
2025-05-02 13:00:35,169 - grottserver - DEBUG: - httpserver_0.0.0.0:5782
2025-05-02 13:00:35,169 - grottserver - DEBUG: - conserver_0.0.0.0:5279
2025-05-02 13:00:35,170 - grottserver - DEBUG: - forward_192.168.50.42:46160
2025-05-02 13:00:35,170 - grottserver - DEBUG: - client_192.168.50.15:43083
2025-05-02 13:00:36,761 - grottserver - DEBUG: handle_readble_socket, input received on socket : <socket.socket fd=6, family=AddressFamily.AF_INET, type=SocketKind.SOCK_STREAM, proto=0, laddr=('192.168.50.42', 5279), raddr=('192.168.50.15', 43083)>
2025-05-02 13:00:36,762 - grottserver - DEBUG: start message buffer processing
2025-05-02 13:00:36,762 - grottserver - DEBUG: Invalid Record (length) detected
2025-05-02 13:01:16,702 - grottserver - DEBUG: handle_readble_socket, input received on socket : <socket.socket fd=6, family=AddressFamily.AF_INET, type=SocketKind.SOCK_STREAM, proto=0, laddr=('192.168.50.42', 5279), raddr=('192.168.50.15', 43083)>
2025-05-02 13:01:16,703 - grottserver - DEBUG: start message buffer processing
2025-05-02 13:01:16,703 - grottserver - DEBUG: Invalid Record (length) detected
2025-05-02 13:01:35,185 - grottserver - DEBUG: Main, available connections:
2025-05-02 13:01:35,191 - grottserver - DEBUG: pconn(fd=4, family=<AddressFamily.AF_INET: 2>, type=<SocketKind.SOCK_STREAM: 1>, laddr=addr(ip='0.0.0.0', port=5782), raddr=(), status='LISTEN')
2025-05-02 13:01:35,191 - grottserver - DEBUG: pconn(fd=5, family=<AddressFamily.AF_INET: 2>, type=<SocketKind.SOCK_STREAM: 1>, laddr=addr(ip='0.0.0.0', port=5279), raddr=(), status='LISTEN')
2025-05-02 13:01:35,192 - grottserver - DEBUG: pconn(fd=6, family=<AddressFamily.AF_INET: 2>, type=<SocketKind.SOCK_STREAM: 1>, laddr=addr(ip='192.168.50.42', port=5279), raddr=addr(ip='192.168.50.15', port=43083), status='ESTABLISHED')
2025-05-02 13:01:35,192 - grottserver - DEBUG: pconn(fd=7, family=<AddressFamily.AF_INET: 2>, type=<SocketKind.SOCK_STREAM: 1>, laddr=addr(ip='192.168.50.42', port=46160), raddr=addr(ip='47.254.130.145', port=5279), status='ESTABLISHED')
2025-05-02 13:01:35,193 - grottserver - DEBUG: Main, available threads:
2025-05-02 13:01:35,193 - grottserver - DEBUG: - MainThread
2025-05-02 13:01:35,193 - grottserver - DEBUG: - httpserver_0.0.0.0:5782
2025-05-02 13:01:35,194 - grottserver - DEBUG: - conserver_0.0.0.0:5279
2025-05-02 13:01:35,194 - grottserver - DEBUG: - forward_192.168.50.42:46160
2025-05-02 13:01:35,195 - grottserver - DEBUG: - client_192.168.50.15:43083
2025-05-02 13:01:36,874 - grottserver - DEBUG: handle_readble_socket, input received on socket : <socket.socket fd=6, family=AddressFamily.AF_INET, type=SocketKind.SOCK_STREAM, proto=0, laddr=('192.168.50.42', 5279), raddr=('192.168.50.15', 43083)>
2025-05-02 13:01:36,875 - grottserver - DEBUG: start message buffer processing
2025-05-02 13:01:36,875 - grottserver - DEBUG: Invalid Record (length) detected
—
Reply to this email directly, view it on GitHub
<#613 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AD56LCWNUJA3N6Z3V42CDKT24NGPDAVCNFSM6AAAAABTHFCXWWVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTGMBRGMYDGOA>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
Beta Was this translation helpful? Give feedback.
-
|
Ok strange. grott(server) 2.7.8 and 2.8.3 should also work for reading and setting inverter information (registers). |
Beta Was this translation helpful? Give feedback.
-
|
But I am glad your testing V3 ;) |
Beta Was this translation helpful? Give feedback.
-
|
Which inverter and datalogger do you use? |
Beta Was this translation helpful? Give feedback.
-
|
the inverter is a Growatt 5000MTL-S I will upgrade (working) grott to version 2.8.3 and try the server again. The problem with grottserver is "api get datalogger" or "api get inverter" is giving the same "datalogger" number. |
Beta Was this translation helpful? Give feedback.
-
|
On tweakers forum others seems have problems with the shinelan as well. I do not not have a shinelan to test with (I have shinelink and shinewifi, and to be honest grott start to be an expensive hobby so buying one will not be an option ). I hope the people with a shinelan will help extensive testing. |
Beta Was this translation helpful? Give feedback.
-
|
Hi Fernando did you sent the log? I have not received it yet (I think).
Op vr 2 mei 2025, 12:35 schreef Fernando Cunha ***@***.***>:
… I increased the log to DEBUGV and will send you the new file. Here's
another catch i think may be relevant
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,582 - grottserver - WARNING: handle_readble_socket, empty read, close connection
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,582 - grottserver - DEBUG: Close Connection for socket: <socket.socket fd=9, family=2, type=1, proto=0, laddr=('192.168.51.103', 5279), raddr=('192.168.51.103', 47976)>
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,582 - grottserver - DEBUGV: retrieve connection information
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,582 - grottserver - DEBUG: Close Connection for socket: <socket.socket fd=9, family=2, type=1, proto=0, laddr=('192.168.51.103', 5279), raddr=('192.168.51.103', 47976)>
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,582 - grottserver - DEBUGV: clean connection queues
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,582 - grottserver - DEBUGV: clean Passthrough connection queues
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,582 - grottserver - DEBUGV: clean connection queues error: cannot access local variable 'pt_qname' where it is not associated with a value
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,582 - grottserver - WARNING: handle_readble_socket, ConnectionResetError exception: cannot access local variable 'pt_socket' where it is not associated with a value
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,582 - grottserver - DEBUG: Close Connection for socket: <socket.socket [closed] fd=-1, family=2, type=1, proto=0>
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,582 - grottserver - DEBUGV: retrieve connection information
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,582 - grottserver - DEBUG: Close Connection for socket: <socket.socket [closed] fd=-1, family=2, type=1, proto=0>
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,582 - grottserver - DEBUGV: clean connection queues
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,582 - grottserver - DEBUGV: clean connection queues error: cannot access local variable 'qname' where it is not associated with a value
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,583 - grottserver - DEBUGV: clean Passthrough connection queues
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,583 - grottserver - DEBUGV: clean connection queues error: cannot access local variable 'pt_qname' where it is not associated with a value
mai 02 11:32:33 note grott[277693]: 2025-05-02 11:32:33,583 - grottserver - WARNING: handle_readble_socket, generic exception: cannot access local variable 'pt_socket' where it is not associated with a value
—
Reply to this email directly, view it on GitHub
<#613 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AD56LCSYYQN327ZB5HFH3OD24NC6XAVCNFSM6AAAAABTHFCXWWVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTGMBRGI3TSNA>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
Beta Was this translation helpful? Give feedback.
-
|
So far no more close_wait connections LOL:) I'll wait this evening and post tomorrow the results. Cheers, |
Beta Was this translation helpful? Give feedback.
-
|
Nevermind... it continues. Looks like the tcp conn buffer is getting filled ( i see netstat increasing mem size every 10 sec for about 3 to 4 times ) and data stays there. I'm trying to find in the code if there's any means for it to fail and stop processing data. I see no errors fortunately of CRC or invalid data in the log. What about creating another timeout to catch an stall connection? doesn't look like the current timeout routines are recognizing it. I'll see if i can make a quick video to show... cheers , |
Beta Was this translation helpful? Give feedback.
-
|
so did the same and included close routines in some of the functions. Now looks fine but still lets get thought the night. |
Beta Was this translation helpful? Give feedback.
-
|
here's the changes: |
Beta Was this translation helpful? Give feedback.
-
|
I'm reverting the changes, still having issues. I'm thinking about implementing a timeout in package processing function or when acquiring the buffer. What if the problem is with the stick that delays sending the package, for ex - never goes to the end and send data in piecemeal? Cheers, |
Beta Was this translation helpful? Give feedback.
-
|
I swapped from beta 2.8.3 to the new alpha version : 3.1.0_20250406 both stand alone installs directly on Linux. Only tried the proxy mode for Alpha so far as I noticed that my energy dashboard in HA was not updating energy used from the battery or energy to the battery. I have tracked it down to differences between versions, see below: Stable and Beta versions: etouser_tod Alpha versions: etousertod FYI I am using a SPH inverter and have swapped back to the beta 2.8.3 for now. Were these changes intended? |
Beta Was this translation helpful? Give feedback.
-
|
Yes I was trying to use the invtype = auto I just wasn't sure if the naming was intended or if not and if I was just the first person to notice these changes as maybe not too many people are using the alpha with sph if they haven't noticed yet. |
Beta Was this translation helpful? Give feedback.
-
|
I designed the layout processing so that it is also possible to have multiple names for the same register. It will be a big job to harmonize the naming while we combine now multiple separate defined Layout definitions (SPH naming might be different then MOD, MIN or SPA at this moment). It will help if I can use multiple names for the same register (only the MQTT message and Influxdb record might become a little bit bigger). |
Beta Was this translation helpful? Give feedback.
-
|
About the values. Beta 2.8.3 - "eacharge_total": 6512 Alpha 3.1.0 - "eacharge_total": 65536 (the highest number that can be stored in a 16bit register). Alpha 3.1.0 - "acchargeenergytotal": 6512 (this number matches up with eacharge_total from Beta 2.8.3) The first Beta 2.8.3 - "eacharge_total" 6512 and the Alpha 3.1.0 - "acchargeenergytotal": 6512 are correct I believe and the Alpha 3.1.0 - "eacharge_total": 65536 I think this one is a bug or wrong register read or something else not right? For the 4 elocalload ones it is hard to tell as the growatt web page doesn't show them accurately for me. But the fact that the first one Beta 2.8.3 - "elocalload_tod": 4294967295 is so high seems wrong as it is a tod (today) amount and it is nearly the same as Beta 2.8.3 - "elocalload_tot": 4294904183 which is a tot (total) amount? If that has changed to allow negative numbers that could make sense with the other inverter that I have a negative local load? Alpha 3.1.0 - "elocalloadtod": -5 - if that is minus 5 kw/h then for that day that could be close (maybe too much but hard to tell)? Alpha 3.1.0 - "elocalloadtot": -63117 - I would need to do a lot of checking to see if that was correct or not? |
Beta Was this translation helpful? Give feedback.
-
|
I am not sure about having more values (old and new naming)? Having lots of names / values that are all for the same thing(s) in the database might not be the best option for data growth and future confusion when dealing with the data later? But I would be ok to change the old names to the new names in the HA and Influx history if the alpha names are the names (or any other names you decided on) that they will staying as going forward. I could always change the local grott code to suit whatever I wanted to have but that could cause problems with keeping up to date with your changes in the future. So I guess once you confirm what you want all the names to be for the future version(s) or what you want to change then the decison might become clearer. |
Beta Was this translation helpful? Give feedback.
-
|
I went back to using v3 and fixed HA to use the new naming and am now just writing v3 values to a new influx database. |
Beta Was this translation helpful? Give feedback.
-
|
@johanmeijer think i got it ! ;) I started debugging all functions, from handle_readable_socket(), process_data() and finally procdata() and I found myself back to influxdb issue. I enabled more debugging and found that the influxdb code was never active when the issue happened. Sometimes MQTT publish.single gets stalled, doesn't return any errors and cause Grott to stop responding for that socket. The timeout routine takes care about it but it stays with close_wait for a great time and they pile up with time. I was not able to destroy the sockets, that would fix the issue. Maybe because the thread is hang waiting for the mqtt publish function to return. I'm trying now to create a threaded call with 1sec timeout and will post the results as soon as I do some more testing. BTW it was a great opportunity for me to learn python and now i think i understand like 15% of your code. Great work Johan! Glad to contribute to a project i've been using for so long Cheers |
Beta Was this translation helpful? Give feedback.
-
|
The challenge with writing to influxdb is that at this moment I have made it a Sync write (so you have to wait for and Grott will fail if it gives an error). With a short update frequency this might causes problems with timing. I can also make this async (probably I will then make the whole data layout processing a separate thread) but then we lose the error detecting (still can send out a message off course) and handling for writing to Inluxdb (and probably also true for MQTT andPVOutput). Atthis moment when the data processing routine fails Grott fails (does not sent a record confirmation tot the datalogger), but.... The failing record is normally being kept in the datalogger and sent out later as a buffered record (true for shinewifi, not sure for shinelink . In this way you will never lose records. So we have to choose what is the best option! Offcourse we can also make this optional (another parameter setting in .ini, we already have to much a I think). Yeah Grott is also great for me to learn python. I am a technical guy (computer Infra/OS/Middleware and electronics/electrical engineering), learn programming in difference languages (including assembler ;) ) but Python was new for me. Probably (Hopefully) you see the change in skills in the difference parts of Grott also (it is empirically developed). |
Beta Was this translation helpful? Give feedback.
-
|
great, let me try this one :) so this night i learnd both MQTT and Influxdb are randomly silent stalling the code. I'm 100% sure of MQTT but not that much of influxdb. I think the pass option in case there's an error in the code might be enough for influxdb but not for MQTT. So I'm concentrating now in MQTT, evening i tried to use concurrent.futures to start another thread and limit the MQTT publish to 1sec. Thread works but guess what? the limitation of 1sec stalled also LoL. I asked some support of grok and he suggested another class to handle MQTT and setting sock timeout. I'm trying it now... I've put Inflxdb call under concurrent.futures threaded also and for it, it looks fine. I'll run the code a few hours and see how it goes. If it goes ok I'll see the next steps to submit a PR. PS. I'm network engineer, but spend most of my life with infra/programming (old days ) and the last 15 years I'm in Information Security... Cheers, Fernando |
Beta Was this translation helpful? Give feedback.
-
|
By the way also something it's intriguing me is why both calls are failing in my system.... do you feel you ever experienced something like that in your system ? has anybody? I'm wondering about these factors:
|
Beta Was this translation helpful? Give feedback.
-
|
Looks stable and solid for the last two hours. I will send the PR the end of afternoon. Cheers, |
Beta Was this translation helpful? Give feedback.
-
|
Still stable, 0 restarts from yesterday and 0 close_wait connections. |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for your hard work, @johanmeijer. For anyone interested, I’ve created a Docker image for the Alpha branch of Grott. Below is the And here’s a basic Note: You must mount the Also, as previously mentioned by @donaldwpage: |
Beta Was this translation helpful? Give feedback.
-
|
Oke. Thanks. I will also publish an image at the "official" Grott docker page soon. https://hub.docker.com/repository/docker/ledidobe/grott/general |
Beta Was this translation helpful? Give feedback.
-
|
I will look at the new settings. I think I do not finished that part of the config yet indeed. |
Beta Was this translation helpful? Give feedback.
-
I take it this didn't happen? |
Beta Was this translation helpful? Give feedback.
-
|
@johanmeijer wanted to provide some feedback. I fired up @egguy alpha channel home-assistant add-on that includes Grott3.0.0. Working great in server mode! SPH5000 here, with SDM630 smart meter. Single phase inverter, three phase meter. Could never get the 2nd and 3rd phase information before but magically started working with the upgrade. Was able to set my interval register (and get it to stick) now and super happy with 1minute interval vs the old 5min. Massive thankyou to @egguy and @johanmeijer for your combined efforts. |
Beta Was this translation helpful? Give feedback.
-
|
Currently testing Alpha 3.x to my Growatt 5000UE. Grott 2.7.8 was working fine, but in Grottserver Alpha i get nothing but invalid record length detected. Sample of the plain data from grott 2.7.8 at the bottom Sample og plain data captured from grott 2.7.8: |
Beta Was this translation helpful? Give feedback.
-
|
will retry tomorrow when my inverter wakes up again |
Beta Was this translation helpful? Give feedback.
-
|
Alas, that didn.t help. It seems to not recognize the record layout T020203 or T02NNNN in Alpha 3.0.0., while the log shows that layout is loaded. |
Beta Was this translation helpful? Give feedback.
-
|
When running in proxy mode, i get an important log line "DEBUG: no matching generic inverter record layout found". I even made a separate json file for the T02NNNN layout (copied from grottconf.py in version 2.7.8), but alas, 'no matching generic inverter record layout found' keeps coming. |
Beta Was this translation helpful? Give feedback.
-
|
OK thanks, try having the json in the grott directory and setting invtype=auto. If it doesn't work, may would be nice to let @johanmeijer take a look as he changed the detection code for this release. |
Beta Was this translation helpful? Give feedback.
-
|
Some progress: this combination works:
and i see the data on screen (like pv voltages and stuff) But, it doesn't work in server mode, this messages gets repeated over and over: How do i let johan know? |
Beta Was this translation helpful? Give feedback.
-
|
@pw-h So it works in proxy mode, not in server mode. Luckily we now have data to test with from the proxy run. I activated the record checking in grottserver again (it was always on in grottproxy) but it does not seem to work well for voor type 02 records. This is an older record format that does not have CRC checking, different from the one I have, but that is ok it should work, Grott should not perform CRC checking. If you specify auto. The record layout is build by Grott (based on the register information in the record). It will ignore the T02NNNN.json layout (to be more specific all Txxxxx.json layouts). I am going to test (and hopefully fix) it. |
Beta Was this translation helpful? Give feedback.
-
|
@pw-h it should be fixed in version 3.2.1 |
Beta Was this translation helpful? Give feedback.
-
|
3.2.1 fixed it for me, server mode is working! Thank you. |
Beta Was this translation helpful? Give feedback.
-
|
I cannot get api calls to work. http://10.0.0.80:5782/inverter gives me: but http://10.0.0.80:5782/inverter?command=regall&inverterid=02 gives me I don't know what to enter for 'inverterid' , "AH24030766" and "RX04020003" don't work either. |
Beta Was this translation helpful? Give feedback.
-
|
This command does work: http://10.0.0.80:5782/datalogger?datalogger=AH24030766&command=register®ister=4 |
Beta Was this translation helpful? Give feedback.
-
|
Been a while since something happened here. Have plans changed? |
Beta Was this translation helpful? Give feedback.
-
|
Been using this code I think for over a year - 0 issues so far. for SPF series inverter i'd say the version qualifies for production. But there are some bits in documentation and other HW to check. Happy to support @johanmeijer to move fw |
Beta Was this translation helpful? Give feedback.
-
|
I used this code now for several months, and also curious what the plans forward are. I however noticed that Johan Meijer is not active anymore since july?
|
Beta Was this translation helpful? Give feedback.
-
|
Primarily asking because I'd like to deploy it for my local system which is a MOD10000TL3-X with APX battery and a SYN backup box, so a bit more complex system setup I don't really want to have to dig through the code to fix things, given Growatts own platform is currently so broken the data doesn't add up. I noticed there were a few forks already adding features. There is apparently still a lot of interest in making this work better than Growatts own platform and removing the whole cloud connectivity nonsense. Was gonna email some of the forks that seemed to be active asking for setup help, since I'm not keen to break my brain over this(already had enough trouble getting the system working in the first place). Not sure where @johanmeijer has gone. Be neat if the forks could consolidate some of the better changes, perhaps development needs to move somewhere if this repo isn't being updated anymore. |
Beta Was this translation helpful? Give feedback.
-
|
Ok I am here (again!?). Last months I have not been able to do a lot with Grott because of other priorities (rebuilding 3 houses ;) ). I have to say for me there was no urgent reason while I am running V3 now for a couple of months with no problems (min 3600TL-XH with ARK batteries). I am planning to start developing / problem solving again the coming weeks with V3 as basis. For me priority is to set battery load priorities so I can make optimal use of the configuration and keeping costs in control. For me this has no real priority while in the Netherlands we have some special rules for netting used against delivered KWH. But this will end in 2027 and my contract with my current fixed contract with VattenFall will end mid 2026 (two moments I have to take into account). I will start with looking at the issues and after that I will try to build a smart control mode in Grott. |
Beta Was this translation helpful? Give feedback.
-
|
I do understand that people moved on. and improved (or at least modified for own use) Grott. As I also stated in past that I am not in favor of adapting pull request of others while Grott is also a project for me to extend my python skills and I really want to be in control of my code to be able to do debugging. But..... I will see what interesting and useful things has been developed and will look if it can be integrated in this original build. |
Beta Was this translation helpful? Give feedback.
-
|
Great to hear you are still looking at this project. Ordered a little box a while ago to install on and see if I could get it to work. Would be up for troubleshooting? It's a MOD TL3-X, SYN, APX setup, so may not work out of the box as I can't seem to find any references to SYN in the code. I'm keen to fix some logic issues with the configuration, mainly battery sleep mode and SOC levels on/off grid. |
Beta Was this translation helpful? Give feedback.
-
|
@johanmeijer Happy new year! Just setup grott alpha with grottserver running as well. Seeing various messages in the debug output. Some apparently ignored cause they aren't long enough. Thing I'm wondering though. Is there any plan to have some form of interface to view system status and manage it like on growatts interface? Cause I can't really tell right now if everything is working. I see on the wiki there is mention for ways to monitor, but I'd like to have something that just locally outputs some sort of grafana or something. Preferably some way to do control changes as well for the various registers and options. What's best for that or do I have to build that myself? I setup in proxy mode now to test if this keeps running long term. According to growatt at least it is connected. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
The new 3.0.0 version of grott is available in the alpha branche (under development!!!)
What is new in 3.0.0
How to use:
If you do not change grott.ini Grott and grottserver can be used in the same way as in previous releases (backwards compatible) !
But....
You have to install PSUTIL package:
New options at startup:
loglevel is python loglevel, see below which are available
option -v will still work and is equal to -l debug
New settings in grott.ini:
[Generic]
Specify mode (sniff/proxy/server/serversa)
New section added to define grottserver settings (works also in grottserver standalone)
[Server]
advanced settings only for debug or advanced tuning!!!!!
Beta Was this translation helpful? Give feedback.
All reactions