Grott 3.0.0 now available in Alpha branche!!!!!

[johanmeijer]

The new 3.0.0 version of grott is available in the alpha branche (under development!!!)

What is new in 3.0.0

• Grott and Grottserver are fully integrated
• Auto invertype detection mode is added (will automatically detect invertype and process data, be aware: not working for SPF type inverters yet)
• In server mode optionally sending data to growatt servers (be aware data form grwatt server will be ignored)
• Started with using default python logging capabilities (instead of print log messages)
• Use more multithreading to enhance performance and create stability
• new grott.ini in example directory (include settings and documentation)

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:

(sudo) pip3 install psutil   

New options at startup:

python grott.py -l _loglevel_ 

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]

loglevel = INFO

• Specify loglevel for extended messaging, uses standard python logging level: DEBUG,INFO,WARNING,ERROR,CRITICAL added DEBUGV (debug verbose+)
• Specify verbose for extended messaging is deprecated >3.0.0 use loglevel instead, verbose = True will be defaulted to loglevel = debug!)

mode = server 

Specify mode (sniff/proxy/server/serversa)

• server: run grott and grottserver integrated
• serversa: run as grottserver

 invtype = auto

• auto detect inverter type
• specify inverter type: default (use standard tl-s type of inverters, auto (>3.0.0, lets grott automatically detect),spf, sph, mod etc use specific inverter types)
• be aware: auto does not detect spf type inverters!!!
• auto will also work with invtypemap parameter to specify multiple inverters (useful when 1 inverter is spf type)

New section added to define grottserver settings (works also in grottserver standalone)

[Server]

serverpassthrough = True

• In server mode, send data also to growatt servers (one way only, data from growatt will not be processed to block commands comming from growatt!)

advanced settings only for debug or advanced tuning!!!!!

##Define ip/port if grottserver is running standalone and/or at other system
serverip = 0.0.0.0
serverport = 5781
httpip = 0.0.0.0
httpport = 5782
##Time to sleep waiting on API response
#serverrespwait = 0.5
##Total time in seconds to wait on Iverter Response
#serverinverterrespwait = 10
##Totaal time in seconds to wait on Datalogger Response
#serverloggerrespwait = 5

[donaldwpage]

new dependency
pip3 install psutil

[johanmeijer]

Oh yes. I will document that! Thans.

[donaldwpage]

Working fine with https://github.com/muppet3000/homeassistant-grott

[tabascoz]

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.

[johanmeijer]

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.

[tabascoz]

Many thanks Johan
Sure, perhaps a clue might be the power parameters setup that differentiates spf3000es from 5000 series. it's accessible from the console in hidden menu but should be available in a register also.

cheers,

[donaldwpage]

I know it is in another thread, but just to make it clear here also, if you set
[Server] serverpassthrough = True
you also need to set the IP address of server.growatt.com

[tabascoz]

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...

[tabascoz]

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.

[johanmeijer]

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 [email protected]?

[tabascoz]

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.
cheers, Fernando

[tabascoz]

It happened now, will send the logs as soon as it gets back reporting. Many thanks, Tabascoz

[johanmeijer]

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.
I have two inverters so hopefully I can recreate your error (by shorten the update interval time).

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.

[tabascoz]

HI there

Question: Did it restart itself or do you have to restart the inverter?

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.

Is the inverter coming back after circa 5 minutes (300 seconds)?

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,

[tabascoz]

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

[johanmeijer]

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.

[tabascoz]

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...

[tabascoz]

Hi there Johan,

I had the issue a few times - didn't have opportunity to post before. Some facts/tests:

• I can ping the inverters with normal times. ping -f goes ok, <1% losses
• When it starts happening, it happens frequently.
• Inverter goes almost instantly reporting if I restart grott.
• Grott process is running fine.

I noticed that i changed also grott.service, in the old one i had these lines to do periodically restarts:

Restart=always
RestartSec=1
StartLimitInterval=10
StartLimitBurst=5
RuntimeMaxSec=604800

I've increased debugging to DEBUGV and will send you the file once it starts to happen.

Thanks once more. Cheers

[johanmeijer]

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).

[tabascoz]

ok thanks, it might be something related to stale connections ? Does it helps to send you further logs?

cheers,

[johanmeijer]

Ok I have issue's as well at this moment. It seems that cleaning the threads does not work that well.
At this moment I have enough information to try to repair it (and I am back to one minute interval again to see if that is more stable).

[tabascoz]

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.