Elli Wallbox not reacheable via mdns -> Timeout #15651
Comments
|
Does 130.2 still work for you? Current nightly? |
0.127.3 is working. Later released Version not. |
|
/cc @DerAndereAndi avahi issue |
|
Well, v0.127.3 is also using avahi. If the mDNS entry provided by the wallbox, does not provide an IPv4 address, then no connecting can be provided. Usually shortly after another mDNS entry is provided which includes an IPv4. I never got a connection with IPv6 working, so I do not try. The log entry is NOT an error, just an info. Also, it is NOT the case that only the HEMS shall initiate a connection, the wallbox shall does so as well, but as far as I know hardly ever does. But please provide a log with eebus on TRACE level to see more details. As a note: previously there was the option to define an IP address to "fix" this problem of the wallbox not reporting the IP address, but I was hoping that using avahi this should be solved now. Looks like it isn't and another bug of the wallbox software. |
|
Could you please run on your linux system where evcc is running the following command and provide the output (may require to install the avahi-utils apt package): avahi-browse -d local _ship._tcp --resolve -tOn macOS you can use the |
I can only provide the log while running 0.127.3: 0.130.2 now crashes with |
Is there more to configure for TRACE-Level than in the config above? |
|
The crash is fixed and will be part of the next release. evcc accidentally included the wrong version of the mdns avahi library. No, set the log level in the configuration file and run evcc. Then grab the output as you did before.
|
|
Could you please try to set add the network interface `` to the eebus configuration in evcc.yaml? See https://docs.evcc.io/docs/reference/configuration/eebus#interfaces eebus:
interfaces:
- eno1This way evcc only uses that network interface for its mDNS announcement and all other IPs are ignored. The log file shows that the IPv4 address is provided for the wallbox, but somehow not visible to the eebus code. I really would need the evcc log with eebus set to level trace and hope I'll see a hint on why it is not reported. |
I retried with the interface 0.130.2 under Linux. eno1 is the right interface (checked ifconfig). |
|
@mbjw77 I have similar issues with my evbox. Do you have ipv6 and ipv4 over the same network ? It could be that the box is advertising on both ... |
|
me too. looking forward to the next release... |
|
Could you guys please run this tool for a minute and post the output? Windows Intel/AMD CPU: mdns_windows_amd.exe zeroconfLinux ARM CPU (run both): mdns_linux_arm avahi
mdns_linux_arm zeroconfLinux Intel/AMD CPU (run both): mdns_linux_amd avahi
mdns_linux_amd zeroconfmacOS ARM CPU: mdns_macos_arm zeroconf |
The box is advertising on both, but does not accept a connection via IPv6. No box does that, so the EEBUS implementation doesn't use it so far. |
|
[eebus ] DEBUG 2024/08/25 17:57:30 connection to ae82347bb2de22aa304c83fb1d8f1d94aeb29b4f failed: dial tcp 192.168.1.230:4712: i/o timeout |
|
@jomach please contact EVBox. Seems they still have a lot of software issues. They are also the ones who developed the Elli Connect/Pro wallboxes. The Wallbox reports the IP address and port and then doesn't answer. There is absolutely nothing that can be done from the outside. Also your problem is very different from the reported issue here. |
|
Ok. Let me know id I can help someone where |
of course. Made with running 0.127.3. |
|
The confusing thing is how that should have worked on 0.127. the only explanation we have sofar is that current version trips different bugs in den FW that lead to this behavior. Fact is: the box has entirely crappy SW, not even featuring a valid certificate to begin with. |
|
here is my output from an Elli Connect wallbox (software release v3.8.12) and evcc 0.127.3: |
|
One reason why this version is working for me is that I can set the IP via the config. So, mdns should not be needed... |
|
Can we verify if mdns leads to different/wrong IPs than manual config? |
|
@sebastianohl some notes:
mDNS is needed anyway, since we need to know the port address. |
|
@mbjw77 Can you try to ping Did you quit the app directly or did you let it run a bit? Especially with |
They ran for about 2 minutes. But I can add some more... running 6 minutes each. |
|
Thanks. Sometimes fe80 addresses work, but not always depending on the network setup. Could you guys please give this updated version a try? The zeroconf option should hopefully now also provide an mDNS entry for the wallbox. |
|
@DerAndereAndi i can ping the fqdn and the ipv6 address. I let the zeroconf run for a few minutes (<5m) |
|
@sebastianohl thanks. could you retry with the new build and post the outputs ? |
|
By New buidl you mean 130.4 or master? I can try tomorrow... |
|
The new build of the test tool attached here: #15651 (comment) |
10 minutes each |
|
Sadly zeroconf still doesn't show the wallbox. I can't reproduce this over here. What kind of setup do you guys have? WLAN, LAN, both, ... ? Is anyone maybe able to provide a wireshark trace of the mDNS packages? |
|
@DerAndereAndi I just tried the version 0.126.1 and I still get the timeout but it works. on the network config from evcc I removed the .local and I only enter the IP address. Then I get this: With 0.130.x it does not work and I get: So something has changed between 0.126.x and 0.130.x |
|
@jomach please create a different ticket, you do have a different issue than the one discussed over here. This gets way too confusing for me to track. |
Wallbox, server running evcc, a FritzBox and the Kostal-stuff are directly connected to an unmanaged gigabit switch via LAN.
I would try to generate one. But I'm not that experienced with Wireshark. Do you have a filter-configuration I could apply? |
|
@mbjw77 Can you disable IPV6 from your FritzBox? can you run |
I disabled IPv6 in my FritzBox
and tried 0.130.4 works :-) |
I see that you have ipv6 still active on your ethernet card. You can try to disable it. So the issue is confirmed to be related to mdns and ipv6 ? |
|
sorry for the delay. here are my logs. also letting it run for some time did not change the logs. |
|
The next nightly build will introduce a few updates:
|
|
Has anyone verified the fix? Can this issue be closed? |
|
I just upgraded my installation to the latest nightly container (sha256:a9dea5584fa4b0990205061fdd9b4575fbc69cd657d73f3bc973e3b7b2eb0407). It is working with an ip in the config. If I remove the IP, it is looking up the announced DNS entry but do not get a reply (and than it is failing). For me the solution with it IP in the config works. So, the ticket can be closed (for my part). |
|
I don't have this Box ... |
|
Vielen Dank für euren Support :-) |
Describe the bug
I have a problem when starting evcc. It seems that my wallbox (ID. Charger Pro) is not reachable.
The problem is displayed as “
mdns: avahi - no IPv4 addresses available”.My last successful configuration was with version 0.127.3. and I would like to update to 0.130.2.
Since 0.127.3 I can no longer reach my ID. Charger Pro. Neither with a previously working evcc.yaml nor after replacing the certificate during configuration. I have tried it with every version.
Is there a way to enter a fixed IP address as before? Or another workaround?
I know that this wallbox is defective, but unfortunately I have to use it. That's why I bought an external electricity meter, hoping that the latest optimization for the Elli wallboxes would solve my problem. But unfortunately not.
Tried macOS, Linux and a native home assistant.
Steps to reproduce
Start evcc
Configuration details
Log details
What type of operating system are you running?
Linux
Version
evcc version 0.130.2
The text was updated successfully, but these errors were encountered: