I buy 5 shelly 2.5 (EM, Dimmer) black friday (a year ago), and i am stuck with it since then. (Network issue with all 5). on my different attempts, I went all version firmware up to 1.9.3 rc5 but the problem persist. Knowing some other user have them working, I was looking (deep) into my network settings.
Before explaining underneath, I want to say before installing the Shelly 2.5 everything work fine including Shelly EM and Shelly Dimmer, and some other Sonoff/tasmota devices (ping<10ms). Routeur is a Draytek vigor 2820Vn, very reliable, Home Network very stable.
Classic Shelly 2.5 install via Shelly 2.5 ssid, (I notice the the 192.168.33.1 page is quiet slow to appear) ping from my phone show 30-40 ms. good enough to be able to look at fw version and set wifi on the phone ssid. reboot again web page from the phone is very slow to come by, but I am able to update firmware that way, reboot, new firmware ok (radio signal -45db).
I change ssid to home routeur, reboot, average ping from wired iMac > 100ms (phone, routeur & shelly are within 2 m of my desk).
I try with dhcp or fixed IP (both in the routeur and in Shelly2.5), no change.
UNTIL you do not try to talk to the shelly, just ping it, the average is > 100ms, BUT if you open a webpage pointing to Shelly IP, it take 2 to 3 minutes (really long) to have an open shelly page with "failed to connect". 100 pings show mostly time out, some packet ping between 10.000 to 15.000 (15 secondes). about 1 mn after closing the webpage, ping fall back to 100ms (and my wife iMac recover internet connection, network was flooded)
factory test page take minutes (http://192.168.2.50/settings/developer?selftest_done=1), but work. (but firmware update via browser failed)
Getting some tools to troubleshoot, upload Wireshark, session with filter shelly2.5 (192.168.2.50) and my Mac (192.168.2.58).
save a short period capture, you can find here -> in my dropbox public DL.
You will see a dialogue of the deaf, with lost packets, resend, duplicated, etc... (and the ping)
I would love if some expert eyes could have a look, to interpret the [RST, ACK],...
look a bit like in this thread -> https://stackoverflow.com/ques…nds-up-in-rst-ack-from-cl -> and the solution was ''It turned out that not-changing emphernal port of the client was the problem. Socket was in
TIME_WAIT state which prevented the client from connecting to the server once again. – Mateusz Stompór Jan 20 at 14:02''
if his finding apply to my case how to change that in shelly2.5 ?
Conclusion, my 5 shelly 2.5 (same batch) are affected. happen in Shelly ssid, phone ssid, home ssid. I have raspberrypi with HA 2020-12, shelly 2.5 are detected but failed to connect during configuration, but EM dimmer configure ok. Shelly 2.5 flood the network if you try to address them.
Well I reach my limits here, need some help ...
Ps: just got this year black friday after 20 days of storage by the courrier NEXIVE. (I went to pick it up myself), will be able to also check with Shelly 1, 1PM, 1L