local network access to the gateway completely unreliable

[crackers8199]

Anyway... I only have 2 SSIDs (a 2.4 and 5 ghz), and I dont have them firewalled off from each other or anything. I have always ment to setup an IoT network but havent ever gotten around to it.

I will also say that I am not a network engineer, so.....

Based on what you have said so far, the general troubleshooting I would likely be advising would be to ask if you have any of those 4 SSIDs firewalled off between each other, or any other specific firewall rules etc. I would ask you if you have been able to narrow down "devices which cant reach" and "devices which can" to any specific subnet or vlan if you have those.

I would also ask if you have a way to either simplify for testing, or put a device on the same SSID / subnet / vlan that the gateway is on, and see if you continue to have connection issues.

i do have an IoT network set up, but it's only wired. all of the wireless networks are connected at this point, except for the guest network...it's my intent to wall off the IoT SSID (which is what i have this connecting to), but i haven't gotten that far yet. that'll eventually get put on the same VLAN that the wired IoT network is on, but like i said right now they're all on the same subnet. all of the devices (the ones where some can and some can't connect) are already on the same subnet as the gateway, which is what makes this so incredibly frustrating.

i do have a set of powerline network adapters that i'm contemplating trying at this point, but i don't know that i'll have any more luck there than i am with the wireless bridge.

 

[crackers8199]

example of what i'm talking about... earlier i posted the ping stats from my HA server. here are the ping stats now from the computer i'm typing this on (which is on the same subnet as both the HA server and the gateway):

~ % ping 192.168.88.225
PING 192.168.88.225 (192.168.88.225): 56 data bytes
ping: sendto: Host is down
ping: sendto: Host is down
Request timeout for icmp_seq 0
ping: sendto: Host is down
Request timeout for icmp_seq 1
ping: sendto: Host is down
Request timeout for icmp_seq 2
^C
--- 192.168.88.225 ping statistics ---
4 packets transmitted, 0 packets received, 100.0% packet loss

 

[jjrandorin]

With the same qualifiers I posted earlier about "sorry for the basic question", have you tried flushing the DNS cache of the device you are on now, and / or your OS equivalent of "ipconfig /release and renew" and trying again?

If some devices can reach it and others cant, at basically the same time, that doesnt describe an issue with the actual gateway (else all devices would be unable to reach it at once). It also doesnt appear to describe a connection drop, because, again, all devices wouldnt be able to connect at the same time.

Can you run a continuous ping from a couple devices and see if there is a simultaneous drop, or not? This feels like "routing" to me, which is above my network skills.

 

[crackers8199]

With the same qualifiers I posted earlier about "sorry for the basic question", have you tried flushing the DNS cache of the device you are on now, and / or your OS equivalent of "ipconfig /release and renew" and trying again?

i have previously, yes.

here's some more fun: i just took the bridge out of the equation and connected the two powerline adapters i mentioned. i can still access the UI now (on both machines), but the gateway says not connected now (even though it obviously is)...good times.

 

[crackers8199]

now the gateway says it's connected again. we'll see how well this holds using the powerline adapters.

 

[scdoubleu]

The issue I used to see with the TEG over WiFi is that it randomly stops responding to ARP requests over WiFi for a period of time. Once the ARP entry expires from your computer's network stack, the TEG appears to be off the network. It's not - it just isn't responding to the new ARP request.

My Tesla devices/cars are on a segregated WLAN connected via a firewall/router to the rest of my network. Adding a static ARP entry for the TEG on the router cured the problem for TEG access from the rest of my network.

I don't know if the problem exists over Ethernet as I can't easily run a cable to the TEG.

 

[gpez]

I read through the thread and may have missed it - @crackers8199 did you have the TEG forget your wifi network? I had to do so in order to get the ethernet connection to work consistently but it has been rock solid ever since.

 

[crackers8199]

I read through the thread and may have missed it - @crackers8199 did you have the TEG forget your wifi network? I had to do so in order to get the ethernet connection to work consistently but it has been rock solid ever since.

I did see that and yes, tried that too. I guess it was just the wifi AP I was using, because at the risk of jinxing it this has been solid since I switched to the powerline adapters.

 

[jjrandorin]

Let us know how it comes out.

As I said earlier in this thread, this subforum has a disproportional amount of people who would be interested in this type of discussion, and invested enough to care what the solution is.

 

[crackers8199]

Let us know how it comes out.

As I said earlier in this thread, this subforum has a disproportional amount of people who would be interested in this type of discussion, and invested enough to care what the solution is.

it has been just less than 24 hours, and i haven't been unable to connect to the TEG web UI once since switching to the powerline adapters. i'm not ready to declare it a complete success yet, but the early returns look very good.

 

[crackers8199]

i think we can declare the powerline adapters as a success. it has been solid for four days now, and i have been able to access it every single time i've tried. hasn't failed once.

 

[gpez]

i think we can declare the powerline adapters as a success. it has been solid for four days now, and i have been able to access it every single time i've tried. hasn't failed once.

This is good news!

Question: if you go off grid while the Powerwall is full and the local frequency goes up, does that screw up your powerline adapters?

I ask because the old TP Link adapters I had seemed to struggle when the local frequency was 62hz+...

 

[crackers8199]

This is good news!

Question: if you go off grid while the Powerwall is full and the local frequency goes up, does that screw up your powerline adapters?

I ask because the old TP Link adapters I had seemed to struggle when the local frequency was 62hz+...

good question, i haven't tried it yet. i'll test tomorrow once the powerwalls fill back up.

actually, nvm, i'll test it right now. i should have enough solar left to get them back up around 98% and get the solar to shut off...

 

[crackers8199]

or not. i'm at 98+ right now and it's still charging from solar. i guess i don't have *enough* solar to get them to turn off (the system probably knows from past charging that i won't get to 100% with what's left today so it's letting it keep going...at least that's my guess). will try again tomorrow.

 

[crackers8199]

This is good news!

Question: if you go off grid while the Powerwall is full and the local frequency goes up, does that screw up your powerline adapters?

I ask because the old TP Link adapters I had seemed to struggle when the local frequency was 62hz+...

didn't seem to be an issue for me...I noticed the grid code override was already in place to only bump my frequency to 62.5, and when I went off grid I had no issues still connecting to the gateway.

I didn't want to stay off grid very long because I want to be sending my excess solar back right now, but at least for the 10-15 mins I was off grid I didn't notice any issues.

 

[120v]

I had similar issues, and using

arp

allows devices that don't work to start working again. For my server that pulls metrics from the Solar Gateway, I have this run when network connectivity completes during boot up:

arp -s 192.168.1.140 cc:88:26:8e:2e:01

With the known IP and MAC address. I believe the command is similar for windows. On ubuntu, this command was put in a new file

/etc/network/if-up.d/001addpowerwallstaticroute

 

[crackers8199]

randomly I've started seeing the "connect powerwalls to wifi" message or whatever it is in the app again, but it doesn't seem to be dropping off the network at all. is this a new thing in a recent update?

 

[Redhill_qik]

I've had issues a few times

randomly I've started seeing the "connect powerwalls to wifi" message or whatever it is in the app again, but it doesn't seem to be dropping off the network at all. is this a new thing in a recent update?

I have had a few issues trying to directly access the TEG using its IP address (fixed IP from the router, WiFi connection), but with no indication from the app. When this happens to me I have found that the issues resolves after connecting directly to the TEG wifi SSID. I've never attempt to "pair" my phone with the TEG as it isn't clear what that actually does or why I would need to do it.

Your comment about any changes that might have happened with a recent app update made me go take a look at the release notes. In June the Powerwall 22.18 firmware has a line for Improved reliability for Gateway network connectivity. That is probably the 22.18.3 version and not the more recent 22.18.5 that was a pushed out to a few people and then withdrawn. I'm on 22.18.3 firmware and the app is 4.11.2.

 

[crackers8199]

I've had issues a few times

I have had a few issues trying to directly access the TEG using its IP address (fixed IP from the router, WiFi connection), but with no indication from the app. When this happens to me I have found that the issues resolves after connecting directly to the TEG wifi SSID. I've never attempt to "pair" my phone with the TEG as it isn't clear what that actually does or why I would need to do it.

Your comment about any changes that might have happened with a recent app update made me go take a look at the release notes. In June the Powerwall 22.18 firmware has a line for Improved reliability for Gateway network connectivity. That is probably the 22.18.3 version and not the more recent 22.18.5 that was a pushed out to a few people and then withdrawn. I'm on 22.18.3 firmware and the app is 4.11.2.

I have wifi disabled on the gateway and have for months. I have it hardwired to a powerline Ethernet adapter.

it was working fine until recently. actually, it's still working fine...was just disconcerting to see that message start popping up again. that used to show all the time before I finally gave up on wifi to begin with and hardwired it.

 

[crackers8199]

wanted to chime in with an update on this: the powerline adapters i've been using work flawlessly, UNLESS something (notably our christmas lights) are plugged into the same outlet as the powerline adapter and turned on (via a z-wave plug-in adapter). once the christmas lights turn on, the connection to the gateway from our local network goes kaput. interestingly enough, it does seem to still be able to communicate out to tesla, maybe via the separate 192.168.1.x network it creates for itself...but accessing it via my own subnet and its local address doesn't work as long as the lights are turned on. once they turn off, the connection is available again.

i don't know if this is the z-wave outlet, the christmas lights, or a combination of both...but it's definitely real. it happens every single time. lights on = connection unavailable, lights off = connection available again. without fail, like clockwork. i noticed connection issues around halloween, when we had other outdoor decorative lights installed, but didn't connect it to the powerwall issues until yesterday, when we put our christmas lights up for the season and i started to see the pattern.

that being said, unifi (which is all of my networking gear in the house except for my pfsense router) has greatly improved wireless meshing in recent updates - to the point where you can control it at the access point level, and even create a wireless bridge with each AP if you so desire. so, as a test, i took one of our three APs and moved it out to the garage, and enabled meshing on both it and the other AP that is nearest to it. that AP now has a wireless uplink to the other AP (and as such the rest of my network), and by plugging the gateway into the LAN port on the PoE adapter, it also has a link to the rest of the network via that wireless bridge.

everything works perfectly, including when my christmas lights are on...and i get roughly 75mbps up and down when i connect my computer via ethernet to that same wireless bridge (we have 500/500 fios). so, not perfect...but that's good enough for what i need it to do in the garage. good enough that i'm probably going to stick with this solution rather than hiring someone to do an ethernet drop to the garage. i'll just order another unifi wireless AP so that i can put my existing one back where we had it, and then leave it alone. the other advantage here is that the connection seems stable enough that i can probably add a switch in the garage as well, and have a strong enough, reliable enough connection to be able to occasionally hook other stuff up out there if need be.