Gateway password resets (v4.1.3) #456
Comments
|
Hi @simonrb2000 - Creating an issue for this. Where are you resetting the password? On the Powerwall gateway itself? Can you give us a sample of the log file you see for pypowerwall? |
|
Thanks Jason, On the gateway IP. Prior to 4.1.3 it would need updating every couple of hours. It only needs reseting to log back into gateway as the Powerwall-Dashboard remains logged in. It's only when I restart Powerwall-Dashboard does it become a problem. Hope that makes sense? Basically, I can't log into the gateway as it comes up with bad credentials and I have to reset the password then redo the setup for Powerwall-Dashboard. If I'm not fussed about logging into the gateway then I can leave it for days and days and its only a problem when I restart Powerwall-Dashboard. Thanks again for opening an issue and apologise for not doing it myself |
|
Thanks for raising this issue @simonrb2000 - others may have the same problem. When you "reset" the password, are you using an old password to log in? If so, i wonder, does it work if you use the old password in Powerwall-Dashboard? |
|
No, I've been using a new one each time. I have to select forgotten password as an option and the last 5 characters from Gateway serial and then toggle the Powerwall switch. Only then do I need to change the password on Powerwall-dashboard |
|
Thanks. I'm sorry for so many question. I would like to understand the exact steps you are taking so we can see if there a way to make it "stick". It feels like this is a problem with the Gateway "forgetting". I'm guessing this is what you are doing - the https://10.0.1.99 being the IP address of the Powerwall:
After you enter a new password (and do the toggle or other instruction?)
|
|
That's exactly it. It only seems to happen when I'm running Powerwall-Dashboard. If I disable it then I don't have the issue.. can be a week later and the password would still work. I'm only putting 2 and 2 together and maybe not getting 4. V4.1.3 has helped but instead of everyday it's every few days it needs resetting now. |
|
Thanks. I'm not able to replicate. Is anyone else seeing this? It might be load from the dashboard, but I have 5 different Powerwall-Dashboard instances running all hitting my gateway all the time, and no issues. Did your system upgrade to Firmware version 24.4.0? |
|
Could it be that it’s asking for solar info and it doesn’t have it? Is there a way for me to stop it polling for solar info? Yes, I reported 24.4.0 to you 😁 although it did happen on the previous version. |
I apologize. My context windows is too small and I have too many concurrent threads / projects. 😊
It polls for these as defined in the file Powerwall-Dashboard/telegraf.conf Lines 101 to 109 in f6d829f You could remove the "strings" one as it is the only one that is exclusive for solar. The rest contain payload data for the powerwall, home and grid. I'm sure you already sent this at some point, but can you paste a sample list of your pypowerwall logs to see anything unusual? |
|
@simonrb2000 there is a new version of pypowerwall about to be released. It has some optimizations and refactoring. It may be interesting to see if it makes a difference for your system. You can give it a try: Edit the pypowerwall:
image: jasonacox/pypowerwall:0.8.0t51-beta3
container_name: pypowerwall
hostname: pypowerwall
restart: unless-stoppedRestart the stack: /compose-dash.sh up -d |
|
I am seeing this problem too for a few weeks. I didn't notice this issue until now. I updated to beta3 of pypowerwall and will report in if the reset problem goes away. |
|
In out preparation for the new pypowerwall release, we did find an issue with how the proxy caches the login session causing it to attempt to log in when pypowerwall restarts. I don't think this is your particular use case, but you could try the updated beta if you are interested: |
|
Sorry, I’ve been away and not had a chance to look at this, I’ll be back soon and I’ll update how it goes. Ta
On 1 Apr 2024, at 05:07, Jason Cox ***@***.***> wrote:
In out preparation for the new pypowerwall release, we did find an issue with how the proxy caches the login session causing it to attempt to log in when pypowerwall restarts. I don't think this is your particular use case, but you could try the updated beta if you are interested: jasonacox/pypowerwall:0.8.1t52-beta1
—
Reply to this email directly, view it on GitHub<#456 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AD2AEAGLTK326CUHYBVL7ITY3DMQLAVCNFSM6AAAAABFF3FXZKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMRZGEZDENBZGU>.
You are receiving this because you were mentioned.Message ID: ***@***.***>
|
|
Updated to |
|
That’s the same error I receive mostly. Occasionally it says bad credentials even though they are correct.
On 1 Apr 2024, at 16:01, Kevin Miller ***@***.***> wrote:
Updated to jasonacox/pypowerwall:0.8.1t52-beta1 today. When I restarted I noticed that the gateway had locked me out again (likely pre update). At the gateway web page I am receiving this error: API Limit Reached. So maybe for my situation pypowerwall is hammering the gateway too hard? Is there a way to access or turn on logging so I can retrieve extra logs for your use?
—
Reply to this email directly, view it on GitHub<#456 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AD2AEAFU4JEYKFLKS2M6GM3Y3FZEVAVCNFSM6AAAAABFF3FXZKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMRZHEYDAMRTGI>.
You are receiving this because you were mentioned.Message ID: ***@***.***>
|
|
Logging output to a file manually via |
|
By chance are you running any other application or scripts that access the Powerwall? |
|
I do not. Only Powerwall-Dashboard. My logs from yesterday were filled with this: |
|
Unfortunately, the error is expected if the Powerwall stops responding. This is very puzzling. The basic pypowerwall setup is designed to cache data to prevent heavy loads on the Powerwall. I can't replicate this even after having 5 instances of pypowerwall running. It would be good to get more debug information to see if there is any other indication on what is happening. If you want to do that: Edit [email protected]
PW_PASSWORD=password
PW_HOST=10.0.1.2
PW_TIMEZONE=America/Los_Angeles
PW_DEBUG=yes
PW_STYLE=grafana-darkRestart and watch logs... |
|
Could this be a powerwall vice powerwall + thing? I have 2 x Powerwall with Gateway 2.
What do you guys have?
On 3 Apr 2024, at 03:55, Jason Cox ***@***.***> wrote:
Unfortunately, the error is expected if the Powerwall stops responding.
This is very puzzling. The basic pypowerwall setup is designed to cache data to prevent heavy loads on the Powerwall. I can't replicate this even after having 5 instances of pypowerwall running. It would be good to get more debug information to see if there is any other indication on what is happening. If you want to do that:
Edit pypowerwall.env and turn PW_DEBUG=yes:
***@***.***
PW_PASSWORD=password
PW_HOST=10.0.1.2
PW_TIMEZONE=America/Los_Angeles
PW_DEBUG=yes
PW_STYLE=grafana-dark
Restart and watch logs...
/compose-dash.sh up -d; docker logs pypowerwall -f
—
Reply to this email directly, view it on GitHub<#456 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AD2AEAEK2DVR3IWPSURS643Y3NVRHAVCNFSM6AAAAABFF3FXZKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMZTGQZTSMJSGE>.
You are receiving this because you were mentioned.Message ID: ***@***.***>
|
|
I have a single Powerwall 2 and gateway. I restarted today (it had stopped working again) with debug on and am watching the logs. I thought it was interesting that there is a connection pool of size 3 connections (if you are avoiding stressing the gateway why hit it 3 ways at a time?) and they seem to churn a lot. Here are the logs from the startup of Pypowerwall and first 30 seconds: The empty responses seem worrisome. Also hitting every route on the gateway every 5 seconds seems aggressive. Is there a setting to dial this back? |
|
Powerwall had stopped working for my login. I reset the password via the gateway. What is interesting is I noticed that the contents of the cached AuthCookie in the
So it appears that even though the account password was cycled the auth cookie didn't need to be refreshed? Likely not related to the problem but very curious. |
|
Thanks @KevM - I have a Powerwall + and Powerwall 2. I agree about the concern about the empty response.
Yes, there is! You can set this in # Disable connection pooling - default is 15
PW_POOL_MAXSIZE=0
# You could increase the cache duration - default is 5s
PW_CACHE_EXPIRE=10Restart and watch logs...
Yes, that is true. It continues to honor the auth until the TTL is reached. Now, if you changed the password in the gateway portal but didn't update the password in |
|
Also, you may switch to this new beta version that adds additional DEBUG logging showing details on cache vs polling: |
|
Running new beta3 with the following .env I have two timezones in there for some reason. Maybe that is from the Will let you know if this fails and send logs. |
|
One day later and we are still good (no forced password reset). I think the lower connections, or the longer expiration may be helping. |
|
That's great news @KevM . I know you have debug on so it is probably filling up your disk fast. But it would be good for you to sample the logs to see if you see anything unusual. |
|
It has been running well now for 4 days. Longest I can remember. Here is a gist of a log sample. |
|
Thanks @KevM - I'm glad you found a fix! Part of me wants to know if it was PW_CACHE_EXPIRE or PW_POOL_MAXSIZE that fixed it, or both? Thanks for the update. Hopefully it will help others seeing the same thing. |
|
My guess it maybe a combination. Could be an API limit and the 5 second interval creeps outside of it. Or maybe the concurrent access triggers an internal limit. Of course, I’m not sure. I’m happy to try just one of the settings. Any suggestions?
…On Mon, Apr 8, 2024, at 10:02 PM, Jason Cox wrote:
Thanks @KevM <https://github.com/KevM> - I'm glad you found a fix! Part of me wants to know if it was PW_CACHE_EXPIRE or PW_POOL_MAXSIZE that fixed it, or both? Thanks for the update. Hopefully it will help others seeing the same thing.
—
Reply to this email directly, view it on GitHub <#456 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAAMHL63YLHGCPHLL73VSDY4NK5XAVCNFSM6AAAAABFF3FXZKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDANBUGA3DQOBWGE>.
You are receiving this because you were mentioned.Message ID: ***@***.***>
|
|
Not sure if I saw this same issue but I just this morning noticed that on my two instances of Powerwall-Dashboard neither had been collecting data for the last 3 days. The verify script show pypowerwall was down and the logs indicated an issue with authentication. Several days ago I had received notification in my Home Assistant Tesla integration that I needed to re-authenticate. I did so with no problem but never thought to check other applications to see if they needed me to re-authenticate. I do use the free iOS app Netzero on my iPad (highly recommend) and check it daily and it never has asked for me to authenticate. |
|
@billraff Did you need to reset your gateway password after the authentication issue? That was what was happening to me. My resolution so far was so slow pypowerwall down (30s) and use less (1) connections in the pool. |
|
I did go to the Gateway and it asked me to login. It did ask me to reset the password but honestly it seems I almost always have to do that there. It's pretty infrequent I log in to it. I'll keep an eye on things for a few days. |
|
Hi @billraff - thanks for the updates. What version of pypowerwall are you running (shows in verify.sh)? Several recent updates have been made to handle the newer Powerwall firmware that removed some API endpoints.
Yes, Netzero, like the Tesla App, uses the Tesla cloud endpoint to get data so it would not be impacted by the local authentication issues. You can switch pypowerwall into cloud mode but you will lose some of the data only available in local mode and the fidelity is less (the cloud caches data for longer periods of time). The Home Assistant integration is probably using local access similar to pypowerwall. |
|
I am running 4.2.1 on both my systems. It happened again yesterday where it appears I was locked out of the gateway. I noticed this because in Home Assistant there was a notification from the Powerwall integration that it needed fixing. It wanted me to provide credentials to the gateway. Before doing so I looked at my two Powerwall-Dashboards and saw neither had been collecting data for about an hour. I fixed the integration issue by providing gateway credentials and then using a browser went to the gateway and logged in there. It wanted me to reset the password which I did and both dashboards began working again. I've disabled the Tesla Powerwall integration (https://www.home-assistant.io/integrations/powerwall) and the Tesla Custom Integration that is installed through HACS. I'll watch for a week or so and then if all is well enable the Powerwall one to observe behavior. If all is well for another week then the custom integration. |
I'm having to reset my gateway password everyday or so. Says Bad Credentials. I have the latest version 4.1.3; did this version have a fix for this? Could the cause be it polling solar_powerwall when I don't have Powerwall+? Cheers
Originally posted by @simonrb2000 in #109 (reply in thread)
The text was updated successfully, but these errors were encountered: