Announcement: HADasboard beta4 - Release Candidate!

Third party integrations HADashboard
62

Ok, guessing still doesn’t work with the dash? Might need someone more familiar with the rpi as I’m on an osx box. I had some issues with certs getting mine up so I thought I might be able to help but it looks like you’re running into some other issues. Here is a small synopsis of my struggles in case anyone stumbles across this.

osx
python3.6
openssl 1.0.2l

Originally had certificate verify failed.
[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed
Could get around that with the no verify option but I didn’t like that. Figured out it had to do with how python can’t get the root certs from the mac os anymore. Had to install certifi (it pulls the os certs into a file) but still had the issue. I had to manually add the missing letsencrypt link cert (#2 in the cert chain) into the CA bundle. Doing this I was able to get past the first certificate error.

Ran into another certificate error farther upstream (same exception as above though). Turns out I had the public key instead of the fullchain as my cert option in HASS. After fixing that I’m able to run appdaemon dash as normal.

appdaemon.yaml:

AppDaemon:
  logfile: STDOUT
  errorfile: STDERR
  threads: 10
  disable_apps: 1
HASS:
  ha_url: https://[myurl]:8123
  ha_key: [myHASSPASS]
HADashboard:
  dash_url: http://[myurl]:5050

This server runs hass/appdaemondash/web/dns etc. it’s my main externally accessible server. I have a dns entry that redirects the external url to the internal server’s ip so that I don’t have to switch mobile devices to use different wan/lan urls/ips.

Kind of a mind dump, but hopefully it helps someone since I was stuck for about two days figuring it out.

63

No, still same error

64

I hate to break it to you, Rene but for many (most?) of us, if the internet goes down, AD won’t matter as a lot of HA will fail as well.

65

Whew, I had to go back and read Rene’s original email to get that quote in context. It had me really scared for a minute. The only reason AD would quit working is if it used an external link to get to HA. As long as the connection is on local network, everything keeps working except for components that use cloud based services. Right.

66

right.
if AD connect to HA locally it will keep working even without internet.

@rpitera do you need internet for your lights, your garage door, your camera’s, your sensors, etc.?
HA shouldnt fail if internet is down, just some functions that are on the net should fail.
everything else should work as normal.

weather prediction, RSS feeds, yeah they break down with the internet.
but if you speak about lights, security, camerás you want them to keep running and you want to be able to access them even when the internet is down.

and that is all possible when you connect AD to HA locally and when you open HA with a local link when you are at home.

67

Just for giggles, I tried to switch from using my DDNS URL to http: and https: plus the IP both with an without a port.\

In all cases, nothing worked until I went back to DDNS. Like @anon35356645, this is the only way I can reach my HA instance.

68

that is strange and not really good. somehow the HA server gets blocked then.
i have NGINX in front of HA and i connect to HA with http://ip:port

69

Well, it was good for me but now I can’t connect at all. And I have everything in the configs set back to what it was originally. I just removed the entry suggested above from my host file and am rebooting again in hopes that things will go back to where they were.

I should have just stuck to what was working instead of comparing my setup to yours. That was my first mistake and I should have known better.

UPDATE: Removing the host entry and rebooting brought me back to normal again. Thank God.

1 Like
70

Removing the host entry and rebooting brought me back to normal again

Normal is boring.

2 Likes
71

Were there any changes to the base Appdaemon libraries (the old non-web related stuff) with this release?

72

When I use https://xxx.duchdns.org:8123 in ha_url, then I get this errror

TypeError: cafile, capath and cadata cannot be all omitted

When I use https://xxx.duchdns.org in ha_url, then I get this error

ConnectionRefusedError: [Errno 111] Connection refused

73

i hope you made a typo twice?
because it is duck and not duch :wink:

i think that your latest changes made things worse.

sorry, but i have no other suggestions at this moment. encryption is really making things very complicated sometimes.

74

It’s odd that you guys have to bounce off the external dns servers and then come back in through from the WAN. I just double checked using nslookup and when inside my network my url resolves to my internal ip using my internal dns server. From outside my LAN it resolves to my outside ip as it should.

75

Last 2 suggestions as I also am out of ideas…

  1. Try removing the password/https from HASS and see if you can get it to connect using http (to prove it out)
  2. Try running appdaemon with --commtype SSE flag enabled. I had issues with websockets and cert issues with lets encrypt until i imported the missing CA in the link as I mentioned above. Before I did that I was able to get it up and running using the SSE flag.
76

How you do number 2??

77

appdaemon -c conf --commtype SSE [-D DEBUG]

I assume SSE stands for server sent events, but I have no idea the repercussions from using one over the other in this case…

78

like this?

pi@hassbian:~/appdaemon_dashboard/appdaemon $ python3 -m appdaemon.appdaemon -c /home/pi/appdaemon_dashboard/appdaemon/conf -commtype SSE -D DEBUG
usage: appdaemon.py [-h] [-c CONFIG] [-p PIDFILE] [-t TICK] [-s STARTTIME]
                    [-e ENDTIME] [-i INTERVAL]
                    [-D {DEBUG,INFO,WARNING,ERROR,CRITICAL}] [-v]
                    [--commtype {SSE,WEBSOCKETS}] [--convertcfg] [-d]
appdaemon.py: error: unrecognized arguments: SSE
pi@hassbian:~/appdaemon_dashboard/appdaemon $
79

I’ve got some problem with my camera widget, I’m generating plot of my temperature every 5 minutes, external and internal on my server http://45.55.16.88/Solar_Temp_Last12HourNegatif.png

My Hass conf have this:

The token is changing in Home Assistant and the widget get blank…
Would it be possible to simply reference the entity in state of the entity_picture ?
My Dash:

80

I wonder if it has anything to do with the base_url: setting in the http: component.

I just commented it out and I’m restarting HA to see what effect it has.

81

Two – next to commtype