OK. Thanks.
I'm going to see if I can figure this out with my container in Unraid.
Hi All,
Is there a fix for a person that is not that tech savy and using a Docker Server on Windows?
Hoping for a simple solution and instructions to update either the app or the docker container.
Thanks,
John
Look here for several options.
For those using esTile.groovy (https://raw.githubusercontent.com/thebearmay/hubitat/refs/heads/main/esTile.groovy) I had to update the code due to an integer overflow on calculating days between refreshes for those attempting to use 90 day refreshes.
yes looking for a patch for the docker container i use as well
Image:
tonesto7/echo-speaks-server:latest@sha256:c3410b4dadeb92072c3dccbef0e58a51874ac3f385b77b3a293813a62a2c2e01
Port forwarding:
8091 → 8091/TCP
Command:
node index.js
See my post 3 above this one. It links to another post with options. There are options for a modified version of Echo Speaks server and an option for Echo Whispers, as well as other options for just modding the ES app code in Hubitat.
the options in your post above are for modding es in hubitat not the docker container and in fact I was the first person to test and modify the refresh request to 90 days in ES on hubitat.
Read through fully. I reference another post in that one with options. The second option is for server. There's also @Bloodtick_Jones Echo Whispers mentioned and is in the post below mine in that thread.
This should make them easier to find:
thanks but looking at that appears to be not a replacment with the fix mentioned above to the original echo speaks docker server.. not sure exactly what it is?
there was mention of a actual fix to the original es server and i was looking for a docker link i can use with that fix intact on my qnap server.
looking at your code fix here ... I
i am not sure looking at the code fix you have that it actually does anything.. it seems to just instead of trying to actually get a new token from amazon just return the existing refresh token. My guess is just like our fix in the es hubitat image eventually the original token/cookie with still time out after say 60 90 120 whatever days and an error will still occur without user intervention..
granted it will get around the error we were seeing actually hitting amazon.. for a while in my estimation..
this unfortnately tends to follow my observations of ai.. it does not really understand the underlying issue/problem and many times proposes weird or incorrect solutions.. i think in this case to work around the error returned rather than figure out what changed with amazon and what really needs to be done (if at all possible) to correctly fix it!
i am not trying to criticize and do applaud your effort to fix this in my estimation the correct way vs the way i and others have worked around it with increasing the timeout in the hubitat hosted code.
is there a log that could answer any questions? if so which one?
What questions?
Your Hubitat logs should show info regarding Echo Speaks. Errors and warnings should show there.
Not sure how to copy or upload logs. I am new to this. This is most current screen shot but someone on forums setup echo speaks to work locally and I was waiting for a fix and said to use hero.. so set that up but apperentally messed up what he set up, but it did break before I went with the forum soultion and he restored my hubitat to a week prior but speak was still broken. He said was a amazon issue. Wait screen shot is from today with the the old...was working version so it shouldnt include hero serve
r, this should be hosted on the hubbitat.
If you are not using Heroku (not Hero), you need to bring up the web page for your Echo Speaks server and use that to log back into Amazon to refresh your Echo Speaks cookie. There is a current issue with something that changed on the Amazon side of things, that is not allowing for automatic cookie refreshes.
Your Echo Speaks server does not get hosted on the Hubitat hub itself. It either runs on Heroku or somewhere else local. It can be running as a docker container somewhere, directly on a device like a Raspberry Pi, etc. There are multiple options, so you need to know how your ES server is configured and what it's access page/URL is, if as you mention above, you are not running it on Heroku.
I have already applied a "work around" that allows me to change the update time. I have updated my time to 90 days. This seems to be working well (see screenshots).
However, is there a known fix for my configurations (Windows Docker Server) that allows automatic retrieval of the cookie without needing to log into Amazon/Alexa?
AFAIK there’s no way to retrieve the cookie without logging into your Amazon account.
No. You must have some way of logging in.
I haven't done anything with my cookie server some time now, but I just never seem to lose my cookie. Just how long can we go without a updating the cookie? I'm at 211 days now, and the update is scheduled for 159 days from now.
Everything is working so I haven't been touching it. I'm using a Rasp PI 3 with the non-docker option for the cookie server, which is on a UPS and it never reboots.
No one really knows at this point. I recall seeing someone go over a year. Things are at a fragile state lately with the way Amazon keep changing things. We are all fortunate to have various ways of keeping ES alive.
Both of your auto cookie refreshes are not working as indicated by the extended "Last Refresh" days you've both posted. This is a known issue of late for some. However, both of your instances of the ES app on your hubs still show that it is OK with the cookie state.
That being said, if everything is working, leave everything alone. Of late, there have been numerous reports where people's instances of Echo Speaks stop working entirely once their cookie expires. Some have even had issues logging back into Amazon to refresh their cookie.