Home of the Squeezebox™ & Transporter® network music players.
Page 4 of 27 FirstFirst ... 2345614 ... LastLast
Results 31 to 40 of 267
  1. #31
    i cant get the lite to find any players on my system

    the tunnel is up and the full blown version works
    the ngrok inspection window shows the query found 20 players, but none show up in a browser or the alexa app - ive logged out of both, relogged in and searched again, it always shows 20 players found in ngrok but nothing in the smart home section of alexa

    any ideas?Name:  Capture.JPG
Views: 250
Size:  55.0 KB

  2. #32
    Senior Member philchillbill's Avatar
    Join Date
    Jan 2019
    Location
    The Netherlands
    Posts
    793

    [Announce] LMS-lite certified Smart Home skill for Alexa

    Quote Originally Posted by dunk View Post
    i cant get the lite to find any players on my system

    the tunnel is up and the full blown version works
    the ngrok inspection window shows the query found 20 players, but none show up in a browser or the alexa app - ive logged out of both, relogged in and searched again, it always shows 20 players found in ngrok but nothing in the smart home section of alexa

    any ideas?Name:  Capture.JPG
Views: 250
Size:  55.0 KB
    That’s likely somehow related to your Amazon account. Forgetting all your discovered devices and then re-discovering can help. So can logging in and out of your account in the app. The full blown skill stores it’s own data in DynamoDB and never runs out of space. Amazon has restricted space for smart home under their control so they pull the strings.

    I saw in the logs that somebody discovered 20 players (congrats!). The JSON returned to Amazon looks fully correct. But it’s then up to your account to do something with that data. The thing about smart home skills is that the returned JSON is never commented upon by Amazon in the logs - you hand it off to them and it's a black hole with no feedback. I kid you not.

    EDIT: Just found a mention that a smart home response can be max 24KB of JSON, which you might be hitting with 20 players. Can you maybe decouple a few from LMS and re-discover and see if that works?
    Last edited by philchillbill; 2020-01-17 at 16:00.

  3. #33
    thanks for the comments

    ive tried the following:

    logged out of the account - logged in on another computer
    i removed 2 of the players and searched - nothing found at all
    forgot all devices and searched - just the existing devices but no LMS players
    went down another 2 players so 16 now - nothing discovered....

    Last edited by dunk; 2020-01-18 at 03:00.

  4. #34
    Senior Member philchillbill's Avatar
    Join Date
    Jan 2019
    Location
    The Netherlands
    Posts
    793
    Quote Originally Posted by dunk View Post
    thanks for the comments

    ive tried the following:

    logged out of the account - logged in on another computer
    i removed 2 of the players and searched - nothing found at all
    forgot all devices and searched - just the existing devices but no LMS players

    Do any of your players fake a MAC address? Might be a collision. Or any accented characters in the player names?

  5. #35
    no faked mac addresses that i know of and nothing but characters and -

    cant be the size of the json, with 16 players it returns 4861 bytes according to the ngrok :4040 portal

    8 of the players are copies of squeezelite though

  6. #36
    Senior Member philchillbill's Avatar
    Join Date
    Jan 2019
    Location
    The Netherlands
    Posts
    793
    Quote Originally Posted by dunk View Post
    no faked mac addresses that i know of and nothing but characters and -

    cant be the size of the json, with 16 players it returns 4861 bytes according to the ngrok :4040 portal

    8 of the players are copies of squeezelite though
    The discovery JSON I return is much much bigger. But I have other skills that return way more data than yours and that works fine, so I doubt if the length is the problem. In retrospect, that reference to 24KB was likely for responses other than discovery responses.

    What locale are you setup for with Amazon?

    EDIT: I think it's related to duplicated endpointIDs being caused by your MAC addresses. You have several MusicServer4Lox-Player types which seem to be spoofing a MAC address. If you decouple those 8 players and rediscover, does it work then? If It does, I can do a simple code change that will prevent this being an issue moving forward, but that would require everybody to rediscover. Not a big deal as no unlink/link needed, just saying 'Discover Devices'. Let me know.
    Last edited by philchillbill; 2020-01-20 at 00:35.

  7. #37
    Senior Member philchillbill's Avatar
    Join Date
    Jan 2019
    Location
    The Netherlands
    Posts
    793
    There was an interesting error in the logs for a US user this morning: 402 Payment Required. It was returned by ngrok and apparently means that a subscription to ngrok needs extending (I guess somebody with a paid tunnel was not on auto-renewal). There is of course no premeditated Amazon smart home error message for this kind of thing, so whoever it was just got the skill's standard error message about a connectivity error (which it clearly was not). Anyway, one more thing to think about and add to your probing checklist if you ever get an error response from the skill. Looking at the ngrok console at :4040 to debug the issue would of course reveal the 402 so please remember to do that if you have issues and are using ngrok with the skill.

  8. #38
    Junior Member
    Join Date
    Feb 2019
    Posts
    24
    Hi,
    I am failing installing LMS-Lite skills to my French Alexa :

    Skill setup is OK : "Your LMS server (v7.9.2) responded correctly. Go ahead and submit!"
    But after submitting : layla.amazon.com issues an <InternalFailure/>

    Using ngrok. PiCorePlayer v6 on RPi4. Alexa Echo v2. Freebox (internet ADSL router). My IP is xxx.xxx.xxx.165.

    I tried with and without -auth "userassword".
    I tried with region=us and region=eu.
    But to me ngrok setup is Ok, as I can access my LMS from outside using the ####.ngrok.io address.
    I tried renaming my Squeezebox to 'Salon' (standard room name).

    Any idea to where to look at next?

  9. #39
    Senior Member philchillbill's Avatar
    Join Date
    Jan 2019
    Location
    The Netherlands
    Posts
    793
    Quote Originally Posted by Ludopi View Post
    Hi,
    I am failing installing LMS-Lite skills to my French Alexa :

    Skill setup is OK : "Your LMS server (v7.9.2) responded correctly. Go ahead and submit!"
    But after submitting : layla.amazon.com issues an <InternalFailure/>

    Using ngrok. PiCorePlayer v6 on RPi4. Alexa Echo v2. Freebox (internet ADSL router). My IP is xxx.xxx.xxx.165.

    I tried with and without -auth "userassword".
    I tried with region=us and region=eu.
    But to me ngrok setup is Ok, as I can access my LMS from outside using the ####.ngrok.io address.
    I tried renaming my Squeezebox to 'Salon' (standard room name).

    Any idea to where to look at next?
    Amazon is currently investigating why some people get <InternalFailure/>. It's basically a 500 internal server error from the Amazon account linking server. You're doing nothing wrong and neither is the skill - it's a bug elsewhere at Amazon. I can see your linking attempts in the logs and the oauth tokens are being properly populated so it all looks just fine. But the automatic device discovery which immediately follows linking is not even being triggered, so my guess is that Amazon's internal error is in that direction. Are you trying this from the Alexa app or from the web console? If the app, log out and back in to your account and retry.

    In fact, if you just force-kill the app or close the web session at the point where you see <InternalFailure/>and verbally ask Alexa to discover devices, there's a chance that they will actually discover OK and that it only looks like something went wrong because of that error message. As I said, in the logs it all looks fine. Maybe you are actually linking OK and it's just Amazon's usual 'You are now account-linked" message that's not returned from the bowels of AWS. Would be worth trying...
    Last edited by philchillbill; 2020-01-25 at 00:15.

  10. #40
    Junior Member
    Join Date
    Feb 2019
    Posts
    24
    Quote Originally Posted by philchillbill View Post
    Amazon is currently investigating why some people get <InternalFailure/>. It's basically a 500 internal server error from the Amazon account linking server. You're doing nothing wrong and neither is the skill - it's a bug elsewhere at Amazon. I can see your linking attempts in the logs and the oauth tokens are being properly populated so it all looks just fine. But the automatic device discovery which immediately follows linking is not even being triggered, so my guess is that Amazon's internal error is in that direction. Are you trying this from the Alexa app or from the web console? If the app, log out and back in to your account and retry.

    In fact, if you just force-kill the app or close the web session at the point where you see <InternalFailure/>and verbally ask Alexa to discover devices, there's a chance that they will actually discover OK and that it only looks like something went wrong because of that error message. As I said, in the logs it all looks fine. Maybe you are actually linking OK and it's just Amazon's usual 'You are now account-linked" message that's not returned from the bowels of AWS. Would be worth trying...
    Thank you so much for this detail answer.
    So, I am not crazy. This kept me pretty busy late at night trying everything.
    I only tried through the Alexa app. on Android (Xiaomi Redmi 3S).
    Will soon try what you say and let you know.
    But your skill never became Activated in the mobile Alexa app if that matters.

    EnvoyÚ de mon Redmi 3S en utilisant Tapatalk

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •