Could not check origin Phoenix LiveView

When accessing the page she tries to load lasts about 10 seconds and reloads again and stays in that loop without stopping. When accessing the server I saw that at each attempt it shows this message in the log:

[error] Could not check origin for Phoenix.Socket transport.
Origin of the request: http://www.autoplacasma.com.br
This happens when you are attempting a socket connection to
a different host than the one configured in your config/
files. For example, in development the host is configured
to “localhost” but you may be trying to access it from
“127.0.0.1”. To fix this issue, you may either:

  1. update [url: [host: …]] to your actual host in the
    config file for your current environment (recommended)
  2. pass the :check_origin option when configuring your
    endpoint or when configuring the transport in your
    UserSocket module, explicitly outlining which origins
    are allowed:
    check_origin: [“https://example.com”, “//another.com:888”, “//other.com”]

prod.exs

config :autoplacas, AutoplacasWeb.Endpoint,
  http: [port: 80],
  url: [host: "autoplacasma.com.br", port: 80],
  check_origin: ["//autoplacasma.com.br"],
  code_reloader: false,
  cache_static_manifest: "priv/static/cache_manifest.json"

releases.exs

config :autoplacas, AutoplacasWeb.Endpoint,
  http: [:inet6, port: 80],
  secret_key_base: "mysecret",
  server: true,
  pubsub_server: Autoplacas.PubSub,
  live_view: [signing_salt: "mysalt"]

Have you tried adding this

to releases.exs?

Is it already hosted at autoplacasma.com.br or is it in your local host?

Are you running it with mix phx.server or are you running the release?

It is in production. It was generated via releases. Hosted on autoplacas.com.br

I made this change and keeps giving the same message

This may seem obvious but which command are you using to run your app?

Service (I don’t use nginx):

[Unit]
Description=Phoenix server for DeployPhoenix app
After=network.target

[Service]
User=root
Group=root
Restart=on-failure
RestartSec=5
WorkingDirectory=/home/phoenix-autoplacas
ExecStart=/home/phoenix-autoplacas/_build/prod/rel/autoplacas/bin/autoplacas start
ExecStop=/home/phoenix-autoplacas/_build/prod/rel/autoplacas/bin/autoplacas stop

[Install]
WantedBy=multi-user.target

releases.exs

config :autoplacas, AutoplacasWeb.Endpoint,
  http: [:inet6, port: 80],
  url: [host: "autoplacasma.com.br", port: 80],
  secret_key_base: "mysecret",
  server: true,
  check_origin: ["//autoplacasma.com.br"],
  pubsub_server: Autoplacas.PubSub,
  live_view: [signing_salt: "mysalt"]

Comparing with my configuration in releases.exs I found the following differences

  1. I do not have the check_origin key
  2. I do not specify the port in the url key
  url: [host: "autoplacasma.com.br"],

instead of

  url: [host: "autoplacasma.com.br", port: 80],

Can I ask what check_origin actually is and what does it do?

I might be totally off track, but I think yours is the key question. Because I was fooled as well I’d like to elaborate:

From doc:

`:check_origin` - if the transport should check the origin of requests 
when the `origin` header is present

Now the “origin” header from first random doc:

indicates the origin (scheme, hostname, and port) that **caused** the request

Caused… which is the website itself you connected to in the first place (ah-ah moment to me… this explains why I find domain names in the “check_origin”!)

BUT assuming you have a config such this

config :mygoodservice, MyGoodService.Endpoint,
  https: [
    port: 8443,
    ...
  ],
  url: [host: "mygoodservice.com"],

What is the proper origin? You are conveniently told that by the console error (which I got with the default check_origin: true config):

[error] Could not check origin for Phoenix.Socket transport.
Origin of the request: https://mygoodservice.com:8443

I suspect that just letting the default value true is not actually defaulting to the origin you are prompted about in the console :man_shrugging: … (maybe it’s forgetting the port? or I’m doing bad somwhere else… dunno).

Anyway setting the whole exact string reported by the console solved for me!
Also: if you are troubleshooting start/stop/restarting your service, I’d close the tab and open a new one… for what is worth I’ve seen strange things happening on firefox (which might well be consequence of me doing this late night).

Cheers

1 Like

I’m late to the game, but using Phoenix v1.7.11 on Fly.io the following showed up in my logs:

This happens when you are attempting a socket connection to a different host than the one configured in your config/files. For example, in development the host is configured
to "localhost" but you may be trying to access it from "127.0.0.1". To fix this issue, you may either:
1. update [url: [host: ...]] to your actual host in the config file for your current environment (recommended)
2. pass the :check_origin option when configuring your endpoint or when configuring the transport in your UserSocket module, explicitly outlining which origins are allowed:
check_origin: ["https://example.com", "//another.com:888", "//other.com"]

I added check_origin: ["https://my_domain.com"] and it worked.