@wordpress/env: Loopback requests fail

[noisysocks]

Describe the bug
WordPress isn't able to complete a loopback request when running under wp-env.

To reproduce

1. npx wp-env start
2. Navigate to http://localhost:8888/wp-admin/site-health.php

There's two errors related to this:

• The REST API encountered an error
• Your site could not complete a loopback request

Expected behavior
There should only be one critical error (the one about debug mode.)

Additional context
WordPress is trying to make a request to http://localhost:8888 from the underlying Docker container but this fails because Apache is listening on port 80. One solution would be to set up the docker-compose.yml config to map localhost to the host machine.

[tbclark3]

I am dealing with the same issue. Wordpress attempts to make a loopback request to its own hostname (not to localhost) on port 443 which fails because the container doesn't provide SSL. If there were a way to force loopback requests to use http instead of https, that might solve the problem, but I haven't found one yet.

[adrianduffell]

This issue seems to impact WP Cron for me. It looks like the scheduled events can't run because it uses loopback requests to create a background process.

I am able to work around this with the ALTERNATE_WP_CRON config.

// .wp.env.json

"config": {
  "ALTERNATE_WP_CRON": true
}

[glendaviesnz]

Two possible solutions to this that I have found are:

• Access wp-env instance through an external domain tunnel like https://ngrok.com/ in which case the callback will be made to that domain rather than localhost.
• Add a hosts entry on the local machine for 127.0.0.1 host.docker.internal and access wp-env on http://host.docker.internal:8888 - within a docker container host.docker.internal resolves to itself.

For both of the above solutions you will need to change your WP config/settings to accept that domain name instead of the default localhost that wp-env sets.

[miminari]

I reproduced this at @wordpress/[email protected] in M1 Mac

[jmichaelward]

This issue seems to impact WP Cron for me. It looks like the scheduled events can't run because it uses loopback requests to create a background process.

I am able to work around this with the ALTERNATE_WP_CRON config.

// .wp.env.json

"config": {
  "ALTERNATE_WP_CRON": true
}

@adrianduffell I'm wondering whether you're able to elaborate on your workaround for this a bit more? I've encountered this same issue while working on a plugin I'm developing that registers a custom cron task, and I've attempted to test the cron using the following commands:

• wp-env run cli bash - to shell into the container
• wp cron test <my_event_name> - to test the event.

This results in this error, whether or not ALTERNATE_WP_CRON is set to true in .wp-env.json:

Warning: The ALTERNATE_WP_CRON constant is set to true. WP-Cron spawning is not asynchronous.
Error: WP-Cron spawn failed with error: cURL error 7: Failed to connect to localhost port 8888 after 0 ms: Connection refused

It would be wonderful if we're able to find a solution for this issue. I've just started using wp-env for the first time and it's a terrific utility.

[ryelle]

This issue also prevents loading editor styles from plugins (or themes if you're not paying attention).

For example, I have a plugin that sets up some font files, and it's added to the editor by add_editor_style( plugins_url( 'style.css', __FILE__ ) );

When the editor styles are loaded, it tries to wp_remote_get the file contents, but it's prevented by "cURL error 7: Failed to connect to localhost port 8888: Connection refused".

I had also added the theme's editor styles with add_editor_style( get_template_directory_uri() . '/build/editor.css' ); causing the same issue, but that's easily fixed by switching to add_editor_style( '/build/editor.css' );

[Sam-Smyth]

@noisysocks It looks like this was never resolved and I'm running into this issue. Your solution of editing the docker-compose.yml file doesn't seem to work since wp-env start overwrites that file, right? That's what I'm finding.

[noisysocks]

Yeah, I was referring to making it so that wp-env writes a docker-compose.yml file that maps localhost to the host machine.

I don't use wp-env anymore (lately I've been trying https://ddev.com) so haven't been prioritising work on it. Feel free to open a PR in this repo containing a fix.

[mikey242]

I was having this issue too and one solution that I settled on was to change the default port number from 8888 to 80. This obviously won't work if you are using port 80 for another service but it is a relatively simple fix to implement.

You can either add "port": 80 to your .wp-env.json file or prepend your wp-env command like this: WP_ENV_PORT=80 wp-env start

[celtislab]

I created a plugin to prevent loopback request errors in WordPress environments that use Docker, such as wp-env.

localhost2host-docker-internal

Give it a try!

[rmorse]

Thanks @celtislab , I dropped this into my mu-plugins and its working great right now.

[BrianHenryIE]

Potentially, wp-env could script some changes to the Apache config as described here: docker-library/wordpress#611 (comment)

I tried to manually do it but it didn't seem to work

sed -i 's/:80/:80 *:8888/g' /etc/apache2/sites-enabled/000-default.conf
sed -i 's/80/80\nListen 8888/g' /etc/apache2/ports.conf

[collinr3]

I got it to work using _ as the delimiter instead of / ie
sed -i 's_80_80\nListen 8888_g' /etc/apache2/ports.conf