# Setting up Bedrock-Ansible multisite with subdomains using the Wiki guide

**URL:** https://discourse.roots.io/t/setting-up-bedrock-ansible-multisite-with-subdomains-using-the-wiki-guide/3824
**Category:** trellis
**Created:** 2015-05-19T07:01:49Z
**Posts:** 3

## Post 1 by @wearecore — 2015-05-19T07:01:50Z

We have several Bedrock-Ansible / Bedrock sites in production, all good, no issues.

However, we seem to be hitting a wall with setting up a multisite using subdomains following the guide in the wiki - [GitHub - roots/trellis: WordPress LEMP stack with PHP 8.1, Composer, WP-CLI and more](https://github.com/roots/bedrock-ansible/wiki/Multisite)

We have not progressed beyond trying to get set up locally so our issues are related to development (local Vagrant setup) only. I’m sure we’ll have other issues when we reach staging / production but this thread is only about development / Vagrant setup.

After config (using the wiki guide) / Vagrant up, which goes smoothly, we log into WP and add a new site, again no issues. However, when hitting up the newly added subdomain we get a 342 error ‘no data received - ERR\_EMPTY\_RESPONSE’

Has anyone else experienced the same difficulties - is there something we are missing? We’ve followed the guide to the letter and the only ambiguity / point we can’t reconcile is from the Wiki docs:

> ensure the following values are placed somewhere in wp-config.php (or config/application.php if you’re using Bedrock) **before running the wordpress-sites role** :

We presume this to mean ‘before running Vagrant up’ or provisioning a remote for the first time - is that correct?

Anyone have any pointers? Anyone got this setup successfully yet?

---

## Post 2 by @wearecore — 2015-05-22T05:33:23Z

Doesn’t anyone have this setup and working? :worried:

Just for confirmation, we have followed notes in previous thread: [Recommended subdomain multisite nginx vhost configuration with new web/ layout](https://discourse.roots.io/t/recommended-subdomain-multisite-nginx-vhost-configuration-with-new-web-layout/1429/2)

Our nginx .conf looks like:

```
server {
    listen 80;

  server_name mu-test.blah *.mu-test.blah;
  access_log /srv/www/mu-test.blah/logs/mu-test.blah.access.log;
  error_log /srv/www/mu-test.blah/logs/mu-test.blah.error.log;

  root /srv/www/mu-test.blah/current/web;
  index index.php;

  charset utf-8;

  # See Virtualbox section at http://wiki.nginx.org/Pitfalls
  sendfile off;
       rewrite ^/(wp-.*.php)$ /wp/$1 last;
       rewrite ^/(wp-(content|admin|includes).*) /wp/$1 last;

 include wordpress.conf;
}
```

Still getting: ERR\_EMPTY\_RESPONSE after service nginx restart

What are we missing?

---

## Post 3 by @wearecore — 2015-06-23T20:01:32Z

My mistake was not doing:

service nginx reload

I had been restarting nginx but it wasn’t until I did reload that it picked up the new config. Schoolboy error.