I noticed that Google Chrome browser caches these redirects, even after clearing the cache and other data, it has to be actually restarted. For testing redirects you could use a command line tool without preserving any state between requests. I use Mozilla Firefox Developer edition for testing complex redirects because it doesn’t cache them until restart.
@strarsis is right; Chrome aggressively and annoyingly caches redirects. But if you’re curling the non-www address and it’s showing a 301 redirect then something still isn’t right: the client request was to remove the non-www spelling from the site entirely right?
Yes, that was their request. I already checked, it’s not just Chrome.
Doing curl -I site.com returns HTTP/1.1 301 Moved Permanently
The servers are on their side, and they say they have no redirect configured on their end. That’s why I’m wondering if there is anything I’m overlooking with my Trellis setup.