500 or ERR_TOO_MANY_REDIRECTS when using *.lndo.site URL

[mattmacneil]

Hello,

I am hoping for a resolution to a problem that has been persisting for a long time now, for me and my coworkers on multiple devices - Multiple Macs (mostly newer M1/M2) but also on a laptop running Windows with WSL. I have only used this recipe for WordPress, so unsure if it applies to Drupal as well. I have used the normal Pantheon WordPress upstream as well as Composer-managed WP from Pantheon and the result is the same on both.

No matter how may times I attempt to fix via lando start or stop / lando destroy / lando rebuild; or restore Docker to default settings or reinstall it completely, the issue always returns. Either the website or wp-admin dashboard (or both) will not load and only return the "white screen of death" (but a browser error message, not a WP error screen). Sometimes a cache/cookie clear-out will fix it, but oftentimes it will not. I've used the Lando-recommended version of Docker as well as the latest version and the result is always the same in the end. All browsers affected in my experience.

I have tried just about every solution I've found from other open issues and posts on outside support forums. Some of these things work for a few days and may get marked as "resolved", but they always return for me after a while.

Unfortunately I'm not super knowledgeable in advanced Docker config, so the best I can do is provide some possible starting points, and hope that someone with skills in this area might recognize something they can fully diagnose.

1 - appserver_1 | NOTICE: PHP message: PHP Warning: Undefined array key "host" in /app/web/wp/wp-includes/canonical.php on line (689|690|698|701). This series of ~4 notices (one per each line) is written to the error log every time the page is refreshed while returning a white "ERR_TOO_MANY_REDIRECTS" screen. Interestingly (I think) this is not the case when refreshing a working page after a fresh rebuild. It seems to only happen after the WSOD has appeared.

2 - Related cookies "capped" at ~36. When clearing site data to get the site working again, chrome will nearly always say 36 cookies (maybe plus or minus 1 or 2?) associated with [sitename].lndo.site. Sometimes I will do a refresh that breaks causes the site to not load, and there will instantly be 36 cookies. Could whatever function that logs the user in or sets that logged in cookie be stuck in a loop as it's set? Is there any significance to the 3 dozen upper limit?

3 - 1000kb cap on container [sitename]_appserver_1 - this is a bit of a stretch but I'm not sure if there's any unintended limitation on that container that can cause an issue, some other containers hold much larger amounts of data yet *_appserver_1 is always at that nice round 1000kb and never anything else. Is this how it is supposed to appear?

I hope that this is somehow helpful in tracking down the issues. If there is any more testing I can do to assist in a diagnosis please let me know. Thanks!

[mattmacneil]

Removed: this additional comment referred to an unrelated issue caused by a missing .env-local file.

[webidextrousdev]

I'm having the same problem. Hoping to get a resolution with it as well.