Replies: 0
Had an issue today. I updated the Yoast SEO plugin (yes, different plugin, but bear with me, it’s not actually about Yoast), tested it on our staging site, worked fine, no conflicts, then I deployed to our production site. The site wouldn’t load, kept on timing out. I checked the logs, saw immediately references to
PHP Fatal error: require(): Failed opening required '/public_html/wp-content/plugins/wordpress-seo/vendor/composer/ClassLoader52.php' (include_path='.:/usr/share/php') in /public_html/wp-content/plugins/wordpress-seo/vendor/composer/autoload_real_52.php on line 10
I ended up deleting the Yoast plugin folder from the command line and the site loaded up again. I then looked further back into the logs and found:
PHP Warning: Redis::connect(): connect() failed: Connection refused in /public_html/wp-content/object-cache.php on line 436
So bear in mind that Yoast SEO works fine on our staging site (fresh — today — copy of production site’s database). SO it’s really not a problem with Yoast, that just happened to be the particular update I was deploying.
Is it possible that Redis was referencing plugin files that no longer existed? I’m somewhat fuzzy what Redis::connect(): connect() means.
But if so, what’s the appropriate way to deal with this? I use deployhq.com to deploy our codebase, and I can run SSH commands before and after the deployment.
Would this be a good idea:
wp redis stop
Run Deployment
wp redis start
?