This is the strangest bug I ever seen.

I'll start a new job at a local magazines editor in a week, and today I passed by to submit some papers they asked me to provide.

When I got there, they told me about this bug they found about recently:

When accessing some links on their WordPress-on-Nginx site from Google's Search Results Page, said links render a blank page, with no HTML at all and 200 status code.

When refreshing said pages, nothing happens — they're blank still.

BUT if selecting the URL on the address bar and hitting ’Enter’ with no modification what so ever, this would make the page load as it should. This behavior is consistent across different browsers, computers/smartphones and networks (DSL and cellular).

Accessing the articles directly from the site's homepage show no sign of blank pages at all.

What's more irritating, is that not all pages are affected. And I failed to detect a pattern among those who are (or aren't).

To make things harder, they have no direct access to the server, so I couldn't check any log files (but the company taking care of their hosting will try to send logs and config files by the next week).

So my question is: what's going on here? And if it's hard to tell, at least, where can I start to look?

  • Do you have a URL we can look at? – Mike Jan 24 '16 at 5:48
  • Unfortunately, I wasn't granted permission to post links here :/ – ahmed Jan 25 '16 at 21:02

It could be the infamous WordPress "White Screen of Death" which occurs when WordPress experiences a fatal error with debugging features disabled, such as to not dump potentially sensitive information to end-users. Refrain from enabling WP_DEBUG in live environments in the course of troubleshooting the issue as it may expose an installation to additional vulnerabilities. The "White Screen of Death" can be delivered with HTTP 200 depending on the circumstances that created it.

If not a general fatal error, it sounds as though something in the installation is acting on the HTTP Referrer header, for whatever reason. It may be worth a try to investigate/refresh the settings pertaining to any SEO and caching plugins, as well as flushing any caches.

I feel your company should consider finding a new host - it's pretty bizarre that they don't provide their clients direct access to any logs, and worse still that their best support effort seems to amount to "We'll try to send you those logs and configs next week." Do they expect your company to figure out the issue on your own and let them know if they need to change a configuration setting?

Checking the Nginx configuration for the site would also be a good idea, but is likely impossible as well given the host's restrictions. Review your company's support contract - given the hosting company's relative level of control, it seems likely that they're contractually obligated to address the problem, or to provide more hands-on assistance in troubleshooting it beyond sending you some files at their leisure.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.