Well, despite what I said yesterday, today was actually that super fun monthly event known as website stats checking day. So I was mainly just doing that. Though in the evening I watched Prometheus with Billesden.
When checking my website error logs, I found three errors that looked like potential true errors. One was to do with an undefined index when posting to xml-rpc on a WP site. I don't use this myself, so it would have been a spam bot / hacker request. But still, there shouldn't be an undefined index. When I checked the particular site, the WP version was out of date. So I just updated WP, and I'll wait and see if the same error appears in the logs next month.
The next error was PHP Fatal error: Call to undefined function get_header() in /wp-content/themes/twentyten/404.php on line 10
. According to this post: How to fix the “PHP Fatal error: Call to undefined function get_header()” error in WordPress, the error is when the file is called directly in the browser, since WP is not loaded when calling the theme file directly. That makes sense, and also means it's an error not worth fixing in my opinion. Only hackerbots are likely to be calling that file directly.
The final error was an upstream response is buffered to a temporary file (file name) while reading upstream
. There's a good article here: Nginx FastCGI response buffer sizes. I've removed the logs from the web server and my dev environment, so I can't run the average / max request size commands they suggest in their article, so I'll have to do that next month. However, I'm fairly confident that I have the buffer size set to a sensible amount already, as the upstream response buffer error messages were all for /feed/
, which is a much larger size (the last 10 posts in full I think) than any other page on the site.
No comments:
Post a Comment