Posted: May 28, 2015 at 3:17 pm
For no apparent reason my website started to show 500 internal server errors nearly all the time. The logs (at the point when the website always gave 500 errors) used to have this: (11)Resource temporarily unavailable: FastCGI: failed to connect to server “…cgi-local/php5.fcgi”: connect() failed … FastCGI: incomplete headers (0 bytes) received from server “cgi-local/php5.fcgi” The hosting company said: “A local request, like a cron job maybe, is hitting the site frequently and generating a large number of fcgi errors/timeouts to the point where a new connection may not have a good process available to service the request. I have reset the fcgi processes. I would try and run those jobs manually to see if you get an error. … the requests I’m seeing are coming from the … server itself. One appears to be a wordpress cron.” Resetting the fcgi processes made the website work (or seem to, at first glance anyway) temporarily. I wonder if this is related to EE4 and in particular this item in the changelog for EE4.6.25: “Don’t generate IPN crons for locked TXNs unless payment method uses a “true” IPN in a separate request”. I would upgrade but can’t log in. To no apparent avail, I have taken the following steps: using a real cron job instead of WP’s default behaviour, increasing PHP_FCGI_MAX_REQUESTS from 50 to 5000, and preventing another website from using the RSS feed. The error logs now show this instead: FastCGI: comm with (dynamic) server “…php5.fcgi” aborted: (first read) idle timeout (180 sec) … FastCGI: incomplete headers (0 bytes) received from server “…php5.fcgi” Occasionally the website does display, but I can’t log in to WP, so can’t upgrade EE (or can I?) Using: EE4.6.24 WP4.2.1 |
|
Hi Jonathan, The “Don’t generate IPN crons for locked TXNs unless payment method uses a “true” IPN in a separate request” fix was added to fix a potential issue with duplicate emails going out for a registration when paying via the Mijireh gateway (one email could be triggered by visiting the payment page and another by the IPN). It would not hurt to update, but it may not fix the issue you’re seeing on your site. Along with that, it turns out you can update via FTP. Or, you can also manually deactivate Event Espresso by renaming its plugin folder (also via FTP), then follow these steps: https://eventespresso.com/wiki/updating-event-espresso/#update_manually |
|
Thanks for your detailed response. Turns it out was because Wordfence updated itself (and presumably there was some sort of conflict) so I just disabled it… |
|
You’re welcome. |
|
The support post ‘500 error – related to item on EE4 changelog?’ is closed to new replies.
Have a question about this support post? Create a new support post in our support forums and include a link to this existing support post so we can help you.