I'm hoping someone can lead me in the right direction of how to track down this problem. I recently had HHVM installed on my server with PHP as a fallback. Since then I've been having issues in WP admin.

When I attempt to post or update a post my connection times out at 2m and in my NGINX error log I see this:

recv() failed (104: Connection reset by peer) while reading response header from upstream, client

What I can't figure out is why it's happening to me but not to my server management service or to my writer who have both tried and not had this problem.

I've done a traceroute and ping to my server and there's no latency issues showing up. I've tried different browsers, I've changed my DNS, and I've tried connecting through a VPN and I still have the same problem. I don't have any problems other than that.

This is a dedicated server with plenty of resources available and it only started after HHVM was installed. The admin panel became very slow until Redis was installed but that didn't help the issue of timing out. I didn't initially time out but now I can't post or update without timing out but the data does get saved to the DB.

Are there any other ways I can track down what the problem might be or any ideas on what the problem could be?


From the symptoms, i would say it's nginx timing out after waiting for a result from hhvm.

When this happens, the client gets an error, but the php proccess executiong that request never stops, thus you find the data saved on the DB.

Can you get some more verbosity from the php logs?

Another thing i remember from worpress is that the admin part a bit heavy on the DB.

BTW, here there is a very similar question:

nginx errors "recv() failed (104: Connection reset by peer) while reading response header from upstream"

  • Thanks for your input. I read that one earlier but I wasn't sure if it applied to my situation since I'm using HHVM. I don't think php logging is turned on and since NGINX & HHVM are new to me I'm not sure how to go about it. What I don't understand is why it only happens to me and not others. – DRK Nov 5 '15 at 19:05
  • You can setup hhvm loggin, check here for more info: github.com/facebook/hhvm/wiki/runtime-options Check the "logging" section, and if applicable, the MySQL slow query log. Enabling wordpress logging can help too: codex.wordpress.org/Debugging_in_WordPress – Fredi Nov 5 '15 at 19:08
  • Awesome thanks. I'll try some of these things and report back my results. – DRK Nov 5 '15 at 23:22
  • Just an update. Wordpress debug didn't show any issues and I found my hhvm error log but the file was too big to open due to a plugin issue that kept kicking back a ton of errors which was unrelated. Of course after I deleted the file and attempted to reproduce the timeout it stopped timing out but still takes almost 2 minutes to update a post. Now I'm just waiting for the problem to occur again. – DRK Nov 7 '15 at 18:11

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.