Non-response

[drozas]

This is the same problem that affects the ERIE site (CRESS-Surrey/erie_website_code#31). Indeed, when it happens, both sites are not reachable. I have been able to minimise its impact by disabling cron in Drupal in both sites. When it happens (very rarely - in ERIE only happened once in the last 3 months approximately), it is solved by restarting Apache as in CRESS-Surrey/erie_website_code#31 (comment) It is definitely something related to the server configuration (I presume Apache's one), since this is not affecting any other servers with the same sites (with exactly the same code and Drupal configuration), and affects only Drupal 8 sites. I will keep on looking at possible sources of this error with regard to the server configuration.

[drozas]

Check https://www.drupal.org/node/553430

[drozas]

MySQL servers? https://forum.mediatemple.net/topic/6558-504-gateway-timeout-errors-for-drupal-site/

[drozas]

http://drupal.stackexchange.com/questions/196929/timeout-in-drupal-8-sites-perhaps-related-to-cron

[drozas]

https://www.drupal.org/node/2700393

[drozas]

A couple of possible configurations regarding this:

• https://www.drupal.org/node/1154448#comment-8951471 (disabling it)
• https://www.drupal.org/node/1154448#comment-10373811 (increasing it)

On 05/04/16 17:47, David Rozas wrote:

Thanks Nigel,

Should we try this?

"""
I have some experience with this It means that APC is out of memory. You can try increasing APCs memory. In my case APC was caching a bunch of stuff I didn't really want it to (a sandbox and staging site - which were taking up the memory). I ameliorated the problem by telling APC to not cache that stuff and monitoring APCs memory and clearing the cache when it was getting full.

In the end, the fix was upgrading PHP to 5.5 which has it's own opcode cache (if I remember correctly).

"""
(https://www.drupal.org/node/2700393#comment-11040843)

[drozas]

No incidences after a week :-)
Keeping on tracking...

[drozas]

Create cron task to reboot automatically every day at 4AM

[drozas]

Checking http://stackoverflow.com/questions/28498400/restart-apache-service-automatically-using-cron-12am-daily

[drozas]

drozas: stop and start apache to solve cache issue with Drupal sites (restarting does not produce the same effect)

0 4 * * * /opt/local/etc/LaunchDaemons/org.macports.apache2/apache2.wrapper stop; /opt/local/etc/LaunchDaemons/org.macports.apache2/apache2.wrapper start

[drozas]

Check results...

[drozas]

Reported again on 02/02/2017

[drozas]

Solved by manually restarting Apache

[drozas]

Solved after Nigel remove old Joomla sites! :-)

[drozas]

Seems to be happening again (reported by Julia) with new CECAN site

[drozas]

It seemed this was done to the simian site. Nigel has put it down again, so I will mark this as closed