| [15:24:57] | * helmo has quit (Ping timeout: 240 seconds) |
| [17:18:39] | * orangey has quit (Ping timeout: 248 seconds) |
| [17:23:53] | * orangey has joined #aegir |
| [17:26:10] | * colan has quit (Ping timeout: 264 seconds) |
| [17:26:44] | * colan has joined #aegir |
| [17:27:22] | * rajnishkumar[m] has quit (Ping timeout: 264 seconds) |
| [17:27:47] | * rajnishkumar[m] has joined #aegir |
| [17:28:05] | * orangey has quit (Ping timeout: 240 seconds) |
| [17:29:43] | * orangey has joined #aegir |
| [17:33:34] | * helmo has joined #aegir |
| [17:34:10] | * orangey has quit (Ping timeout: 260 seconds) |
| [17:36:39] | * orangey has joined #aegir |
| [19:16:44] | * reaper013 has joined #aegir |
| [19:42:49] | * ybabel has joined #aegir |
| [20:28:50] | * oluabbeys has joined #aegir |
| [20:56:23] | * oluabbeys has quit (Quit: oluabbeys) |
| [21:04:34] | * reaper013 has quit (Ping timeout: 260 seconds) |
| [21:27:21] | <helmo42[m]> | I granted Antonio Barrera(antonioG4) the developer access he requested for hosting_https , I've mailed him about moving issues back to D.o |
| [23:57:40] | <helmo42[m]> | jonpugh: I created a tag for you on https://drupal.stackexchange.com/questions/tagged/opendevshop |
| [00:07:32] | * shaneonabike1 has joined #aegir |
| [02:29:30] | * TaoStyle has joined #aegir |
| [03:03:43] | <memtkmcc> | I'm back online, Happy New Year everyone! At least, it's still rather new, I hope.. |
| [03:04:06] | <memtkmcc> | Did I miss any Nginx related issues I should take care of? |
| [03:07:04] | <bgm> | memtkmcc: happy new year :) thanks for commenting on the issue. I wasn't sure if it was better to have 2 locations, or 1 more complicated. I'll update the PR :) |
| [03:12:21] | <memtkmcc> | bgm: thanks! |
| [03:18:48] | <bgm> | memtkmcc: although, for the second location block, it still has to use "location ~ ^/\w\w/civicrm", so while it might speed up a bit requests on /civicrm, it still slows nginx in general, for other pages? |
| [03:20:10] | <bgm> | i.e. something like this: http://paste.debian.net/hidden/4265c856/ ? |
| [03:23:09] | <gusaus[m]> | https://opencollective.com/aegir has not been activated - so people can send you $$ |
| [03:24:14] | <bgm> | tangentially, should we move these location blocks to hosting_civicrm? |
| [03:24:50] | <memtkmcc> | bgm: There will be no difference for i18n URIs, but users w/o i18n will get their requests processed faster, since ^~ is a full stop for any further regex checks, including ~* ^/\w\w/civicrm |
| [03:25:44] | <bgm> | memtkmcc: yep, but any non-civicrm request will get the regex check? |
| [03:25:48] | <memtkmcc> | how you would like to move these locations to hosting_civicrm? |
| [03:26:58] | <memtkmcc> | Yes, which only means that ~* ^/\w\w/civicrm should be added not directly below ^~ but as far further as possible |
| [03:28:15] | <memtkmcc> | you can't inject it effectively via module, because it belongs within the existing config include stream |
| [03:28:24] | <bgm> | I notice that hosting_civicrm doesn't have a `hook_provision_nginx_vhost_config` (it does for Apache), but I suspect it's because the nginx implementation already includes the necessary blocks |
| [03:29:13] | <bgm> | isn't this just included from the server block of the vhost? |
| [03:30:05] | <memtkmcc> | not only that, but rather because it would be ineffective to do this via module, b/c you need this location to selectively disable (micro)caching in Nginx |
| [03:31:07] | <bgm> | we're talking about a provision module, right? |
| [03:31:07] | <memtkmcc> | so has to be withing existing config, not appeneded |
| [03:31:14] | <bgm> | oh |
| [03:31:24] | <memtkmcc> | oh typos |
| [03:32:34] | <bgm> | is it because the "location" appended by a module would be too far below the microcache location block%? |
| [03:33:28] | <bgm> | although modules inject before the "include /var/aegir/config/includes/nginx_vhost_common.conf" |
| [03:33:28] | <memtkmcc> | yes, this location is used only to `set $nocache_details "Skip";` which has to be done before the req chain hits location @drupal within existing config |
| [03:34:28] | <memtkmcc> | that would be even worse, because it would force everyone to use the extra location, and too early, while ^~ are processed first |
| [03:35:05] | <memtkmcc> | this is why I would suggest to add it as late as possible, in the existing config |
| [03:35:24] | <memtkmcc> | should I propose a patch? |
| [03:36:47] | <memtkmcc> | this may all seem crazy a bit, but we Nginx users are pretty picky when it comes to config optimization -- it all adds up |
| [03:37:27] | <bgm> | i understand, and I appreciate that of the nginx config |
| [03:37:37] | <bgm> | if you don't mind, please suggest a patch |
| [03:37:58] | <bgm> | to be honest, I'm a bit confused because looking at other locations, there are things such as "location ~* ^/sites/.*/files/.*\.json$" |
| [03:38:39] | <memtkmcc> | ok, will do |
| [03:40:38] | <memtkmcc> | for some things you can't avoid regex locations, that is why we are using sto-here locations to limit that stuff where possible, and even place some wildcards within parent stop-here locations, if possible |
| [03:43:17] | <bgm> | if there was a way to detect hosting_civicrm when the nginx_vhost_common is generated, people could globally opt-out of those location blocks.. but.. too complicated, for now :) |
| [03:45:35] | <memtkmcc> | yeah, too much effort for too little gain, I suppose, also, not really feasible, since _common is used by all sites anyway, not generated per site |
| [03:45:59] | <memtkmcc> | but first, dinner! bbl |
| [03:46:12] | <bgm> | :) |
| [04:08:11] | <jonpugh1> | The open collective donation page has been activated. https://opencollective.com/aegir |
| [04:09:31] | <jonpugh1> | Apparently you can now sponsor Aegir :D |
| [04:54:00] | * shaneonabike1 has left #aegir ("PART #drupal-commerce :PING 1515779639") |
| [05:34:29] | * anto_ has quit (*.net *.split) |
| [05:34:37] | * anto has joined #aegir |
| [05:37:30] | * mlhess has quit (Ping timeout: 260 seconds) |
| [05:41:22] | * mlhess has joined #aegir |
| [08:04:36] | <colan> | ! |
| [08:29:40] | * ybabel has quit (Quit: ybabel) |
| [10:56:35] | * mengi has quit (Read error: Connection reset by peer) |
| [10:57:47] | * mengi has joined #aegir |
