| [11:35:10] | * cmcintosh has quit (Ping timeout: 240 seconds) |
| [13:21:55] | * bgm has quit (Quit: Reconnecting) |
| [13:21:56] | * bgm_ has joined #aegir |
| [13:22:15] | * bgm_ has quit (Changing host) |
| [13:22:15] | * bgm_ has joined #aegir |
| [13:22:25] | * bgm_ is now known as bgm |
| [14:26:13] | * cmcintosh has joined #aegir |
| [17:32:18] | * bgm has quit (Ping timeout: 245 seconds) |
| [17:33:10] | * bgm has joined #aegir |
| [17:33:10] | * bgm has quit (Changing host) |
| [17:33:10] | * bgm has joined #aegir |
| [18:05:01] | * viashimo has quit (Ping timeout: 260 seconds) |
| [18:11:45] | * viashimo has joined #aegir |
| [18:15:28] | * orangey has left #aegir ("Part") |
| [18:33:33] | * ybabel has joined #aegir |
| [19:00:16] | * boshtian has joined #aegir |
| [19:08:34] | * reaper013 has joined #aegir |
| [21:05:36] | * oluabbeys has joined #aegir |
| [21:13:18] | * oluabbeys has quit (Quit: oluabbeys) |
| [22:23:14] | * cmcintosh has quit (Quit: Leaving...) |
| [00:10:52] | * oluabbeys has joined #aegir |
| [00:20:25] | * cmcintosh has joined #aegir |
| [00:48:39] | * v20th has joined #aegir |
| [01:08:59] | <colan> | see ergonlogic ? |
| [01:09:06] | <ergonlogic> | colan: hi |
| [01:09:25] | <colan> | ergonlogic: hey! haven't seen you around in a bit. |
| [01:09:55] | <ergonlogic> | no, vacation, then flu/bronchitis, plus a couple out-of-town hockey tournaments for Aidan |
| [01:10:07] | <ergonlogic> | getting back in the saddle now |
| [01:10:32] | <ergonlogic> | how've you been? |
| [01:11:39] | <colan> | can't complain. been doing a good amount of client work, which has been keeping me away from some of the fun stuff. ;) |
| [01:12:19] | <colan> | started to pull away a bit to keep things balanced. didn't go anywhere, was just working. |
| [01:12:27] | <ergonlogic> | yeah, I'm gearing up for some of that over the coming months... some should be Kubernetes-focused, though, so that's cool |
| [01:12:40] | <ergonlogic> | I hear you |
| [01:12:52] | <ergonlogic> | Thanks for merging that aegirNG MR |
| [01:13:19] | <ergonlogic> | fwiw, I got blocked on trying to run a queue worker as a service in Docker on gitlab-ci |
| [01:14:02] | <ergonlogic> | systemd and docker don't play nice unless you're running a privileged container |
| [01:14:07] | <colan> | been trying to stay on top of Aegir HTTPS, but could use more help. *nudge* *nudge* *wink* *wink* |
| [01:14:14] | <ergonlogic> | hah |
| [01:14:21] | <ergonlogic> | ok, how can I help? |
| [01:14:27] | <colan> | basically, it works, so folks aren't too worried about the other stuff that should get done. |
| [01:14:45] | <ergonlogic> | yeah, LE is deffinitely the killer feature there |
| [01:14:50] | <colan> | ergonlogic: there's some stuff in the issue queue... |
| [01:15:33] | <colan> | some of those need to get done before a beta can be released. |
| [01:16:11] | <colan> | ergonlogic: but if you're working on aegir4, that's just as good, if not better. ;) |
| [01:16:21] | <ergonlogic> | since I've been having health issue going back to early December, I haven't been billing enough. So I'm going to have to focus on that for a bit. However, NDI has expressed interest in LE, so I'll see if I can get some hours on it for my next SOW |
| [01:17:12] | <ergonlogic> | well, I'd prefer to move the Kubernetes stuff over to aegir4, if I can get it far engough along |
| [01:18:23] | <colan> | right. the issues i'm thinking about in HTTPS are of the type: "Oh, don't do that or this bad thing will happen, do this other thing instead" |
| [01:18:53] | <ergonlogic> | let me take a look at the queue |
| [01:19:14] | <ergonlogic> | there's also an ongoing issue with installing Aegir3 on 16.04/PHP7 |
| [01:19:32] | <ergonlogic> | I think it's just some missing/complicated dependencies |
| [01:20:48] | <colan> | yeah, i commented in there that i'm not affected b/c i install/configure all of that stuff before i install aegir. |
| [01:21:08] | <ergonlogic> | yeah, same here, since I use my ansible role |
| [01:21:15] | <ergonlogic> | still, the .deb should just work |
| [01:21:18] | <colan> | agreed. |
| [01:21:39] | <ergonlogic> | also, I guess Precise will ne EOL soon (12.04) |
| [01:21:45] | <colan> | ergonlogic: the HTTPS issues that should be fixed are the ones i tagged "To Do" |
| [01:21:51] | <ergonlogic> | so that might present an opportunity to clean up a bit |
| [01:22:01] | <ergonlogic> | Apache 2.2, for example |
| [01:22:09] | <colan> | yeah, folks are really starting to move to xenial. |
| [01:22:33] | <colan> | basically, i didn't trust the deb so it worked for me. :) |
| [01:23:06] | <ergonlogic> | right, it really is *only* some missing dependencies, afaict |
| [01:23:09] | * oluabbeys has quit (Read error: Connection reset by peer) |
| [01:23:26] | <colan> | ah, ok. yeah, you shouldn't have to read the debian control file to get it work properly. |
| [01:23:29] | * oluabbeys has joined #aegir |
| [01:24:15] | <colan> | took me a while to figure out how not to install a DB b/c i've got a remote one, etc. |
| [01:24:45] | * MrAdamJohn[m] has quit (Remote host closed the connection) |
| [01:25:09] | <ergonlogic> | hm... yeah, I was working on a related issue, to allow better db mgmt for the hostmaster site |
| [01:25:44] | <ergonlogic> | currently, we can't start with a local db, then migrate the hostmaster site to a remote DB |
| [01:25:49] | <ergonlogic> | for example |
| [01:26:00] | <ergonlogic> | anyway, that's a bit of an edge-case |
| [01:26:14] | <ergonlogic> | probably best not to waste time on it, and do it right in aegir4 |
| [01:26:39] | <colan> | in my mind, that was one of those: "It would be nice it that worked, but I'm assuming it doesn't." :) |
| [01:27:32] | <colan> | agreed on aegir4. if someone really wants to get it done in aegir3, he/she can work on it. |
| [01:28:44] | <ergonlogic> | I banged my head on it for a couple dayds straight |
| [01:28:53] | <colan> | it would be better to have 4 working with an upgrade path. then it doesn't matter. |
| [01:28:57] | <ergonlogic> | it *ought* to work |
| [01:29:06] | <ergonlogic> | but yeah |
| [01:29:07] | <colan> | yes. |
| [01:29:23] | <ergonlogic> | an upgrade path will be a challenge |
| [01:29:34] | <ergonlogic> | it'll likely be a migration path again |
| [01:30:23] | <ergonlogic> | if we can assume an Aegir3 backup as the input, we should be fine |
| [01:30:24] | <colan> | yeah, or just remotely import the old sites. doesn't need to be fancy. |
| [01:30:30] | <ergonlogic> | exactly |
| [01:30:47] | <colan> | that's what i've been doing when stuff breaks across 3. |
| [01:31:18] | <colan> | one of the reasons i got that module working again. |
| [01:32:08] | <colan> | what would be really useful for migration would be multiple sites: https://www.drupal.org/node/1950452 |
| [01:32:09] | <hefring> | https://www.drupal.org/node/1950452 => Allow batch remote imports of multiple sites or entire platforms [#1950452] => 3 comments, 1 IRC mention |
| [01:32:15] | <colan> | you can only do one at a time now. |
| [01:32:47] | <ergonlogic> | yes, that would be nice :) |
| [01:33:29] | <ergonlogic> | as long as they'd all be going to the same (new) platform, I don;t think it should be all that hard to implement |
| [01:34:06] | <ergonlogic> | except that we provide an opportunity to change the site name currently, which'd be hard to do UI-wise for a batch import |
| [01:35:25] | <colan> | having to click through stacked forms would be too bad to start. |
| [01:36:42] | <ergonlogic> | hmm, that's an option... |
| [01:36:52] | <ergonlogic> | it'd alleviate the single target platform thing too |
| [01:37:34] | <ergonlogic> | but it wouldn't really be a batch operation, per se |
| [01:37:45] | <ergonlogic> | it'd just speed things up some |
| [01:37:54] | <ergonlogic> | which is really the point, after all |
| [01:38:31] | <ergonlogic> | colan: I commented on the 3 issues marked "todo" |
| [01:38:57] | <ergonlogic> | but I don;t know how helpful it was |
| [01:46:56] | <colan> | yes, thanks. right. could be exposing a core bug there or some such. we've already done that once before on this project. |
| [01:50:23] | * MrAdamJohn[m] has joined #aegir |
| [01:52:10] | <colan> | ergonlogic: do you think we can merge this, or do you feel there's more we should be doing in the deb? https://github.com/aegir-project/documentation/pull/32 |
| [02:10:52] | <ergonlogic> | merged |
| [02:11:34] | <ergonlogic> | re. https, I don't recall adding any views integration, which is why it seems unlikely that the probelm is really in that code |
| [02:11:45] | <ergonlogic> | also, that the views return eventually... |
| [02:12:59] | <bgm> | I didn't read the full backlog, so apologies if non-sense, but I ran into that views bug on another site. Forcing to flush the views cache fixed it. |
| [02:13:20] | <ergonlogic> | bgm: was it running hosting_https? |
| [02:13:27] | <bgm> | (from admin -> structure -> views -> advanced -> basement -> tiger warnings -> flush cache) |
| [02:13:30] | <bgm> | ergonlogic: yes |
| [02:13:38] | <ergonlogic> | lol |
| [02:13:42] | <bgm> | after enabling hosting_https, all views disappeared |
| [02:13:50] | <ergonlogic> | hmm |
| [02:14:15] | <ergonlogic> | maybe we need to trigger such a views cache clear in our install hook? |
| [02:16:16] | <bgm> | maybe, I can try to debug further next time. before this it didn't seem very systematic, but now I think it is. |
| [02:16:29] | * memtkmcc has joined #aegir |
| [02:17:39] | <ergonlogic> | it might be worth trying enabling hosting_ssl too, to see if it's some other change in core that's causing it. |
| [02:17:43] | <colan> | yeah, i think debugging is the way to go here. cache clearing in the install hook won't do much, I believe, as it happens on every cache clear. |
| [02:18:06] | <colan> | it's only fixed by rebuilding the registry or time (another mystery). |
| [02:18:11] | <ergonlogic> | there's a views cache that doesn't get triggered by normal cache clears |
| [02:18:28] | <bgm> | doing "flush all caches" from admin_menu would not fix the issue. I had to go in the views UI for it. |
| [02:18:30] | <colan> | ergonlogic: oh? that would do it then. |
| [02:18:41] | <ergonlogic> | I believe so, anyway |
| [02:18:48] | <colan> | ergonlogic: so drush cc all doesn't cover it? |
| [02:19:35] | <ergonlogic> | I've seen other circumstances where a `cc all` doesn't do anything, but digging into the Views settings form to clear the Views cache did |
| [02:20:13] | <bgm> | ^^ that's what happened to me |
| [02:20:17] | <colan> | should be renamed to "drush cc all-except-one-of-the-views-caches" |
| [02:20:18] | <ergonlogic> | but I haven't looked at the code, or anything. |
| [02:22:58] | <colan> | still doesn't explain what views has to do with the module, so does sound like we're triggering a lower-level problem. |
| [02:25:48] | <ergonlogic> | http://cgit.drupalcode.org/views/tree/views.module#n956 |
| [02:26:15] | <ergonlogic> | the "full" views cc sets a menu router rebuild |
| [02:26:44] | <ergonlogic> | and also triggers hook_views_invalidate_cache() |
| [02:27:50] | * colan nods |
| [02:30:45] | <ergonlogic> | when we enable the extension, could we be doing something to the menu router that isn't allowing Views to respond at those paths? |
| [02:30:51] | <memtkmcc> | colan: cache_views should be cleared when you run drush cc all, I think: https://api.drupal.org/api/views/drush%21views.drush.inc/function/views_... |
| [02:31:44] | <ergonlogic> | yep: http://cgit.drupalcode.org/views/tree/views.module#n925 |
| [02:32:19] | <memtkmcc> | right |
| [02:32:38] | <ergonlogic> | but views_invalidate_cache() appears to take additional steps |
| [02:32:57] | <ergonlogic> | and *those* appear to work |
| [02:34:41] | <ergonlogic> | it appears to clear the Views cache (not the View data cache, yet), then run a `cc all`, which'd clear the views data cache too, then sets the menu router to be rebuilt, and finally invokes hook_views_invalidate_cache() |
| [02:34:52] | <memtkmcc> | yeah, it forces menu rebuild plus https://api.drupal.org/api/drupal/includes%21module.inc/function/module_...('https://api.drupal.org/api/views/views.api.php/function/hook_views_inval...); // Allow modules to respond to the Views cache being cleared. — https://api.drupal.org/api/views/views.module/function/views_invalidate_... |
| [02:34:55] | <memtkmcc> | yep |
| [02:35:40] | <ergonlogic> | since all our Views just disappear, I suspect it has something to do with the menu router |
| [02:36:17] | <memtkmcc> | and maybe also to // Allow modules to respond to the Views cache being cleared. |
| [02:36:17] | <memtkmcc> | https://api.drupal.org/api/drupal/includes%21module.inc/function/module_...('https://api.drupal.org/api/views/views.api.php/function/hook_views_inval...); |
| [02:36:31] | <ergonlogic> | Does the menu router get rebuilt on all cache clears? |
| [02:36:52] | <memtkmcc> | no |
| [02:37:04] | <ergonlogic> | I didn't think so |
| [02:37:12] | <memtkmcc> | https://api.drupal.org/api/drupal/includes%21cache.inc/function/cache_cl... |
| [02:37:50] | <ergonlogic> | are there any implementations of that hook in our codebase? |
| [02:38:20] | <memtkmcc> | the could be both https://api.drupal.org/api/drupal/includes%21bootstrap.inc/function/vari...('menu_rebuild_needed', TRUE); and module_invoke_all('views_invalidate_cache') |
| [02:39:08] | <memtkmcc> | ops, it keeps converting copy/paste to URL |
| [02:40:05] | <ergonlogic> | I don't see any implementations of that hook |
| [02:40:54] | <memtkmcc> | we rely on cache_clear_all() only |
| [02:41:10] | <memtkmcc> | so it is not enough |
| [02:42:02] | * cmcintosh has quit (Remote host closed the connection) |
| [02:42:44] | <ergonlogic> | I bet it's related to https://gitlab.com/aegir/hosting_https/issues/7 |
| [02:44:27] | <ergonlogic> | can anyone confirm whether this is on enabling hosting_https alone, or in concert with one or more submodules? |
| [02:51:50] | * reaper013 has quit (Quit: Page closed) |
| [03:02:43] | <v20th> | ergonlogic, enabled hosting_https module on a fresh aegir installation. don't see any problems |
| [03:03:05] | <ergonlogic> | v20th: ok, how about one of the submodules? |
| [03:06:43] | <v20th> | enabled all of them in the submodules/ directory - all good |
| [03:07:10] | <ergonlogic> | so, your sites/platforms/servers lists are all working properly? |
| [03:08:47] | <v20th> | yes, I can see all created platforms just like before |
| [03:09:46] | <v20th> | we are talking about aegir3 and drupal7 here, right? |
| [03:10:18] | <ergonlogic> | yes |
| [03:10:40] | * ybabel has quit (Remote host closed the connection) |
| [03:10:47] | <ergonlogic> | I've update the issue with relevant stuff from these discussions. |
| [03:10:52] | * ybabel has joined #aegir |
| [03:11:06] | <colan> | v20th: if you "drush cc all", and then click on the Sites, Platforms, and Servers tabs, the lists show up? |
| [03:15:48] | <v20th> | uninstalled, reenabled hosting_https and hosting_apache_https, cleared cache with drush and with UI - none of three lists ever disappeared |
| [03:17:10] | * boshtian has quit (Quit: boshtian) |
| [03:20:03] | <colan> | v20th: OS? |
| [03:20:39] | <colan> | could be a xenial/php7 thing. |
| [03:21:01] | * oluabbeys has quit (Quit: oluabbeys) |
| [03:22:45] | <v20th> | I'm testing on 16.04.1 LTS (Xenial Xerus) with PHP 7.0.13 |
| [03:23:23] | <v20th> | but.. I'll be installing aegir on 14.04 later today. can test there as well |
| [03:24:35] | <ergonlogic> | v20th: please report any findings here or in https://gitlab.com/aegir/hosting_https/issues/24 |
| [03:24:43] | <ergonlogic> | and thanks for testing! |
| [03:25:38] | <v20th> | ergonlogic, will do |
| [05:31:53] | * hestenet has quit (Ping timeout: 245 seconds) |
| [07:00:37] | * christefano has joined #aegir |
| [07:06:44] | <v20th> | ergonlogic, tested 'hosting_https' on a fresh installation on 14.04 with PHP5.5, same results. enable modules, disable, clear cache - don't see no problems with the views |
| [07:28:26] | * ybabel has quit (Quit: ybabel) |
| [07:28:37] | <ergonlogic> | hmm, ok, thanks |
| [09:01:25] | * memtkmcc has quit (Quit: Leaving.) |
| [09:22:19] | * v20th has quit (Quit: Leaving) |
| [09:27:57] | * christefano has quit (Quit: christefano) |
| [10:03:54] | * cmcintosh has joined #aegir |
