| [18:40:31] | * ybabel has joined #aegir |
| [21:45:43] | * stefan_ has joined #aegir |
| [23:57:13] | * not_saved has joined #aegir |
| [00:01:18] | <not_saved> | Good morning, I am trying to use the upgrade.sh script to upgrade a 6.x-2.1 aegir installation to the current 7.x-3.120 and I am getting a very unhelpful error. Is there any way to tell exactly what is going wrong? Running on CentOS 6.5/apache 2.2.15 |
| [00:02:36] | <not_saved> | I get the error while I think the script is trying to move /var/aegir/platforms |
| [00:03:10] | <helmo42[m]> | not_saved: that's not something it should be doing ... have you read http://docs.aegirproject.org/en/3.x/install/upgrade/#major-upgrade-from-... ? |
| [00:04:26] | <not_saved> | Sorry, it's not trying to move them, it changes permissions of /var/aegir/platforms to 755 then throws an application error. |
| [00:04:33] | <not_saved> | Yes I followed those steps |
| [00:04:44] | <helmo42[m]> | I should also not that that most upgrade testing was done from the later 2.x versions ... 2.1 is fairly old :( |
| [00:05:01] | <helmo42[m]> | s/not/note/ |
| [00:05:13] | <not_saved> | Hmm |
| [00:06:46] | <not_saved> | is 6.x-2.4 the version upgrade testing was performed on |
| [00:10:12] | <helmo42[m]> | 2.5 I guess, that's the last one |
| [00:12:57] | <not_saved> | Alright, I'll bring it up to 2.5 then try again |
| [00:13:34] | <not_saved> | By the way whoever made the upgrade script, I would buy you more than a few beers. |
| [00:15:46] | <helmo42[m]> | You can also add --verbose and or --debug to the MIGRATE_OPTIONS var in that script ... be sure to have some console logging though .. it will be a lot. |
| [00:20:01] | <helmo42[m]> | Beer is always apreciated, many have helped with that script over the years. http://www.aegirproject.org/donate |
| [00:24:22] | * not_saved has quit (Quit: Page closed) |
| [00:34:52] | * not_saved has joined #aegir |
| [00:35:48] | <not_saved> | Well, at least the 6.x-2.5 gives me a more descriptive error. |
| [00:36:30] | <not_saved> | fileperms(): stat failed for [warning] /var/aegir/hostmaster-7.x-3.120/sites/xx.com/local.settings.php FileSystem.php:148 copy(/var/aegir/hostmaster-7.x-3.120/sites/xx.com/local.settings.php.bak): failed to[warning] open stream: No such file or directory FileSystem.php:19 Failed to restore backup of local.settings.php. |
| [00:37:12] | <not_saved> | I checked these files, there shouldn't be any permission issues.. |
| [00:38:20] | <helmo42[m]> | does the old site have a local.settings.php file? |
| [00:38:36] | <not_saved> | Yes |
| [00:39:12] | <not_saved> | drush |
| [00:39:16] | <not_saved> | whoops |
| [00:40:47] | <not_saved> | I think I may have figured it out. |
| [00:51:41] | <not_saved> | It may actually be something with a db connection error. |
| [00:52:32] | <not_saved> | It throws a generic drush can't bootstrap drupal db after it runs all the normal db upgrades |
| [00:54:14] | <not_saved> | and for some reason it screwed up my drush version |
| [00:54:20] | <not_saved> | Very interesting |
| [01:21:31] | <not_saved> | This is probably a really dumb question, but the very first line AEGIR_VERSION, should that be current (old) version or desired upgrade version? |
| [01:21:40] | <not_saved> | in upgrade.sh |
| [01:33:23] | <not_saved> | Nevermind that was indeed a dumb question. |
| [01:59:32] | * theMusician has joined #aegir |
| [02:11:39] | * gusaus has joined #aegir |
| [02:15:54] | * gusaus has quit (Ping timeout: 246 seconds) |
| [02:25:05] | * gusaus has joined #aegir |
| [02:57:48] | * theMusician_ has joined #aegir |
| [02:58:34] | * theMusician has quit (Ping timeout: 264 seconds) |
| [02:58:34] | * theMusician_ is now known as theMusician |
| [03:10:12] | <not_saved> | is it possible to remove an aegir installation and reinstall it without affecting the platforms/sites it manages? and reimporting them properly? |
| [03:10:33] | <not_saved> | my installation is all fucked up now :( |
| [03:11:25] | <colan> | not_saved: why not just set up a new aegir installation with the latest stable release, import the sites from the old one, and then take down the old site? https://www.drupal.org/project/hosting_remote_import |
| [03:13:05] | <colan> | i really hope you were messing around on dev/staging, and didn't break prod. |
| [03:13:11] | <not_saved> | that's a good idea, but my client retains control of the actual hosting hardware and for now I'm limited to a single server |
| [03:14:24] | <not_saved> | welllll I only broke the aegir installation, the sites themselves are still fine :* |
| [03:14:47] | <not_saved> | the frontend I believe |
| [03:15:39] | <colan> | not_saved: i would make the business case that at least 2 are necessary (Dev/Staging & Prod), but 3 are ideal (Dev, STaging & Prod). Anything less than 2 (to test things before running on Prod) is insane. |
| [03:16:20] | <colan> | your situation is exactly why i started doing this. :) |
| [03:16:24] | <not_saved> | I still have old versions of the aegir frontend on there, is it possible to move the hostmaster site to a downgrade hostmaster platform? |
| [03:16:36] | <not_saved> | Right right, this client is kind of "old school".. |
| [03:16:57] | <colan> | educate 'em! |
| [03:17:24] | <colan> | if they're ok with blowing up Production, I suppose it's fine. ;) |
| [03:18:07] | <colan> | i don't think any such downgrade process would be supported. maybe now they'll let you set up a new server, and import the sites. |
| [03:19:24] | <not_saved> | hopefully |
| [03:19:55] | <not_saved> | well maybe I can just fix the error |
| [03:20:17] | <not_saved> | I was following the install process swimmingly until hostmaster-install |
| [03:20:19] | <not_saved> | and I get this error |
| [03:20:33] | <not_saved> | Error: syntax error, unexpected '[', expecting ')' in /var/aegir/hostmaster-7.x-3.121/profiles/hostmaster/modules/aegir/hosting_civicrm/drush/install.provision.inc |
| [03:21:26] | <not_saved> | using drush 7.4.0 |
| [03:24:44] | <colan> | i feel like an issue like that just came up that was fixed, but I can't find it now. helmo42 ? |
| [03:25:21] | <helmo42[m]> | colan: that was in eldir |
| [03:28:15] | * theMusician has quit (Quit: theMusician) |
| [03:36:11] | <colan> | no idea. maybe i'm confusing it with something else. |
| [03:38:00] | <helmo42[m]> | not_saved: The last error you pasted... wasn't there a line number with it? |
| [03:40:54] | <not_saved> | line 107 |
| [03:41:29] | <not_saved> | \Drupal::service('module_installer')->install(['civicrm']); |
| [03:42:05] | <helmo42[m]> | ah, yes that has the new_ish array syntax :( |
| [03:42:09] | * theMusician has joined #aegir |
| [03:42:52] | <helmo42[m]> | can be replaced with install (array('civicrm')) |
| [03:44:44] | <ergonlogic[m]> | that syntax was added in PHP 5.4 |
| [03:45:05] | <ergonlogic[m]> | PHP 5.3 isn't supported any more: http://php.net/supported-versions.php |
| [03:45:26] | <ergonlogic[m]> | it'd be worthwhile to upgrade |
| [03:45:52] | <colan> | ergonlogic: doing a demo today? if not, i may skip the meeting. ;) |
| [03:46:09] | <ergonlogic[m]> | sure, if you like |
| [03:46:30] | <colan> | have been curious on progress, is all. :) |
| [03:46:46] | <ergonlogic[m]> | I got sidetracked with some paid work, so NG isn't quite as far along as it might otherwise be |
| [03:46:59] | <ergonlogic[m]> | but then, I get to keep paying my mortgage, so... |
| [03:47:19] | <colan> | meh, housing. what's it good for? |
| [03:47:21] | <ergonlogic[m]> | most of the core components are in place now |
| [03:48:20] | <ergonlogic[m]> | I'm at the point of starting some backend components (the actual task Ansible roles), to see how it holds up |
| [04:02:11] | <jonpugh[m]> | colan[m]: I've got some things to demo if you'd like ;) |
| [04:03:28] | <colan> | @freenode_jonpugh[m]:matrix.org: sure, curious about that too! |
| [04:06:45] | * jonpugh[m] is now known as jonpugh[m]1 |
| [04:07:18] | * jonpugh[m]1 is now known as jonpugh[m]2 |
| [04:08:53] | * jonpugh[m]2 has left #aegir ("User left") |
| [04:09:24] | * jonpug4 has joined #aegir |
| [04:09:33] | <jonpug4> | hmm the nick trick isn't working for me |
| [04:11:33] | * jonpug4 is now known as jonpug6 |
| [04:12:33] | <jonpug6> | ok, I had to manually message NickServ ... |
| [04:12:37] | <jonpug6> | Does this look better now? |
| [04:31:41] | <colan> | @freenode_jonpugh[m]:matrix.org: |
| [04:31:59] | <jonpug6> | nope guess not :) |
| [04:32:36] | <jonpug6> | you see me as @freenode_jonpugh%5Bm%5D:matrix.org: huh |
| [04:32:55] | <colan> | @freenode_jonpugh[m]:matrix.org: forget about all that as we have a proper matrix room irc is bridged with. just join the matrix room |
| [04:33:22] | <colan> | #drupal-aegir:matrix.org |
| [04:33:25] | * theMusician has quit (Quit: theMusician) |
| [04:33:52] | <colan> | then it'll use whatever screenname you chose. |
| [04:34:07] | <jonpug6> | wondering how I appear to IRC then |
| [04:34:08] | <colan> | jonpugh: yay! |
| [04:34:28] | <colan> | well, how do i appear? |
| [04:34:54] | * jonpug6 uploaded an image: Screenshot from 2017-10-26 13-34-39.png (16KB) <https://matrix.org/_matrix/media/v1/download/matrix.org/BqTrVYpnFhrGhCpb... |
| [04:35:12] | <jonpug6> | you look fine (with an [m] but meh) |
| [04:35:25] | <jonpug6> | I've got that freenode stuff at the front |
| [04:36:14] | <jonpug6> | This is why I think we should just use #aegir on IRC, I don't see the need for a matrix room especially since it's bridged anyway. What benefits are there? |
| [04:36:53] | <colan> | you may be in either room more than once? |
| [04:37:30] | <colan> | well, we can do the extra stuff that matrix provides. |
| [04:38:20] | <colan> | i could live without it, but just went along with https://www.drupal.org/node/2885060 |
| [04:38:21] | <hefring> | https://www.drupal.org/node/2885060 => Drupal Chat Bridges IRC <-> Matrix <-> Slack [#2885060] => 54 comments, 5 IRC mentions |
| [04:38:53] | * theMusician has joined #aegir |
| [04:40:07] | <colan> | jonpugh: i'm seeing 3 of you in this room. this native one, and two others from freenode. |
| [04:40:18] | <colan> | try removing the other 2? |
| [04:41:04] | * jonpugh_ has joined #aegir |
| [04:41:08] | <jonpug6> | I tried, I only used irccloud, but I thought I disabled that... |
| [04:41:16] | <colan> | @freenode_jonpugh[m]:matrix.org and @freenode__neb_github_=40jonpugh=3amatrix.org[m]:matrix.org |
| [04:41:29] | * jonpugh_ has left #aegir () |
| [04:41:50] | <colan> | jonpugh: if you want, i may be able to kick them out. |
| [04:41:58] | <jonpug6> | the github one is from riot, used for notifying of commits |
| [04:42:56] | <colan> | oh, you're already an admin. was going to ops you. |
| [04:42:56] | * jonpug6 has left #aegir ("User left") |
| [04:44:41] | * theMusician has quit (Quit: theMusician) |
| [05:00:04] | * theMusician has joined #aegir |
| [05:02:57] | <helmo42[m]> | Hi all, ergonlogic bgm jonpugh colan cweagans gboudrias memtkmcc .. and any one else ... Scrum time?. |
| [05:03:10] | * jonpug6 has joined #aegir |
| [05:03:11] | * jonpug6 waves |
| [05:03:15] | <ergonlogic[m]> | Hi helmo42 |
| [05:04:40] | <helmo42[m]> | I have two issues for jonpugh to review ... |
| [05:04:53] | <helmo42[m]> | https://www.drupal.org/node/2693795 and can we RTBC? https://www.drupal.org/node/2915001 |
| [05:04:54] | <hefring> | https://www.drupal.org/node/2693795 => mkdir(): Permission denied provision_logs.drush.inc:20 [#2693795] => 14 comments, 2 IRC mentions |
| [05:04:54] | <hefring> | https://www.drupal.org/node/2915001 => Cleanup git meta data when nolonger detecting a repo [#2915001] => 4 comments, 1 IRC mention |
| [05:06:35] | <jonpug6> | oops, didn't realize I had assignments! :D |
| [05:10:55] | <jonpug6> | hang on, committing some changes... |
| [05:11:36] | <helmo42[m]> | jonpugh, gusaus Any news from opencollective? |
| [05:18:17] | <gusaus> | oh hey - I forgot I turned IRC on :) helmo42[m] you wondering if Aegir is set up on OpenCollective? I think jonpug6 was going to do that? I’d be happy to help if needed |
| [05:19:30] | <jonpug6> | I really need to stop volunteering for things! :) sorry yall, been a crazy autumn so far, I'm very distracted by other things |
| [05:23:21] | <helmo42[m]> | gusaus: Why does the github login require 'read and write all public and private repository data'. That seems to lock me out... I'm not at liberty to share access to a repo owned by my clients. |
| [05:25:27] | <gusaus> | helmo42[m]: There’s another way to set it up without first linking your repo - the OpenCollective folks can also tweak the permissions I believe |
| [05:26:20] | <helmo42[m]> | For the AegirCoop organisation I have a 'grant' button which seems to give me a choice ... the client ones have a green checkmark indicating their OK with it but I doubt they ever did that knowingly. |
| [05:28:01] | <helmo42[m]> | gusaus: You mean the 'Apply to create a collective' form? |
| [05:28:17] | <jonpug6> | helmo42: I think they just read the number of stars, they state a requirement for 100 but clearly we don't have that yet |
| [05:28:30] | <jonpug6> | but gus can talk to them and switch it on anyway |
| [05:28:44] | <jonpug6> | I have 102 with opendevshop, but I don't want to start a collective for that, just for Aegir |
| [05:28:47] | <gusaus> | helmo42[m]: ya https://opencollective.com/apply there’s actually a Cooperative option - or you can just choose ‘other’ |
| [05:29:10] | <jonpug6> | ah even better |
| [05:29:23] | <helmo42[m]> | ok I'll fill it in |
| [05:30:43] | <gusaus> | all the ones I’ve set up were just by choosing other - then it takes about 5 mins to fill in the details - and now… with there new release, there’s a lot more flexability in terms of editing and customizing levels |
| [05:31:16] | <gusaus> | I can go over any of that with jonpug6 ergonlogic[m] or others that are in OpenCollective slack |
| [05:31:48] | <gusaus> | easier with the entire OpenCollective team there to help :) |
| [05:33:52] | <gusaus> | that would be great if Aegir was set up - then we can coordinate Drupal Dojo sprints and sessions on any Aegir related project |
| [05:34:25] | <gusaus> | Drupal Dojo is already on OpenCollective so we could actually txfer funds between collectives |
| [05:37:07] | <jonpug6> | Ok, update on Provision 4.x: |
| [05:37:46] | <jonpug6> | Context saving for all 3 types works. |
| [05:38:04] | <jonpug6> | `provision verify` triggers writing of new Configuration classes. |
| [05:38:39] | <jonpug6> | Added apache server config and part of vhost.tpl |
| [05:39:58] | <jonpug6> | Lots of stuff is done, easier to demo than to type descriptions |
| [05:54:32] | * hefring has joined #aegir |
| [05:56:08] | <ergonlogic[m]> | I think I'll put off the AegirNG demo too, since most of the work recently has bee API-focused |
| [05:57:24] | <ergonlogic[m]> | If anyone's curious, I'm happy to go over it, but the code remains the most interesting part atm |
| [05:58:17] | <ergonlogic[m]> | fwiw, entity revisions changed quite a bit in Drupal 8.4.0, in case you weren't aware |
| [05:58:18] | <helmo42[m]> | yes lets do that next week |
| [05:58:28] | <ergonlogic[m]> | so I had to refactor more than I'd thought |
| [05:58:59] | <ergonlogic[m]> | they're good improvements, but I don't think the core committers really get how backward compatibility is supposed to work |
| [06:00:41] | <gusaus> | is AegirNG D8? |
| [06:00:51] | <ergonlogic[m]> | yes |
| [06:01:17] | <ergonlogic[m]> | with an Ansible backend by default |
| [06:02:32] | <gusaus> | how’s it progressing? would it help if there were additional contirbutors? |
| [06:02:52] | <colan> | yeah, they definitely don't get it. otherwise we wouldn't have 3 branches now. |
| [06:03:36] | <ergonlogic[m]> | gusaus: very well, actually. Colan and I spec'd out a high-level architecture last month |
| [06:03:46] | <ergonlogic[m]> | https://gitlab.com/aegir/aegir/wikis/architecture |
| [06:05:08] | <ergonlogic[m]> | and since I'm between major contracts, I've had the time to make some significant progress. |
| [06:06:09] | <gusaus> | that’s good to hear :) |
| [06:06:12] | <ergonlogic[m]> | I should have a prototype that works end-to-end next week |
| [06:06:32] | <gusaus> | wow - that soon |
| [06:06:54] | <ergonlogic[m]> | at which point it'll become easier for additional contributors to engage |
| [06:07:54] | <ergonlogic[m]> | well, I'm borrowing largely off the prototype I built for the Aegir Summit a couple years ago |
| [06:08:39] | <ergonlogic[m]> | along with a bunch of base Aegir entity/profile stuff I started last winter |
| [06:08:58] | <ergonlogic[m]> | but now it's getting pretty sollid |
| [06:09:51] | <helmo42[m]> | I've gotta go too ... see you all next week |
| [06:09:59] | <ergonlogic[m]> | take care |
| [06:10:18] | <gusaus> | ergonlogic[m]: possibly the setup that’s currently on your AWS could be updated and be sort of a working prototype - then we can run Dojo sprints to raise awareness and get folks invovled? also could funnel a % of funds to the Aegir OpenCollective :) |
| [06:11:29] | <ergonlogic[m]> | yeah, a number of core members plan to run alphas once it's ready |
| [06:11:59] | <gusaus> | there are a couple new mentoring groups that also could be a source for contributors |
| [06:14:08] | <ergonlogic[m]> | the core components should be significantly simpler than Aegir 3.x, since we're leveraging the Drupal (8) API more fully, and using a full-blown task queue (Celery), and Ansible for the heavy lifting |
| [06:15:20] | <gusaus> | ok - I’m definitely up for making our Aegir setup a proof of concept and a way to direct resources back to sustaining the project |
| [06:15:57] | <ergonlogic[m]> | for the moment, I think you're better off moving forward with proven systems (i.e. Aegir 3.x) |
| [06:16:50] | <ergonlogic[m]> | Once we have AegirNG alpha/beta releases, you could run those in parallel for awhile |
| [06:17:21] | <ergonlogic[m]> | also, Jon is working on Aegir 4.x which is a bridge solution decoupling the back-end from Drush |
| [06:18:09] | <ergonlogic[m]> | also, being able to import platforms and sites from existing Aegir installations will be a priority |
| [06:18:40] | <ergonlogic[m]> | so it's not like you'd get stuck on 3.x if you use that pending NG being ready to go |
| [06:19:13] | <gusaus> | ergonlogic[m]: sure - we could also use the current setup to direct resrouces to making it easier for site builders like myself to develop and manage the platform; also think we could help out with docs and training |
| [06:19:31] | * theMusician has quit (Quit: theMusician) |
| [06:20:03] | <ergonlogic[m]> | as for contributions, we're architecting AegirNG to greatly simplify adding new functionality in a modular fashion, and allow re-use of existing tasks more easily |
| [06:22:07] | <ergonlogic[m]> | so, the basic idea is that we now have "Task" and "Operation" entities. You build a task bundle by giving it a name, attaching fields for whatever variables it needs, export that as a Feature, then write a small Ansible role to go along with it. |
| [06:25:28] | <ergonlogic[m]> | for example, I have a task to write an Nginx vhost. It has a couple fields in the front-end: URL and URL aliases. On the backend, I have a simple Ansible role to write the vhost and then reload Nginx. |
| [06:28:08] | <ergonlogic[m]> | This gets combined with other tasks into an Operation, which combines the various task fields using Inline Entity Form, marshals all those variables, and passes it into the queue, along with a list of tasks. These are then picked up by a queue worker that writes a temporary Ansible playbook referencing the task-roles and injecting the variables passed in from the front-end. |
| [06:28:49] | <ergonlogic[m]> | this then gets run through ansible-playbook, and the output posted back to the front-end |
| [06:29:24] | <ergonlogic[m]> | tasks are meant to be relatively granular, and so easily re-usable |
| [06:30:13] | <ergonlogic[m]> | operations are more application-specific, but are really nothing more than a collection of tasks |
| [06:30:44] | <ergonlogic[m]> | so almost everything it just done through the FieldUI, and exported as Features |
| [06:31:40] | <ergonlogic[m]> | as a result, adding new functionality should be pretty easy. At least, relative to Aegir 3 |
| [06:32:34] | <ergonlogic[m]> | so, that's where most contributors will likely to be engaging |
| [06:33:13] | <ergonlogic[m]> | with AegirNG core mostly focused on how Tasks and Operations function |
| [06:33:37] | <ergonlogic[m]> | and just shipping with some default supported applications, like Drupal, CiviCRM, etc. |
| [06:37:03] | <ergonlogic[m]> | the idea is to push all the app-specific stuff to the edges. That is, there'll be a set of operations for managing a Drupal site, and another set of operations for managing WordPress sites. They'll likely share some common tasks, like writing vhosts and provisioning databases, but also have some app-specific stuff like running `drush site-install` |
| [06:38:02] | <ergonlogic[m]> | but the app-specific stuff will simply leverage the Aegir API, rather than being intertwined with it, as it is now |
| [06:42:01] | <ergonlogic[m]> | anyway. gotta get back to it |
| [06:43:13] | * ybabel has quit (Quit: ybabel) |
| [06:51:17] | * theMusician has joined #aegir |
| [06:55:47] | * theMusician has quit (Client Quit) |
| [06:58:49] | <gusaus> | ergonlogic[m] I gotta jump off, but it would be great to followup on how the Dojo and related mentoring groups can direct resources back to Aegir; also would be good to followup on https://github.com/OpenProducer/openproducer/issues/1 and some of the related issues that are still blockers for the current setup. Setting up an easier to use development environment for the current setup may also be good source material for a Dojo session or sprint |
| [06:59:27] | * theMusician has joined #aegir |
| [07:54:45] | * stefan_ has quit (Ping timeout: 260 seconds) |
| [08:53:41] | * theMusician has quit (Ping timeout: 255 seconds) |
| [09:12:17] | * gusaus has quit (Ping timeout: 248 seconds) |
| [09:59:06] | * realityloop has joined #aegir |
| [09:59:37] | * realityloop has quit (Client Quit) |
| [10:02:13] | * realityloop has joined #aegir |
