IRC logs for #aegir, 2016-12-15 (GMT)

2016-12-14
2016-12-16
TimeNickMessage
[11:06:49]* christefano has quit (Quit: christefano)
[11:08:49]* realityloop has quit (Quit: Leaving..)
[11:52:50]* theMusician has quit (Quit: theMusician)
[15:01:31]* g1i7ch has joined #aegir
[15:14:30]* g1i7ch has quit (Ping timeout: 250 seconds)
[18:18:55]* boshtian has joined #aegir
[19:04:15]* ybabel has joined #aegir
[19:24:45]* reaper013 has joined #aegir
[19:59:21]* ybabel has quit (Read error: Connection reset by peer)
[19:59:40]* ybabel has joined #aegir
[20:01:35]* oluabbeys has joined #aegir
[20:26:00]* oluabbeys has quit (Ping timeout: 256 seconds)
[20:26:04]* oluabbeys1 has joined #aegir
[20:28:25]* oluabbeys1 is now known as oluabbeys
[22:07:10]* oluabbeys has quit (Ping timeout: 268 seconds)
[22:47:37]* ybabel1 has joined #aegir
[22:49:15]* ybabel has quit (Ping timeout: 252 seconds)
[22:49:16]* ybabel1 is now known as ybabel
[23:05:11]* fatguylaughing has quit (Read error: Connection reset by peer)
[23:05:18]* fatguylaughing_ has joined #aegir
[23:08:15]* reaper013 has quit (Quit: Page closed)
[23:28:43]* boshtian has quit (Quit: boshtian)
[23:43:52]* g1i7ch has joined #aegir
[23:52:36]* g1i7ch has quit (Ping timeout: 250 seconds)
[03:16:56]* theMusician has joined #aegir
[03:39:48]* g1i7ch has joined #aegir
[03:50:16]<colan>ergonlog1c: https://github.com/kubernetes/charts/tree/master/stable/drupal
[03:51:09]<colan>anarcat: Would be cool if that were possible with OpenStack intead of AWS. Actually, looks like that's being worked on: https://www.mirantis.com/blog/mirantis-iron-io-bringing-serverless-compu...
[03:52:05]* g1i7ch has quit (Ping timeout: 260 seconds)
[03:54:16]<anarcat>i guess that's what the pantheon people were doing all the way back then eh
[03:55:03]<colan>unless they weren't doing that at the beginning.
[03:55:44]<colan>these mirantis folks say they're open source / prevent vendor lock-in, but i can't find any code.
[04:03:03]* g1i7ch has joined #aegir
[04:09:52]<ergonlog1c>colan: yep, charts are pretty close to the resources/app stuff we did with hosting_kubernetes
[04:25:00]<ergonlog1c>colan, it looks like Mirantis is mostly an implementer of OpenStack, rather than building their own software
[04:25:17]<ergonlog1c>iron.io appears to be oss
[04:27:24]<ergonlog1c>"serverless" sounds suspiciously like rebranding of "shared hosting"
[04:28:06]<ergonlog1c>no doubt with better process isolation, etc. but still...
[04:31:40]* g1i7ch has quit (Ping timeout: 250 seconds)
[04:33:33]* theMusician has quit (Ping timeout: 252 seconds)
[05:16:28]* noecc has joined #aegir
[05:21:18]* g1i7ch has joined #aegir
[05:43:21]* gusaus has joined #aegir
[05:47:59]* g1i7ch has quit (Ping timeout: 258 seconds)
[06:00:40]<jonpugh>helmo: Another new devshop-inspired patch: https://www.drupal.org/node/2836185
[06:00:42]<hefring>https://www.drupal.org/node/2836185 => Allow "force-reinstall" drush option for "provision-install" command. [#2836185] => 4 comments, 1 IRC mention
[06:07:56]<colan>helmo: what can we do to get https://www.drupal.org/node/2698027 rolling?
[06:07:57]<hefring>https://www.drupal.org/node/2698027 => Minor release to update core? [#2698027] => 22 comments, 8 IRC mentions
[06:08:53]<helmo>HI all, Scrum time right?
[06:09:10]<helmo>jonpugh: that patch seems to include a numver of other changes ;)
[06:13:17]<helmo>colan: the thing we can do is try the release script ... and abort before pushing ... tis is not something travis can test :(
[06:15:35]<colan>helmo: can we get instructions for that, or do you just want to commit it? :)
[06:16:04]<colan>The code looks good to me, but not sure how to test it via apt.
[06:16:27]<colan>i.e. the debian package.
[06:20:00]<helmo>I'm looking at it now... I'll see how far I get without publishing stuff ...
[06:28:35]* memtkmcc has joined #aegir
[06:28:53]<colan>jonpugh: will there be a front-end task for force-reinstall?
[06:33:10]<jonpugh>That's my plan for devshop, Might be nice in a separate module in aegir
[06:33:48]<jonpugh>I'm not going to submit a UI patch until the backend option is accepted. Don't want to ruffle too many feathers if I can improve backend stuff.
[06:35:20]* g1i7ch has joined #aegir
[06:36:16]<colan>jonpugh: fair enough, but would be good to get this into hostmaster eventually.
[06:36:31]<colan>(once the backend is nailed down)
[06:36:52]<colan>as in, it's useful outside of devshop, in core.
[06:37:13]<colan>i've done the delete, reinstall thing too many times myself.
[06:37:16]<jonpugh>Yeah, actually, I can add my UI as a patch to hosting
[06:37:31]<colan>thumbs up
[06:38:19]<jonpugh>yeah, I have to build a confirm button anyway....
[06:42:38]* g1i7ch has quit (Ping timeout: 264 seconds)
[06:42:47]* christefano has joined #aegir
[06:43:09]* g1i7ch has joined #aegir
[06:47:06]<jonpugh>helmo: sorry about that. New patch :)
[06:47:07]<jonpugh>https://www.drupal.org/files/issues/2836185-force-reinstall_0.patch
[06:54:36]<ergonlog1c>there's already some re-install stuff in contrib. here, for example: http://cgit.drupalcode.org/hosting_dev/tree/modules/reinstall/drush/prov...
[06:55:54]<helmo>well for that matter there's also https://www.drupal.org/project/hosting_reinstall ... just D6 though
[06:58:03]<ergonlog1c>right, that's what I'd ported to Aegir3/D7 as a base for hosting_dev
[06:58:45]* g1i7ch has quit (Ping timeout: 260 seconds)
[07:00:31]<jonpugh>that's funny, I thought i searched pretty well
[07:01:31]<jonpugh>Never heard of hosting_dev :)
[07:02:21]<colan>same, but it appears to be on the contrib page. :)
[07:04:01]<helmo>I think I'm gonna do a 3.9-beta1 release. I'll have to make sure not to push it to the stable Debian repo though... but that's the only way I can really test //www.drupal.org/node/2698027
[07:04:22]<jonpugh>So purge_site() is a delete site task that works no matter what ergonlog1c ?
[07:04:28]<colan>helmo++
[07:04:33]<helmo>I'll do that for hostmaster, hosting and provision
[07:06:52]<jonpugh>dang I wish I knew about this this morning :)
[07:07:44]<ergonlog1c>jonpugh: not yet, but that's the idea
[07:08:14]<jonpugh>do you mean it doesn't work?
[07:08:23]<ergonlog1c>the re-install stuff does
[07:08:35]<ergonlog1c>I don't recall where the other parts are in dev
[07:09:07]<jonpugh>hmm
[07:10:00]<jonpugh>Another quick Q: the purge_platform module, that is designed to delete all sites on the platform as well?
[07:10:12]<ergonlog1c>I believe so, yes
[07:10:16]<ergonlog1c>it's been a while
[07:10:45]<jonpugh>ok, I'm going to include this in devshop then. I need all of that.
[07:10:53]<ergonlog1c>the idea, iirc, was to break down the normal delete task into separate steps
[07:11:12]<ergonlog1c>I think it had blocked on the fix_perms stuff
[07:11:18]<ergonlog1c>but I've implemented that since
[07:11:25]<jonpugh>yeah, that would help clarify what's going on
[07:11:46]<jonpugh>I'll see about including it in devshop and using the purge tasks instead of delete, and the reinstall task too.
[07:11:53]<ergonlog1c>at this point, it's got a bunch of TODOs for the very specific steps required
[07:12:02]<ergonlog1c>it'll need more work
[07:12:24]<ergonlog1c>take a look at these: http://cgit.drupalcode.org/hosting_dev/tree/modules/purge_site/drush/inc...
[07:13:00]<ergonlog1c>those are the various purge steps that get called form http://cgit.drupalcode.org/hosting_dev/tree/modules/purge_site/drush/pro...
[07:13:50]<ergonlog1c>the idea is to make purge extensible, so that it'd be easy to (for example) delete a solr core, if one had been provisioned for a site too
[07:14:45]<jonpugh>makes me wish it was simple to do that with services :(
[07:15:05]<jonpugh>i mean, technically it already is extensible, it's just really confusing to implement :)
[07:15:11]<jonpugh>I wrote hosting_solr a few years back
[07:15:25]<ergonlog1c>right, this makes explicit the various steps
[07:15:38]<ergonlog1c>rather than post_hooks all over the place
[07:15:52]<jonpugh>In a perfect world wouldn't they just be methods on a class next to "create" methods
[07:15:53]<jonpugh>?
[07:16:07]<jonpugh>Technically you did add more hooks there ;)
[07:16:34]<ergonlog1c>:p
[07:16:44]<jonpugh>I like it, though. at least we can see everything we are doing in one place
[07:17:10]<ergonlog1c>it's for aegir3. aegirng will be a different beast, and should make all this kind of thing much cleaner
[07:17:51]<ergonlog1c>like having entity interfaces that can enforce implementations of 'delete' methods, for example
[07:18:08]<jonpugh>I'm not ready to pull in the purge stuff into devshop, but reinstall is needed now.
[07:18:29]<jonpugh>I already got it to work today, using the install tasks instead of creating a new type of task
[07:19:19]<ergonlog1c>yeah, but that seems to just complicate things further, imo
[07:20:10]<ergonlog1c>I won't block the patch, if others prefer that approach
[07:20:39]<ergonlog1c>but it seems pretty straight-forward the way that helmo (iirc) had done it in the original hosting_reinstall
[07:23:33]* noecc has left #aegir ("pax")
[07:25:52]<jonpugh>yes, since you figured out how to save the context :)
[07:27:17]<jonpugh>Perhaps just adding the hosting_reinstall module to hosting_task_extras...
[07:27:33]<jonpugh>helmo: what would you think of that? as a separate module
[07:30:29]<ergonlog1c>jonpugh: I'm pretty sure it was helmo who figured that part out :)
[07:36:09]<helmo>jonpugh: I don't mind to include it in tasks extra
[07:36:23]<jonpugh>I'm already working on the patch :)
[07:36:38]<helmo>ergonlog1c: the credits for hosting_reinstall should go to Stuart Clark though... It wasn't me
[07:37:37]* fatguylaughing_ has quit (Quit: fatguylaughing_)
[07:37:52]<jonpugh>huh, just now seeing fix_ownership and fix_permission ... wondering why these are separate modules?
[07:38:00]* fatguylaughing has joined #aegir
[07:38:23]<helmo>3.9-beta1 tags have been pushed
[07:39:04]<helmo>jonpugh: ergonlog1c had a reason for it
[07:45:19]<ergonlog1c>also, the 'purge' stuff is actually pretty close to complete, fwiw
[07:46:23]<ergonlog1c>all that should be left is finding and calling the functions in core that we aleardy use for various deletions
[07:46:55]<ergonlog1c>it's sprinkled with TODOs for everything that should be left
[08:00:24]<jonpugh>Hmm. Methinks "install_method" needs to be a site context property then. Drush options won't pass through this
[08:08:08]* ybabel has quit (Quit: ybabel)
[08:09:19]<helmo>3.9-beta1 debian packages are in our unstable repo... and normal packages on D.o
[08:09:37]<helmo>My first dev server upgraded smoothly
[08:10:03]<helmo>But somehow travis is now hanging on a sudo call : https://travis-ci.org/aegir-project/provision/builds/184346670
[08:16:40]<colan>that doesn't seem related?
[08:17:28]<helmo>I'm not sure what could be causing that ... maybe jonpugh has some idea...
[08:17:55]<helmo>but for me it's time to close my eyes
[08:19:33]<colan>helmo: goodnight! thanks for working on this.
[08:43:50]* memtkmcc has quit (Quit: Leaving.)
[10:02:50]* fatguylaughing has quit (Quit: fatguylaughing)