IRC logs for #aegir, 2016-05-25 (GMT)

2016-05-24
2016-05-26
TimeNickMessage
[10:22:59]* Yaazkal has joined #aegir
[10:42:57]* bgm has quit (Changing host)
[10:42:57]* bgm has joined #aegir
[11:00:04]* Egyptian[Home] has quit (Quit: Leaving.)
[11:04:40]* mstenta has quit (Ping timeout: 260 seconds)
[11:05:26]* gusaus has quit (Quit: gusaus)
[11:13:15]* Egyptian[Home] has joined #aegir
[11:16:26]<bgm>anyone else seeing this in system logs? "drush[4629]: mysqli_close() expects parameter 1 to be mysqli, null given [warning]"
[11:16:30]<bgm>drush[4629]: hosting_queued.drush.inc:227
[11:44:26]* Yaazkal has quit ()
[11:46:31]* ceaucari has quit (Quit: My Mac has gone to sleep. ZZZzzz…)
[11:52:55]* orangey has quit (Ping timeout: 244 seconds)
[11:55:57]* orangey has joined #aegir
[12:03:59]* gusaus has joined #aegir
[12:05:28]<gboudrias>bgm: Yep, same bug
[12:05:33]<gboudrias>(Since 3.5)
[12:06:03]<gboudrias>Meaning, I the same error in my logs, but only on the 3.5 servers
[12:18:24]* TommyCox has joined #aegir
[12:23:30]* TommyCox has quit (Ping timeout: 276 seconds)
[12:30:52]* gusaus has quit (Quit: gusaus)
[12:32:57]* ceaucari has joined #aegir
[12:33:20]* gusaus has joined #aegir
[13:07:51]* gusaus has quit (Quit: gusaus)
[13:19:56]* TommyCox has joined #aegir
[13:24:01]* TommyCox has quit (Ping timeout: 240 seconds)
[13:52:23]* Egyptian[Home] has quit (Quit: Leaving.)
[14:05:40]* maestrojed has joined #aegir
[14:23:49]* maestrojed has quit (Quit: Textual IRC Client: www.textualapp.com)
[15:44:55]* darthsteven has joined #aegir
[16:08:47]* boshtian has joined #aegir
[16:11:12]* darthsteven has quit (Quit: darthsteven)
[16:12:52]* boshtian has quit (Client Quit)
[16:49:11]* ybabel has joined #aegir
[17:25:22]* TommyCox has joined #aegir
[17:30:29]* TommyCox has quit (Ping timeout: 244 seconds)
[17:56:37]* ceaucari has quit (Quit: My Mac has gone to sleep. ZZZzzz…)
[17:58:53]* ceaucari has joined #aegir
[18:01:31]* TommyCox has joined #aegir
[18:05:50]* TommyCox has quit (Ping timeout: 260 seconds)
[18:11:08]* darthsteven has joined #aegir
[18:30:47]* ceaucari has quit (Quit: My Mac has gone to sleep. ZZZzzz…)
[18:55:03]* darthsteven has quit (Quit: darthsteven)
[19:09:58]* darthsteven has joined #aegir
[19:13:28]* boshtian has joined #aegir
[19:14:11]* rominronin has joined #aegir
[20:03:11]* TommyCox has joined #aegir
[20:08:49]* TommyCox has quit (Ping timeout: 260 seconds)
[20:56:43]* rominronin has quit ()
[20:57:43]* Egyptian[Home] has joined #aegir
[21:02:30]* rominronin has joined #aegir
[21:05:45]* TommyCox has joined #aegir
[21:10:34]* TommyCox has quit (Ping timeout: 244 seconds)
[22:00:57]* Egyptian[Home] has quit (Quit: Leaving.)
[22:05:58]* noecc has joined #aegir
[22:06:56]* TommyCox has joined #aegir
[22:08:05]* boshtian has quit (Quit: boshtian)
[22:11:03]* TommyCox has quit (Ping timeout: 240 seconds)
[22:12:59]* rominronin has quit ()
[22:15:03]* rominronin has joined #aegir
[23:02:05]* mstenta has joined #aegir
[23:07:44]* TommyCox has joined #aegir
[23:12:51]* TommyCox has quit (Ping timeout: 244 seconds)
[23:16:38]<ergonlogic>gboudrias bgm: This commit moved around that part of the code: http://cgit.drupalcode.org/hosting/commit/?id=f392622ad52a8fd2dd248d0434...
[23:17:13]<ergonlogic>perhaps file an issue, and consider reverting it, until we've had a chance to look closer at what's going on there.
[23:18:42]<bgm>ok thanks, will file an issue for now
[23:19:14]<bgm>wasn't sure if I was not noticing it before, since I was ignoring my logs for the past weeks (I use logstash, but also logcheck)
[23:24:23]<ergonlogic>re-opened https://www.drupal.org/node/2150787#comment-11228687
[23:24:24]<hefring>https://www.drupal.org/node/2150787 => Queue daemon fails to restart itself [#2150787] => 20 comments, 1 IRC mention
[23:24:28]<ergonlogic>bgm: ^^^
[23:24:50]<bgm>ah thx
[23:59:24]* boshtian has joined #aegir
[00:09:06]* TommyCox has joined #aegir
[00:13:49]* TommyCox has quit (Ping timeout: 244 seconds)
[00:31:52]* boshtian has quit (Quit: boshtian)
[00:43:08]* hestenet has joined #aegir
[01:00:35]<colan>jonpugh: looks like nobody replied, but yes, i think a session wouldn't be a bad idea. if it's too late, a BoF would be the next option.
[01:00:56]* ceaucari has joined #aegir
[01:01:02]<ergonlogic>I submitted a couple sessions under the wire
[01:01:08]<jonpugh>Ok good :)
[01:01:08]<colan>sweet.
[01:01:15]<ergonlogic>neither were specific to the coop
[01:01:29]<colan>BoF for that one then ;)
[01:01:32]<jonpugh>oh :)
[01:01:32]<ergonlogic>one on drush make, the other on advanced Aegir capabilities
[01:01:42]<ergonlogic>and I'll mention the coop in both
[01:01:56]* colan is definitely going to the advanced one as he's still a newbie
[01:01:56]<ergonlogic>if either are accepted, of course
[01:02:02]<ergonlogic>:)
[01:02:31]<colan>although i did start playing with the "advanced" git stuff.
[01:03:08]<ergonlogic>we made some good progress on letsencrypt on Monday, so we should be in pretty good shape to get something functional, if we plan to sprint for a day at DrupalNorth
[01:03:46]<ergonlogic>and that's using the Aegir Services API, so that's about as advanced as it comes
[01:03:58]<ergonlogic>*not* hosting_services
[01:04:26]<ergonlogic>this one: http://api.aegirproject.org/api/Hosting/example%21hosting_example.module...
[01:04:56]<jonpugh>I'll talk to alex, see if we can sneak one in.
[01:05:09]<jonpugh>I think it would be a good PR opportunity for the Coop to do one specifically on that
[01:05:22]<jonpugh>It didn't block me, so I just submitted one on that.
[01:05:38]<jonpugh>added you as a speaker, ergonlogic :)
[01:06:00]<ergonlogic>cool :)
[01:06:23]<colan>ergonlogic: great, re: letsencrypt. so we can get the full update tomorrow?
[01:06:40]<ergonlogic>sure, if you want to wait :p
[01:07:05]<colan>can do sooner too ;)
[01:07:37]<ergonlogic>so, first off, started an 'aegir' group on gitlab.com
[01:07:41]<ergonlogic>then posted https://gitlab.com/aegir/hosting_certificate
[01:08:51]<ergonlogic>so, basically, this is intended to replace the current SSL cert generation in core, starting in 4.x
[01:09:06]<ergonlogic>but we can use it as contrib in 3.x
[01:09:33]<jonpugh> fantastic
[01:09:45]<ergonlogic>hosting_certificate is very slim, and just defines the "Certificate" service type
[01:10:07]* TommyCox has joined #aegir
[01:10:10]<ergonlogic>it's a dependency of the included hosting_letsencrypt module
[01:10:21]<ergonlogic>where we define the service implementation
[01:10:43]<ergonlogic>I think this is a model we should adopt in core too, btw
[01:11:13]<ergonlogic>right now, the base http class, and it's default implementation, Apache, are bundled together
[01:11:28]<ergonlogic>so you can't have an Nginx-only install, for example
[01:12:01]<ergonlogic>we've got notes for next steps, etc. here: https://pad.riseup.net/p/gbXnp4pDNO6J_le
[01:13:13]<ergonlogic>so, far, all it does is provide the api boilerplate: writing config into the server alias/context, hooking into verify tasks, etc.
[01:13:46]<ergonlogic>and it creates the /var/aegir/config/letsencrypt and letsencrypt.d dirs
[01:14:57]<ergonlogic>we've included letsencrypt.sh directly for the time being, to get to a working prototype
[01:15:18]* TommyCox has quit (Ping timeout: 244 seconds)
[01:15:53]<ergonlogic>but we're likely to wither have to move to a proper acme library, or split the script in 2, to be able to get the verification challenge to work properly
[01:16:11]<ergonlogic>especially in a remote/cluster setup
[01:17:02]<ergonlogic>there are a few recent commits to core that needed to be reverted, as they were hard-coding the path to the .well-known dir
[01:17:38]<ergonlogic>and were added in support of the hosting_le module, which will be deprecated once this is working
[01:18:37]<ergonlogic>anyway, the idea is that when we enable the service on a server, it'll write a global location block (or th nginx equivalent) to the http server config
[01:19:01]<ergonlogic>this'll then get sync'd out to the remotes, if necessary
[01:19:52]<ergonlogic>that way, wherever the DNS entries for the site are pointing, those servers will be able to respond to the letsencrypt challenge
[01:20:20]<ergonlogic>the plan is to add an option to opt-out of this behaviour, in case the DNS mechanism is preferred
[01:21:22]<ergonlogic>we aren't trying to handle the DNS method at all, btw, just allow for it, with minimal cruft in vhosts
[01:22:12]<ergonlogic>apparently there's a challenge with this strategy, since nginx doesn't support bare location blocks (outside of a vhost)
[01:22:36]<ergonlogic>but bgm is working on a solution to that
[01:23:20]<ergonlogic>anyway, once the servers are verified and the .well-known/challenge stuff is in place, the rest ought to be pretty simple
[01:24:20]<ergonlogic>basically, we call the letencrypt.sh script, which (iirc) generates the private key, triggers the challenge, then provides the certificate
[01:25:31]<ergonlogic>once Aegir's default SSL code has deployed the self-signed SSL certs, we'll replace them with our letsencrypt ones
[01:26:15]<ergonlogic>we plan to alter the site certificate field, to hardcode the generation of a self-signed cert using the site uri
[01:28:16]<ergonlogic>and then implement a queue to regularly check the letsencrypt certs' validity dates, and trigger a new verify to generate new ones, when we're getting close
[01:29:06]<ergonlogic>anyway, that's more or less the plan, as it currently stands
[01:30:26]<ergonlogic>we'll wrap the call to the letsencrypt.sh in a helper method to make it easy to swap in a native acme implementation, once we have a working prototype
[01:31:50]<ergonlogic>once the model proves itself, we can look at implementing "self-signed" and "user-provided" implementations of the certificate service
[01:31:52]* rominronin has quit ()
[01:32:29]<ergonlogic>the self-signed one will basically incorporate what we already do in core, so that we can deprecate that code, for easy removal in 4.x
[01:33:35]<ergonlogic>and the "user-provided" would provide elements on the site form like https://www.drupal.org/node/2466977
[01:33:36]<hefring>https://www.drupal.org/node/2466977 => [META] SSL refactor [#2466977] => 8 comments, 2 IRC mentions
[01:33:58]<ergonlogic>https://www.drupal.org/files/issues/AddListener_SSLCertificate.png
[01:34:47]<ergonlogic>both of those implementations ought to be pretty straight-forward, once we've worked out the kinks in the letsencrypt one
[01:36:37]<ergonlogic>hefring: log-pointer
[01:36:43]<ergonlogic>hefring: log?
[01:36:48]<ergonlogic>hefring: log pointer?
[01:36:48]<hefring>http://hefring.mig5.net/bot/log/aegir/2016-05-25#T622051
[02:00:12]* ergonlogic just updated the relevant issues
[02:08:21]* darthsteven has quit (Quit: darthsteven)
[02:23:34]* freiheit has joined #aegir
[02:23:49]* freiheit has quit (Client Quit)
[02:24:15]* freiheit has joined #aegir
[02:28:54]<colan>ergonlogic: sounds good! (sorry, got distracted earlier)
[02:29:05]<ergonlogic>no worries :)
[02:29:32]<colan>looking forward to digging into this in MTL.
[02:29:33]<ergonlogic>say... hook_requirements()... I'm pretty sure it only runs on install/update, and on the status page, no?
[02:29:52]<colan>not sure, haven't actually used it. ;)
[02:30:57]<colan>just seemed like a stop-gap measure for the lack of more useful dependency-like checking.
[02:31:08]<ergonlogic>looks like it's only implemented in those places. We could use it ourselves, I suppose, but I'm not sure there's a benefit to that, over our own hook
[02:31:30]* colan nods
[02:32:03]<ergonlogic>yeah... it's a weird dependency too
[02:32:05]* freiheit has left #aegir ()
[02:33:00]<ergonlogic>at least one service implementation has to be enabled... but really that's only for web and db, and really only for the server(s) that @hostmaster is installed on
[02:33:15]<ergonlogic>though, of course, the underlying features/modules need to be enabled too
[02:34:55]<ergonlogic>I don't think we need a general solution here, actually
[02:35:33]<ergonlogic>because that would mean that once a service feature was enabled, you'd never be able to fully disable it
[02:35:56]<ergonlogic>it's really *only* for web & db
[02:37:26]<ergonlogic>so we should probably just treat them as special cases, using hook_system_info_alter(), probably
[02:38:56]<ergonlogic>and set a dependency on whichever is running on the @hostmaster site's servers
[03:04:35]* zombiebeard has joined #aegir
[03:35:55]* gusaus has joined #aegir
[03:40:28]* ceaucari has quit (Quit: My Mac has gone to sleep. ZZZzzz…)
[03:40:53]* ceaucari has joined #aegir
[04:06:51]* Einherjer has quit (Remote host closed the connection)
[04:09:47]<memtkmcc>ergonlogic: ping
[04:10:19]<ergonlogic>memtkmcc: hi Adam :)
[04:10:34]<memtkmcc>ergonlogic: Hi, do you have a few seconds to chat about omega8cc request?
[04:10:41]<ergonlogic>sure
[04:11:39]<ergonlogic>is there a particular issue you'd like to review, or is it something new?
[04:12:27]<memtkmcc>ergonlogic: well, it is not really issue, it is about that t-shirt she asked for :)
[04:12:36]<ergonlogic>ah, ok
[04:12:51]<ergonlogic>have you received them?
[04:12:57]<ergonlogic>I sent them out on Friday
[04:15:06]<memtkmcc>ergonlogic oh, thanks! not yet, but I just learned about this today and was about to ask you to send it only for Grace, of course. She probably thinks a bit too fast, sometimes :)
[04:16:46]<ergonlogic>oh, well, it's no problem
[04:16:53]<ergonlogic>and it's already sent
[04:17:08]<ergonlogic>feel free to give it to someone else, if you prefer
[04:18:35]<ergonlogic>but either way, omega8.cc has been a long-time supporter of Aegir, so it wasn't a problem to send 2
[04:20:31]<memtkmcc>ergonlogic: She made me feel rather uncomfortable by this surprise, but, well, this is how Grace works, I suppose :) It will be great to have one, thank you! I just don't think I have earned it yet.
[04:21:00]<ergonlogic>well, then you owe us :p
[04:21:29]<memtkmcc>haha, thanks
[04:21:36]<ergonlogic>She said you'd be taking over maintenance of Aegir's nginx support. That's more than worth a t-shirt :)
[04:22:18]<ergonlogic>Also, I let the rest of the core team know about Grace's retirement, and to expect patches from you, moving forward
[04:22:19]<memtkmcc>I would be happy to, but I'm a newbie (to the community), so it will take some time, I think
[04:22:35]<ergonlogic>no problem, we'll help however we can
[04:22:41]<memtkmcc>Oh, nice, thank you
[04:23:30]<ergonlogic>also, a couple resources you'll probably find useful: http://docs.aegirproject.org/en/3.x/ and http://api.aegirproject.org/
[04:23:37]<ergonlogic>in case you didn't already know about them
[04:24:32]<ergonlogic>the documentation can always be improved, though, to feel free to submit PRs/patches on the docs too
[04:25:20]<memtkmcc>thanks, I'm a bit familiar with some Aegir internals already since 1.x or rather 0.9 I think, but I still need to learn more
[04:26:37]<ergonlogic>yeah, it's a pretty big code-base, so I'm often learning new stuff about it too
[04:27:48]<ergonlogic>with Grace's retirement from the core team, *none* of the core members remain from when I started with Aegir
[04:28:18]<memtkmcc>it is good to see a renewed energy in the project, by the way, I like it!
[04:28:44]<memtkmcc>yeah, that's how life goes
[04:29:33]<ergonlogic>I had taken a job that kept me distracted from core for too long. Being a consultant again, I have more flexibility
[04:30:16]<ergonlogic>with the Aegir Cooperative coming together, I hope we'll be able to actually fund some core maintenance/dev directly
[04:32:26]<memtkmcc>yes, it is a good idea
[04:33:21]<memtkmcc>the project needs some stability in this area (funding the ongoing work etc)
[04:35:06]<gusaus>that’s an issue w/ many large projects - and there’s no clear way to do that
[04:36:09]<memtkmcc>we always send people to the Aegir donate page, when they ask how they could help with BOA, but it is rather a gesture and I don't think it could produce any meaningful stream of funding: https://github.com/omega8cc/boa/blob/master/README.txt#L204
[04:37:49]<colan>especially given that there's some paperwork to do to get paypal to re-enable that. ;) we can probably get this done once the co-op is in place.
[04:38:37]<memtkmcc>colan: oh, it i blocked? wtf
[04:39:01]* darthsteven has joined #aegir
[04:40:43]<gusaus>having an entity to collect / distribute funds should definitely help
[04:42:11]<colan>memtkmcc: it's understandable. they're trying to prevent money laundering & such.
[04:42:43]<memtkmcc>well, paypal never changes, I guess
[04:44:36]<memtkmcc>paypal is just way too aggressive in these policies, so we need both the entity and payment gateway alternatives
[04:57:49]* TommyCox has joined #aegir
[05:01:06]* TommyCox has quit (Remote host closed the connection)
[05:16:13]* colan nods
[05:24:56]* noecc has left #aegir ("pax")
[05:45:47]<ergonlogic>gusaus: fwiw, that's why we invited Richard Stallman to last year's Aegir Summit. His suggestion was basically "offer paid support".
[05:46:44]<ergonlogic>now, approaching the 1-year anniversary, we're planning the founding general assembly at the 2nd Aegir Summit (also in NYC) July 10th
[05:50:14]<gusaus>that was great advice - and really a good thing to move forward with
[06:09:47]* ybabel has quit (Quit: ybabel)
[06:17:30]<gusaus>still think paid product development and support for distibutions that aegir companies could provide as a service could be another way to support the project
[06:46:21]* drakythe is now known as zz_drakythe
[06:47:12]<ergonlogic>right, new feature development is relatively easy to fund. After all, it's easy for a client to justify spening $X on some new feature that will bring $Y revenues, or save $Z dollars. The problem is that it's much harder to justify on-going maintenance, bug-fixing, re-factoring, re-architecture, etc.
[06:47:18]* zz_drakythe is now known as drakythe
[06:47:49]<ergonlogic>that is "harder to justify spending on such work"
[06:48:08]<ergonlogic>as there isn't some shiny new thing to point to at the end of the project
[06:54:05]<ergonlogic>some of us are likely to start an Aegir cloud hosting service too, but that'll essentially be separate from the coop. We don't want to compete with the existing community/eco-system.
[06:54:28]<ergonlogic>more grow a new niche within it.
[06:55:29]<ergonlogic>we figure that the "vendor" support will probably drive organizational adoption, which will grow the market for Aegir services in general
[06:56:01]<ergonlogic>"a rising tide lifts all boats" kinda thing
[06:56:32]<gusaus>ergonlogic: so there will be a way to facilitate paid product development and support for distibutions?
[06:56:57]<ergonlogic>on the flip side, we want to avoid some of the complaints we've heard about Acquia, for example, basically swallowing up large swaths of the market
[06:57:14]<ergonlogic>sure
[06:57:50]<ergonlogic>we plan to have a partner program to allow for referrals
[06:58:15]<ergonlogic>both referrals from partners, for support services, and referring dev work to partners
[06:59:40]<ergonlogic>gusaus: I may not be fully understanding what you mean by "paid product development and support for distibutions"
[07:02:44]<gusaus>ergonlogic: ha - well.. it was about a year ago i started talking to gboudrias and praxis about doind a saas offering of my project - http://getopenproducer.com/
[07:03:33]<gusaus>but it’s not yet done ‘cuz i’m the only one working on it :(
[07:04:54]<ergonlogic>ah
[07:05:51]<ergonlogic>well, I know there's been a recent spike in the development and refinement of SaaS-supporting modules.
[07:07:22]<ergonlogic>so, from an infrastructure stand-point, we should be getting pretty close to out-of-the-box SaaS systems
[07:08:58]<ergonlogic>when I was with Praxis, I always thought the white-box Aegir-SaaS would make the best service
[07:09:10]<ergonlogic>at least as far as supporting ongoing aegir development
[07:09:30]<gusaus>ergonlogic: yes - it’s really not an aegir issue but more of a drupal problem - there’s no clear and easy way to build and sustain complex projects like distributions - so mainly that only ones that exist are built and funded by a handful of large companies
[07:10:28]<ergonlogic>right, the hosted Saas application, on the other hand, generally requires both in-depth knowledge of Drupal, as well as broad domain-specific expertise, and a clear product vision
[07:11:27]<ergonlogic>believe it or not, in a previous life, I used to be a product marketing manager
[07:11:35]<ergonlogic>for microsoft, no less
[07:11:56]<ergonlogic>anyway...
[07:12:11]<gusaus>came from the dark side
[07:12:45]<ergonlogic>I got into Free Software because I thought it was the only viable competition to MS
[07:12:52]<ergonlogic>that was a loooong time ago
[07:13:15]<ergonlogic>before ubuntu, oracle's purchase of sun/mysql, etc.
[07:13:49]<gusaus>i’ve been sold on aegir even before it was called that
[07:14:11]<ergonlogic>but as for Drupal-based products, I'd think D8's CMI would solve a number of issues with D7
[07:14:28]<ergonlogic>what did it used to be called?
[07:14:55]<ergonlogic>you mean the python scripts back at Rain City?
[07:15:03]<gusaus>basicially the thing i’m trying to help make a reality is the hosted band service bryght was trying to launch 10 years ago
[07:15:27]<ergonlogic>ok
[07:30:14]* darthsteven has quit (Quit: darthsteven)
[07:38:33]* Egyptian[Home] has joined #aegir
[07:44:44]<memtkmcc>it was way ahead of its time, back then: http://www.prweb.com/releases/raincity-studios/acquires-bryght/prweb5712... We are in the golden age now, in comparison, even if there is still a ton things to do, but as it is with software, you first make it work, then you make it properly build, then you make it fast, then you make it shiny -- and we are still between 1 and 2, perhaps, so there is a Bright Futur
[07:45:20]* Egyptian[Home] has quit (Ping timeout: 258 seconds)
[08:03:31]* zombiebeard has quit (Quit: zombiebeard)
[08:27:34]* Egyptian[Home] has joined #aegir
[08:38:06]* TommyCox has joined #aegir
[08:42:57]* hestenet has quit (Remote host closed the connection)
[08:52:29]* TommyCox has quit (Remote host closed the connection)
[08:55:04]* Egyptian[Home] has quit (Ping timeout: 264 seconds)
[09:04:08]* Yaazkal has joined #aegir
[09:21:06]* freiheit1 has joined #aegir
[09:21:16]* freiheit1 is now known as freiheit
[09:52:59]* ceaucari has quit (Quit: My Mac has gone to sleep. ZZZzzz…)