| [10:12:40] | * xurizaemon is now known as xurizaemon|afk |
| [10:13:08] | * xurizaemon|afk has quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| [10:18:07] | * mstenta has joined #aegir |
| [10:46:12] | * jonpugh has quit (Quit: Connection closed for inactivity) |
| [10:59:02] | * xurizaemon has joined #aegir |
| [11:34:42] | * DecipheredAFK is now known as Deciphered |
| [11:34:43] | * Deciphered has quit (Changing host) |
| [11:34:43] | * Deciphered has joined #aegir |
| [12:02:40] | * gandhiano has quit (Ping timeout: 272 seconds) |
| [12:14:59] | * mstenta has quit (Quit: Leaving.) |
| [12:22:15] | * titanous has quit () |
| [12:41:29] | * cweagans has quit (Quit: Gone.) |
| [13:10:01] | * xurizaemon is now known as xurizaemon|afk |
| [13:11:29] | * xurizaemon|afk has quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| [13:14:56] | * xurizaemon has joined #aegir |
| [13:50:14] | * xurizaemon is now known as xurizaemon|afk |
| [13:51:43] | * xurizaemon|afk has quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| [13:55:19] | * xurizaemon has joined #aegir |
| [15:33:12] | * xurizaemon is now known as xurizaemon|afk |
| [15:34:25] | * xurizaemon|afk has quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| [15:38:20] | * xurizaemon has joined #aegir |
| [15:40:38] | * xurizaemon has quit (Client Quit) |
| [16:06:46] | * David_Hernandez has joined #aegir |
| [17:08:03] | * Deciphered is now known as DecipheredAFK |
| [17:11:08] | * boshtian has joined #aegir |
| [18:13:05] | * ratioweb has joined #aegir |
| [18:20:10] | * e-anima has joined #aegir |
| [18:29:05] | * Egyptian[Home] has joined #aegir |
| [18:33:19] | * Egyptian[Home] has quit (Client Quit) |
| [18:34:11] | * Egyptian[Home] has joined #aegir |
| [18:54:43] | * xurizaemon has joined #aegir |
| [18:55:02] | * xurizaemon is now known as xurizaemon|afk |
| [18:59:07] | * xurizaemon|afk is now known as xurizaemon |
| [18:59:08] | * xurizaemon is now known as xurizaemon|afk |
| [19:23:28] | * xurizaemon|afk has quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| [19:51:02] | * xurizaemon has joined #aegir |
| [19:55:54] | * xurizaemon has quit (Ping timeout: 265 seconds) |
| [19:58:30] | * neoxavier has quit (Ping timeout: 240 seconds) |
| [20:02:15] | * neoxavier has joined #aegir |
| [20:12:54] | * xurizaemon has joined #aegir |
| [20:25:17] | * xurizaemon has quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| [20:28:10] | * e-anima has quit (Ping timeout: 240 seconds) |
| [20:28:11] | * e-anima has joined #aegir |
| [20:56:01] | * gandhiano has joined #aegir |
| [21:02:28] | * stijnvbrande has joined #aegir |
| [21:08:35] | * xurizaemon has joined #aegir |
| [21:55:53] | * gandhiano has quit (Quit: Saindo) |
| [22:12:10] | * mstenta has joined #aegir |
| [22:13:34] | * David_Hernandez has quit (Quit: :wq!) |
| [22:31:54] | * zombiebeard has joined #aegir |
| [22:48:46] | * xurizaemon has quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| [23:26:34] | * fatguylaughing has quit (Quit: fatguylaughing) |
| [23:30:59] | * ratioweb has quit (Ping timeout: 256 seconds) |
| [23:32:12] | * ratioweb has joined #aegir |
| [23:56:33] | * titanous has joined #aegir |
| [00:26:45] | * cweagans has joined #aegir |
| [00:29:59] | <cweagans> | aw crap, sorry guys. I missed the checkin |
| [00:30:03] | <cweagans> | hefring: log pointer? |
| [00:30:03] | <hefring> | http://hefring.mig5.net/bot/log/aegir/2015-08-06#T581923 |
| [00:32:28] | <ergonlogic> | I think all the rest of us did too |
| [00:32:53] | <ergonlogic> | cweagans: how about we just chat here in IRC, rather than a separate hangout |
| [00:33:31] | <ergonlogic> | at first, at least |
| [00:37:12] | * ergonlogic In an effort to better coordinate our efforts, espcially towards AegirNG, anyone working on Aegir core and related projects is welcome to post what they've been up to recently, and what they have planned for the coming week(s). |
| [00:38:13] | <ergonlogic> | For my part, I've been working up a proposal to the business@ mailing list for a consortium to provide packaged support services for Aegir |
| [00:39:17] | <ergonlogic> | In addition, I've been preparing to release Rán (0.0.1), to invite community engagement on that |
| [00:40:17] | <ergonlogic> | Also, fighting a bit with the Drush maintainers, as they plan to remove support for stand-alone external applications built atop Drush |
| [00:40:48] | <ergonlogic> | fixed a bug or two in Drush and Aegir core |
| [00:43:44] | <ergonlogic> | cweagans helmo *others* anything on your parts that you'd like to share |
| [00:44:30] | <gboudrias> | In the last two weeks I opened a lot of issues for Aegir3 core and contrib, and helped solve some of them (helmo solved a lot of them) |
| [00:44:45] | <gboudrias> | So there is still stuff to do on that front :) |
| [00:45:01] | <gboudrias> | I think hosting_site_backup_manager is the "golden" module with the most work needed |
| [00:46:09] | <ergonlogic> | thanks gboudrias |
| [00:46:10] | <gboudrias> | I don't know if we have a release schedule but I think there's enough stuff for a new bugfix release |
| [00:46:20] | <ergonlogic> | I've seen that activity |
| [00:46:30] | <gboudrias> | I'm curious if anyone has been working on other AegirNG stuff? |
| [00:46:38] | <cweagans> | I've been mostly focused on Aegir 4 - looking into Kubernetes and Openshift, both of which are pretty promising. Mostly just thinking things through, but not a lot of actual work done at this point. I'm going to take a break from that and try to push forward on the SF Conservancy application, as I think that's more or less a prerequisite to getting Aegir funded. |
| [00:47:08] | <cweagans> | I'd also like some assistance putting together user stories for the new Aegirproject.org website. I've kind of dropped the ball on that, as I got wrapped up in a bunch of other things. |
| [00:47:15] | <ergonlogic> | I also added a 'restructure' branch to our docs: http://aegir.readthedocs.org/en/restructure/, to take advantage of mkdocs' deeper menu depths |
| [00:48:15] | <ergonlogic> | well, we also switched www.aegirproject.org to github-pages, and you created a static archive of the community site |
| [00:48:42] | <ergonlogic> | we added some mailing lists to help coordinate community efforts: |
| [00:48:55] | <gboudrias> | cweagans: This is actually something my non-dev Praxis colleague could help with, I'll try and get him to email you |
| [00:49:11] | <ergonlogic> | discuss@aegirproject.org, develop@aegirproject.org, coord@aegirproject.org and business@aegirproject.org, so far |
| [00:49:48] | <cweagans> | gboudrias: that'd be awesome. Let's do it on the coord list, though. That way it's an open discussion. |
| [00:49:51] | <ergonlogic> | I'll tweet them out, to raise awareness, but we should probably also list them on www.aegirproject.org, etc. |
| [00:50:34] | <gboudrias> | cweagans: 10-4 |
| [00:51:00] | <gboudrias> | ergonlogic: Oh I just wanted to mention it's not immediately obvious how to subscribe just from the mailing list addresses |
| [00:51:15] | <cweagans> | ^ |
| [00:51:31] | <cweagans> | if we want people to subscribe and participate, it should be really straightforward. |
| [00:51:43] | <ergonlogic> | fwiw, while I think the conservancy is definitely worthwhile, I doubt that personal donations will fund the project long-term |
| [00:51:57] | <ergonlogic> | right, suggestions? |
| [00:52:08] | <cweagans> | Right. I think that companies will be more likely to jump in if there's a nonprofit where the funds are landing. |
| [00:52:18] | <cweagans> | It looks much better for their accounting. |
| [00:53:04] | <gboudrias> | I guess we just need to share the links at the same time if possible, I think Twitter shortens the URL automatically so should be possible |
| [00:53:10] | <cweagans> | ergonlogic: If I could send a message to the list that said "Subscribe" to sign up |
| [00:53:14] | <ergonlogic> | ok, so links to the subscribe pages? https://listes.koumbit.net/cgi-bin/mailman/listinfo/develop-aegirproject... |
| [00:53:17] | <cweagans> | that would be awesome |
| [00:53:23] | <cweagans> | but links to the pages would work too |
| [00:53:29] | <ergonlogic> | I suspect that mailman supports that, but I'll need to double check |
| [00:54:09] | <ergonlogic> | so, funding-wise, I kinda see at least 2 strategies being req'd |
| [00:54:56] | <ergonlogic> | in my experience, new features are easy enough to get funding for, as larger orgs can see the return on investment |
| [00:55:22] | * mwisner has joined #aegir |
| [00:55:28] | <ergonlogic> | the un-sexy bug-fixing and general maintenance, much less so |
| [00:55:48] | <ergonlogic> | hence the upcoming support consortium proposal |
| [00:56:08] | <ergonlogic> | which has largely been hashed out here in #aegir already, or at the summit |
| [00:57:36] | <ergonlogic> | donations would be a good way to fund events and travel expenses of the core team, etc. IMO |
| [00:59:04] | <ergonlogic> | AegirNG is a bit harder to see, as it's more of a gamble |
| [01:00:18] | <ergonlogic> | but if we can get OpenShift working in a minimalist fashion, I know Poetic and others will be interested in the container isolation, etc. |
| [01:01:08] | <ergonlogic> | I *think* that'll be mostly a scratching our own itches kinda project until there's something beta-quality or better |
| [01:01:56] | <ergonlogic> | at which point larger clients could be convinced to back the last mile |
| [01:02:19] | * ergonlogic isn't sure that all makes sense |
| [01:02:28] | <ergonlogic> | but it's how I see things atm |
| [01:02:49] | <gboudrias> | Well, I definitely see what you mean |
| [01:02:52] | <ergonlogic> | anyway, we can take that offline, and onto the discuss and/or business lists |
| [01:03:20] | <ergonlogic> | don't get me wrong, I'd love to see real funding of AegirNG |
| [01:03:41] | <ergonlogic> | but until it's past the vapour-ware stage, I think it'll be hard to find backers |
| [01:04:04] | <ergonlogic> | considering how much OpenShift does for us, I don't think that very far off, though |
| [01:04:36] | <ergonlogic> | cweagans: I don't mean to rain of the parade, just trying to be realistic |
| [01:04:47] | <ergonlogic> | if you think otherwise, I'd love to hear it |
| [01:05:04] | <ergonlogic> | mlhess' prototype on kubernetes might already put us there, actually |
| [01:05:11] | <cweagans> | ergonlogic: nah, it's cool. The conservancy process will take some time, though, so if we want to have somebody fund us later, we should be moving forward with the application now. |
| [01:05:13] | <ergonlogic> | had anyone contacted him recently |
| [01:05:23] | <cweagans> | No, I've been crazy busy. |
| [01:05:23] | <ergonlogic> | agreed |
| [01:05:37] | <cweagans> | I think we're on the same page re funding. |
| [01:06:03] | <ergonlogic> | how about I reach out to Michael and ask for a demo of what he's done so far, and open discussions around openshift? |
| [01:06:23] | <ergonlogic> | what he'd shown was neat, but apprently broken the next day |
| [01:06:31] | <ergonlogic> | and I think he was going to investigate ceph |
| [01:07:00] | <cweagans> | Yeah, sounds good! |
| [01:07:25] | <ergonlogic> | he seems to be a fan of shell scripts, which can work for me, considering that we can just call those from Ansible/Rán/Vagrant or whatever |
| [01:08:11] | <ergonlogic> | would simply adopting what he's done so far be a reasonable way forward, if it provided a working prototype? |
| [01:08:35] | <cweagans> | Maybe. I'm not sure about the full extent of everything that it does. |
| [01:08:36] | <ergonlogic> | I guess it depends how he's built it |
| [01:08:50] | <ergonlogic> | yeah, so let's take a look |
| [01:08:53] | <cweagans> | However, there's still a *ton* of stuff that I'm almost 100% certain it doesn't do that we'd need (which Openshift also provides) |
| [01:09:05] | <ergonlogic> | right there, he's got a large university backing that effort |
| [01:09:18] | <cweagans> | (for instance, the push to deploy bit, which requires a git receiver, private docker registry, etc) |
| [01:09:19] | <cweagans> | true |
| [01:09:24] | <cweagans> | that's a big plus |
| [01:09:45] | <ergonlogic> | right, but if we can start of container orchestration, etc. with something working, it'll be much easier to get others on board |
| [01:10:16] | <ergonlogic> | rather than just hacking together their own kubernetes deployments or whatever |
| [01:10:40] | <ergonlogic> | like Pronovix, Wunderkraut, etc. |
| [01:11:11] | <ergonlogic> | cweagans: is NBC still planning to move to Kubernetes? (if you're allowed to say) |
| [01:12:28] | <cweagans> | I just asked on this call I'm on. The ops team is leaning toward Kubernetes, but nothing is set in stone yet. |
| [01:13:10] | <ergonlogic> | ok, so from my reading openshift adds to the kubernetes API, rather than trying to wrap it |
| [01:17:04] | <ergonlogic> | I think the priority is probably to get a couple of us bootstrapped into Kubernetes (at least) and Openshift |
| [01:17:30] | * boshtian has quit (Ping timeout: 245 seconds) |
| [01:17:35] | <ergonlogic> | ideally documenting and automating the install |
| [01:17:43] | <ergonlogic> | then re-evaluate |
| [01:18:34] | <cweagans> | That sounds reasonable. I've been able to get Kubernetes up locally with the instructions in their documentation, but haven't been able to successfully deploy Openshift. |
| [01:20:25] | <ergonlogic> | I got part ways to Kubernetes back at the summit, but haven't had a chance to revisit it since |
| [01:20:37] | <ergonlogic> | I'll try to carve out some time |
| [01:24:25] | <ergonlogic> | I think we could reasonably target a single production deployment environment to keep things simple, up front too |
| [01:24:31] | <ergonlogic> | like AWS, or something |
| [01:25:14] | <ergonlogic> | their EFS allow us to defer the shared storage issue, for example |
| [01:25:30] | <ergonlogic> | in the short-term only, of course |
| [01:25:48] | <cweagans> | That's fair, though we should be able to spin all this stuff up locally really easily too |
| [01:26:24] | <ergonlogic> | true, but I mean to get something production worthy in the shortest time-frame |
| [01:26:47] | <cweagans> | I think it's reasonable to say "Step 0 is to go install Kubernetes and configure one of the persistent file storage drivers", and then have our install docs on top of that. |
| [01:27:02] | <cweagans> | because all of that stuff is a) well documented and b) supported by somebody else. |
| [01:27:51] | <ergonlogic> | true, but if we could re-use their Ansible scripts/roles, and make it a single line install, all the better :) |
| [01:27:59] | <cweagans> | Also true :) |
| [01:28:26] | * thunderWilly has joined #aegir |
| [01:28:52] | <ergonlogic> | as well-documented as these are, there are still a lot of docs involved |
| [01:29:14] | <ergonlogic> | making some opinionated decisions seem justified, at this point, to help bootstrap the process |
| [01:29:38] | <ergonlogic> | as long as we're aware that's what we're doing, and plan to allow for more flexibility down the road |
| [01:29:58] | * e-anima has quit (Ping timeout: 260 seconds) |
| [01:30:30] | <ergonlogic> | also, automating that deployment is something I feel that I can help with fairly immediately |
| [01:30:43] | <ergonlogic> | so, I'm not asking anyone else to do it |
| [01:30:51] | <cweagans> | Fair enough :) |
| [01:30:55] | <ergonlogic> | just annoincing my own interest is participating at that level |
| [01:31:13] | <ergonlogic> | I figure I'll get much more familiar with Kube's internals in the process |
| [01:32:11] | <ergonlogic> | so, I guess, to wrap this up, I plan to do some more organizational stuff around financing Aegir3 support, etc. |
| [01:32:31] | <ergonlogic> | contact mlhess for a demo (anyon interested in participating, let me know) |
| [01:32:42] | * cweagans is interested. |
| [01:32:44] | <ergonlogic> | but I'll see if he's open to recording it via hangout |
| [01:32:59] | <ergonlogic> | cweagans: I assumed so :p |
| [01:33:34] | <ergonlogic> | I'll release Rán, with all it's prototype warts |
| [01:34:04] | <ergonlogic> | and start looking to bootstrap kubernetes single command install via Vagrant or something |
| [01:34:43] | <ergonlogic> | that last one will be ongoing, and so probably not done in the coming week |
| [01:42:41] | <ergonlogic> | hefring: log pointer? |
| [01:42:41] | <hefring> | http://hefring.mig5.net/bot/log/aegir/2015-08-06#T582045 |
| [01:45:58] | * fatguylaughing has joined #aegir |
| [02:20:40] | * fatguylaughing has quit (Quit: fatguylaughing) |
| [02:48:53] | * cweagans is now known as cweagans|afk |
| [02:57:49] | * fatguylaughing has joined #aegir |
| [03:01:11] | * cweagans|afk has quit (Quit: Computer has gone to sleep.) |
| [03:22:44] | * mwisner has quit (Remote host closed the connection) |
| [03:27:52] | * stijnvbrande has quit (Quit: Connection closed for inactivity) |
| [03:31:25] | * mwisner has joined #aegir |
| [03:46:21] | * mwisner has quit (Remote host closed the connection) |
| [03:46:37] | * mwisner has joined #aegir |
| [03:51:00] | * gusaus has joined #aegir |
| [03:54:21] | * mwisner has quit (Remote host closed the connection) |
| [03:54:37] | * mwisner has joined #aegir |
| [04:24:03] | * thunderWilly has quit (Quit: reallife not found) |
| [04:24:11] | * e-anima has joined #aegir |
| [04:47:25] | * cweagans has joined #aegir |
| [04:59:36] | * mwisner has quit (Remote host closed the connection) |
| [04:59:50] | * mlhess has joined #aegir |
| [05:00:08] | <mlhess> | ergonlogic cweagans hello |
| [05:00:15] | <cweagans> | ohai! |
| [05:06:14] | <ergonlogic> | Hi! |
| [05:06:14] | <hefring> | what's up |
| [05:09:49] | * mwisner has joined #aegir |
| [05:10:31] | * _formatC_vt has joined #aegir |
| [05:14:21] | * formatC^vt has quit (Ping timeout: 252 seconds) |
| [05:18:19] | <mwisner> | I've been unable to find much information on this... Wondering if anyone has done anything with running the Aegir task daemon on a separate vm? Or hopefully adding workers across multiple vms? |
| [05:18:54] | <cweagans> | I certainly haven't seen that. ergonlogic might know? |
| [05:21:38] | * xurizaemon has joined #aegir |
| [05:23:07] | * niccolox has joined #aegir |
| [05:24:14] | <ergonlogic> | mlhess: have you ever worked with https://www.drupal.org/project/guardr ? It appears to put a keen focus on security |
| [05:24:30] | <ergonlogic> | mwisner: what'd you have in mind? |
| [05:25:03] | <ergonlogic> | the daemon isn't really designed for that |
| [05:25:20] | <ergonlogic> | Aegir currently works on a hub-and-spoke model |
| [05:25:58] | <ergonlogic> | so the daemon usually only runs on the master (hub) server |
| [05:26:21] | <mwisner> | Okay thanks, figured I ask. I wrote a little module that uses advancedqueue for task processing and was able to use that + supervisor to get multiple workers on one machine. Seems to be working okay, just would like to increase the number of workers I've got going, and help reduce the load on the primary aegir box. |
| [05:27:06] | <cweagans> | mwisner: that sounds cool! Are you going to open source that? I know somebody that might be interested. |
| [05:27:10] | <ergonlogic> | what workloads are you talking about, exactly? |
| [05:27:35] | <ergonlogic> | I'm interested, for one! |
| [05:28:22] | <ergonlogic> | but I'm curious wat benefit you're seeing |
| [05:29:01] | <ergonlogic> | the daemon just polls the task queue, and triggers drush/provision commands |
| [05:29:57] | <mwisner> | cweagans: I have it up as a sandbox right now in my drupal profile. However I haven't gone through the "becoming a real module" process yet. Wasn't sure if anyone would actually use it. https://www.drupal.org/sandbox/mwisner/2480511 still a WIP but we've been using it. |
| [05:30:35] | <mwisner> | We're coming out with a new SaaS product that uses aegir + d7. |
| [05:32:17] | <mwisner> | We have one client right now with a network of 4.4k~ sites on one drupal platform and ended up rolling our own non-aegir based production environment. Which is not very awesome. For our next project we are taking full advantage of aegir but what we've learned is that it takes a long time to drush updb 4.4k websites. |
| [05:33:13] | <cweagans> | mwisner: have you seen any of the Aegir 4 planning that we've been working on? It might be of interest to you. |
| [05:33:28] | <cweagans> | it won't speed things up on that many sites, but it might resolve some of your headaches |
| [05:33:48] | * xurizaemon has quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| [05:34:31] | <mwisner> | Yeah I've been following the docs on hackpad. |
| [05:35:35] | <gboudrias> | Wow, I don't think I know anyone else hosting 4k websites, you might be in a pretty unique position :) |
| [05:37:28] | <mwisner> | Aegir 4 looks interesting, I've been working on reading up on the proposed tech. I'll admit I'm not up to par with my Docker solutions =( |
| [05:37:57] | <ergonlogic> | to run remotely, wouldn't you need Provision installed on remotes and contexts (drush aliases) replicated? |
| [05:39:15] | <mwisner> | ergonlogic: Yeah that's what I was going to experiment with. Maybe sharing those dirs using NFS across multiple vms. |
| [05:39:59] | <cweagans> | ergonlogic: doesn't the pack module have you mount /var/aegir on all the systems in the pack? |
| [05:40:04] | <ergonlogic> | mwisner: this might be of interest: https://www.drupal.org/node/2037993 |
| [05:40:05] | <hefring> | https://www.drupal.org/node/2037993 => Make remote servers "smart" [#2037993] => 3 comments, 1 IRC mention |
| [05:40:20] | <ergonlogic> | I *think* webpack only mounts platforms that way |
| [05:40:25] | <ergonlogic> | I'd have to check |
| [05:40:41] | <cweagans> | IIRC, it's the entire /var/aegir dir, cause that's why we didn't use it @ Stanford |
| [05:41:03] | <ergonlogic> | hmm, you may be right |
| [05:41:16] | <ergonlogic> | we need ~/config for the vhosts, etc |
| [05:41:41] | <ergonlogic> | let me see if I can track down a webpack deployment, so I can poke around |
| [05:43:41] | <mwisner> | Yeah that issue seems pretty spot on |
| [05:43:43] | <niccolox> | if Aegir could federate, using smart remotes |
| [05:43:43] | <hefring> | niccolox: 4 days 6 hours ago <ergonlogic> tell niccolox slack is slick and all, but I prefer irc too |
| [05:44:55] | <niccolox> | you could create a Drupal site local, pass it to a small vps, then push it to koumbit, and if the site becomes internet famous could be pushed to a cweagons Aegir NG planet scale container app |
| [05:45:46] | <niccolox> | it might suit the Aegir ecosystem if there was federated trust remotes between hosters of various scales |
| [05:45:50] | <ergonlogic> | so, no, it looks like only /var/aegir/platforms is mounted via nfs |
| [05:45:50] | <niccolox> | crazy idea |
| [05:46:11] | <ergonlogic> | but /var/aegir/config also exists on webpack nodes |
| [05:46:16] | * anarcat waves hi |
| [05:46:25] | <ergonlogic> | mwisner: so it could work |
| [05:46:30] | <ergonlogic> | hey, anarcat |
| [05:46:38] | <ergonlogic> | we were just talking webpack |
| [05:46:53] | <ergonlogic> | mwisner: has come up with a remote queue daemon |
| [05:50:26] | <ergonlogic> | how much load is actually php though? I've found that i/o on mysql calls to often be the bottleneck |
| [05:51:09] | <ergonlogic> | at least, back when we were regularly running into queue daemon issues |
| [05:55:03] | <mwisner> | Well that's a good point, I haven't got to testing the mysql load yet. But on our server we currently have 4 queue workers running, I tried 8 but with 8 I kept getting "maximum amount of memory usage is above 50%, restarting process" drush messages and I would guess that the migrations took about 2x as long. |
| [05:55:08] | * e-anima has quit (Ping timeout: 260 seconds) |
| [05:56:53] | * e-anima has joined #aegir |
| [05:57:04] | <ergonlogic> | since we moved to nice/ionice for the daemon, we haven't seen the behaviour we used to, where subsequent backups would eat up the i/o and eventually crash mysql |
| [05:59:46] | <ergonlogic> | the queue daemon wasn't intended to be run concurrently, as we'd prioritized reliability over speed |
| [06:00:11] | <ergonlogic> | and it is general purpose |
| [06:00:48] | <ergonlogic> | an optimized solution for huge site farms like you're describing would be a welcome addition |
| [06:09:23] | <mwisner> | Yeah I couldn't get standard queue daemon to work concurrently. The advancedqueue module seems to work okay for concurrent tasks. I'm not an approved module maintainer but perhaps I will clean it up and go through the module approval process. |
| [06:09:40] | <mwisner> | Anyway, thanks for the help! I'll dig in and see what I can come up with. |
| [06:10:15] | <ergonlogic> | mwisner: please keep us in the loop |
| [06:10:42] | <mwisner> | Will do. |
| [06:12:04] | * neoxavier has quit (Ping timeout: 260 seconds) |
| [06:14:36] | <cweagans> | mwisner: if you submit a project application, let me know. I'll take a look. I just approved like 200 of them at NYCcamp :P |
| [06:31:26] | * mstenta has quit (Ping timeout: 240 seconds) |
| [06:41:24] | * mstenta has joined #aegir |
| [06:42:42] | * mstenta has quit (Client Quit) |
| [06:43:22] | * mstenta has joined #aegir |
| [06:55:26] | * cweagans has quit (Quit: Gone.) |
| [07:30:45] | * mwisner has quit (Remote host closed the connection) |
| [07:54:10] | * ratioweb has quit (Ping timeout: 245 seconds) |
| [08:01:02] | * realityloopAFK is now known as realityloop |
| [08:02:00] | * fatguylaughing has quit (Quit: fatguylaughing) |
| [08:02:34] | * zombiebeard has quit (Quit: zombiebeard) |
| [08:25:19] | * mwisner has joined #aegir |
| [08:29:03] | * gusaus has quit (Quit: gusaus) |
| [08:48:09] | * gusaus has joined #aegir |
| [09:13:31] | <gboudrias> | helmo: Any chance we can put the workaround in https://www.drupal.org/node/2347557 in a bash script? I don't know enough about Debian packages to make a patch |
| [09:13:32] | <hefring> | https://www.drupal.org/node/2347557 => No systemd service file for queue runner [#2347557] => 11 comments, 3 IRC mentions |
| [09:14:53] | * e-anima has quit (Read error: Connection reset by peer) |
| [09:49:31] | * Egyptian[Home] has quit (Ping timeout: 246 seconds) |
| [09:57:31] | * realityloop is now known as realityloopAFK |
