| *** weezS has quit IRC | 00:02 | |
| *** thorst has joined #openstack-ansible | 00:03 | |
| *** rueben_ has quit IRC | 00:09 | |
| *** sacharya has quit IRC | 00:09 | |
| *** serverascode has quit IRC | 00:17 | |
| *** bryan_att has quit IRC | 00:17 | |
| *** sigmavirus24 has quit IRC | 00:17 | |
| *** erikmwilson has quit IRC | 00:18 | |
| *** wadeholler has joined #openstack-ansible | 00:18 | |
| *** serverascode has joined #openstack-ansible | 00:19 | |
| *** bryan_att has joined #openstack-ansible | 00:20 | |
| *** erikmwilson has joined #openstack-ansible | 00:24 | |
| *** sigmavirus24 has joined #openstack-ansible | 00:24 | |
| *** kstev has quit IRC | 00:32 | |
| *** jamesdenton has quit IRC | 00:34 | |
| *** sacharya has joined #openstack-ansible | 00:44 | |
| *** smatzek has joined #openstack-ansible | 00:55 | |
| *** hybridpollo has quit IRC | 01:02 | |
| *** weezS has joined #openstack-ansible | 01:04 | |
| *** hybridpollo has joined #openstack-ansible | 01:13 | |
| *** shanemcd has joined #openstack-ansible | 01:13 | |
| *** shanemcd has left #openstack-ansible | 01:13 | |
| *** hybridpollo has quit IRC | 01:21 | |
| *** sdake has joined #openstack-ansible | 01:25 | |
| *** jorge_munoz has quit IRC | 01:26 | |
| *** wadeholler has quit IRC | 01:29 | |
| *** mummer has quit IRC | 01:29 | |
| *** mummer has joined #openstack-ansible | 01:30 | |
| *** mummer has quit IRC | 01:30 | |
| *** klamath has joined #openstack-ansible | 01:32 | |
| *** klamath_ has quit IRC | 01:32 | |
| *** klamath has quit IRC | 01:37 | |
| *** klamath has joined #openstack-ansible | 01:38 | |
| *** weezS has quit IRC | 01:49 | |
| *** weezS has joined #openstack-ansible | 01:49 | |
| *** jthorne_ has joined #openstack-ansible | 02:00 | |
| *** jthorne has quit IRC | 02:00 | |
| *** smatzek has quit IRC | 02:13 | |
| *** weezS has quit IRC | 02:25 | |
| *** cloudtrainme has joined #openstack-ansible | 02:44 | |
| *** cloudtrainme has quit IRC | 02:49 | |
| openstackgerrit | Darren Chan proposed openstack/openstack-ansible: [docs] Move minor upgrade content https://review.openstack.org/328056 | 03:03 |
|---|---|---|
| *** markvoelker has quit IRC | 03:14 | |
| *** zerda2 has joined #openstack-ansible | 03:16 | |
| *** saneax is now known as saneax_AFK | 03:27 | |
| openstackgerrit | Alexey Dushechkin proposed openstack/openstack-ansible-os_horizon: Add Nova extensions blacklisting https://review.openstack.org/327230 | 03:34 |
| *** iceyao has joined #openstack-ansible | 03:36 | |
| *** sdake has quit IRC | 03:41 | |
| *** iceyao has quit IRC | 03:42 | |
| *** asettle has joined #openstack-ansible | 03:50 | |
| *** thorst has quit IRC | 03:51 | |
| *** thorst has joined #openstack-ansible | 03:51 | |
| *** asettle has quit IRC | 03:55 | |
| *** thorst has quit IRC | 04:00 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-os_horizon: Consistency for multi-os in the includes https://review.openstack.org/324514 | 04:15 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-os_horizon: Remove pip_lock_down dependency https://review.openstack.org/313888 | 04:15 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-os_sahara: Remove pip_lock_down dependency https://review.openstack.org/326820 | 04:18 |
| openstackgerrit | Merged openstack/openstack-ansible-security: Add release note for V-38524 implementation https://review.openstack.org/326467 | 04:20 |
| *** saneax_AFK is now known as saneax | 04:30 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_magnum: Adding api-paste.ini https://review.openstack.org/327940 | 04:31 |
| *** sdake has joined #openstack-ansible | 04:45 | |
| *** thorst has joined #openstack-ansible | 04:57 | |
| openstackgerrit | Merged openstack/openstack-ansible-security: Add check/audit to gate testing https://review.openstack.org/324482 | 04:57 |
| openstackgerrit | Merged openstack/openstack-ansible-os_horizon: Consistency for multi-os in the includes https://review.openstack.org/324514 | 04:58 |
| openstackgerrit | Merged openstack/openstack-ansible-security: Restart auditd after running augenrules https://review.openstack.org/327863 | 05:03 |
| *** thorst has quit IRC | 05:04 | |
| *** winggundamth has quit IRC | 05:06 | |
| openstackgerrit | Darren Chan proposed openstack/openstack-ansible: [docs] Migrate ops and use case content https://review.openstack.org/328074 | 05:09 |
| *** markvoelker has joined #openstack-ansible | 05:15 | |
| *** asettle has joined #openstack-ansible | 05:19 | |
| *** markvoelker has quit IRC | 05:20 | |
| *** asettle has quit IRC | 05:24 | |
| *** javeriak has joined #openstack-ansible | 05:28 | |
| *** jamielennox is now known as jamielennox|away | 05:29 | |
| *** winggundamth has joined #openstack-ansible | 05:48 | |
| odyssey4me | o/ | 05:51 |
| *** javeriak has quit IRC | 05:55 | |
| *** asettle has joined #openstack-ansible | 05:59 | |
| *** thorst has joined #openstack-ansible | 06:02 | |
| *** asettle has quit IRC | 06:03 | |
| *** thorst has quit IRC | 06:10 | |
| odyssey4me | mcarden mrda If you could review https://review.openstack.org/327104 it'd be great to have some more feedback. | 06:12 |
| *** lunarlamp has joined #openstack-ansible | 06:17 | |
| mcarden | odyssey4me: Can do... er maybe after you make Alex happy. :) | 06:18 |
| openstackgerrit | Darren Chan proposed openstack/openstack-ansible: [doc] Migrate deployment configuration options https://review.openstack.org/328095 | 06:21 |
| odyssey4me | mcarden I'm more interested in feedback on the content's meaning than the grammar. :) | 06:26 |
| odyssey4me | The grammar is easy enough to fix. Whether the content makes sense, is the right stuff and has value to the role development process is important. | 06:27 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Docs: Add role development maturity guidelines https://review.openstack.org/327104 | 06:31 |
| *** openstackgerrit has quit IRC | 06:32 | |
| *** openstackgerrit has joined #openstack-ansible | 06:33 | |
| *** pcaruana has joined #openstack-ansible | 06:40 | |
| *** bsv has joined #openstack-ansible | 06:40 | |
| mrda | odyssey4me: ack | 06:40 |
| bsv | moinmoin | 06:41 |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build: Consistency for multi-os in the includes https://review.openstack.org/324520 | 06:44 |
| *** alikins has quit IRC | 06:46 | |
| *** javeriak has joined #openstack-ansible | 06:48 | |
| odyssey4me | adreznec With regards to the mirror support for ppc64le the only reason the rackspace mirror is used is because we picked up rate limiting on the more official mirrors which killed us on gate checks. If there's another mirror we can use then it'll be fine for a change in default, or perhaps a default specifically for the platform. I'd also be happy to explore the option for Xenial to simply use built-in packages | 06:50 |
| odyssey4me | from Xenial if that's an option. | 06:50 |
| *** jiteka has joined #openstack-ansible | 06:52 | |
| *** jiteka has quit IRC | 06:57 | |
| mcarden | odyssey4me: kk - content vs grammar. | 07:00 |
| odyssey4me | adreznec I've added some feedback to https://etherpad.openstack.org/p/osa-multi-arch-repo | 07:01 |
| odyssey4me | mcarden How's your exploration of Designate going? Are you getting anywhere near making the role actually work yet? | 07:03 |
| mcarden | odyssey4me: For very small values of 'work'... yes. I see 5 of the 4 'core | 07:04 |
| bsv | i found the last 'sink bug' in my designate playbook last night \o/ | 07:04 |
| mcarden | ' processes running | 07:04 |
| odyssey4me | mcarden that sounds like progress | 07:05 |
| mcarden | I have more luck with the ansible role than I do with manually seting it up. | 07:05 |
| odyssey4me | bsv has his own playbooks, but can probably point you to them and help you make stuff work - hopefully it can feed back into our role | 07:05 |
| bsv | core designate was pretty easy to get up, it was the auto-generating a-records that cause me many tears ;) | 07:06 |
| mcarden | odyssey4me: I confess I have been looking at other things and not concentrated on designate for a couple of days | 07:06 |
| *** thorst has joined #openstack-ansible | 07:07 | |
| mcarden | Nolan gave me some valuable insight on the network naming issue in os_tempest and i have been looking at that | 07:08 |
| * mcarden heads off to keep family on side | 07:09 | |
| bsv | mcarden: if you want to share info on the designate part, give me a poke. | 07:10 |
| *** permalac has joined #openstack-ansible | 07:15 | |
| *** thorst has quit IRC | 07:15 | |
| *** markvoelker has joined #openstack-ansible | 07:16 | |
| *** markvoelker has quit IRC | 07:21 | |
| *** asettle has joined #openstack-ansible | 07:21 | |
| *** sacharya has quit IRC | 07:22 | |
| *** jthorne has joined #openstack-ansible | 07:30 | |
| *** jthorne_ has quit IRC | 07:31 | |
| *** agireud has quit IRC | 07:32 | |
| *** openstackgerrit has quit IRC | 07:33 | |
| *** openstackgerrit has joined #openstack-ansible | 07:33 | |
| *** asettle has quit IRC | 07:34 | |
| *** agireud has joined #openstack-ansible | 07:34 | |
| *** jamielennox|away is now known as jamielennox | 07:36 | |
| *** Oku_OS-away is now known as Oku_OS | 07:41 | |
| openstackgerrit | Merged openstack/openstack-ansible-security: Add initial support for Red Hat Enterprise Linux 7 https://review.openstack.org/325421 | 07:43 |
| openstackgerrit | Merged openstack/openstack-ansible-security: Add key fields to audit rules https://review.openstack.org/327872 | 07:45 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Remove libvirt bootstrapping from AIO https://review.openstack.org/327276 | 07:55 |
| *** karimb has joined #openstack-ansible | 07:56 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_gnocchi: Replace all uses of run_once https://review.openstack.org/327873 | 08:01 |
| evrardjp | good morning everyone | 08:10 |
| *** jiteka has joined #openstack-ansible | 08:11 | |
| *** thorst has joined #openstack-ansible | 08:13 | |
| *** markvoelker has joined #openstack-ansible | 08:17 | |
| *** thorst has quit IRC | 08:20 | |
| *** markvoelker has quit IRC | 08:21 | |
| *** sdake has quit IRC | 08:21 | |
| *** javeriak has quit IRC | 08:22 | |
| *** sacharya has joined #openstack-ansible | 08:22 | |
| *** javeriak has joined #openstack-ansible | 08:25 | |
| *** sacharya has quit IRC | 08:27 | |
| odyssey4me | o/ evrardjp | 08:38 |
| *** javeriak has quit IRC | 08:41 | |
| *** javeriak has joined #openstack-ansible | 08:41 | |
| evrardjp | will read yesterday's meeting notes, sorry for my absence | 08:45 |
| *** javeriak has quit IRC | 08:46 | |
| *** javeriak has joined #openstack-ansible | 08:46 | |
| *** asettle has joined #openstack-ansible | 08:52 | |
| *** asettle has quit IRC | 08:52 | |
| *** asettle has joined #openstack-ansible | 08:53 | |
| *** kamsz has joined #openstack-ansible | 08:56 | |
| kamsz | hi there, is ubuntu 16.04 supported with stable/mitaka? | 08:57 |
| evrardjp | kamsz: it shouldn't | 09:01 |
| evrardjp | most of the 16.04 work is done during this newton cycle | 09:01 |
| evrardjp | these changes are too big IMO to be backported | 09:02 |
| evrardjp | but I guess you have to talk about that with odyssey4me :D | 09:02 |
| kamsz | evrardjp: bah, okay | 09:02 |
| kamsz | i guess trying to do it with 16.04 will be painful road | 09:03 |
| evrardjp | for mitaka? | 09:03 |
| evrardjp | yes it will | 09:03 |
| odyssey4me | kamsz yes, evrardjp is right - we're targeting 16.04 support for Newton... 16.04 support is a lot of work thanks to the switch from init to systemd | 09:03 |
| evrardjp | but nothing prevents you from running master | 09:03 |
| evrardjp | it's just not stable | 09:03 |
| kamsz | :D | 09:04 |
| evrardjp | but working | 09:04 |
| evrardjp | michaelgugino: There are a few repos that have issues with their tempest testing, I'm starting to work on this, but I don't know much about tempest... | 09:11 |
| openstackgerrit | Darren Birkett proposed openstack/openstack-ansible: Make ansible managed line in swift.conf static https://review.openstack.org/326374 | 09:11 |
| openstackgerrit | Darren Birkett proposed openstack/openstack-ansible: Make ansible managed line in swift.conf static https://review.openstack.org/326374 | 09:14 |
| *** thorst has joined #openstack-ansible | 09:18 | |
| asettle | evrardjp: how are you feeling?! | 09:20 |
| *** javeriak has quit IRC | 09:25 | |
| *** thorst has quit IRC | 09:25 | |
| odyssey4me | we're getting a lot of integrated gate timeouts lately - it looks like it started on the 6th and escalated on the 8th, although that may related to the number of tests escalating on the 8th | 09:36 |
| *** thorst has joined #openstack-ansible | 09:37 | |
| evrardjp | when did 2.1 merged? | 09:40 |
| evrardjp | the 7th | 09:41 |
| evrardjp | so not really the source but could be the trigger to the escalation | 09:41 |
| evrardjp | like any other commit | 09:41 |
| evrardjp | do you have an example commit I can check odyssey4me? | 09:42 |
| *** javeriak has joined #openstack-ansible | 09:43 | |
| openstackgerrit | Darren Birkett proposed openstack/openstack-ansible-os_swift: Make ansible managed line in swift.conf static https://review.openstack.org/326369 | 09:44 |
| evrardjp | does openstack have an url shortener service? | 09:49 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible: Fail Fast when trying to upgrade with LBaaS v1 enabled https://review.openstack.org/324728 | 09:55 |
| *** thorst has quit IRC | 09:56 | |
| *** thorst has joined #openstack-ansible | 09:57 | |
| *** thorst has quit IRC | 10:01 | |
| *** iceyao has joined #openstack-ansible | 10:04 | |
| *** sacharya has joined #openstack-ansible | 10:05 | |
| *** iceyao has quit IRC | 10:09 | |
| *** sacharya has quit IRC | 10:10 | |
| odyssey4me | evrardjp 4 rechecks, across multiple providers, all timed out: https://review.openstack.org/273793 | 10:13 |
| evrardjp | ok will look at this | 10:14 |
| odyssey4me | we've never had timeouts on rax providers, but we're now getting them - so something is really off | 10:14 |
| evrardjp | but I got issues with these before the 4th | 10:14 |
| evrardjp | 6th* | 10:14 |
| evrardjp | with rechecks* | 10:14 |
| odyssey4me | before it was usually OVH providers | 10:14 |
| evrardjp | true | 10:14 |
| odyssey4me | it might relate to us adding UCA by default in the nova role, I can try disabling that temporarily while we arrange to have it archived in the mirrors | 10:15 |
| evrardjp | we have more and more roles to the integrated build too, but that's another question :p | 10:16 |
| evrardjp | I'm reading the log | 10:16 |
| evrardjp | just a sec | 10:16 |
| evrardjp | it didn't even reach tempest... yes that's bad | 10:17 |
| evrardjp | I'm afraid it's something else | 10:18 |
| *** markvoelker has joined #openstack-ansible | 10:18 | |
| evrardjp | you remember you had stats per tasks? | 10:18 |
| evrardjp | did the setup to gather facts take one minute in the past? | 10:19 |
| evrardjp | because that starts to add up | 10:19 |
| evrardjp | uca + nova apt takes 83 seconds, I don't know how much it was before | 10:21 |
| *** markvoelker has quit IRC | 10:22 | |
| odyssey4me | evrardjp sure, but we're only executing a limited set of them | 10:23 |
| openstackgerrit | Darren Birkett proposed openstack/openstack-ansible-os_swift: Make ansible managed line in swift.conf static https://review.openstack.org/326369 | 10:23 |
| odyssey4me | but I did see a build executing some ironic stuff, so I need to check on that one | 10:23 |
| evrardjp | that's the one I thought about :D | 10:23 |
| openstackgerrit | Darren Birkett proposed openstack/openstack-ansible: Make ansible managed line in swift.conf static https://review.openstack.org/326374 | 10:24 |
| evrardjp | but anyway it shouldn't be a issue, we had room IIRC | 10:24 |
| evrardjp | I'm just parsing the logs right now, and I am asking myself these questions above, I'm not having a general overview right now of what could be the cause... Just still parsing :D | 10:24 |
| *** johnmilton has quit IRC | 10:25 | |
| evrardjp | the install pip packages should be better instrumented, this way we know each package takes x seconds | 10:26 |
| evrardjp | there is one that takes 18 seconds on its own, I wonder which one it is, and why (compared to the others) http://logs.openstack.org/93/273793/20/check/gate-openstack-ansible-dsvm-commit/b95622b/console.html#_2016-06-10_08_37_17_639 | 10:27 |
| openstackgerrit | Darren Birkett proposed openstack/openstack-ansible: Make ansible managed line in swift.conf static https://review.openstack.org/326374 | 10:28 |
| openstackgerrit | Darren Birkett proposed openstack/openstack-ansible-os_swift: Make ansible managed line in swift.conf static https://review.openstack.org/326369 | 10:29 |
| odyssey4me | evrardjp unfortunately even on the same provider the timings aren't always the same - and a lot of the install times vary due to underlying IO issues | 10:29 |
| evrardjp | that's the feeling I have when I read the logs | 10:30 |
| evrardjp | here I see that installing the python-keystoneclient on the host takes 10 seconds | 10:30 |
| evrardjp | or 30 seconds sometimes | 10:32 |
| evrardjp | I'll check other logs | 10:33 |
| bsv | dont get me started on osa install time :) | 10:34 |
| evrardjp | It's not about that, I'm just parsing the logs, and I find patterns on the log, I'll then check across multiple logs to see if the patterns match | 10:35 |
| evrardjp | just curious | 10:35 |
| evrardjp | did we enable apache cache on repo_server ? | 10:39 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Temporarily disable UCA usage in OpenStack-CI https://review.openstack.org/328234 | 10:42 |
| odyssey4me | evrardjp the repo server uses nginx | 10:42 |
| odyssey4me | evrardjp ^ that's a tester to recheck a few times to see if it helps | 10:43 |
| evrardjp | ok | 10:43 |
| evrardjp | good | 10:43 |
| odyssey4me | evrardjp another one that's just not getting through https://review.openstack.org/327262 | 10:44 |
| *** admin0 has joined #openstack-ansible | 10:45 | |
| evrardjp | I was on this one: https://review.openstack.org/#/c/327276/ | 10:45 |
| odyssey4me | evrardjp perhaps you meant https://review.openstack.org/327262 ? | 10:54 |
| evrardjp | I'm briefly checking them all | 10:57 |
| evrardjp | when you see this, you know it's not good: http://logs.openstack.org/69/327269/1/check/gate-openstack-ansible-dsvm-commit/a0043f6/console.html#_2016-06-08_20_01_05_678 | 10:58 |
| evrardjp | 1 minute to get facts, really | 10:58 |
| evrardjp | and it's like that everywhere | 10:58 |
| evrardjp | I wonder if we shouldn't create a playbook that gather facts when all the containers are created, and boom skip the setup part later | 10:59 |
| evrardjp | it's super dangerous | 10:59 |
| odyssey4me | yeah, I'm not a fan - is that some sort of side-effect of upgrading to Ansible 2.1? | 11:02 |
| odyssey4me | on the OSIC nodes we seem to still be getting the same sort of time results, so this may actually be a resource contention issue in the cloud providers | 11:02 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Actually remove Ironic container creation from AIO preparation https://review.openstack.org/328240 | 11:08 |
| odyssey4me | evrardjp ^ found the reason that ironic was still being deployed in the gate | 11:09 |
| evrardjp | I thought it was on purpose | 11:11 |
| *** smatzek has joined #openstack-ansible | 11:11 | |
| evrardjp | why shouldn't we remove both env.d and conf.d in the gates? it seems cleaner | 11:13 |
| odyssey4me | evrardjp the env.d file is necessary for the groups to exist in the inventory | 11:14 |
| evrardjp | I know :D | 11:14 |
| odyssey4me | there's no harm in the group being in the inventory | 11:14 |
| evrardjp | but we don't need that ever if never run a playbook that targets the group | 11:14 |
| odyssey4me | it'll mean that the playbook can run, but the group will have no members | 11:14 |
| *** smatzek has quit IRC | 11:14 | |
| odyssey4me | we will be running setup-openstack.yml eventually, which will run os-ironic-install, which targets the group | 11:15 |
| evrardjp | yes indeed, but sometimes don't we do operation based on group existance, instead of group length > 0 ? | 11:15 |
| evrardjp | just asking, not sure | 11:15 |
| odyssey4me | we shouldn't be | 11:16 |
| *** johnmilton has joined #openstack-ansible | 11:16 | |
| *** bsv has quit IRC | 11:19 | |
| *** McMurlock1 has joined #openstack-ansible | 11:27 | |
| *** deadnull_ has joined #openstack-ansible | 11:28 | |
| *** javeriak has quit IRC | 11:31 | |
| *** javeriak has joined #openstack-ansible | 11:32 | |
| *** javeriak_ has joined #openstack-ansible | 11:40 | |
| *** javeriak has quit IRC | 11:40 | |
| *** vnogin has quit IRC | 11:43 | |
| *** pcaruana has quit IRC | 11:51 | |
| *** karimb has quit IRC | 12:05 | |
| *** pcaruana has joined #openstack-ansible | 12:07 | |
| *** zerda2 has quit IRC | 12:07 | |
| *** karimb has joined #openstack-ansible | 12:08 | |
| mhayden | morning folks | 12:10 |
| *** pcaruana has quit IRC | 12:13 | |
| *** pcaruana has joined #openstack-ansible | 12:13 | |
| *** weshay has joined #openstack-ansible | 12:18 | |
| *** markvoelker has joined #openstack-ansible | 12:18 | |
| *** afred312 has quit IRC | 12:23 | |
| odyssey4me | mhayden thanks for handling the meeting yesterday | 12:23 |
| odyssey4me | I had actually asked automagically to cover for me, but it looks like he missed it :/ | 12:23 |
| odyssey4me | mhayden would you be ok with taking point for the thu meetings from now on, just to spread some of my workload a bit? | 12:24 |
| mhayden | i can probably do that | 12:24 |
| odyssey4me | you can arrange for an alternative chair whenever you're not available | 12:24 |
| odyssey4me | or organise a rotation if you like - up to you | 12:25 |
| mhayden | sure, i'll need help for the week of the 27th | 12:25 |
| odyssey4me | mhayden add it for the meeting, then someone can volunteer then | 12:29 |
| odyssey4me | afk for a bit | 12:31 |
| *** javeriak_ has quit IRC | 12:33 | |
| *** psilvad has joined #openstack-ansible | 12:36 | |
| *** dmsimard is now known as dmsimard|afk | 12:37 | |
| *** karimb has quit IRC | 12:37 | |
| *** psilvad has quit IRC | 12:44 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible: [PoC] avoid gathering facts in the run_playbooks/gate https://review.openstack.org/328274 | 12:47 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible: [PoC] Speed up gate: avoid gathering facts https://review.openstack.org/328274 | 12:48 |
| evrardjp | odyssey4me: this is a PoC, don't worry if it doesn't respect your other changes, it could be easily adapted | 12:50 |
| *** kylek3h has joined #openstack-ansible | 12:50 | |
| *** psilvad has joined #openstack-ansible | 12:51 | |
| *** qba73 has joined #openstack-ansible | 12:52 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Restart auditd after running augenrules https://review.openstack.org/328279 | 12:55 |
| *** psilvad has quit IRC | 12:55 | |
| *** kstev has joined #openstack-ansible | 12:56 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-galera_client: [WIP] Add dependency to the openstack-ansible-plugin role https://review.openstack.org/321582 | 12:56 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-galera_client: [WIP] Add dependency to the openstack-ansible-plugin role https://review.openstack.org/321582 | 12:56 |
| *** deadnull_ is now known as _deadnull | 12:57 | |
| *** psilvad has joined #openstack-ansible | 12:59 | |
| *** messy has joined #openstack-ansible | 13:00 | |
| *** BjoernT has joined #openstack-ansible | 13:00 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Add key fields to audit rules https://review.openstack.org/328288 | 13:00 |
| automagically | Morning all | 13:03 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible: [PoC] Speed up gate: avoid gathering facts https://review.openstack.org/328274 | 13:04 |
| automagically | odyssey4me: Sorry for missing the meeting yesterday, work took over | 13:04 |
| *** javeriak has joined #openstack-ansible | 13:09 | |
| *** rljefferson has joined #openstack-ansible | 13:10 | |
| *** _deadnull is now known as deadnull_ | 13:16 | |
| *** pcaruana has quit IRC | 13:20 | |
| *** Zucan has joined #openstack-ansible | 13:21 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-galera_client: [WIP] Add dependency to the openstack-ansible-plugin role https://review.openstack.org/321582 | 13:23 |
| *** adrian_otto has joined #openstack-ansible | 13:26 | |
| *** adrian_otto has quit IRC | 13:29 | |
| *** qba73 has quit IRC | 13:35 | |
| *** javeriak has quit IRC | 13:37 | |
| *** klamath has quit IRC | 13:38 | |
| *** klamath has joined #openstack-ansible | 13:38 | |
| *** pcaruana has joined #openstack-ansible | 13:39 | |
| *** ajo_ is now known as ajo | 13:41 | |
| *** klamath has quit IRC | 13:41 | |
| *** klamath has joined #openstack-ansible | 13:42 | |
| *** chandanc has joined #openstack-ansible | 13:46 | |
| *** chandanc has quit IRC | 13:47 | |
| *** chandanc has joined #openstack-ansible | 13:47 | |
| *** afred312 has joined #openstack-ansible | 13:47 | |
| *** jmckind has joined #openstack-ansible | 13:48 | |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-memcached_server: Cleanup/standardize usage of tags https://review.openstack.org/328309 | 13:49 |
| openstackgerrit | Merged openstack/openstack-ansible-os_horizon: Remove pip_lock_down dependency https://review.openstack.org/313888 | 13:51 |
| *** jthorne has quit IRC | 13:52 | |
| *** jthorne has joined #openstack-ansible | 13:52 | |
| *** jmckind_ has joined #openstack-ansible | 13:54 | |
| *** jmckind has quit IRC | 13:56 | |
| *** zerda2 has joined #openstack-ansible | 13:56 | |
| *** KLevenstein has joined #openstack-ansible | 13:57 | |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-memcached_server: DOC - Adopting the common role documentation pattern https://review.openstack.org/328311 | 13:58 |
| *** psilvad has quit IRC | 13:59 | |
| *** ametts has joined #openstack-ansible | 14:00 | |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-memcached_server: Update meta to show support for Xenial and CentOS https://review.openstack.org/328312 | 14:01 |
| *** saneax is now known as saneax_AFK | 14:02 | |
| *** javeriak has joined #openstack-ansible | 14:02 | |
| *** Mudpuppy has joined #openstack-ansible | 14:02 | |
| *** javeriak has quit IRC | 14:05 | |
| openstackgerrit | Alexey Dushechkin proposed openstack/openstack-ansible-os_horizon: Add support for CustomLog format modification https://review.openstack.org/327808 | 14:05 |
| *** sacharya has joined #openstack-ansible | 14:06 | |
| odyssey4me | mhayden can you peek at https://review.openstack.org/328240 quickly? | 14:09 |
| *** sacharya has quit IRC | 14:11 | |
| automagically | odyssey4me: Thanks for your thoughtful feedback on https://review.openstack.org/327962 | 14:15 |
| odyssey4me | automagically :) I want to solve that problem too. Thanks for raising it. It needs to stay on the radar and I'm very happy that this has come up as feedback from a deployer, instead of me just thinking up crazy stuff. | 14:16 |
| *** jthorne has quit IRC | 14:18 | |
| *** spotz_zzz is now known as spotz | 14:21 | |
| *** jmckind_ has quit IRC | 14:23 | |
| evrardjp | automagically: I agree with odyssey4me, and to be honest, I don't think it's a good idea to decide to separate OSA roles from the rest inside OSA. The deploy can chose himself by setting an appropriate ansible.cfg | 14:24 |
| evrardjp | also if we handle dependencies correctly, it will simplify things out | 14:24 |
| evrardjp | that's my opinion | 14:25 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible: Fail Fast when trying to upgrade with LBaaS v1 enabled https://review.openstack.org/324728 | 14:28 |
| automagically | evrardjp: Thanks. Can you add it to the review so we recall your opinion if we revisit the topic | 14:28 |
| evrardjp | not needed, odyssey4me has already expressed it in the commit :p I was glad to speak about it here to have a discussion, but it's fine for me if we don't have one too :p | 14:30 |
| automagically | ah, okay | 14:30 |
| *** jthorne has joined #openstack-ansible | 14:32 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-galera_client: Add dependency to the openstack-ansible-plugin role https://review.openstack.org/321582 | 14:32 |
| *** jorge_munoz has joined #openstack-ansible | 14:33 | |
| evrardjp | odyssey4me: what do you think of this? https://review.openstack.org/#/c/328274/ | 14:34 |
| *** jmckind has joined #openstack-ansible | 14:35 | |
| *** chandanc_ has joined #openstack-ansible | 14:36 | |
| *** iceyao has joined #openstack-ansible | 14:39 | |
| *** raddaoui has joined #openstack-ansible | 14:42 | |
| *** iceyao has quit IRC | 14:44 | |
| openstackgerrit | Amy Marrich (spotz) proposed openstack/openstack-ansible-os_heat: Implement Xenial Support https://review.openstack.org/324733 | 14:47 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_ceilometer: Cleanup/standardize usage of tags https://review.openstack.org/328330 | 14:52 |
| *** saneax_AFK is now known as saneax | 14:52 | |
| odyssey4me | evrardjp clever idea, although I'm not a huge fan of doing something so intrusive to the normal execution process - I do wonder how much it really saves | 14:53 |
| evrardjp | I'll do a few rechecks | 14:56 |
| evrardjp | for stats | 14:56 |
| evrardjp | but I know it's a question of minutes | 14:56 |
| evrardjp | not seconds | 14:56 |
| *** javeriak has joined #openstack-ansible | 14:56 | |
| evrardjp | every time we do it, we have 60 seconds lost | 14:56 |
| evrardjp | and it's not really intrusive because the default behavior for users stay the same (except for the run-playbooks, which could be moved to gate-check-commit if you prefer... it's just easier to read that way) | 14:57 |
| evrardjp | gather_facts: True is default behavior | 14:58 |
| evrardjp | or behaviour I don't know | 14:58 |
| *** javeriak has quit IRC | 14:58 | |
| *** Ashana has joined #openstack-ansible | 15:12 | |
| *** pcaruana has quit IRC | 15:13 | |
| *** adrian_otto has joined #openstack-ansible | 15:14 | |
| openstackgerrit | Merged openstack/openstack-ansible: Reduce and organize group vars https://review.openstack.org/324173 | 15:15 |
| openstackgerrit | Merged openstack/openstack-ansible: Remove the AIO metadata checksum fix from run-playbooks https://review.openstack.org/327262 | 15:15 |
| *** weezS has joined #openstack-ansible | 15:19 | |
| *** sacharya has joined #openstack-ansible | 15:25 | |
| *** Drago has joined #openstack-ansible | 15:26 | |
| *** Drago has joined #openstack-ansible | 15:26 | |
| *** Ashana has quit IRC | 15:32 | |
| openstackgerrit | Christopher Hultin proposed openstack/openstack-ansible-plugins: Adding ability to set 'properties' via Glance https://review.openstack.org/328354 | 15:40 |
| *** rljefferson has quit IRC | 15:40 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Docs: Add role development maturity guidelines https://review.openstack.org/327104 | 15:43 |
| mhayden | odyssey4me: just now saw your msg -- looking | 15:44 |
| mancdaz | odyssey4me evrardjp mhayden automagically jmccrory hughsaunders andymccr I can haz merge https://review.openstack.org/#/c/326374/ https://review.openstack.org/#/c/326369/ | 15:44 |
| mancdaz | I have obeyed the laws of release note court | 15:44 |
| automagically | haha mancdaz | 15:44 |
| * automagically looking | 15:44 | |
| mhayden | mancdaz: can you use double `` rather than single ` in the reno? | 15:46 |
| mhayden | or i can just fix it up for ya | 15:46 |
| palendae | mhayden, Should be able to | 15:46 |
| palendae | Just so long as the list items don't start with it | 15:46 |
| * mancdaz slumps into a corner and starts to softly weep | 15:46 | |
| palendae | It's just ReST | 15:46 |
| evrardjp | can't help you merge, but can help you get closer :p | 15:47 |
| palendae | mancdaz, don't ask for a merge, poor form ;) | 15:47 |
| evrardjp | what about mitaka? | 15:47 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_swift: Make ansible managed line in swift.conf static https://review.openstack.org/326369 | 15:48 |
| mhayden | oh no, i edited it after W+1 | 15:48 |
| * mhayden failed | 15:48 | |
| *** rueben_ has joined #openstack-ansible | 15:48 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Make ansible managed line in swift.conf static https://review.openstack.org/326374 | 15:48 |
| evrardjp | thanks major | 15:49 |
| * mhayden needs to just quit computers | 15:49 | |
| mancdaz | thanks mhayden | 15:49 |
| mhayden | mancdaz: sorry :( | 15:50 |
| odyssey4me | thanks for fixing that mhayden - it was bugging me, but I'm trying to be less fixated on minor details | 15:50 |
| * mhayden gets back in the corner for a timeout | 15:50 | |
| mhayden | oh man, i am so torn on 327962 | 15:50 |
| mancdaz | odyssey4me ahaha I can imagine you reading it and shuddering | 15:50 |
| mhayden | not as torn as getting Go into openstack | 15:50 |
| mhayden | but torn nonetheless | 15:51 |
| automagically | I wanted to see Golang in OpenStack so badly | 15:51 |
| evrardjp | what's the decision on this? I didn't read the ML for days now | 15:51 |
| mhayden | mancdaz: this is what odyssey4me does when he sees that -> https://gfycat.com/LividInsistentChimpanzee | 15:51 |
| mancdaz | evrardjp they said no | 15:51 |
| automagically | Not gonna happen | 15:51 |
| automagically | ;( | 15:51 |
| evrardjp | that sucks | 15:51 |
| odyssey4me | mhayden please remove the vote from https://review.openstack.org/326374 - it needs work and the master patch has not yet merged | 15:52 |
| evrardjp | politics | 15:52 |
| mhayden | got it, odyssey4me | 15:52 |
| palendae | automagically, It can get in if swift makes it a dependency and develops that depedency outside the openstack namespace | 15:52 |
| automagically | Well, yeah, but…. | 15:52 |
| palendae | Kind of like how nova relies on hypervisors but those will never (and should never) be in the namespace | 15:52 |
| odyssey4me | well, as far as I see it it's a little different than nova/hypervisors | 15:53 |
| palendae | odyssey4me, I'm sure it is :) I didn't follow the discussions because it seemed kind of pointless | 15:53 |
| odyssey4me | when nova was started there were already hypervisors, and nova is an API to the hypervisor | 15:54 |
| evrardjp | odyssey4me: It has the same changeid | 15:54 |
| odyssey4me | when swift was started, there was no underlying service to use | 15:54 |
| palendae | rsync | 15:54 |
| evrardjp | odyssey4me: but the file is indeed different, it matters for reno, or it was the first part of the name that counts? | 15:54 |
| odyssey4me | the use of the tooling in openstack should be more focused around who the development community is around a product | 15:54 |
| *** ametts has quit IRC | 15:55 | |
| mancdaz | odyssey4me https://review.openstack.org/#/q/Id00cd79636f43776aaaa51e2ae6d7c5eacff8d34,n,z | 15:55 |
| openstackgerrit | Christopher Hultin proposed openstack/openstack-ansible-plugins: Adding ability to set 'properties' via Glance https://review.openstack.org/328354 | 15:55 |
| mancdaz | am I missing something about the cnageid | 15:55 |
| odyssey4me | in this case it's a bit silly to develop it outside of openstack tooling because the only developers for it are openstack developers | 15:55 |
| mancdaz | changeid | 15:55 |
| mancdaz | odyssey4me the change for master goes directly into the swift role | 15:55 |
| odyssey4me | mancdaz my bad, the release note name is not right then | 15:55 |
| evrardjp | mancdaz: maybe the file name for the reno? | 15:56 |
| mancdaz | they are in different repos | 15:56 |
| mancdaz | why does it matter?? | 15:56 |
| odyssey4me | yep, and the backport process copies the reno file from the independent repo to the main one | 15:56 |
| evrardjp | that was my question a few lines above :D | 15:56 |
| evrardjp | is the full filename important, or just the part until the unique thingy | 15:56 |
| odyssey4me | it will help if we ever change how the release notes are published - we need ot identify by that reno change ID which reno's are the same | 15:56 |
| odyssey4me | the whole thingy | 15:57 |
| evrardjp | ok | 15:57 |
| odyssey4me | and I meant the sha bump process, not the backport process | 15:57 |
| evrardjp | I thought it was the name not the full name | 15:57 |
| palendae | evrardjp, Full name + sha, yeah | 15:57 |
| odyssey4me | but the reno file names must match to make comparisons and updates easier | 15:57 |
| odyssey4me | and the rule still stands - a fix must merge in master first before being backported | 15:57 |
| evrardjp | mancdaz: you can also backport to mitaka in the meantime :D | 15:57 |
| odyssey4me | please don't | 15:57 |
| evrardjp | or wait for master to merge | 15:58 |
| mancdaz | evrardjp no I can't | 15:58 |
| odyssey4me | wait for the merge | 15:58 |
| mancdaz | ^^ | 15:58 |
| evrardjp | yeah | 15:58 |
| evrardjp | that's better | 15:58 |
| evrardjp | but you can test without submitting ! | 15:58 |
| mhayden | odyssey4me: i added my £0.02 on that review | 15:58 |
| mancdaz | there's nothing to test | 15:58 |
| mhayden | 𝘣𝘶𝘵 𝘪 𝘢𝘮 𝘴𝘰 𝘵𝘰𝘳𝘯 𝘰𝘯 𝘵𝘩𝘢𝘵 𝘳𝘦𝘷𝘪𝘦𝘸 | 15:59 |
| mancdaz | evrardjp it's not just better to wait for master, you HAVE to wait for master to merge to do a clean backport unless you're 100% convinced that your patch will not change | 16:00 |
| odyssey4me | mhayden which one? | 16:00 |
| evrardjp | jmccrory: thanks for the group vars change btw, I love it :D | 16:00 |
| evrardjp | mancdaz: I know | 16:00 |
| mhayden | odyssey4me: on whether mancdaz should write a release note | 16:00 |
| * mhayden is just kidding | 16:00 | |
| mhayden | odyssey4me: i was talking about https://review.openstack.org/#/c/327962/ | 16:00 |
| odyssey4me | evrardjp I think that quite a few of the vars in all could actually be removed - if you feel inclined, please feel free to do so. :) | 16:00 |
| * mhayden notices that the coffee and decongestants are working well together today | 16:01 | |
| evrardjp | I will go further and create folders before that :p | 16:01 |
| evrardjp | for having overrides :D | 16:01 |
| *** admin0 has quit IRC | 16:01 | |
| odyssey4me | mhayden yeah, that'll be a fun discussion to have when we get to it properly at a summit | 16:01 |
| *** jamesdenton has joined #openstack-ansible | 16:02 | |
| odyssey4me | mhayden still waiting for your nod on https://review.openstack.org/328240 which hopefully will address the build timeout frequency increase | 16:02 |
| evrardjp | mancdaz: to be honest I've been bitten before :p as I thought nobody would change their mind | 16:02 |
| mhayden | odyssey4me: that one looks reasonable to me | 16:04 |
| evrardjp | odyssey4me: what about orchestrating the container create better and have it as an async task? | 16:05 |
| evrardjp | bad idea | 16:05 |
| palendae | evrardjp, It would have to be synchronous at some point, since SSH has to happen after they're up | 16:06 |
| evrardjp | that was a bad idea anyway | 16:06 |
| *** jamesden_ has joined #openstack-ansible | 16:06 | |
| palendae | It'd be nice :( | 16:07 |
| evrardjp | palendae: it ends on the playbook level at latest, so I'm not sure how it work with nested playbooks, but probably not great | 16:07 |
| palendae | But yeah, builds can only go as fast as the slowest host anywya | 16:07 |
| * mhayden embarks on a lunchtime adventure | 16:07 | |
| evrardjp | palendae: you can have a task that polls and makes it sync afterwards | 16:07 |
| evrardjp | but that's not what we are looking for | 16:08 |
| *** ametts has joined #openstack-ansible | 16:08 | |
| evrardjp | palendae: not really with async, we can do that better, but then we'll still limited by the playbook | 16:08 |
| evrardjp | anyway, bad idea | 16:08 |
| palendae | Even with async the task won't be done til the slowest host is :p | 16:08 |
| evrardjp | yes, but you can run other tasks | 16:09 |
| evrardjp | that's the idea :D | 16:09 |
| palendae | True, true | 16:09 |
| *** jamesdenton has quit IRC | 16:10 | |
| *** weshay is now known as weshay_br | 16:10 | |
| *** weshay_br is now known as weshay_brb | 16:10 | |
| odyssey4me | evrardjp something that may be worth doing is devolving the container create into the service install playbooks | 16:13 |
| evrardjp | that starts to be interesting | 16:13 |
| odyssey4me | that means that the highly IO intensive container create process will be spread out a bit more | 16:14 |
| evrardjp | but it's not enough | 16:14 |
| odyssey4me | but I think it would be far better for us to look at implementing a different file system for the containers | 16:14 |
| evrardjp | yes, but it's nothing | 16:14 |
| zerda2 | shouldn't the note about backporting release notes with intact filenames be in the contributor guidelines? Liberty will be around for ~half a year from now | 16:14 |
| evrardjp | odyssey4me: oh you mean have CoW? | 16:14 |
| odyssey4me | evrardjp yeah, something like that | 16:15 |
| evrardjp | true that could help | 16:15 |
| odyssey4me | zerda2 it's only ever a problem when people prematurely backport | 16:15 |
| odyssey4me | which is already against the guidelines | 16:16 |
| evrardjp | zfs is included in 16.04 by default now ? | 16:16 |
| evrardjp | odyssey4me: :D | 16:16 |
| zerda2 | oh ok | 16:16 |
| odyssey4me | evrardjp I hear so. Realistically we could use just about anything for the container FS, because we should not have to care about the FS contents for the container. | 16:16 |
| odyssey4me | It should simply be replacable, and any data we care about should be on the host. | 16:17 |
| evrardjp | odyssey4me: yes you told me | 16:17 |
| evrardjp | writes on disk should still be reliable, for containers like galera | 16:17 |
| evrardjp | but bind mounts etc | 16:18 |
| evrardjp | I know the drill | 16:18 |
| evrardjp | I'll do a poc of async task with nested playbooks, just to see if that works | 16:18 |
| *** afred312 has quit IRC | 16:18 | |
| *** mummer has joined #openstack-ansible | 16:18 | |
| *** mummer has quit IRC | 16:18 | |
| evrardjp | I'm pretty sure it doesn't | 16:19 |
| *** jamesden_ has quit IRC | 16:19 | |
| odyssey4me | sigmavirus24 / stevelle please review https://review.openstack.org/328354 (small glance module patch) when you have a moment | 16:19 |
| odyssey4me | yeah, if we implement the container create in the service install flow instead of in its own playbook then I think we could probably play more liberally with different execution strategies | 16:20 |
| chris_hultin | odyssey4me: I do agree that it needs an overhaul - specifically, it could probably stand to do more detailed checking to see if the image has the same properties, etc., rather than just checking if the name exists | 16:20 |
| odyssey4me | I'd like to actually unwind what's common and different between openstack_hosts and lxc_hosts and see if we can't optimise there too. I kinda think that lxc_container_create is probably a mess and could do with an entire re-do with a different approach. | 16:21 |
| *** weshay_brb is now known as weshay | 16:21 | |
| odyssey4me | chris_hultin yeah, it'd be nice if it verified that the image actually is there too :p | 16:21 |
| v1k0d3n | evrardjp: hey man...have a moment to assist me with the haproxy stuff again? | 16:24 |
| v1k0d3n | i have changed the net to reflect what is needed now (external side). | 16:25 |
| evrardjp | v1k0d3n: sure | 16:25 |
| v1k0d3n | and i'm at a place where i can rebuild...so in a good place i think. | 16:25 |
| evrardjp | could you send me the latest version of your user_variables and openstack_user_config ? | 16:25 |
| evrardjp | this way I understand | 16:26 |
| openstackgerrit | Darren Birkett proposed openstack/openstack-ansible: Make ansible managed line in swift.conf static https://review.openstack.org/326374 | 16:26 |
| evrardjp | one picture is better than a thousand words | 16:26 |
| mancdaz | ok peeps have a great weekend. I'm off to 'release' a beer and make some tasting 'notes' | 16:27 |
| mancdaz | I'll see myself out... | 16:27 |
| *** Drago has quit IRC | 16:27 | |
| evrardjp | enjoy mancdaz | 16:28 |
| evrardjp | enjoy your beer review -d | 16:28 |
| v1k0d3n | yes evrardjp...one sec. changing for a couple of updated exclusions. | 16:29 |
| v1k0d3n | will ping in a sec. | 16:30 |
| evrardjp | no worries | 16:31 |
| *** zerda2 has quit IRC | 16:35 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-galera_client: Add dependency to the openstack-ansible-plugin role https://review.openstack.org/321582 | 16:35 |
| *** asettle has quit IRC | 16:35 | |
| *** jamesdenton has joined #openstack-ansible | 16:38 | |
| odyssey4me | evrardjp the switch to 2.1 is apparently later, I'd like that as a topic at the mid cycle | 16:39 |
| evrardjp | yes, that's what we've discussed | 16:39 |
| evrardjp | if that is still happening this cycle, we could clean up too | 16:40 |
| evrardjp | but yes, I guess testing 2.1 makes more sense | 16:40 |
| evrardjp | I'm on the fence | 16:41 |
| odyssey4me | well, I think the general feeling is that we need to get Xenial right, and that's costing us quite a bit of effort | 16:42 |
| odyssey4me | there's also OVS which is high priority for both automagically and michaelgugino's organisations | 16:43 |
| odyssey4me | I think that combo of features is very likely to tie us up and introducing more refactoring for Ansible 2.1 can possibly distract us and place risk on achieving those two goals. | 16:44 |
| palendae | ^ | 16:44 |
| odyssey4me | But I'd like us to have a decent discussion around it all, and to corral around the top priorities. | 16:44 |
| *** jiteka has quit IRC | 16:46 | |
| *** KLevenstein has quit IRC | 16:47 | |
| v1k0d3n | evrardjp: https://gist.github.com/v1k0d3n/90225bfe3b2d2ec9305eedd8a90ffa36 | 16:49 |
| v1k0d3n | and evrardjp: https://gist.github.com/v1k0d3n/d4bfb52f1b095506322650060b6dd7ee | 16:49 |
| *** sdake has joined #openstack-ansible | 16:50 | |
| v1k0d3n | i know these are not correct where they stand today....some issues are kind of glaring because i was just trying nearly anything. i want to get to a "working" version and then work my way back to understanding it more. | 16:50 |
| v1k0d3n | i currently dont have aohd or ceilometer prepped either. i wanted to ask if there's a method of having that deployed that already exists or if that's a manual process right now? | 16:51 |
| evrardjp | wow 6 hosts | 16:51 |
| evrardjp | for lb | 16:51 |
| evrardjp | that's an interesting edge case :D | 16:52 |
| evrardjp | you'll have two issues | 16:52 |
| v1k0d3n | if that's wrong (most likely) be harsh...i'm ok with that :) | 16:52 |
| odyssey4me | v1k0d3n aodh and ceilometer can be deployed, but mongodb deployment is not something we're doing | 16:52 |
| evrardjp | one is the keepalived priorities, the other is the ip configured for it | 16:52 |
| v1k0d3n | ok | 16:53 |
| v1k0d3n | not surprised, that makes sense. | 16:53 |
| evrardjp | I'll leave the aodh/ceilometer discussion to someone else, I'll focus on keepalived/haproxy | 16:53 |
| v1k0d3n | i can work and edit this gist in real time to try and fix (for visual). | 16:54 |
| evrardjp | first question, do you really need 6 hosts for redundancy? | 16:54 |
| v1k0d3n | ok...that sounds good. | 16:54 |
| v1k0d3n | no. there are three infra hosts...so we can just use those for redundancy...does that clean things up a bit? this was sort of my initial though anyway at first. | 16:54 |
| evrardjp | yes it will | 16:56 |
| evrardjp | so let's clean up a little | 16:56 |
| v1k0d3n | evrardjp: so i updated...those are the infra only boxes. | 16:57 |
| *** jamesdenton has quit IRC | 16:57 | |
| *** jamesden_ has joined #openstack-ansible | 16:57 | |
| evrardjp | ok | 16:58 |
| evrardjp | so now you have a cleaner thingy | 16:58 |
| evrardjp | you'll have to run inventory-manage to remove the remnants | 16:58 |
| evrardjp | remove these nodes from the haproxy group | 16:58 |
| evrardjp | haproxy_hosts | 16:58 |
| evrardjp | group | 16:58 |
| evrardjp | but I have a few bad news for you | 17:00 |
| evrardjp | you're gonna need to run plenty of playbooks :s | 17:00 |
| v1k0d3n | i've been deleting the inventory on re-runs, and using inventory manage for listing hosts...i suppose there is a better way. | 17:00 |
| evrardjp | v1k0d3n: the issue when you do that, is you don't remove the containers that are already created | 17:01 |
| evrardjp | and because you don't have your inventory anymore, you can't target them to remove them | 17:01 |
| v1k0d3n | well, i ran the teardown script. | 17:01 |
| evrardjp | so you'll have to manually destroy the containers | 17:01 |
| evrardjp | ok | 17:01 |
| evrardjp | so you can destroy your containers now ? | 17:01 |
| palendae | fyi the teardown script is gone in newton | 17:01 |
| v1k0d3n | and then i destroy what's remaining in /run/lxc | 17:01 |
| v1k0d3n | yeah, seems to have some issues i think with this release too? | 17:02 |
| evrardjp | palendae: there is still a lxc playbook to remove containers I think | 17:02 |
| odyssey4me | v1k0d3n yeah, it was removed because it was perpetually incomplete | 17:02 |
| evrardjp | v1k0d3n: teardown script was just a hack | 17:02 |
| v1k0d3n | for some reason things may be out of order? now sure...but it complains when it tries to access facts for a host, when the facts actually are already removed :( | 17:02 |
| odyssey4me | v1k0d3n you'll have to clear your fact cache too | 17:03 |
| v1k0d3n | yeah, i heard it's a hack before too. works better in aio dev scenarios? | 17:03 |
| evrardjp | let's clean your environment first | 17:03 |
| v1k0d3n | ah! odyssey4me that makes sense why i'm getting strangeness!! | 17:03 |
| evrardjp | we mostly destroy and recreate environments, it is faster | 17:03 |
| odyssey4me | v1k0d3n http://docs.openstack.org/developer/openstack-ansible/install-guide/ops-troubleshooting.html#cached-ansible-facts-issues | 17:03 |
| evrardjp | yes facts cache is in /etc/openstack_deploy/facts | 17:03 |
| evrardjp | iirc | 17:03 |
| evrardjp | ansible_facts | 17:04 |
| evrardjp | damn! | 17:04 |
| evrardjp | anyway | 17:04 |
| *** weezS has quit IRC | 17:04 | |
| v1k0d3n | yeah, evrardjp i was thinking about that, but every time these servers reboot...it takes 8 minutes to come up. painful. IBM AMM is pretty bad :( | 17:04 |
| evrardjp | yes I understand :D | 17:04 |
| evrardjp | so | 17:04 |
| v1k0d3n | oh, ok...i was deleting the right thing then :) | 17:04 |
| evrardjp | after you've cleaned your hosts of containers, check on /etc/hosts on each of your hosts | 17:04 |
| evrardjp | just to make sure that's clean too | 17:05 |
| v1k0d3n | if facts are only cached and gathered there...i think i'm good. | 17:05 |
| evrardjp | v1k0d3n: yes, that's where our fact cache is | 17:05 |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift: Make ansible managed line in swift.conf static https://review.openstack.org/326369 | 17:05 |
| evrardjp | odyssey4me: you do the backport of this to mitaka? | 17:05 |
| v1k0d3n | yup...^^^ hosts makes sense too. and even clearing out ~/.ssh/known_hosts help too i've found in some strange situations when trying to access the containers again. | 17:05 |
| odyssey4me | evrardjp would you like to backport https://review.openstack.org/326369 on behalf of mancdaz to help get it in quicker? | 17:06 |
| evrardjp | https://review.openstack.org/#/c/328393/ | 17:07 |
| odyssey4me | thanks evrardjp | 17:07 |
| odyssey4me | ok all, I'm out for the w/end - cheerio! | 17:07 |
| evrardjp | I did nothing :p | 17:07 |
| evrardjp | enjoy the weekend odyssey4me! | 17:07 |
| evrardjp | so v1k0d3n | 17:07 |
| evrardjp | when everything is clean on your hosts side | 17:08 |
| evrardjp | (make sure by running lxc-ls -f too) | 17:08 |
| evrardjp | we can clean up the deploy side | 17:08 |
| evrardjp | but before running any playbook we will edit some variables | 17:08 |
| evrardjp | v1k0d3n: are the ip 10.2.31.8 and 10.2.32.8 taken? | 17:10 |
| evrardjp | in your env | 17:10 |
| *** Oku_OS is now known as Oku_OS-away | 17:11 | |
| *** weezS has joined #openstack-ansible | 17:13 | |
| v1k0d3n | ok | 17:13 |
| evrardjp | I'm waiting for your answer :D | 17:13 |
| v1k0d3n | so... 10.2.31.8 and 10.2.32.8 will be taken at some point... | 17:13 |
| v1k0d3n | sorry...was thinking. | 17:13 |
| v1k0d3n | .7 isn't if that's what you want to use. | 17:13 |
| evrardjp | ok | 17:14 |
| evrardjp | fine | 17:14 |
| v1k0d3n | or even .5 if we want to keep things "clean" | 17:14 |
| v1k0d3n | do we want to claim an unused IP then? | 17:14 |
| evrardjp | https://gist.github.com/evrardjp/228b3bcaee953c5b5c4f38965428d24b | 17:14 |
| evrardjp | Ok I'll take .5 | 17:15 |
| v1k0d3n | ah | 17:15 |
| v1k0d3n | ok... | 17:15 |
| evrardjp | https://gist.github.com/evrardjp/228b3bcaee953c5b5c4f38965428d24b | 17:15 |
| openstackgerrit | Merged openstack/openstack-ansible: Actually remove Ironic container creation from AIO preparation https://review.openstack.org/328240 | 17:15 |
| v1k0d3n | so those do need to be virtual and not assigned to anything. i get it, that makes sense.... | 17:15 |
| evrardjp | yup | 17:15 |
| evrardjp | https://gist.github.com/evrardjp/affbb9681760709b2a35529a577f5eaf | 17:16 |
| v1k0d3n | perfect...done | 17:16 |
| v1k0d3n | so there are no running hosts. | 17:16 |
| evrardjp | what do you mean? | 17:16 |
| evrardjp | everything is clean? | 17:17 |
| evrardjp | no containers, etc? | 17:17 |
| evrardjp | that's what you mean? | 17:17 |
| v1k0d3n | i ran teardown...it puked....i then deleted anything in run/lxc...etc, etc. then deleted inventory, ansible_facts is gone. | 17:17 |
| v1k0d3n | just containers. | 17:17 |
| v1k0d3n | sorry... | 17:17 |
| v1k0d3n | so there are no running lxc contianers is what i mean. | 17:17 |
| v1k0d3n | just the base host systems. that's all. | 17:18 |
| evrardjp | ok | 17:18 |
| v1k0d3n | i can clean up /etc/hosts.. | 17:18 |
| evrardjp | and the /etc/hosts + the ssh known hosts are fine | 17:18 |
| evrardjp | ok | 17:18 |
| evrardjp | good | 17:18 |
| v1k0d3n | actually already cleaned up. | 17:18 |
| *** deadnull_ has quit IRC | 17:19 | |
| v1k0d3n | ok. those are "cleaned up" | 17:19 |
| v1k0d3n | what else? | 17:19 |
| evrardjp | cool | 17:20 |
| evrardjp | nothing that comes to mind right now | 17:20 |
| evrardjp | maybe check the ifconfig | 17:20 |
| evrardjp | to see if there is no weird interface that should be removed | 17:21 |
| v1k0d3n | hmm | 17:21 |
| evrardjp | and something in your lxc cache | 17:21 |
| evrardjp | just in case | 17:21 |
| evrardjp | anyway | 17:21 |
| evrardjp | last question for you | 17:21 |
| v1k0d3n | ah, no teardown seemed to clean that up too. | 17:21 |
| evrardjp | do you plan to have your public endpoints with a domain name instead of an IP? | 17:22 |
| *** sdake has quit IRC | 17:22 | |
| v1k0d3n | great question... | 17:23 |
| v1k0d3n | so dns for this lab is not quite there yet. | 17:23 |
| v1k0d3n | so these hosts are not defined yet, and i have suggested going dns name route. | 17:24 |
| v1k0d3n | this won't be the last rebuild. the goal is dns names. | 17:24 |
| *** afred312 has joined #openstack-ansible | 17:25 | |
| v1k0d3n | btw...i probably need to nuke the openstack_hostnames_ips.yml file...is that correct? | 17:25 |
| evrardjp | yes | 17:28 |
| evrardjp | an *inventory.json | 17:28 |
| evrardjp | and ansible_facts/* | 17:29 |
| evrardjp | because you are starting clean | 17:29 |
| evrardjp | so remember, when you'll have to work on dns, you'll have to edit your global_overrides > external_lb_vip_address and run everything again | 17:30 |
| evrardjp | because you'll change your endpoints | 17:30 |
| openstackgerrit | Merged openstack/openstack-ansible-plugins: Adding ability to set 'properties' via Glance https://review.openstack.org/328354 | 17:30 |
| v1k0d3n | that makes sense. | 17:31 |
| v1k0d3n | so users will access the external haproxy lb address, the .5, correct? | 17:31 |
| evrardjp | ok | 17:31 |
| evrardjp | yup | 17:31 |
| v1k0d3n | i tried this before, and for some reason it wasn't working...but we also didn't have the external side either yet. | 17:31 |
| v1k0d3n | have another question...maybe you could help me with? | 17:32 |
| evrardjp | you had an issue with your config | 17:32 |
| evrardjp | you used .0/cidr | 17:32 |
| evrardjp | which is difficult to bind on | 17:32 |
| evrardjp | :p | 17:32 |
| evrardjp | sure go ahead | 17:32 |
| v1k0d3n | well, that endpoint was actually the .7 before :( | 17:32 |
| evrardjp | in user_variables I mean | 17:33 |
| v1k0d3n | yeah, it was definitely misconfigured. | 17:33 |
| v1k0d3n | so now, everthing that you see on your side for haproxy is gtg? | 17:33 |
| evrardjp | I don't see the nic vworks-float01 on each of your controllers | 17:34 |
| v1k0d3n | and i don't need to do anything with conf.d/haproxy.yml? | 17:34 |
| evrardjp | because I don't have the network interfaces | 17:35 |
| evrardjp | but I guess the rest is good to go | 17:35 |
| evrardjp | do you have a conf.d/haproxy.yml ? | 17:35 |
| v1k0d3n | so the vworks-float01 is 10.2.32.9-11 for each of the infra boxes. | 17:35 |
| evrardjp | ok | 17:35 |
| v1k0d3n | i also have that tied to neutron flat...is that acceptable? | 17:35 |
| evrardjp | good | 17:35 |
| evrardjp | nope | 17:35 |
| v1k0d3n | lol! | 17:35 |
| v1k0d3n | crap | 17:36 |
| v1k0d3n | that probably explains why neutron is unhappy post deploy. | 17:36 |
| evrardjp | I missed that | 17:36 |
| v1k0d3n | yeah, look for vworks-float01 defined in openstack_user_config.yml | 17:36 |
| evrardjp | yeah, you can do that with a veth if you prefer | 17:36 |
| v1k0d3n | oh. hmm | 17:37 |
| v1k0d3n | so maybe i can explain how we're networked up? | 17:37 |
| evrardjp | but you'd probably don't want that, because I guess you'd prefer to isolate API traffic of the users/tenant traffic | 17:37 |
| evrardjp | Sorry I missed that | 17:37 |
| openstackgerrit | Ian Cordasco proposed openstack/openstack-ansible: Use correct keystone auth parameters https://review.openstack.org/328430 | 17:37 |
| v1k0d3n | bond1 has a few vlans. one is br-mgmt (2200), the other is named vworks-float01 (2203). | 17:38 |
| evrardjp | you can explain | 17:38 |
| openstackgerrit | Ian Cordasco proposed openstack/openstack-ansible: Use correct keystone auth parameters https://review.openstack.org/328432 | 17:38 |
| evrardjp | ok | 17:38 |
| v1k0d3n | the what i wanted to do was configure dhcp service and public routable space on that vworks-float01 and have neutron use that. | 17:39 |
| evrardjp | you have bond0 ? | 17:39 |
| evrardjp | and bond 2 ? | 17:39 |
| evrardjp | I'm greedy | 17:39 |
| v1k0d3n | bond0 is entirely backend management for pxe boot. | 17:39 |
| *** jamesden_ has quit IRC | 17:39 | |
| v1k0d3n | yes, and bond 2 has a br-storage (2098) vlan association. | 17:39 |
| evrardjp | ok let's bond2 alone | 17:40 |
| evrardjp | let's leave* | 17:40 |
| v1k0d3n | so br-mgmt is where we'd run internal API's (obviously). | 17:40 |
| evrardjp | yeah | 17:40 |
| evrardjp | last question do you have other nics? | 17:41 |
| v1k0d3n | so...let me show you something. this will make some sense here... | 17:41 |
| v1k0d3n | https://github.com/att-esst/osa-prep | 17:41 |
| v1k0d3n | sorry...that's really messy | 17:42 |
| v1k0d3n | one sec evrardjp | 17:42 |
| v1k0d3n | https://github.com/att-esst/osa-prep/tree/master/roles/osa-deploy-networking/templates | 17:42 |
| openstackgerrit | Christopher Hultin proposed openstack/openstack-ansible-os_magnum: Adding function to upload bay images to Glance https://review.openstack.org/327939 | 17:42 |
| v1k0d3n | so i will explain, but you have the general config templates.... | 17:42 |
| v1k0d3n | the only difference is that i've added an external side that is supposed to be used for floating ranges. | 17:43 |
| v1k0d3n | and that runs in it's own route table on the host (separate userland space and cleaner route tables). | 17:44 |
| evrardjp | v1k0d3n: FYI ansible has JunOS modules | 17:44 |
| v1k0d3n | yeah actually trying to convince our team to use https://github.com/napalm-automation/napalm for the network part. | 17:45 |
| v1k0d3n | that's been a really tough sell though, for some reason. | 17:45 |
| *** rljefferson has joined #openstack-ansible | 17:45 | |
| evrardjp | I've seen, but if they don't want full fledge stuff, there are directly usable JunOS modules | 17:45 |
| evrardjp | which is far easier | 17:45 |
| openstackgerrit | Amy Marrich (spotz) proposed openstack/openstack-ansible-os_heat: Implement Xenial Support https://review.openstack.org/324733 | 17:46 |
| evrardjp | http://docs.ansible.com/ansible/junos_config_module.html | 17:46 |
| evrardjp | http://docs.ansible.com/ansible/junos_netconf_module.html | 17:46 |
| evrardjp | ... | 17:46 |
| evrardjp | anyway | 17:46 |
| v1k0d3n | we're working on that. just not quite there yet. | 17:46 |
| evrardjp | :D | 17:46 |
| evrardjp | oh you have two route tables? | 17:47 |
| v1k0d3n | we're slowly working towards "self-service" everything. | 17:47 |
| *** woodard has joined #openstack-ansible | 17:47 | |
| v1k0d3n | yes 3 route tables now, to separate route tables for management, float and default on the liux hosts. | 17:48 |
| v1k0d3n | the networking seems to be fine, and neutron was deploying fine at one point. but i want to know what needs to be used for the host_bind_override variable. | 17:49 |
| *** woodard has quit IRC | 17:49 | |
| *** woodard has joined #openstack-ansible | 17:50 | |
| evrardjp | in fact what you're doing is you give to neutron the interface you're passing in the openstack_user_config | 17:51 |
| evrardjp | but you want to keep that also on the host to have it working for keepalived/haproxy | 17:52 |
| v1k0d3n | well, i have vworks-float01. that's incorrect? i'm probably misunderstanding the answer a bit.. | 17:52 |
| *** jamesdenton has joined #openstack-ansible | 17:52 | |
| *** kstev has quit IRC | 17:53 | |
| evrardjp | I'm thinking what's the best way in your case | 17:53 |
| evrardjp | because I hate having a nic that is dual assigned, I hate unnecessary veths | 17:53 |
| *** permalac has quit IRC | 17:54 | |
| *** Ashana has joined #openstack-ansible | 17:54 | |
| v1k0d3n | hmmm. ok that makes sense. so how do i correct this? | 17:55 |
| evrardjp | https://gist.github.com/evrardjp/f2ef647a6f349d2b88d4ffcdb78b596b | 17:57 |
| stevelle | v1k0d3n: if you are going to deploy aodh you should make a note to not use mongodb for it. I have a task slated to update the aodh playbook to make that easier in master and we might be able to backport that but won't be changing the default in liberty/mitaka at this point. | 17:58 |
| *** asettle has joined #openstack-ansible | 17:58 | |
| evrardjp | something like this | 17:58 |
| evrardjp | and then use host_bind_override: "eth12" | 17:58 |
| *** catintheroof has joined #openstack-ansible | 17:59 | |
| chris_hultin | automagically: Mind looking at this again when you get a chance? https://review.openstack.org/327939 | 17:59 |
| *** javeriak has joined #openstack-ansible | 18:00 | |
| *** asettle has quit IRC | 18:00 | |
| evrardjp | v1k0d3n: you'll have to adapt this to your config ofc | 18:00 |
| evrardjp | but I'm tired, I'll stop for today sorry | 18:01 |
| evrardjp | can we discuss that another day? | 18:01 |
| automagically | chris_hultin: Done. Minor feedback | 18:01 |
| evrardjp | don't hesitate to ask your questions here | 18:01 |
| v1k0d3n | evrardjp: ok...think my head is starting to hurt. trying to connect dots with yours and ours now. | 18:02 |
| v1k0d3n | one sec...let me process. | 18:02 |
| *** LiftedKilt has quit IRC | 18:02 | |
| v1k0d3n | and stevelle...really interested in this. let me ping you back in a sec on aohd? definitely want it, but running into issues trying to deploy and have questions for you with it. | 18:02 |
| openstackgerrit | Christopher Hultin proposed openstack/openstack-ansible-os_magnum: Adding function to upload bay images to Glance https://review.openstack.org/327939 | 18:03 |
| chris_hultin | automagically: Good catch, thanks | 18:03 |
| stevelle | v1k0d3n: when you have space in front of you for that step, ping me | 18:03 |
| v1k0d3n | stevelle: awesome...thanks for the help! :) i need a lot today hahaha | 18:03 |
| evrardjp | this is just a way (that I find ugly btw) to have a fake link that you can assign for neutron (br-vlan-veth) and another end on something else for keepalived | 18:04 |
| evrardjp | IIRC | 18:04 |
| evrardjp | If it's not that way, sorry, I'm really tired | 18:04 |
| evrardjp | I used to dedicate nic for that | 18:04 |
| evrardjp | this way I can isolate both traffic | 18:04 |
| v1k0d3n | ok...so i'm fine with a dedicated nic. | 18:04 |
| evrardjp | anyway | 18:04 |
| evrardjp | if you have a dedicated nic for keepalived, you should forget that | 18:05 |
| v1k0d3n | i can gist it, if that helps. you usually leave about this time? i hate keeping you...this has been extremely helpful. | 18:05 |
| evrardjp | you just assign your br-vlan to neutron, and you have another bridge on your hosts for the external part | 18:05 |
| evrardjp | another bridge/nic | 18:05 |
| evrardjp | It's far too late for me, because I'm sick :p | 18:06 |
| evrardjp | that's why I say "don't hesitate to leave message on the chan" | 18:06 |
| *** Ashana has quit IRC | 18:06 | |
| evrardjp | but I stay because I want to help :D | 18:06 |
| evrardjp | I leave 2 hours early I'd say | 18:07 |
| v1k0d3n | ouch that sucks evrardjp hope you feel better soon | 18:07 |
| evrardjp | yup I hope too ! | 18:07 |
| evrardjp | sorry I didn't want to bother you, let's forget that! | 18:07 |
| evrardjp | but I'm still leaving for today | 18:07 |
| evrardjp | :p | 18:07 |
| openstackgerrit | Ian Cordasco proposed openstack/openstack-ansible-os_neutron: Use default metadata_backlog value from Neutron https://review.openstack.org/328441 | 18:08 |
| evrardjp | enjoy your week-end everyone! | 18:08 |
| *** LiftedKilt has joined #openstack-ansible | 18:08 | |
| automagically | Later evrardjp Hope you feel better soon | 18:08 |
| evrardjp | thanks automagically | 18:09 |
| *** jamesdenton has quit IRC | 18:09 | |
| *** jamesdenton has joined #openstack-ansible | 18:09 | |
| *** sacharya has quit IRC | 18:13 | |
| v1k0d3n | take care and thanks again evrardjp | 18:13 |
| *** adreznec has quit IRC | 18:15 | |
| *** errr has quit IRC | 18:15 | |
| *** adreznec has joined #openstack-ansible | 18:17 | |
| *** jamesdenton has quit IRC | 18:26 | |
| *** galstrom_zzz is now known as galstrom | 18:27 | |
| *** tiagogomes has quit IRC | 18:27 | |
| *** sacharya has joined #openstack-ansible | 18:31 | |
| *** woodard has quit IRC | 18:32 | |
| *** chandanc_ has quit IRC | 18:33 | |
| *** chandanc has quit IRC | 18:33 | |
| *** KLevenstein has joined #openstack-ansible | 18:37 | |
| *** KLevenstein_ has joined #openstack-ansible | 18:39 | |
| *** Drago has joined #openstack-ansible | 18:41 | |
| *** KLevenstein has quit IRC | 18:42 | |
| *** KLevenstein_ is now known as KLevenstein | 18:42 | |
| *** errr has joined #openstack-ansible | 18:42 | |
| openstackgerrit | Sudarshan Acharya proposed openstack/openstack-ansible: Adding missing / to swift rabbit connection string https://review.openstack.org/325413 | 18:47 |
| openstackgerrit | Sudarshan Acharya proposed openstack/openstack-ansible: Adding missing / to swift rabbit connection string https://review.openstack.org/325413 | 18:48 |
| *** niknakpaddywak has quit IRC | 18:48 | |
| v1k0d3n | stevelle: hey man...i'm a bit better, can you share about mongo and aohd? | 18:51 |
| v1k0d3n | is there an "easy" way to deploy other than having to manually prep aohd and ceilometer? | 18:51 |
| v1k0d3n | i'd like to have swift too, but that seems manual in the guide as well. wondering if there's a similar bolt-on solution (i'm sure there is an i just haven't found it). | 18:52 |
| stevelle | v1k0d3n: you have to manually work through your drives for swift, that is true | 18:53 |
| v1k0d3n | man, was hoping there was an lxc deployment for that too :) haha | 18:54 |
| v1k0d3n | i suppose i was going to have the infra boxes be swift storage hosts (at least temporary for now). i'm sure that people run those on separate storage servers though? | 18:54 |
| stevelle | dedicated storage hosts is kind of standard | 18:55 |
| v1k0d3n | someone is working on a ceph back-end right now, but we'll have internal "customers" that will want the full swift API's. | 18:55 |
| stevelle | you can get started using the infra boxen for swift storage, you just have to do vol mgmt | 18:56 |
| stevelle | for aodh the mongo storage is deprecated in favor of having that use galera. our playbook and role needs an update before it makes that easy for you and I'm hoping to get that work started in the next week but I can see about starting today if that is coming up for you fast | 18:58 |
| stevelle | but for ceilometer you will want to put in mongo for now, and in newton gnocchi will be available. | 18:58 |
| stevelle | gnocchi will offer a better alternative to mongo | 18:58 |
| v1k0d3n | makes sense. | 18:59 |
| v1k0d3n | but mongo is manual now too, right? | 19:00 |
| stevelle | yes, or you can find a galaxy role to help | 19:00 |
| stevelle | but we have not done any evaluation on those. | 19:00 |
| v1k0d3n | ah, i was wondering how to really add galaxy roles honestly...haven't done that yet, but very interested. i thought there was something about ceph build during a deployment too...? but not sure. | 19:00 |
| stevelle | yes, ceph provides roles you can use for that | 19:01 |
| v1k0d3n | didn't really understand clearly by reading through the docs how much is done for you or manual | 19:01 |
| *** hybridpollo has joined #openstack-ansible | 19:01 | |
| stevelle | each role you want to add should have their own instructions but in many cases there is little to do but configure and run the role. | 19:02 |
| *** johnmilton has quit IRC | 19:03 | |
| stevelle | the key concept is to use only one role to manage one thing (service) across your inventory. because OSA doesn't do anything to ceph or mongo you could use another role from galaxy | 19:03 |
| *** aludwar has quit IRC | 19:04 | |
| stevelle | it's a little more nuanced than that but basically that should do | 19:04 |
| v1k0d3n | ok, interesting. very different from an aio deployment which does everything for the user. | 19:05 |
| openstackgerrit | Amy Marrich (spotz) proposed openstack/openstack-ansible-os_heat: Implement Xenial Support https://review.openstack.org/324733 | 19:06 |
| stevelle | v1k0d3n: with more power comes more responsibility I guess :) | 19:07 |
| v1k0d3n | good way of looking at it, for sure. | 19:08 |
| v1k0d3n | stevelle: are you guys using ceph? | 19:12 |
| v1k0d3n | and if so, how are you deploying...using galaxy roles along with the OSAD? | 19:12 |
| sigmavirus24 | Anyone here remember keystone v2's behaviour around the admin service? Can service roles authenticate against that? (I think the answer is no) | 19:12 |
| sigmavirus24 | v1k0d3n: ansible-galaxy let's you point at git repositories as well as roles on galaxy.ansible.com | 19:13 |
| sigmavirus24 | You can use both at the same mtime | 19:13 |
| sigmavirus24 | Because galaxy.ansible.com requires someone to authenticate with GitHub to upload the role, we can't do that because no one can login as the OpenStack organization for us | 19:13 |
| v1k0d3n | are there more detailed instructions for doing this by chance? | 19:13 |
| *** Mudpuppy has quit IRC | 19:14 | |
| v1k0d3n | sigmavirus24: this is good to know. what i've always thought too. | 19:14 |
| sigmavirus24 | v1k0d3n: openstack-ansible has docs for doing an OSA deployment | 19:14 |
| v1k0d3n | sigmavirus24: yeah, but the ceph part along with osad? | 19:15 |
| sigmavirus24 | v1k0d3n: you could look at how https://github.com/rcbops/rpc-openstack does ceph + OSA | 19:15 |
| v1k0d3n | sigmavirus24: thanks. i did that...what's throwing me off is the Juno/Kilo support messages. | 19:16 |
| v1k0d3n | are you doing this yourself? | 19:16 |
| sigmavirus24 | v1k0d3n: juno and kilo are dead in OpenStack | 19:16 |
| sigmavirus24 | The fact that there are still juno/kilo branches there does not mean very much | 19:16 |
| sigmavirus24 | liberty is presently the oldest still supported stable version of openstack | 19:17 |
| stevelle | v1k0d3n: have you seen this? https://www.openstackfaq.com/openstack-ansible-ceph/ | 19:22 |
| stevelle | the author / maintainer isn't in here right now and I'm not sure if it is totally current but it seemed a good start | 19:22 |
| stevelle | changes from liberty to mitaka shouldn't bee too hard to work out | 19:23 |
| *** adrian_otto has quit IRC | 19:26 | |
| openstackgerrit | Steve Lewis (stevelle) proposed openstack/openstack-ansible-os_ceilometer: Implement Ubuntu 16.04 support with SystemD https://review.openstack.org/327900 | 19:27 |
| *** jwagner is now known as jwagner[away] | 19:28 | |
| chris_hultin | sigmavirus24: Mind looking at this one again? https://review.openstack.org/327939 | 19:31 |
| sigmavirus24 | chris_hultin: much better | 19:37 |
| chris_hultin | sigmavirus24: There's not really any reason we can't move it into the post_install.yml file. | 19:39 |
| sigmavirus24 | it just sticks out | 19:39 |
| sigmavirus24 | and that kind of thing feels like something you do once you have magnum installed | 19:40 |
| * sigmavirus24 shrugs | 19:40 | |
| sigmavirus24 | I didn't -1 because I'll defer to the rest of the cores on that | 19:40 |
| chris_hultin | stevelle: Is there a specific location you feel that documentation should go? | 19:41 |
| stevelle | chris_hultin: README.rst in the os_magnum role | 19:43 |
| stevelle | I'd add a secion for Optional Variables | 19:43 |
| automagically | chris_hultin: Typical role doc: http://docs-draft.openstack.org/11/328311/1/check/gate-openstack-ansible-memcached_server-docs/d9583b6//doc/build/html/ | 19:44 |
| automagically | You’ll note the newer pattern is to let the README.rst include defaults/main.xml in the Default Variables section | 19:44 |
| stevelle | ^ that is even better I guess. | 19:44 |
| *** bsv has joined #openstack-ansible | 19:46 | |
| chris_hultin | automagically: It looks like the README is set to do that already. | 19:47 |
| automagically | Perfect | 19:47 |
| automagically | http://docs-draft.openstack.org/39/327939/4/check/gate-openstack-ansible-os_magnum-docs/bf5369e//doc/build/html/ <—— That seems to render pretty well | 19:47 |
| stevelle | give me a sec to look that over again | 19:48 |
| openstackgerrit | Christopher Hultin proposed openstack/openstack-ansible-os_magnum: Adding function to upload bay images to Glance https://review.openstack.org/327939 | 19:48 |
| chris_hultin | automagically: OK, went ahead and moved that to post_install | 19:48 |
| *** Zucan has quit IRC | 19:48 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-ops: Add script to diff projects in OSA releases https://review.openstack.org/328469 | 19:49 |
| *** sdake has joined #openstack-ansible | 19:52 | |
| *** bsv has quit IRC | 19:53 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-ops: Add script to diff projects in OSA releases https://review.openstack.org/328469 | 19:54 |
| mhayden | so i don't know what the long term plan is for openstack-ansible-ops, but i put something in there :P | 19:54 |
| mhayden | perhaps it should go into scripts/ in openstack-ansible, but i am not sure | 19:55 |
| automagically | It should go somewhere…its really useful | 19:56 |
| automagically | We’ll figure out where next week no doubt | 19:56 |
| openstackgerrit | Merged openstack/openstack-ansible-os_magnum: Adding function to upload bay images to Glance https://review.openstack.org/327939 | 19:57 |
| *** rueben_ has quit IRC | 19:57 | |
| *** ametts has quit IRC | 20:00 | |
| mhayden | automagically: should i propose it to openstack-ansible instead? | 20:04 |
| mhayden | pygithub3 should be the only new requirement for it to work | 20:05 |
| automagically | I don’t have firm opinions on where it belongs | 20:05 |
| automagically | but where you have it now makes sense | 20:05 |
| openstackgerrit | Amy Marrich (spotz) proposed openstack/openstack-ansible-os_heat: Implement Xenial Support https://review.openstack.org/324733 | 20:05 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-ops: Add script to diff projects in OSA releases https://review.openstack.org/328469 | 20:07 |
| openstackgerrit | Steve Lewis (stevelle) proposed openstack/openstack-ansible-os_gnocchi: Add dependency on openstack-ansible-plugin role https://review.openstack.org/328479 | 20:08 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-ops: Add script to diff projects in OSA releases https://review.openstack.org/328469 | 20:10 |
| mhayden | got the project names sorted now | 20:11 |
| mhayden | not sure why i didn't do that before | 20:11 |
| spotz | why ask why mhayden? | 20:12 |
| *** adrian_otto has joined #openstack-ansible | 20:12 | |
| *** ametts has joined #openstack-ansible | 20:15 | |
| automagically | See you all on Monday. Enjoy your weekends | 20:22 |
| spotz | Seeya automagically | 20:25 |
| *** cloader89 has joined #openstack-ansible | 20:26 | |
| openstackgerrit | Cameron Loader proposed openstack/openstack-ansible: Destroy repo containers before upgrade https://review.openstack.org/328485 | 20:28 |
| *** aludwar has joined #openstack-ansible | 20:28 | |
| *** johnmilton has joined #openstack-ansible | 20:29 | |
| *** rljefferson has quit IRC | 20:29 | |
| *** ametts has quit IRC | 20:30 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Add release file prep script https://review.openstack.org/324736 | 20:37 |
| *** johnmilton has quit IRC | 20:37 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Remove deleted override vars from inventory https://review.openstack.org/325380 | 20:37 |
| *** jwagner[away] is now known as jwagner | 20:40 | |
| *** bsv has joined #openstack-ansible | 20:42 | |
| mhayden | enjoy automagically | 20:42 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Update lists of skipped security role tasks https://review.openstack.org/327732 | 20:47 |
| *** johnmilton has joined #openstack-ansible | 20:49 | |
| mhayden | cloader89: install the 'reno' python module and run 'reno new add-a-slug-here' | 20:57 |
| *** messy has quit IRC | 20:57 | |
| mhayden | so something like: reno new rebuild-repo-servers-during-upgrade (or something like that) | 20:57 |
| openstackgerrit | Steve Lewis (stevelle) proposed openstack/openstack-ansible-os_ceilometer: Implement Ubuntu 16.04 support with SystemD https://review.openstack.org/327900 | 20:58 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Add check/audit to gate testing https://review.openstack.org/328497 | 21:04 |
| mhayden | cloader89: here's an example with docs and a release note -> https://review.openstack.org/#/c/320993/ | 21:05 |
| *** galstrom is now known as galstrom_zzz | 21:08 | |
| *** itlinux has quit IRC | 21:13 | |
| *** McMurlock1 has quit IRC | 21:14 | |
| *** kylek3h has quit IRC | 21:18 | |
| *** javeriak has quit IRC | 21:25 | |
| *** KLevenstein has quit IRC | 21:25 | |
| openstackgerrit | Merged openstack/openstack-ansible: Adding missing / to swift rabbit connection string https://review.openstack.org/325413 | 21:26 |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron: Use default metadata_backlog value from Neutron https://review.openstack.org/328441 | 21:27 |
| openstackgerrit | Cameron Loader proposed openstack/openstack-ansible: Destroy repo containers before upgrade https://review.openstack.org/328485 | 21:27 |
| *** jorge_munoz has quit IRC | 21:28 | |
| *** itlinux has joined #openstack-ansible | 21:30 | |
| openstackgerrit | Ian Cordasco proposed openstack/openstack-ansible: Use default metadata_backlog value from Neutron https://review.openstack.org/328502 | 21:33 |
| openstackgerrit | Ian Cordasco proposed openstack/openstack-ansible: Use default metadata_backlog value from Neutron https://review.openstack.org/328503 | 21:34 |
| *** galstrom_zzz is now known as galstrom | 21:34 | |
| *** sdake has quit IRC | 21:36 | |
| *** sdake has joined #openstack-ansible | 21:41 | |
| openstackgerrit | Ian Cordasco proposed openstack/openstack-ansible: Use correct keystone auth parameters https://review.openstack.org/328430 | 21:43 |
| openstackgerrit | Ian Cordasco proposed openstack/openstack-ansible: Use correct keystone auth parameters https://review.openstack.org/328432 | 21:43 |
| *** jwitko has quit IRC | 21:43 | |
| *** galstrom is now known as galstrom_zzz | 21:43 | |
| *** sdake_ has joined #openstack-ansible | 21:43 | |
| *** sdake has quit IRC | 21:45 | |
| openstackgerrit | Amy Marrich (spotz) proposed openstack/openstack-ansible-os_heat: Implement Xenial Support https://review.openstack.org/324733 | 21:47 |
| openstackgerrit | Merged openstack/openstack-ansible-os_horizon: Add support for CustomLog format modification https://review.openstack.org/327808 | 21:49 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 21:49 | |
| *** itlinux has quit IRC | 21:50 | |
| *** iceyao has joined #openstack-ansible | 21:54 | |
| *** jmckind has quit IRC | 21:56 | |
| *** jmckind_ has joined #openstack-ansible | 21:56 | |
| *** iceyao has quit IRC | 21:59 | |
| *** cloader89 has quit IRC | 22:04 | |
| *** catintheroof has quit IRC | 22:06 | |
| openstackgerrit | Amy Marrich (spotz) proposed openstack/openstack-ansible-os_heat: Implement Xenial Support https://review.openstack.org/324733 | 22:07 |
| openstackgerrit | Steve Lewis (stevelle) proposed openstack/openstack-ansible-os_aodh: Convert default from MongoDB to MySQL storage https://review.openstack.org/328510 | 22:08 |
| *** bsv has quit IRC | 22:14 | |
| openstackgerrit | Steve Lewis (stevelle) proposed openstack/openstack-ansible-os_aodh: Convert default from MongoDB to MySQL storage https://review.openstack.org/328510 | 22:19 |
| openstackgerrit | Steve Lewis (stevelle) proposed openstack/openstack-ansible-os_aodh: Convert default from MongoDB to MySQL storage https://review.openstack.org/328510 | 22:22 |
| *** jamesdenton has joined #openstack-ansible | 22:22 | |
| openstackgerrit | Merged openstack/openstack-ansible: Make ansible managed line in swift.conf static https://review.openstack.org/326374 | 22:22 |
| *** adrian_otto has quit IRC | 22:22 | |
| openstackgerrit | Steve Lewis (stevelle) proposed openstack/openstack-ansible-os_aodh: Convert default from MongoDB to MySQL storage https://review.openstack.org/328510 | 22:23 |
| stevelle | best way to spend Friday night on dad/grad weekend :D | 22:26 |
| *** jmckind_ has quit IRC | 22:30 | |
| *** klamath has quit IRC | 22:38 | |
| *** jamesdenton has quit IRC | 22:42 | |
| *** scarlisle has quit IRC | 22:44 | |
| *** spotz is now known as spotz_zzz | 22:46 | |
| *** BjoernT has quit IRC | 22:46 | |
| *** weezS has quit IRC | 23:12 | |
| *** adrian_otto has joined #openstack-ansible | 23:16 | |
| openstackgerrit | Steve Lewis (stevelle) proposed openstack/openstack-ansible-os_aodh: Convert default from MongoDB to MySQL storage https://review.openstack.org/328510 | 23:19 |
| *** kjw3 has joined #openstack-ansible | 23:26 | |
| *** kjw3 has quit IRC | 23:27 | |
| *** weshay has quit IRC | 23:29 | |
| *** elo has quit IRC | 23:31 | |
| *** elo has joined #openstack-ansible | 23:33 | |
| *** sacharya has quit IRC | 23:34 | |
| *** zerda2 has joined #openstack-ansible | 23:36 | |
| *** Drago has quit IRC | 23:38 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!