| *** bradjones has joined #kolla | 00:00 | |
| *** bradjones has joined #kolla | 00:00 | |
| *** jasonsb has quit IRC | 00:00 | |
| *** fangfenghua has joined #kolla | 00:01 | |
| *** dims has quit IRC | 00:03 | |
| *** dims has joined #kolla | 00:03 | |
| *** daneyon_ has quit IRC | 00:30 | |
| *** dwalsh has joined #kolla | 00:45 | |
| fangfenghua | morning folks | 01:10 |
|---|---|---|
| *** bmace_ has joined #kolla | 01:11 | |
| *** bmace has quit IRC | 01:12 | |
| *** bmace_ is now known as bmace | 01:13 | |
| *** tobe has joined #kolla | 01:18 | |
| *** dims has quit IRC | 01:23 | |
| *** erkules_ has joined #kolla | 01:29 | |
| *** erkules has quit IRC | 01:32 | |
| *** dwalsh has quit IRC | 01:40 | |
| *** bmace has quit IRC | 01:48 | |
| *** bmace has joined #kolla | 01:49 | |
| *** dims has joined #kolla | 02:12 | |
| *** dims has quit IRC | 02:23 | |
| fangfenghua | SamYaple: did docker 1.7 dev had support mount propagation ? | 02:28 |
| *** vinkman has quit IRC | 02:43 | |
| *** vinkman1 has joined #kolla | 02:43 | |
| *** bradjones has quit IRC | 02:52 | |
| *** bradjones has joined #kolla | 02:53 | |
| *** bradjones has joined #kolla | 02:53 | |
| *** dolpher has joined #kolla | 02:59 | |
| *** jasonsb has joined #kolla | 03:01 | |
| SamYaple | fangfenghua: docker 1.7 has all the features we have been waiting for, including allowing mount propogation | 03:16 |
| dolpher | Does kolla support running on ubuntu host now? | 03:38 |
| SamYaple | dolpher: the containers should run on Ubuntu hosts now after some fixes | 03:48 |
| bmace | hey dolpher, i think a lot of developers run it on ubuntu. i know all the build stuff works ok on ubuntu since i did it just the other day. at least 15.04. didn't have great luck with older ubuntu versions. | 03:48 |
| *** dolpher has quit IRC | 03:51 | |
| *** dolpher has joined #kolla | 03:52 | |
| dolpher | OK, I'm trying Kolla with ubuntu 14.04, and met a few issues. So I just want to make sure it supports ubuntu before I troubleshooting them further | 03:54 |
| bmace | i had a lot of issues with 14.04 as well dolpher which is why i did the upgrade.. i think the version of aufs is a problem, and there were several others. | 03:55 |
| bmace | if you really want to stick with 14.04 you will probably need to move to using devicemapper or somehow update the aufs past the version they have in the 14.04 repos | 03:56 |
| dolpher | Thanks bmace. I'm using 14.04 just because it's a LTS version, I don't have to stick with it. I can move to 15.04, even CentOS for the try | 04:02 |
| bmace | i have tried several between bare metal / vms and so far 15.04 has been the smoothest, at least for me. | 04:03 |
| bmace | i think sdake said he does his devel on whatever the latest fedora rev is. | 04:04 |
| bmace | so that is likely to work well also | 04:04 |
| dolpher | I'll try with 15.04 first, thanks | 04:06 |
| bmace | my pleasure dolpher. i have been through the dev start up process about half a dozen times in the last couple of weeks so if you hit issues hopefully it has been seen, perhaps very recently :) | 04:07 |
| *** bmace has quit IRC | 04:21 | |
| *** bmace has joined #kolla | 04:32 | |
| *** dolpher has quit IRC | 04:35 | |
| SamYaple | anyone around to talk about environment and bindmounted configs? | 04:42 |
| bmace | well, i think you and i were on the same page.. so if you want a multi-sided discussion i may not help :) | 04:43 |
| SamYaple | nope you will. so new ideas on how we can all play together | 04:44 |
| SamYaple | so the thought ive had with some discussion | 04:45 |
| bmace | fantastic. as long as it doesn't add a ton of complexity / maintenance i think more is better as far as config options, imho.. | 04:45 |
| SamYaple | we bindmount in a "config" folder. when the entrypoint script starts for keystone (as an example) it looks and says does /opt/kolla/configs/keystone.conf exist? | 04:46 |
| SamYaple | if it _does_ exist we cp it into the container | 04:46 |
| SamYaple | if not, environment variables | 04:46 |
| bmace | that feels ok to me.. better than leaving a straight up file being used during execution, certainly | 04:49 |
| SamYaple | well it also means we can, should we choose, validate the file before launching keystone | 04:49 |
| SamYaple | it also allows us to just not use the external config file at all | 04:49 |
| bmace | sure, if we determine the file contents are borked we can either fall back to env or fail (probably the better of the two options) | 04:50 |
| bmace | sanity checking the config sounds like a great idea. | 04:50 |
| SamYaple | i dont like the fallback idea, then you may be running with an unexpected config | 04:51 |
| SamYaple | i would prefer container stop | 04:51 |
| bmace | agreed, if their intent was to run from the config, and the config is borked, ending up having it running using some other set of values (from env for example) would be pretty confusing | 04:52 |
| vinkman1 | So that would mean all the images will bind mount /opt/kolla ? | 04:55 |
| bmace | or /opt/kolla/config if there is some other data under /opt/kolla that isn't needed by containers. | 04:56 |
| SamYaple | vinkman1: perhaps, but it wouldn't be hard to make a bind mount per service | 04:56 |
| SamYaple | /opt/kolla/keystone, /opt/kolla/glance, etc | 04:56 |
| SamYaple | i dont want one container to have access to all configs | 04:56 |
| *** bmace_ has joined #kolla | 04:59 | |
| *** bmace has quit IRC | 04:59 | |
| *** bmace_ is now known as bmace | 04:59 | |
| vinkman1 | So how would that work if I run for example 2 or 3 keystone containers? | 04:59 |
| SamYaple | vinkman1: you cant do that on the same host anyway | 04:59 |
| bmace | i think the assumption is no more than one instance of any given service on a single node. | 05:00 |
| vinkman1 | OK, so the orchestration tool will disallow that since Docker does not... | 05:00 |
| SamYaple | its not an assupmtion, it uses the host net namespace. two keystone containers cant bind ot the same port | 05:00 |
| vinkman1 | Ahh…kk… | 05:00 |
| bmace | yes, one way or another one of them would fail even if you tried to bring up 2 on a single host, but i do believe the ansible orchestration is supposed to not have that even tried in the first place. | 05:01 |
| *** vinkman1 has quit IRC | 05:02 | |
| SamYaple | well the whole architecture is built around not doing that | 05:02 |
| SamYaple | i can't think of a valid reason to run multiple keystone o nthe same host other than testing | 05:02 |
| SamYaple | it really comlicates he design. im not sure ansible will be able to handle that even if we did implement it | 05:03 |
| bmace | masochism ;) | 05:03 |
| *** vinkman has joined #kolla | 05:03 | |
| *** nihilifer_ has joined #kolla | 05:32 | |
| *** prad has quit IRC | 05:39 | |
| *** prad_ is now known as prad | 05:39 | |
| *** vinkman has quit IRC | 05:43 | |
| *** bmace has quit IRC | 05:52 | |
| *** bmace has joined #kolla | 05:53 | |
| *** bmace has quit IRC | 05:59 | |
| *** bmace has joined #kolla | 06:01 | |
| *** prad_ has joined #kolla | 06:48 | |
| *** shardy has joined #kolla | 07:01 | |
| *** jmccarthy has joined #kolla | 07:07 | |
| *** bmace_ has joined #kolla | 07:31 | |
| *** bmace has quit IRC | 07:31 | |
| *** bmace_ is now known as bmace | 07:33 | |
| *** athomas has joined #kolla | 07:52 | |
| *** alisonholloway has quit IRC | 08:11 | |
| *** fangfenghua has quit IRC | 08:18 | |
| *** shardy_ has joined #kolla | 08:18 | |
| *** shardy has quit IRC | 08:20 | |
| *** shardy_ has quit IRC | 08:23 | |
| *** shardy has joined #kolla | 08:24 | |
| *** walters has quit IRC | 08:27 | |
| *** walters has joined #kolla | 08:29 | |
| *** shardy_ has joined #kolla | 08:32 | |
| *** shardy has quit IRC | 08:34 | |
| *** shardy_ has quit IRC | 08:38 | |
| *** shardy has joined #kolla | 08:39 | |
| *** mickt has joined #kolla | 08:40 | |
| *** bradjones has quit IRC | 08:40 | |
| *** bradjones has joined #kolla | 08:42 | |
| *** bradjones has joined #kolla | 08:42 | |
| mickt | test | 08:43 |
| SamYaple | ack | 08:43 |
| mickt | :) | 08:44 |
| *** jtriley has joined #kolla | 09:06 | |
| *** jmccarthy has quit IRC | 09:06 | |
| *** jtriley has quit IRC | 09:11 | |
| *** jmccarthy has joined #kolla | 09:15 | |
| *** jmccarthy has quit IRC | 09:20 | |
| *** dims has joined #kolla | 09:27 | |
| *** dims has quit IRC | 09:31 | |
| *** alisonholloway has joined #kolla | 09:33 | |
| *** jmccarthy has joined #kolla | 09:40 | |
| *** dims has joined #kolla | 09:51 | |
| *** dims_ has joined #kolla | 09:52 | |
| *** dims has quit IRC | 09:56 | |
| *** jmccarthy has quit IRC | 10:00 | |
| *** jmccarthy has joined #kolla | 10:20 | |
| *** athomas has quit IRC | 10:37 | |
| *** vbel has joined #kolla | 10:38 | |
| *** athomas has joined #kolla | 10:43 | |
| *** absubram has joined #kolla | 10:59 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Symlink all non Dockerfile resources in images https://review.openstack.org/190117 | 11:07 |
| SamYaple | pdb: did you catch the irc snippet earlier for a way to allow configs managed from the host as well env vars? | 11:20 |
| pdb | SamYaple: I did, but let me read it again :) | 11:20 |
| SamYaple | were you in the "manage from host" or "manage from container" camp? | 11:20 |
| pdb | mostly manage from host. though I see sdake's side too. I think your idea above sounds very good | 11:23 |
| SamYaple | ok. so environment varible named "legacy" or similar. If it is set, pull in the config file from the host bind mount. | 11:24 |
| SamYaple | do you have a suggestion other than legacy? | 11:25 |
| pdb | legacy isn't the best term imo | 11:27 |
| SamYaple | traditional? | 11:27 |
| SamYaple | i agree about legacy | 11:27 |
| pdb | let me break out the thesauraus :) | 11:28 |
| SamYaple | shazam | 11:28 |
| pdb | all these terms in some way suggest that we dont recommend this methhod | 11:29 |
| pdb | in which case why are we adding it | 11:29 |
| pdb | the simplest to me would just be something like CONFIG_STRATEGY=bind|env | 11:30 |
| pdb | and document the ins and outs of each | 11:30 |
| jmccarthy | astral ;) | 11:34 |
| SamYaple | pdb: i like it. time to convince | 11:38 |
| SamYaple | ill add areview | 11:39 |
| SamYaple | pdb: so we are on the same page. ansible will generate the config and save it to a location on the host (/opt/kolla/keystone/keystone.conf). We bind /opt/kolla/keystone/ into the container. in the container we cp keystone.conf /etc/keystone/keystone.conf | 11:44 |
| SamYaple | we make changes externally to the config. on container restart those changes are pulled in | 11:45 |
| SamYaple | if ansible runs again, htose changes are stomped out | 11:45 |
| *** tobe has quit IRC | 11:47 | |
| *** jruano has joined #kolla | 12:07 | |
| *** dwalsh has joined #kolla | 12:12 | |
| pdb | yes | 12:19 |
| *** bradjones has quit IRC | 12:26 | |
| *** bradjones has joined #kolla | 12:27 | |
| *** bradjones has joined #kolla | 12:27 | |
| *** prad_ has quit IRC | 12:34 | |
| *** prad_ has joined #kolla | 12:35 | |
| *** pdb has quit IRC | 12:42 | |
| *** pdb has joined #kolla | 12:42 | |
| *** nihilifer_ has quit IRC | 13:06 | |
| *** jmccarthy has quit IRC | 13:10 | |
| *** jmccarthy1 has joined #kolla | 13:10 | |
| *** kevsi has joined #kolla | 13:11 | |
| *** jmccarthy1 has quit IRC | 13:18 | |
| *** jmccarthy has joined #kolla | 13:18 | |
| *** jmccarthy has quit IRC | 13:21 | |
| *** jmccarthy has joined #kolla | 13:21 | |
| *** jtriley has joined #kolla | 13:50 | |
| pdb | SamYaple: did you have anything in mind as an alternative to the .buildconf files? | 13:53 |
| pdb | I was just putting in a bp to add an arg to build-docker-image, but then it struck me the original intention was probably so you dont have to specify this | 13:54 |
| SamYaple | i dont want to get rid of .buildconf | 14:06 |
| SamYaple | the way we source them actually allows us to execute code | 14:06 |
| SamYaple | (like pulling down a tarball for a specific source install) | 14:06 |
| SamYaple | maybe rename it to extras or something | 14:07 |
| SamYaple | pdb: check multi-node spec again, added a comment elaborating what we were talking about | 14:09 |
| jmccarthy | Any folks actively looking into swift at all at the moment ? | 14:09 |
| SamYaple | jmccarthy: i want to get it in, i am not actively looking at it. It has the wierdness of needing disks and mounts so it is likely oging to be up for debate | 14:10 |
| SamYaple | I would prefer to see swift backed by ceph personally since that requires only the swift proxies | 14:10 |
| jmccarthy | Ok - would a mention of it come under 'Liberty Priority Review' at all ? I'm new to these meetings - interested to help out here if possible | 14:12 |
| SamYaple | i dont know. sdake handles the coordination. i code monkey things | 14:13 |
| *** jmccarthy has quit IRC | 14:13 | |
| *** jmccarthy has joined #kolla | 14:13 | |
| jmccarthy | Ok I'll ping rhallisey and see what's cooking | 14:21 |
| pdb | SamYaple: not so sure about the whole ansible setting the modes part | 14:21 |
| SamYaple | pdb: ansible would always be setting the modes | 14:22 |
| SamYaple | it created the environment variables | 14:22 |
| pdb | I assumed that would be set by the operator prior to deploy | 14:22 |
| pdb | so its clear which method is in use | 14:22 |
| SamYaple | in an ansible config | 14:22 |
| SamYaple | pdb: there will be a single point config to make all of these changes, config type would reside there. it will be clear | 14:24 |
| pdb | right it just sounds like you're describing it as some sort of state machine | 14:24 |
| SamYaple | im not sure how else to work it, but it is late I am tired. You can add a comment there to clarify? | 14:25 |
| pdb | sure | 14:25 |
| *** nihilifer has quit IRC | 14:28 | |
| pdb | SamYaple: replied | 14:37 |
| SamYaple | ah pdb i understand the confusion | 14:37 |
| SamYaple | the env: and bind: were suppose to only run when you chose that method | 14:38 |
| SamYaple | they dont both run | 14:38 |
| jmccarthy | SamYaple: I sent rhallisey a note via irc and launchpad, do you have an idea what might be outstanding or what are the challenges relating to swift offhand ? (i.e. It doesn't appear under this list on github 'The Kolla developers build images in the kollaglue namespace for the following services') | 14:38 |
| SamYaple | pseudo code is pseudo code, mine is no good | 14:38 |
| SamYaple | jmccarthy: yea, the biggest will be the disks | 14:38 |
| pdb | yeah I had a feeling it was just Gerrit messing up the formatting | 14:39 |
| SamYaple | you have to somehow specify disks, format them and mount them automatically | 14:39 |
| SamYaple | i did this with my ceph implementation, and the same could be done for swift, but it was entirely ansible | 14:39 |
| SamYaple | plus i dont like anything about swift but the proxy | 14:39 |
| SamYaple | thats just personally though | 14:39 |
| jmccarthy | Ok, maybe - I was just going to mention ansible, I think maybe some of that voodoo could be handled by initial config via ansible possibly ? Thinking out loud | 14:40 |
| *** mstachow has joined #kolla | 14:40 | |
| *** openstackgerrit has quit IRC | 14:41 | |
| *** openstackgerrit has joined #kolla | 14:41 | |
| mstachow | hello (: | 14:41 |
| SamYaple | hoesntly, thats a big discussion. you could draw up a rough draft of the implementation. then we can discuss | 14:41 |
| SamYaple | mstachow: yo | 14:41 |
| jmccarthy | Will see what Ryan comes back with, and have a think =) | 14:42 |
| jmccarthy | mstachow: hiya | 14:43 |
| bmace | you still here SamYaple? | 14:50 |
| SamYaple | zombishly bmace | 14:55 |
| bmace | lol. i can add it on the spec review, but i had one point on your conversation earlier with pdb about the config file stuff. you mentioned that the config file would be over written with another ansible run. | 14:57 |
| SamYaple | please mention it here, but definetely add it on the spec too for others | 14:58 |
| SamYaple | that could be variable, no reason it MUST be overwritten | 14:58 |
| bmace | i was of the thought that when in this bind mod we would not push out the configs.. at that point we are saying, the configs are yours, do with them as you please.. do your own management of their values / distribution. or maybe we need a separate value / bit if we want to enable to them to do that, and then we, on the ansible side would just handle actual restart of containers, etc. | 14:59 |
| SamYaple | with the option of not overridding it would be what you describe | 14:59 |
| bmace | right, i think if we specifically call out that it is indeed variable, and provide the variable to control that behavior it would be excellent. | 14:59 |
| bmace | i'll check out the latest spec to see what the wording is in there around this. | 15:00 |
| SamYaple | i want to push the defaults personally as they will be the bare minimum to get openstack running | 15:00 |
| SamYaple | we can for sure (its literally one ansible "if" statement) | 15:00 |
| bmace | sure. i think the flow i described in one of my comments for a sort of 'common' scenario was to, for the first run, let kolla do its thing to get all the files with their defaults in place, and then the customer could take control over all the file contents from there. | 15:01 |
| SamYaple | configs definetely should be the only part that kolla doesnt _NEED_ to touch, i like hte enduser having full control over it | 15:01 |
| SamYaple | just so were all on the same page, I am a power to the people kinda guy here. Im an admin/engineer at my job so thats the perspective I am viewing this project from | 15:02 |
| bmace | yup. lets all those puppet / chef / whatever existing folks continue to use most of their current config control systems. | 15:02 |
| *** ccrouch has joined #kolla | 15:03 | |
| SamYaple | theyll likely have to tweak it, but sure why not | 15:03 |
| *** dims_ has quit IRC | 15:03 | |
| SamYaple | honestly it makes it easier to consume the containers with things other than ansible too | 15:03 |
| bmace | agreed. nice of us to have our own orchestration if people want it, but the least odd way for people used to traditional open stack administration to get started with the containerized services in their own solutions. | 15:05 |
| *** jmccarthy has quit IRC | 15:08 | |
| *** dims has joined #kolla | 15:08 | |
| *** dims has quit IRC | 15:08 | |
| *** jmccarthy has joined #kolla | 15:08 | |
| *** dims has joined #kolla | 15:08 | |
| *** rhallisey|pto is now known as rhallisey | 15:09 | |
| *** ChanServ sets mode: +o rhallisey | 15:10 | |
| rhallisey | jmccarthy, hey, I just emailed you back regarding swift | 15:11 |
| jmccarthy | Ok cool - lemme check | 15:11 |
| *** mstachow has quit IRC | 15:25 | |
| *** inc0 has joined #kolla | 15:27 | |
| *** vinkman has joined #kolla | 15:27 | |
| inc0 | afternoon guys | 15:27 |
| SamYaple | afternoon inc0 | 15:31 |
| SamYaple | hey rhallisey | 15:32 |
| rhallisey | SamYaple, hey! | 15:32 |
| *** sdake has joined #kolla | 15:33 | |
| inc0 | rhallisey, how's cinder going?:) | 15:34 |
| sdake | morning folks | 15:34 |
| sdake | rhallisey you gtg for running the team meeting? | 15:34 |
| rhallisey | sdake, ya all set | 15:34 |
| inc0 | I'm eager to try tripleo + kolla for compute nodes | 15:34 |
| inc0 | good morning sdake | 15:35 |
| rhallisey | inc0, cinder is almost done | 15:35 |
| rhallisey | very close | 15:35 |
| rhallisey | inc0, that patch is also close! | 15:35 |
| jmccarthy | inc0: afternoon, sdake: morning | 15:35 |
| inc0 | rhallisey, anything I can help you with? I'm almost done with keepalived | 15:35 |
| rhallisey | inc0, no I think I'm good, just on PTO so I haven't touched it this week | 15:36 |
| inc0 | cool, let me know when you'll have something plz:) | 15:37 |
| rhallisey | sure! | 15:37 |
| sdake | rhallisey nice i my be around for the first few minutes | 15:38 |
| rhallisey | ok cool | 15:38 |
| sdake | i'm supposed tobe at a demo pod but my ass is dragging | 15:38 |
| jmccarthy | SamYaple: I'm the proud owner of a swift bp, go me ! | 15:39 |
| sdake | so if I drop out in the meeting, thats why :) | 15:39 |
| sdake | direlection of duties :( | 15:39 |
| SamYaple | jmccarthy: go you! | 15:39 |
| *** blahRus has joined #kolla | 15:40 | |
| SamYaple | rhallisey: sdake: can we put config management on the agenda? there is a conversation to be had | 15:42 |
| rhallisey | SamYaple, ya sure | 15:43 |
| sdake | samyaple there is an entire section devoted to ansible in the agenda | 15:44 |
| sdake | not guaranteeeing we can get it all out on the table in the time allotted :) | 15:44 |
| *** ccrouch has quit IRC | 15:44 | |
| sdake | but its there | 15:44 |
| SamYaple | good deal | 15:44 |
| *** ccrouch has joined #kolla | 15:45 | |
| *** mstachow has joined #kolla | 15:56 | |
| sdake | hey guys does this link sound familiar: http://t.co/tGuWf1eI22 | 15:57 |
| *** mickt has quit IRC | 15:58 | |
| rhallisey | meeting in 1 minute #openstack-meeting-4 | 15:59 |
| *** mickt has joined #kolla | 15:59 | |
| *** Kupo24z has joined #kolla | 15:59 | |
| sdake | channel up to 52 peoplz | 15:59 |
| sdake | new record :) | 15:59 |
| pdb | sdake: yup that one is doing the rounds :) | 15:59 |
| bmace | lol, love that page sdake. | 16:00 |
| sdake | "mongodb is web scale | 16:00 |
| inc0 | looks kinda like propaganda | 16:00 |
| sdake | "mongodb is shit" | 16:00 |
| sdake | thats the best part | 16:00 |
| inc0 | do you know mango db? | 16:00 |
| SamYaple | its sounds delicious | 16:00 |
| pdb | I didnt get the part where he suggests OpenStack | 16:01 |
| pdb | ew | 16:01 |
| pdb | ? | 16:01 |
| inc0 | https://github.com/dcramer/mangodb more predictible mongo | 16:01 |
| sdake | ya me either | 16:01 |
| inc0 | and faster! | 16:01 |
| sdake | that web page would be so much better if he threw in magnum | 16:01 |
| bmace | lol | 16:01 |
| *** bmace has left #kolla | 16:02 | |
| rhallisey | bmace, meeting? | 16:02 |
| sdake | then he would be on fire ;) | 16:02 |
| *** loth has quit IRC | 16:02 | |
| *** bmace has joined #kolla | 16:02 | |
| *** Kupo24z is now known as loth | 16:02 | |
| sdake | bmace that page would be so much better if he threw in magnum | 16:03 |
| sdake | then he would be on fire ;) | 16:03 |
| bmace | some day sdake.. some day :) | 16:03 |
| *** loth1 has joined #kolla | 16:03 | |
| *** daneyon has joined #kolla | 16:06 | |
| *** jtriley has quit IRC | 16:27 | |
| *** jtriley has joined #kolla | 16:28 | |
| *** prad has quit IRC | 16:36 | |
| *** mickt has quit IRC | 16:41 | |
| *** mickt has joined #kolla | 16:41 | |
| *** prad_ is now known as prad | 16:42 | |
| *** jtriley has quit IRC | 16:55 | |
| SamYaple | first | 16:59 |
| bmace | lol | 17:00 |
| jmccarthy | you ran the light ! | 17:00 |
| *** juggler has joined #kolla | 17:00 | |
| inc0 | rhallisey, admins wants to keep their configs as it is | 17:00 |
| SamYaple | rhallisey: option 2 needs to exist because as an operator i need to turn on debug for one service on one host without reruning ansible | 17:00 |
| juggler | o/ | 17:00 |
| inc0 | also they want to keep their fail procedures as it is | 17:00 |
| jasonsb | inc0: is a kolla-provided upgrade strategy a big enough carrot to lure them? | 17:01 |
| rhallisey | SamYaple, I see :), sounds like devstack | 17:01 |
| jasonsb | inc0: IE: do it this way and upgrades with matching configs is handled nicely | 17:01 |
| SamYaple | ok, I would like to pose a questions to al lthe people who DONT want option 2, is there a valid reason _not_ to include it? | 17:01 |
| SamYaple | rhallisey: i want kolla to replace devstack, so thats not an insult ;) | 17:01 |
| sdake | which was option #2 again? | 17:02 |
| rhallisey | SamYaple, oh I agree | 17:02 |
| SamYaple | bindmount | 17:02 |
| SamYaple | sdake: ^ | 17:02 |
| inc0 | jasonsb, it is, but I'm sure it will makes things harder for them | 17:02 |
| jasonsb | inc0: true true | 17:02 |
| rhallisey | SamYaple, not meant to be an insult, I suggested the same thing to sdake, I just didn't think operators would want it | 17:02 |
| sdake | I have two objections, loss of immutability and two sources of truth | 17:02 |
| SamYaple | rhallisey: its all operators want, immutability be damned | 17:02 |
| sdake | wheither that outweighs the value it bringss, open question | 17:02 |
| SamYaple | sdake: those are objections ot making it default, and valid ones | 17:03 |
| SamYaple | those arent objections to including it | 17:03 |
| SamYaple | its an option for the well informed user | 17:03 |
| inc0 | value it brings: it will be easy to deploy kolla on existing clusters with existing config management tools | 17:03 |
| sdake | samyaple agree | 17:03 |
| sdake | inc0 we are not solving brownfeild deployment problems ;) | 17:03 |
| rhallisey | I like default #1 and I'm fine with #2 and #3 as additional methods | 17:03 |
| pdb | same | 17:04 |
| sdake | with ansible #3 or #2 would have to be teh default | 17:04 |
| rhallisey | #3 I want to use to build up our stack of cinder/neutron containers | 17:04 |
| pdb | its just the mechanism of #3 I dont like | 17:04 |
| rhallisey | stock* | 17:04 |
| SamYaple | sdake: #3 as the default for ansible, #2 is the optinos | 17:04 |
| SamYaple | crudini is not available for ansible | 17:04 |
| rhallisey | ok | 17:04 |
| sdake | samyaple ya I'm pretty sure I just said that :) | 17:04 |
| rhallisey | in that case I'd vote default #3 and #2 as an option | 17:05 |
| SamYaple | sdake: yea lots of people are saying lots of things | 17:05 |
| harmw | guys, (completely different topic) since most of us are still here I wanted to bringt up a slight change to the meetingschedule | 17:05 |
| rhallisey | harmw, I did earlier in the meeting | 17:05 |
| harmw | you did? | 17:05 |
| rhallisey | 1600 -> 1700 time change | 17:05 |
| sdake | during announcements | 17:05 |
| SamYaple | harmw: i am not going to be able to make 1700, this is way late for me as it is | 17:05 |
| rhallisey | is that what you were referring to? | 17:05 |
| harmw | rhallisey: yep, that would mean we'd be having the meeting *now*, which works a lot better :) | 17:06 |
| SamYaple | can we meet halfway with 1630-1730? | 17:06 |
| SamYaple | dont know if we can schedule that | 17:07 |
| rhallisey | does 16:30 work or even 15:30? | 17:07 |
| rhallisey | sdake, ^ you ok with that? | 17:07 |
| rhallisey | I think those times don't bother you | 17:07 |
| sdake | maling list on the timing guys | 17:07 |
| sdake | i can roll with whatever | 17:07 |
| rhallisey | ok cool | 17:07 |
| juggler | curious can meetbot do half hours? | 17:07 |
| sdake | sure | 17:08 |
| sdake | meetbot is not aware of "slots" :) | 17:08 |
| harmw | worst case for me, I'd miss the first 30mins (it's already an improvement as it is) | 17:08 |
| rhallisey | harmw, ya post your concern on the thread | 17:08 |
| juggler | or is it manual set | 17:08 |
| rhallisey | we can work it out | 17:08 |
| juggler | sdake ah :) | 17:08 |
| sdake | rhallisey i think you have an action to follow up on the thread with the results of the meeting that the time change was nacked :) | 17:08 |
| jasonsb | rhallisey: sdake: if you can accomodate 3 of us from hds would like to attend mid-cycle meetup | 17:10 |
| jasonsb | rhallisey: sdake: i'll get the other two to fill out the doodle | 17:10 |
| harmw | the mid-cycle meetup was in Amsterdam, right? | 17:10 |
| sdake | what is hds | 17:10 |
| rhallisey | harmw, San Jose | 17:10 |
| sdake | the midcycle is open to anyone to attend | 17:10 |
| harmw | is that in NL? :p | 17:10 |
| bmace | lol | 17:11 |
| inc0 | don't you want Amsterdam? Its cool city with lot's of funny stuff there:P | 17:11 |
| jasonsb | amsterdam sounds nice | 17:11 |
| *** pdb-mobile has joined #kolla | 17:11 | |
| harmw | jeez, you guys sound like genuine tourists :p | 17:11 |
| jasonsb | hitachi data systems | 17:12 |
| sdake | jasonsb oh i thought you meant some kind of acronym that wasn't a company name :) | 17:12 |
| sdake | jasonsb dont care which company you want to attend from your welcome to attend :) | 17:13 |
| jasonsb | sdake: thank you sir | 17:13 |
| jasonsb | sdake: i didn't want to take too many seats if space was limited | 17:13 |
| inc0 | if I'd guess, we won't run out of space | 17:13 |
| sdake | jasonsb if 100 people sign up I'm going tobe in trouble :) | 17:14 |
| sdake | but I expect 10-25 | 17:14 |
| juggler | spillover at a dennys lol | 17:14 |
| jasonsb | sounds good | 17:14 |
| sdake | not sure on logistics yet | 17:14 |
| sdake | need dates so i can pass them on to the facilities folks | 17:14 |
| jmccarthy | They have Dennys in Amsterdam ? ;) | 17:14 |
| SamYaple | dates so i can get time off | 17:15 |
| *** loth has quit IRC | 17:15 | |
| jasonsb | sdake: i'm also going to pitch to tailgaters to use kolla | 17:15 |
| jasonsb | but its early days for tailgate still | 17:15 |
| SamYaple | lets not let anyone use kolla yet.... were still in alpha | 17:15 |
| SamYaple | :P | 17:15 |
| jasonsb | but you have SamYaple | 17:16 |
| jasonsb | bam bam all done | 17:16 |
| SamYaple | i am here | 17:16 |
| sdake | ok time to get to the conference before daneyon fires me! :) | 17:18 |
| SamYaple | sdake: get me them dates | 17:18 |
| sdake | samyaple there is a 1 week clock on signup | 17:19 |
| sdake | I am not sure what the turnaround on confirmation from facilities is, but I'll let you know the proposed dates | 17:19 |
| SamYaple | i need 30 days notice.... gonna be close. ill tell them about it now so theyll be flexible im sre | 17:19 |
| sdake | ok got to jet | 17:21 |
| sdake | thanks for saying up samyaple ;) | 17:21 |
| *** daneyon has quit IRC | 17:21 | |
| harmw | re:ansible, what I don't see is why there's an additional option avaialbe with that augmented thing | 17:21 |
| SamYaple | me too | 17:21 |
| juggler | cya sdake | 17:21 |
| SamYaple | the sleep awaits | 17:21 |
| harmw | and who's working on galera again? | 17:21 |
| harmw | inc0 was that you? | 17:21 |
| juggler | have a good sleep SamY | 17:21 |
| sdake | harmw the reason is for cusotmization of deployments | 17:21 |
| bmace | nite Sam | 17:21 |
| mstachow | harmw That's me | 17:22 |
| harmw | sdake: ah, kinda like openstack.env customizations upon stuff inside start.sh (in a very, very rough way) | 17:22 |
| harmw | mstachow: ok | 17:22 |
| rhallisey | I'm on PTO for the next 3 days I'll be in and out | 17:22 |
| sdake | yes sort of harmw | 17:22 |
| sdake | hamw i gotta grab a showe rand get to the conference | 17:23 |
| rhallisey | I'll respond to emails | 17:23 |
| harmw | sdake: think I got it now :) | 17:23 |
| *** rhallisey is now known as rhallisey|PTO | 17:23 | |
| *** pdb-mobile has quit IRC | 17:24 | |
| harmw | SamYaple: I think you objected on merging galera into mariadb container, right? | 17:24 |
| harmw | oh shit, he's asleep | 17:24 |
| SamYaple | harmw: i still get pings on my phone though! | 17:24 |
| harmw | anyway, it's more mariadb 5.5 vs mariadb 10.1 | 17:24 |
| harmw | lol | 17:24 |
| SamYaple | harmw: i only partly objected | 17:24 |
| SamYaple | i mentioned that an external repo must be used to get meriadb with wsrep/galera support | 17:25 |
| harmw | wsprep was your concern | 17:25 |
| harmw | yea | 17:25 |
| SamYaple | that might be a no go for rdo stuff | 17:25 |
| SamYaple | dunno | 17:25 |
| harmw | mariadb 10.1 comes with wsprep included, so we'll get that automatically when we go to Fedora22 | 17:26 |
| harmw | rdo on Centos7 is mariadb5.5, so thats different indeed | 17:26 |
| harmw | but perhps worth to look into, albeight after finishing the initial work | 17:26 |
| SamYaple | harmw: there is also the same repo with 5.5 and wsrep | 17:26 |
| SamYaple | centos7 mariadb does not have wsrep (afaik) | 17:27 |
| harmw | probably not | 17:27 |
| harmw | anyway, it comes with mariadb10.1 so when we can get a hold of that it's only a matter of applying the right config bits | 17:28 |
| juggler | need to head out folks. it was nice meeting you all | 17:30 |
| harmw | btw, why JSON and no YAML? | 17:30 |
| harmw | bye! | 17:30 |
| harmw | (even though they share a lot) | 17:30 |
| juggler | later harmw | 17:30 |
| juggler | :) | 17:30 |
| jmccarthy | juggler: laters | 17:30 |
| SamYaple | harmw: mariadb 10.1 is a beta/dev thing | 17:31 |
| SamYaple | dont use that | 17:31 |
| harmw | agreed, didn't know that | 17:31 |
| *** sdake has quit IRC | 17:31 | |
| harmw | I thought it was ready for prime, but perhaps I'm mistaken with 10.0 | 17:31 |
| harmw | since the announcemnt I got this info from was from last year | 17:32 |
| *** rhallisey|PTO sets mode: +o SamYaple | 17:32 | |
| *** ChanServ sets mode: -o SamYaple | 17:32 | |
| SamYaple | harmw: i think 10.1 will always be the dev version | 17:33 |
| SamYaple | they commit there and do unstable things there | 17:33 |
| *** jasonsb has quit IRC | 17:33 | |
| SamYaple | then pull them in | 17:33 |
| harmw | ah, they follow a scheme for that | 17:33 |
| SamYaple | my understanding at least | 17:33 |
| harmw | ok | 17:33 |
| *** SamYaple has left #kolla | 17:34 | |
| *** jasonsb has joined #kolla | 17:34 | |
| *** SamYaple has joined #kolla | 17:34 | |
| harmw | is there an opinion about container logs btw? | 17:34 |
| *** inc0 has quit IRC | 17:34 | |
| SamYaple | > /dev/null | 17:35 |
| SamYaple | logging is.... tricky. thats a whole other beast | 17:35 |
| harmw | yea | 17:35 |
| harmw | I'm just after some opinions, before getting my own thoughts into some new BP | 17:36 |
| jmccarthy | Gotta jet, laters ! =) | 17:37 |
| harmw | getting all logs into 1 datacontainer, rsyslog, perhps logstash, something kibana | 17:37 |
| harmw | bye! | 17:37 |
| SamYaple | so most/all? of the openstack services can send to rsyslog directly i believe | 17:37 |
| SamYaple | we should og that route in my opinion | 17:37 |
| harmw | I'm only just scratching the surface :) | 17:38 |
| SamYaple | dont you know it! | 17:38 |
| harmw | if there are great options available, we should go with those | 17:38 |
| SamYaple | (had to deal with this issue for the past year) :( | 17:38 |
| *** jasonsb has quit IRC | 17:38 | |
| *** jmccarthy has quit IRC | 17:39 | |
| harmw | Ill give a thought some other day again | 17:39 |
| *** daneyon has joined #kolla | 17:40 | |
| *** Slower has quit IRC | 17:44 | |
| *** Slower has joined #kolla | 17:45 | |
| diga | Hi SamYaple | 17:47 |
| diga | When is out meeting today ?? | 17:47 |
| diga | 1700 UTC ?? | 17:47 |
| bmace | 1600, it is over | 17:48 |
| SamYaple | ill never sleep! | 17:48 |
| bmace | or not the kolla meeting? :) | 17:48 |
| SamYaple | diga: the kolla meeting was at 1600 | 17:48 |
| SamYaple | if we had something scheduled, I am not remembering it | 17:48 |
| bmace | it is past 1700 as well utc anyway | 17:49 |
| diga | oops | 17:49 |
| diga | SamYaple: Sorry I missed it today I thought It is at 1700 | 17:49 |
| SamYaple | to sum up "configs configs configs, configs config config. CONFIGS! Config." | 17:50 |
| bmace | pulling a ballmer there? :) https://www.youtube.com/watch?v=8To-6VIJZRE | 17:51 |
| SamYaple | ballmer is life | 17:51 |
| *** jtriley has joined #kolla | 17:52 | |
| bmace | well, life is crazy and therefore ballmer is life.. sure, i can go with that. | 17:52 |
| SamYaple | he brought apple from the edge of bankrupcy | 17:55 |
| SamYaple | too bad it wasn't hist company | 17:55 |
| vinkman | Thankfully it was not… Otherwise Apple would suck almost as much as Win 8 :D | 17:56 |
| *** dwalsh has quit IRC | 17:56 | |
| SamYaple | doesnt it already though? | 17:57 |
| * bmace makes some popcorn. | 17:58 | |
| *** jtriley has quit IRC | 17:58 | |
| *** blahrus_ has joined #kolla | 17:59 | |
| * mstachow makes some popcorn too :O | 17:59 | |
| SamYaple | why? this fight is already over | 17:59 |
| SamYaple | they gave me an Apple macbook thing at work | 18:00 |
| SamYaple | ugh | 18:00 |
| bmace | i actually wouldn't mind one. seems like some nice hardware. it is the price points and extreme vendor lock-in that annoys me. | 18:01 |
| mstachow | I don't like this whole "IT fight". I'm respecting everyone with their choices. Moreover - I really like popcorn ;) | 18:02 |
| bmace | lol, agreed about the popcorn + being good with peoples choices mstachow. i'm not particularly a zealot in either direction. | 18:03 |
| mstachow | http://static.celebuzz.com/uploads/2013/06/14/request-five.gif bmace ;) | 18:05 |
| *** shardy is now known as shardy_afk | 18:05 | |
| bmace | lol. nice msachow. :) | 18:07 |
| *** daneyon has quit IRC | 18:14 | |
| *** daneyon has joined #kolla | 18:16 | |
| *** jasonsb has joined #kolla | 18:20 | |
| *** bradjones has quit IRC | 18:23 | |
| *** jtriley has joined #kolla | 18:24 | |
| *** bradjones has joined #kolla | 18:24 | |
| *** bradjones has joined #kolla | 18:24 | |
| *** daneyon has quit IRC | 18:26 | |
| SamYaple | i like drama. i dont care | 18:27 |
| *** walters has quit IRC | 18:32 | |
| *** walters has joined #kolla | 18:33 | |
| *** dwalsh has joined #kolla | 18:42 | |
| *** nihilifer has joined #kolla | 19:07 | |
| mstachow | bye :) ! | 19:21 |
| *** prad has quit IRC | 19:23 | |
| *** mstachow has quit IRC | 19:24 | |
| *** nihilifer has quit IRC | 19:35 | |
| *** prad has joined #kolla | 19:47 | |
| *** athomas has quit IRC | 19:52 | |
| *** prad_ has joined #kolla | 19:53 | |
| *** prad has quit IRC | 19:54 | |
| *** prad_ is now known as prad | 19:54 | |
| openstackgerrit | Jeff Peeler proposed stackforge/kolla: Fix Heat container env vars and dependencies https://review.openstack.org/189974 | 19:56 |
| *** sdake has joined #kolla | 20:06 | |
| sdake | hey folks | 20:08 |
| vinkman | That was a short conference :D | 20:11 |
| bmace | i expect sdake just did it at 2x speed and told them to watch the replay in slow mo if they can't follow ;) | 20:28 |
| *** Slower has quit IRC | 20:29 | |
| *** jruano has quit IRC | 20:29 | |
| *** sdake has quit IRC | 20:37 | |
| *** Slower has joined #kolla | 20:45 | |
| *** daneyon has joined #kolla | 20:56 | |
| openstackgerrit | Jeff Peeler proposed stackforge/kolla: Fix Heat container env vars and dependencies https://review.openstack.org/189974 | 21:10 |
| *** dwalsh has quit IRC | 21:19 | |
| *** shardy_afk has quit IRC | 21:19 | |
| *** sdake has joined #kolla | 21:22 | |
| *** jtriley has quit IRC | 21:36 | |
| loth1 | sdake: you familar with rdo repos vs cloudsig? I'm not sure what the current recommended is to use for kilo/el7 | 21:47 |
| *** loth1 is now known as loth | 21:47 | |
| sdake | i dont think there is a vs there | 21:52 |
| sdake | rdo repos are the ones to use i think | 21:52 |
| sdake | i wasn't aware there was overlap | 21:52 |
| *** daneyon has quit IRC | 22:06 | |
| jpeeler | i guess that's the same thing as maintaining packages in fedora, except for centos. i guess it makes sense to have packages in the base distribution (assuming i'm understanding their setup correctly) | 22:07 |
| *** daneyon has joined #kolla | 22:44 | |
| *** daneyon has quit IRC | 22:45 | |
| *** daneyon has joined #kolla | 22:45 | |
| *** daneyon has quit IRC | 23:04 | |
| *** daneyon has joined #kolla | 23:07 | |
| *** sdake has quit IRC | 23:10 | |
| *** daneyon has quit IRC | 23:19 | |
| *** daneyon has joined #kolla | 23:19 | |
| *** daneyon has quit IRC | 23:25 | |
| *** jasonsb has quit IRC | 23:26 | |
| *** jasonsb has joined #kolla | 23:27 | |
| *** blahrus_ has quit IRC | 23:34 | |
| *** blahRus has quit IRC | 23:44 | |
| *** jasonsb has quit IRC | 23:50 | |
| *** jasonsb has joined #kolla | 23:50 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
