| SamYaple | diogogmt: is your '~/kolla/.gitreview' file correct? | 00:03 |
|---|---|---|
| SamYaple | does '~/.gitconfig' have a gitreview section? | 00:03 |
| SamYaple | are you using the global git-review.conf? | 00:03 |
| diogogmt | SamYaple http://pastebin.com/gnit7SR4 | 00:07 |
| kfox1111_ | Any updates on the galera recovery guide thing? | 00:07 |
| diogogmt | i even tried using https and i get the same problem, does it take a while for my account to be activated? | 00:07 |
| kfox1111_ | want to deploy kolla in the next few days but can't without it. | 00:07 |
| asalkeld | diogogmt: it might be that | 00:07 |
| diogogmt | i’ll give it a shot tomorrow then, i just register my account 30 mins ago so it might be that | 00:08 |
| SamYaple | kfox1111_: we talked about lightsout at the summit that wil lfix rabbitmq | 00:10 |
| SamYaple | kfox1111_: the galera stuff I can write, but since rabbitmq will be changing I cannot write those docs | 00:10 |
| SamYaple | good news is it will not require operator intervention with the new way ive implemented it | 00:10 |
| SamYaple | so rabbitmq auto recovers | 00:11 |
| *** jasonsb has quit IRC | 00:11 | |
| kfox1111_ | ah. cool. | 00:11 |
| *** jasonsb has joined #kolla | 00:11 | |
| SamYaple | that change will not land until the 1.1 release of Kolla unfortunately | 00:11 |
| SamYaple | i think sdake plans two weeks until then | 00:12 |
| SamYaple | sdake_: ^ | 00:12 |
| kfox1111_ | for the first phase of deployment, I'm just do keystone/dashboard/galera for keystone. | 00:12 |
| kfox1111_ | this will support all of our planned regions. | 00:12 |
| kfox1111_ | so I think we can live without rabbit on that. | 00:12 |
| kfox1111_ | but since its so central, the recovery piece is really critical. | 00:12 |
| diogogmt | switching gears a bit, any ideas where this error would be coming from? http://pastebin.com/XRmzrFeZ | 00:13 |
| diogogmt | i have all the other service images already built | 00:13 |
| SamYaple | sounds interesting. let me know how that works out. | 00:13 |
| kfox1111_ | yup. :) | 00:13 |
| diogogmt | it just fails at the lokolla/centos-binary-openstack-base:latest | 00:13 |
| SamYaple | kfox1111_: after the summit discussion i will need to tweak the mariabd container so let me get on that now | 00:13 |
| kfox1111_ | k. thx. | 00:13 |
| kfox1111_ | I was really hoping that keystone federation would be finished. :/ | 00:14 |
| kfox1111_ | then we could do a keystone for authentication, and a keystone per region that is federated. | 00:14 |
| kfox1111_ | but dashboard support still doesnt work!?! :( | 00:14 |
| kfox1111_ | federation without dashboard isn't very useful. :/ | 00:14 |
| SamYaple | in kolla? | 00:14 |
| SamYaple | we havent done much with federation in kolla. where you see limitations please file bugs | 00:15 |
| kfox1111_ | no. in openstack-dashboard. it onlyh knows how to login to one keystone and doesn't support federation at all yet. | 00:15 |
| SamYaple | oh ok | 00:15 |
| kfox1111_ | it needs to know how to authenticate to one keystone, then ask that keystone for a token for the federated to keystone. | 00:15 |
| SamYaple | yup | 00:16 |
| kfox1111_ | not a thing yet. :/ | 00:16 |
| SamYaple | itll get there | 00:16 |
| kfox1111_ | yeah. but when public clouds' are selling you "you can federate to us", its not quite true. | 00:16 |
| kfox1111_ | its more like, if your an api developer, you can federate to us. | 00:16 |
| kfox1111_ | if your an end user, not so much. :/ | 00:16 |
| kfox1111_ | hopefully mitaka will finish it up. | 00:17 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Add Docker in Docker container https://review.openstack.org/241444 | 00:31 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Remove unused setup_fedora.sh https://review.openstack.org/241498 | 00:31 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Restrucutre gate scripts https://review.openstack.org/241499 | 00:31 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Adjust gate btrfs setup https://review.openstack.org/241465 | 00:31 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 00:31 |
| *** achanda has quit IRC | 00:32 | |
| SamYaple | kfox1111_: so you know how to recover from a normal galera lightout event? | 00:41 |
| SamYaple | or is this all new? | 00:41 |
| kfox1111_ | I do, but my management doesn.t so if you can easily copy/paste in those bits, that woudl be cool. if its much more expedient to just link to the other docs, thats cool too. | 00:42 |
| *** kelkoobenoitr has quit IRC | 00:42 | |
| *** diogogmt has quit IRC | 00:42 | |
| SamYaple | ok ill link a few docs around and post the general idea | 00:42 |
| SamYaple | im going to try and write a playbook to automate the steps involved | 00:42 |
| *** nihilifer has quit IRC | 00:44 | |
| kfox1111_ | awesome. thanks. :) | 00:44 |
| *** kelkoobenoitr has joined #kolla | 00:46 | |
| kfox1111_ | should we be standing up our own docker registry or using stuff streight off the hub, and should we be building containers ourselves, or using the kollaglue ones? | 00:52 |
| SamYaple | your call on both accounts | 00:52 |
| SamYaple | kolla will be providing vanilla images, but you can provide our own as well | 00:53 |
| kfox1111_ | ok. if you were me? :) | 00:53 |
| SamYaple | well pushing to dockerhub isnt an option for me, so i have a local registry | 00:53 |
| kfox1111_ | ah. | 00:53 |
| SamYaple | but if it were prod, it would be in dockerhub | 00:53 |
| SamYaple | it just slows down dev too much | 00:53 |
| kfox1111_ | yeah, I could see that. | 00:54 |
| kfox1111_ | k. thanks. | 00:54 |
| SamYaple | for the company i work at I will definetely be building my own images, but i will also be applying custom patches on nova and neutron | 00:54 |
| SamYaple | so i have those forked | 00:54 |
| kfox1111_ | is the repo location all or nothing? | 00:54 |
| SamYaple | using Kolla-team build images makes you dependant on us to deliever updates, and that may not be acceptable to your SLAs | 00:55 |
| SamYaple | no, you can mix and match | 00:55 |
| SamYaple | but if the base layer does match you end up with different libs and sutff in the containers | 00:55 |
| kfox1111_ | ok. so I may just pull kollaglue ones all local, just to make it easy, then I can switch out some for others later if I need to. | 00:55 |
| SamYaple | yes that should work just fine | 00:56 |
| kfox1111_ | cool. thanks. :) | 00:56 |
| SamYaple | but from a purist standpoint you introduce the risk of different libs and that mamy cause problems | 00:56 |
| SamYaple | just keep it in mind | 00:56 |
| kfox1111_ | yeah. always a potential issue with docker containers in general. | 00:57 |
| SamYaple | no more so than bare metal | 00:57 |
| SamYaple | infact less so than bare metal | 00:57 |
| kfox1111_ | right. :) | 00:57 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 00:57 |
| kfox1111_ | hmm.... | 00:58 |
| kfox1111_ | https://hub.docker.com/r/kollaglue/centos-rdo-base/tags/ | 00:58 |
| kfox1111_ | don't see a liberty or a latest. | 00:58 |
| kfox1111_ | what's the tagging scheme? | 00:58 |
| SamYaple | there should be a liberty-1.0.0 image, but perhaps no one pushed it. i dont use centos so i dont know who was taking up the reighs on that one | 00:59 |
| SamYaple | sdake should know | 00:59 |
| kfox1111_ | k. thx. | 00:59 |
| *** sdake_ has quit IRC | 01:12 | |
| *** sdake has joined #kolla | 01:14 | |
| *** ssurana has quit IRC | 01:14 | |
| kfox1111_ | sdake: not seeing a liberty-1.0.0 in kollaglue/centos-rdo-base. know when those will be pushed? | 01:15 |
| sdake | kofx they werednt psuehed because 1.0.0 was doa | 01:23 |
| *** asalkeld is now known as asalkeld_afk | 01:25 | |
| *** tzn has quit IRC | 01:25 | |
| bmace | eh? i know a while back there was talk about when to push images up to dockerhub and tag them, etc. were the images non-function as of the 1.0.0 tag? | 01:26 |
| SamYaple | bmace: they were fine when 1.0.0 was tagged | 01:27 |
| SamYaple | there was an update to xtrabackup that broke centos (only centos) | 01:27 |
| SamYaple | so since the images werent pushed after they were tested for the tag | 01:27 |
| SamYaple | 1.0.0 can't be built anymore for centos | 01:27 |
| SamYaple | ubuntu still works | 01:28 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Add Docker in Docker container https://review.openstack.org/241444 | 01:28 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Remove unused setup_fedora.sh https://review.openstack.org/241498 | 01:28 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Restrucutre gate scripts https://review.openstack.org/241499 | 01:28 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Adjust gate btrfs setup https://review.openstack.org/241465 | 01:28 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 01:28 |
| *** ByPasS has quit IRC | 01:28 | |
| bmace | seems like it would be awfully nice to have at least some set of working images out there, for some distro. | 01:29 |
| kfox1111_ | oh. bummer.... | 01:29 |
| *** ByPasS has joined #kolla | 01:29 | |
| kfox1111_ | whem will 1.1 be released? | 01:29 |
| SamYaple | sdake: ^ | 01:30 |
| bmace | no point in even having a space on dockerhub if the images are always useless :) | 01:30 |
| SamYaple | kfox1111_: there is a list of things needed to be backported/fixed for 1.1.0 and we have a timeframe to do so | 01:30 |
| kfox1111_ | I need to get a keystone/galera cluster up by next week. :/ | 01:30 |
| SamYaple | bmace: no the problem was the images were never pushed | 01:30 |
| SamYaple | bmace: they _would_ be useful if they were pushed when they were tested | 01:30 |
| SamYaple | but you can blame dockerhub for alot of this | 01:31 |
| SamYaple | pushes take like 8 hours | 01:31 |
| SamYaple | stupid 2.0 | 01:31 |
| bmace | yeah, that is downright silly | 01:31 |
| kfox1111_ | yeah. shouldn't take that long. :/ | 01:31 |
| SamYaple | from my house I have to let it run for 14 hours for a full push | 01:31 |
| SamYaple | im only pushing up 2GB of raw data :( | 01:32 |
| kfox1111_ | wow. | 01:32 |
| kfox1111_ | hmmm... | 01:32 |
| SamYaple | should* only be pushing up 2GB | 01:32 |
| kfox1111_ | would it be possible to just upload a bunch of generated Docker files to github repo's that are tracked by hub autobuilding repo's? | 01:33 |
| kfox1111_ | (Bad hack....) | 01:33 |
| SamYaple | the 2.0 registry has a problem in that it makes you push the same layer up multiple times | 01:33 |
| kfox1111_ | ah. | 01:33 |
| SamYaple | kfox1111_: yes i was just looking into that :) | 01:33 |
| SamYaple | it may be possible yes | 01:33 |
| *** tpot has joined #kolla | 01:33 | |
| *** diogogmt has joined #kolla | 01:33 | |
| kfox1111_ | the one issue I've run into with that path, is you can auto build, or you can upload tagged stuff, | 01:33 |
| kfox1111_ | but you cant tag autobuilt stuff without uploading. :/ | 01:34 |
| kfox1111_ | so you can't do an autobuild, test it, then tag the tested thing as "stable" :/ | 01:34 |
| SamYaple | you can.... but you have to download it and repush it up P | 01:34 |
| kfox1111_ | I wasn't able to push to an autobuilt repo. it was blocking it. | 01:34 |
| kfox1111_ | you have to create a stable repo. :/ | 01:34 |
| kfox1111_ | and then you almost don't want to bother with the autobuilding one. :/ | 01:34 |
| SamYaple | oh then nevermind | 01:35 |
| SamYaple | ok i gotta go get food for the family | 01:35 |
| SamYaple | bbiab | 01:35 |
| kfox1111_ | k. have a good one. | 01:35 |
| *** dolpher has joined #kolla | 01:40 | |
| *** tzn has joined #kolla | 01:46 | |
| *** dims has joined #kolla | 01:49 | |
| *** tzn has quit IRC | 01:56 | |
| *** jasonsb has quit IRC | 02:02 | |
| sdake | kfox i think tagging 1.0.1 and dpsuhing images is a reasonable course of faction | 02:03 |
| sdake | kfox1111_ ^^ | 02:03 |
| sdake | i just woke up from 36 hours of sleep - so something clearly is busted on my end | 02:03 |
| sdake | and my mom is afu in the hospital atm so i'm a bit defocused from kolla atm | 02:04 |
| *** bmace_ has joined #kolla | 02:08 | |
| *** tfukushima has joined #kolla | 02:08 | |
| SamYaple | hope your mom gets better | 02:08 |
| *** alisonh1 has joined #kolla | 02:09 | |
| *** stvnoyes1 has joined #kolla | 02:10 | |
| *** alisonh has quit IRC | 02:10 | |
| *** stvnoyes has quit IRC | 02:11 | |
| *** bmace__ has joined #kolla | 02:11 | |
| *** bmace has quit IRC | 02:11 | |
| *** alisonh has joined #kolla | 02:12 | |
| *** stvnoyes has joined #kolla | 02:13 | |
| *** alisonh1 has quit IRC | 02:14 | |
| *** stvnoyes1 has quit IRC | 02:14 | |
| *** bmace_ has quit IRC | 02:14 | |
| *** tfukushima has quit IRC | 02:17 | |
| *** asalkeld_afk is now known as asalkeld | 02:18 | |
| *** tzn has joined #kolla | 02:23 | |
| *** tfukushima has joined #kolla | 02:25 | |
| sdake | SamYaple wsass the team meeting held this mroning? | 02:25 |
| SamYaple | i was asleep | 02:26 |
| SamYaple | looking at the logs.... no | 02:26 |
| SamYaple | :) timezone confusion | 02:26 |
| *** achanda has joined #kolla | 02:27 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Add Docker in Docker container https://review.openstack.org/241444 | 02:27 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Remove unused setup_fedora.sh https://review.openstack.org/241498 | 02:27 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Restrucutre gate scripts https://review.openstack.org/241499 | 02:27 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Adjust gate btrfs setup https://review.openstack.org/241465 | 02:27 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 02:27 |
| docaedo | what's the regularly scheduled time for the meeting? I need to put it on my calendar (for weeks when there's less jet lag, time zone confusion and sicknesses of course) | 02:27 |
| SamYaple | docaedo: its in the topic | 02:27 |
| SamYaple | 1630 UTC | 02:27 |
| SamYaple | wednesday | 02:27 |
| *** tzn has quit IRC | 02:28 | |
| docaedo | SamYaple: ahahaha, yep, right in front of me - but who looks at topics?? | 02:28 |
| docaedo | SamYaple: thanks :) | 02:28 |
| SamYaple | not me! | 02:28 |
| *** tzn has joined #kolla | 02:28 | |
| *** tzn has quit IRC | 02:29 | |
| docaedo | ah damn, math tells me that translates to a local time which finds me in the car taking a kid to a thing every wednesday :/ | 02:29 |
| SamYaple | use yer phone | 02:29 |
| * docaedo can't wait for robot cars to arrive, so he can join while in transit :) | 02:30 | |
| *** _tzn has joined #kolla | 02:30 | |
| *** vilobhmm11 has quit IRC | 02:35 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 02:47 |
| *** tfukushima has quit IRC | 02:51 | |
| *** tfukushima has joined #kolla | 02:55 | |
| *** sdake_ has joined #kolla | 03:02 | |
| *** sdake has quit IRC | 03:03 | |
| *** cloudnautique has joined #kolla | 03:10 | |
| *** cloudnautique has quit IRC | 03:11 | |
| *** sdake_ has quit IRC | 03:16 | |
| *** cloudnautique has joined #kolla | 03:26 | |
| *** cloudnautique has quit IRC | 03:26 | |
| *** cloudnautique has joined #kolla | 03:27 | |
| *** tobe has joined #kolla | 03:27 | |
| *** jasonsb has joined #kolla | 03:27 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 03:30 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Allow disabling of sysctl values https://review.openstack.org/241856 | 03:30 |
| *** tfukushima has quit IRC | 03:34 | |
| *** jasonsb has quit IRC | 03:34 | |
| *** _tzn has quit IRC | 03:35 | |
| *** cemmason has joined #kolla | 03:43 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 03:43 |
| *** tfukushima has joined #kolla | 03:47 | |
| *** unicell has joined #kolla | 04:02 | |
| *** unicell1 has quit IRC | 04:02 | |
| *** Slower has joined #kolla | 04:05 | |
| *** tfukushima has quit IRC | 04:06 | |
| *** vilobhmm11 has joined #kolla | 04:14 | |
| *** achanda has quit IRC | 04:22 | |
| *** tzn has joined #kolla | 04:25 | |
| *** tzn has quit IRC | 04:30 | |
| *** jasonsb has joined #kolla | 04:38 | |
| *** tfukushima has joined #kolla | 05:07 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Add Docker in Docker container https://review.openstack.org/241444 | 05:11 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Allow disabling of sysctl values https://review.openstack.org/241856 | 05:11 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Restrucutre gate scripts https://review.openstack.org/241499 | 05:11 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 05:11 |
| *** tfukushima has quit IRC | 05:11 | |
| *** achanda has joined #kolla | 05:22 | |
| *** achanda has quit IRC | 05:30 | |
| *** nihilifer has joined #kolla | 05:39 | |
| *** athomas has quit IRC | 05:41 | |
| *** achanda has joined #kolla | 05:42 | |
| *** dwalsh has joined #kolla | 05:55 | |
| *** tfukushima has joined #kolla | 05:58 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 06:06 |
| *** tpot has quit IRC | 06:14 | |
| *** vilobhmm111 has joined #kolla | 06:20 | |
| *** vilobhmm11 has quit IRC | 06:23 | |
| *** tzn has joined #kolla | 06:27 | |
| *** tzn has quit IRC | 06:29 | |
| *** tzn has joined #kolla | 06:29 | |
| openstackgerrit | Kuo-tung Kao proposed openstack/kolla: using registry v2 in image-building doc https://review.openstack.org/241900 | 06:33 |
| *** rajalokan has joined #kolla | 06:37 | |
| *** rajalokan has quit IRC | 06:38 | |
| nihilifer | Kuo-tung, are you in IRC? ;) | 06:38 |
| kjelly | nihilifer: Yes | 06:47 |
| nihilifer | kjelly: hi | 06:47 |
| kjelly | nihilifer: I reproduct the problem I meet. And post error message to you. | 06:47 |
| nihilifer | kjelly: oh, ok | 06:48 |
| nihilifer | kjelly: I wanted to ask about v2 registry | 06:48 |
| nihilifer | does it work for you fine? | 06:48 |
| kjelly | nihilifer: Yes, I use v2 and work fine. | 06:49 |
| kjelly | [+1]1d [user@user-GA-6PXSV3:~/github/kolla] master 1 ± sudo docker run -d -p 3000:5000 --restart=always registry | 06:49 |
| kjelly | [sudo] password for user: | 06:49 |
| kjelly | 8043877990ce455280046d48d4ab86b9270b683410d95566015280fb6876fd92 | 06:49 |
| kjelly | [+1]1d [user@user-GA-6PXSV3:~/github/kolla] master ± sudo tools/build.py --registry 172.22.2.81:3000 -b ubuntu -t source --push data | 06:49 |
| kjelly | INFO:__main__:data:Building | 06:49 |
| kjelly | INFO:__main__:data:Step 0 : FROM ubuntu:latest | 06:49 |
| kjelly | INFO:__main__:data: ---> 1d073211c498 | 06:49 |
| kjelly | INFO:__main__:data:Step 1 : MAINTAINER Kolla Project (https://launchpad.net/kolla) | 06:49 |
| kjelly | INFO:__main__:data: ---> Using cache | 06:49 |
| kjelly | INFO:__main__:data: ---> a1d87cb7dd4a | 06:49 |
| kjelly | INFO:__main__:data:Step 2 : CMD /bin/sleep infinity | 06:49 |
| nihilifer | nooooo | 06:49 |
| kjelly | INFO:__main__:data: ---> Using cache | 06:49 |
| kjelly | INFO:__main__:data: ---> 09f6ae18edc6 | 06:49 |
| kjelly | INFO:__main__:data:Successfully built 09f6ae18edc6 | 06:49 |
| kjelly | INFO:__main__:data:Built | 06:49 |
| kjelly | ERROR:__main__:unable to ping registry endpoint https://172.22.2.81:3000/v0/ | 06:49 |
| kjelly | v2 ping attempt failed with error: Get https://172.22.2.81:3000/v2/: EOF | 06:49 |
| nihilifer | please not paste it here!!! | 06:49 |
| kjelly | v1 ping attempt failed with error: Get https://172.22.2.81:3000/v1/_ping: EOF | 06:50 |
| kjelly | INFO:__main__:Images that failed to build | 06:50 |
| kjelly | INFO:__main__:=========================== | 06:50 |
| kjelly | ERROR:__main__:data Failed with status: error | 06:50 |
| nihilifer | omg... | 06:50 |
| kjelly | Well, So many message. | 06:50 |
| kjelly | nihilifer: Sorry | 06:50 |
| nihilifer | paste.openstack.org ;) | 06:50 |
| nihilifer | plz | 06:50 |
| nihilifer | do you use vagrant or you deployed it somewhere else? | 06:50 |
| kjelly | nihilifer: http://paste.openstack.org/show/478051/ | 06:51 |
| nihilifer | ok, thx | 06:51 |
| kjelly | nihilifer: I am not use vagrant. And the issue happen when building image then pushing. | 06:51 |
| kjelly | nihilifer:wait | 06:52 |
| nihilifer | sorry, I'm so confused now... | 06:53 |
| kjelly | nihilifer: wait a moment. | 06:53 |
| nihilifer | you say that registry v2 is working fine for you and in the same time you cannot ping to it or push? | 06:53 |
| nihilifer | or are these things separate? | 06:53 |
| nihilifer | ok, I'll wait | 06:54 |
| kjelly | nihilifer: I found I can't reproduce the error I meet before. | 07:00 |
| kjelly | nihilifer: In 11/03, I have error about the docker registry too old. But I don't save the message. | 07:00 |
| kjelly | nihilifer: I abandon the patch now. And try to reproduct the problem I meet before. If I can reproduct it, I will reopen the patch. | 07:01 |
| nihilifer | well, if registry v2 will work for someone, it would be great to make such a change to kolla | 07:02 |
| kjelly | nihilifer: Now I can push image to registry v2. But I don't test it using ansible. | 07:04 |
| kjelly | nihilifer: I will test it using ansible in all in one mode. | 07:04 |
| bmace__ | that was a docker version specific issue, there is a known ansible bug related to it. | 07:07 |
| bmace__ | i believe the problem was with docker 1.8.3 with v1 registries | 07:07 |
| kjelly | nihilifer: I recall what the problem I meet. http://paste.openstack.org/show/478053/ | 07:08 |
| kjelly | nihilifer: I have the problem when deploying instead of building | 07:08 |
| SamYaple | the problem with docker 1.8.3 is registries and ansible module | 07:14 |
| SamYaple | blame docker for chaning the api without chanigng the api version | 07:14 |
| SamYaple | i brought it to thier attention | 07:14 |
| kjelly | SamYaple: http://paste.openstack.org/show/478053/ Is the problem docker issue? | 07:17 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Add Docker in Docker container https://review.openstack.org/241444 | 07:17 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Allow disabling of sysctl values https://review.openstack.org/241856 | 07:17 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Restrucutre gate scripts https://review.openstack.org/241499 | 07:17 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 07:17 |
| SamYaple | kjelly: kinda docker changed some stuff (you seem to be rrunning >1.8.2 docker) in 1.8.3 | 07:18 |
| SamYaple | kjelly: because of this some outputs changed. the ansible docker module has to update to correct for these changes | 07:19 |
| SamYaple | the issue is two fold, firstly ansible will not be releasing another 1.9.x tag | 07:19 |
| SamYaple | so onto ansible 2.0 we go | 07:19 |
| SamYaple | second, ansible will not be catching this return (without additional pressure) forcing you to use the v2 registry | 07:20 |
| SamYaple | so basically we are stuck with 1.9.4 ansible and docker 1.8.2 as a requirement until ansible 2.0 lands | 07:20 |
| SamYaple | then we will have to force ansible 2.0 and regisry v2 across the board | 07:21 |
| SamYaple | when docker 1.10 drops we will also have a hard requirement on that since it has some features we will be using | 07:21 |
| *** sdake has joined #kolla | 07:22 | |
| kjelly | SamYaple: Yes, I use Docker version 1.8.3. I reinstall my computer in monday. | 07:23 |
| kjelly | So I hit the issue. | 07:23 |
| *** sdake_ has joined #kolla | 07:24 | |
| SamYaple | Kolla has a hard <=1.8.2 requitement | 07:25 |
| SamYaple | requirement* | 07:25 |
| SamYaple | its in the documentation | 07:25 |
| kjelly | Ok. I saw. Thank for SamYaple, nihilifer | 07:27 |
| *** sdake has quit IRC | 07:27 | |
| SamYaple | sorry about the trouble | 07:29 |
| SamYaple | I have been seriously considering building a docker module for us | 07:29 |
| SamYaple | im tired of this change all the time stuff | 07:29 |
| kjelly | SamYaple: It seems good. | 07:31 |
| sdake_ | what type of module would oyo ubuild samyaple | 07:32 |
| bmace__ | i assume a replacement of the ansible docker stuff | 07:34 |
| SamYaple | sdake_: what bmace__ said | 07:34 |
| sdake_ | soundsgood go forth and prospoer ;) | 07:34 |
| SamYaple | its overly bloated and it has to have support for all versions of all docker | 07:34 |
| SamYaple | i hate breaking away from the community though | 07:35 |
| bmace__ | SamYaple except instead it has broken support for many versions? :) | 07:35 |
| SamYaple | still seems like it may resolve more problems for us as time goes on | 07:35 |
| SamYaple | yea i know | 07:35 |
| SamYaple | if there was a better way..... | 07:35 |
| SamYaple | it feels dirty | 07:35 |
| sdake_ | feels good to me | 07:35 |
| sdake_ | it woudl be differenet in ansiblee guys would cut a 1.9.5 | 07:35 |
| sdake_ | which they wil not do | 07:35 |
| * SamYaple makes note: Never cross sdake_ | 07:36 | |
| sdake_ | just being pragmatic here ;) | 07:36 |
| SamYaple | me too its why i said it | 07:36 |
| SamYaple | it would help us get new features quickly and fix bugs as well | 07:37 |
| SamYaple | the docker module is the most volitale module we use from them... and the most critical | 07:37 |
| SamYaple | maybe it is the right call | 07:37 |
| sdake_ | imo get a review up and ssee what the cr team says | 07:40 |
| SamYaple | it would allow us to keep using registry v1 until they fix v2 | 07:41 |
| sdake_ | boy my time is afu | 07:45 |
| sdake_ | slept 36 hours about 6 hours ago | 07:45 |
| sdake_ | ready to go to bbed again :( | 07:45 |
| kjelly | Good night, sdake. | 07:46 |
| sdake_ | need to stay up | 07:46 |
| *** vilobhmm111 has quit IRC | 07:47 | |
| SamYaple | sdake_: ive been the same way | 07:50 |
| *** slotti has joined #kolla | 07:56 | |
| *** zigo has joined #kolla | 08:00 | |
| zigo | Hi there! | 08:00 |
| zigo | Anyone awake already over here? | 08:00 |
| asalkeld | hi zigo | 08:00 |
| zigo | Hi ! :) | 08:00 |
| zigo | asalkeld: I'm starting the initial Debian packaging of Kolla. | 08:00 |
| asalkeld | nice | 08:00 |
| zigo | asalkeld: Can you help by replying to a few questions? | 08:01 |
| zigo | Easy questions ... | 08:01 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Add Docker in Docker container https://review.openstack.org/241444 | 08:01 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Allow disabling of sysctl values https://review.openstack.org/241856 | 08:01 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Restrucutre gate scripts https://review.openstack.org/241499 | 08:01 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 08:01 |
| asalkeld | as best as i can, others here | 08:01 |
| asalkeld | can help too | 08:01 |
| zigo | I need to know what are the names of daemons running. | 08:01 |
| SamYaple | zigo: awesome. i tried to make that as simple as possible | 08:01 |
| zigo | I haven't seen them "documented" in the setup.cfg. | 08:01 |
| asalkeld | zigo: no deamons | 08:01 |
| zigo | Ah? :) | 08:01 |
| *** inc0 has joined #kolla | 08:01 | |
| zigo | So, just python code and some config files? | 08:01 |
| asalkeld | mostly dockerfiles and ansible | 08:02 |
| zigo | Hum... | 08:02 |
| asalkeld | would you normally package ansible stuff? | 08:02 |
| zigo | I haven't done so yet, but anything can be packaged. | 08:02 |
| asalkeld | k | 08:03 |
| zigo | I guess it'd go in /usr/share/kolla/ansible, no? | 08:03 |
| asalkeld | i'd guess | 08:03 |
| asalkeld | zigo: i am truely not familar with packaging of ansible and docker files | 08:04 |
| zigo | Does the setup.py install everything as it should already? | 08:04 |
| inc0 | good morning | 08:04 |
| asalkeld | hi inc0 | 08:04 |
| zigo | The data_files seems to say so. | 08:04 |
| *** shardy has joined #kolla | 08:04 | |
| sdake_ | zigo yes we think so | 08:04 |
| zigo | (within setup.cfg, I mean) | 08:04 |
| zigo | Ok, easy enough then. | 08:05 |
| kjelly | inc0： morning, inc0 | 08:05 |
| zigo | I'll try. | 08:05 |
| sdake_ | zigo note we kind of guessed on the packaging but i think its correct | 08:05 |
| zigo | Would it make sense to have just 2 packages, one called python-kolla, and one called just kolla then? | 08:05 |
| asalkeld | i don't see an include of anisble | 08:05 |
| asalkeld | in manifest.ini | 08:05 |
| sdake_ | asalkeld that is handled in setup.cfg | 08:05 |
| asalkeld | ok, cool | 08:06 |
| inc0 | we includes ansible in ansible modules at least | 08:06 |
| zigo | asalkeld: There's data_files = share/kolla/ansible = ansible/* in the setup.cfg, that's enough. | 08:07 |
| sdake_ | zigo persoanlly i woul go with one package but if python needs to be an a separate packagey ou can ddo it that way | 08:07 |
| inc0 | anyway, I'm out of context, what seems to be an issue? | 08:07 |
| zigo | sdake_: It does, because we need to allow later move from python2 (ie: python-kolla) to python3 (ie: python3-kolla) | 08:07 |
| asalkeld | inc0: zigo is doing debian packaging of kolla | 08:07 |
| SamYaple | hello inc0 | 08:08 |
| inc0 | you are aware that ansible has slightly different licence | 08:08 |
| SamYaple | how is the life? | 08:08 |
| inc0 | got my visa;) | 08:08 |
| inc0 | in 2 weeks from now I'll be in US | 08:08 |
| SamYaple | w00t w00t | 08:08 |
| inc0 | less than 2 week really | 08:08 |
| inc0 | that means I'll be pretty inactive in following days, for which I'm terribly sorry | 08:09 |
| inc0 | moving across ocean will consume all my attention | 08:09 |
| SamYaple | i have multinode gating stuff in the works :D | 08:09 |
| inc0 | cool | 08:09 |
| asalkeld | guys could you imagine that *if* we split kolla-anisble, that the config templates would go into a common place? | 08:10 |
| asalkeld | i want to rely on those | 08:10 |
| *** dwalsh has quit IRC | 08:10 | |
| SamYaple | asalkeld: the jinja2 templates in the ansible/*/templates folders? | 08:10 |
| inc0 | what kind of config do you refer to? I don't recall any overlap | 08:10 |
| asalkeld | yip | 08:10 |
| inc0 | ah | 08:10 |
| inc0 | I would expect it will... | 08:11 |
| asalkeld | so i am driving the jinja for those files | 08:11 |
| SamYaple | hmm im not sure how that would work | 08:11 |
| SamYaple | im sure we can get it to work out though | 08:11 |
| asalkeld | (partially redoing some of wat ansible does) | 08:11 |
| SamYaple | thats to be expected | 08:11 |
| inc0 | we can symlink stuff... | 08:11 |
| asalkeld | seems a waist to totlly copy that stuff | 08:11 |
| SamYaple | inc0: >:( | 08:11 |
| *** egonzalez__ has joined #kolla | 08:12 | |
| inc0 | just saying | 08:12 |
| *** egonzalez__ has quit IRC | 08:12 | |
| *** egonzalez90 has joined #kolla | 08:12 | |
| asalkeld | SamYaple: how is "hostvars" genterated by ansible? do you know? | 08:12 |
| SamYaple | asalkeld: ansible has a "setup" module that fetchs all that stuff | 08:12 |
| inc0 | asalkeld, facts? | 08:12 |
| SamYaple | strictly and ansible thing | 08:13 |
| SamYaple | asalkeld: i think the real trouble will come into is that if we split off kolla-ansible then people deploying with kolla-ansible wont be cloning kolla unless they are building images | 08:13 |
| asalkeld | SamYaple: is it runtime calcultated? | 08:13 |
| SamYaple | asalkeld: most of the time | 08:13 |
| SamYaple | though ansible supports fact caching and database facts | 08:13 |
| SamYaple | all outside of kollas control | 08:13 |
| inc0 | when you run ansible-playbook you'll see "gathering facts" step | 08:13 |
| asalkeld | ok, i need to convert that into zookeeper variables i think | 08:13 |
| inc0 | that's the place | 08:13 |
| SamYaple | inc0: only if not doing fact caching, its not a requirement | 08:14 |
| inc0 | that's general idea | 08:14 |
| asalkeld | ok, that's enough for the moment - thanks guys | 08:14 |
| SamYaple | asalkeld: ill have to see an implementation. my 5 minute brainstrom says it wont work out | 08:15 |
| SamYaple | bubt im not opposed | 08:15 |
| *** openstackgerrit has quit IRC | 08:16 | |
| *** openstackgerrit has joined #kolla | 08:16 | |
| *** akwasnie has joined #kolla | 08:17 | |
| asalkeld | :-( | 08:17 |
| *** egonzalez90 has quit IRC | 08:18 | |
| *** egonzalez90 has joined #kolla | 08:18 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Restrucutre gate scripts https://review.openstack.org/241499 | 08:18 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 08:18 |
| kjelly | If you use ubuntu 15.10, it seems that you can't downgrade the docker version to 1.8.2 | 08:20 |
| SamYaple | kjelly: whats the reason? | 08:20 |
| SamYaple | no that cant be right | 08:20 |
| SamYaple | im running 1.8.2 on 15.10 right now | 08:20 |
| kjelly | I use the command to list version, "sudo apt-cache policy docker-engine". I saw "1.9.0-0~wily 0" only. | 08:20 |
| SamYaple | you can still pin | 08:21 |
| SamYaple | try uninstalling docker entirely then `apt-get install docker-engine=1.8.2 | 08:21 |
| kjelly | SamYaple: Ok, try it. | 08:22 |
| SamYaple | kjelly: http://apt.dockerproject.org/repo/dists/ubuntu-wily/main/binary-amd64/Packages | 08:22 |
| SamYaple | its there | 08:22 |
| SamYaple | `apt-get intall docker-engine=1.8.2-0~wily` | 08:22 |
| SamYaple | fyi policy only shows the installed version and most recent version available per repo | 08:23 |
| kjelly | SamYaple: E: Version '1.8.2-0~wily' for 'docker-engine' was not found | 08:23 |
| SamYaple | kjelly: did you add the docker repo? | 08:25 |
| *** achanda has quit IRC | 08:25 | |
| kjelly | SamYaple: I rerun the command, 'curl -sSL https://get.docker.io | bash'. Now it found the package. | 08:26 |
| SamYaple | wierd | 08:26 |
| SamYaple | glad its wokring | 08:27 |
| kjelly | SamYaple: ok, it works. | 08:27 |
| SamYaple | $ cat /etc/apt/preferences.d/docker | 08:27 |
| SamYaple | Package: docker-engine | 08:27 |
| SamYaple | Pin: version 1.8.2* | 08:27 |
| SamYaple | Pin-Priority: 550 | 08:27 |
| SamYaple | fyi kjelly ^^ that makes the pi permanent | 08:27 |
| *** mfalatic has quit IRC | 08:27 | |
| SamYaple | pin* | 08:27 |
| *** mfalatic has joined #kolla | 08:28 | |
| kjelly | SamYaple: Ok, does this should be document ? Or I just write it in my note? | 08:29 |
| SamYaple | if you feel its important enough to document, go for it | 08:30 |
| zigo | What's the point of empty config files in etc/kolla ? | 08:30 |
| SamYaple | zigo: they allow overriding of any of the openstack conf file values | 08:30 |
| zigo | SamYaple: So, anything that an operator would put there would override what Kolla does automatically? | 08:31 |
| SamYaple | in relation to the config files (glance-api.conf, nova.conf, etc) yes | 08:31 |
| SamYaple | it actually merges, not strictly overrides | 08:32 |
| SamYaple | anything in /etc/kolla gets preference, but you only need ot put in your override values | 08:32 |
| SamYaple | [libvirt] | 08:32 |
| SamYaple | virt_type=qemu | 08:32 |
| *** openstack has joined #kolla | 08:35 | |
| *** egonzalez90 is now known as egonzalez | 08:39 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Restrucutre gate scripts https://review.openstack.org/241499 | 08:42 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 08:42 |
| openstackgerrit | Merged openstack/kolla: Remove unused setup_fedora.sh https://review.openstack.org/241498 | 08:45 |
| *** egonzalez is now known as egonzalez90 | 08:51 | |
| *** egonzalez90 is now known as Guest89281 | 08:51 | |
| *** Guest89281 is now known as egonzalez | 08:52 | |
| *** sdake_ has quit IRC | 08:53 | |
| *** egonzalez is now known as egonzalez90 | 08:57 | |
| *** dolpher has quit IRC | 08:57 | |
| *** egonzalez90 is now known as egonzalez | 08:58 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Restrucutre gate scripts https://review.openstack.org/241499 | 09:02 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 09:02 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Add additional ntp information around ceph https://review.openstack.org/236758 | 09:11 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Add additional ntp information around ceph https://review.openstack.org/236758 | 09:14 |
| openstackgerrit | Merged openstack/kolla: Tidy up quickstart.rst https://review.openstack.org/241927 | 09:14 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Add additional ntp information around ceph https://review.openstack.org/236758 | 09:17 |
| openstackgerrit | Merged openstack/kolla: Adjust package name for Ubuntu Ironic https://review.openstack.org/238640 | 09:20 |
| inc0 | SamYaple, is there such thing as 2 space after period?;) never seen it | 09:21 |
| SamYaple | inc0: its an older thing and he is quite consistent about it | 09:21 |
| inc0 | ok, seems legit | 09:22 |
| inc0 | for obvious reasons I will not argue about proper english stuff | 09:22 |
| *** mbound has joined #kolla | 09:22 | |
| SamYaple | most of that quickstart doc is actually two spaced | 09:22 |
| *** jmccarthy has joined #kolla | 09:23 | |
| *** gfidente has joined #kolla | 09:24 | |
| *** diga has joined #kolla | 09:28 | |
| *** sdake has joined #kolla | 09:34 | |
| *** tzn has quit IRC | 09:40 | |
| openstackgerrit | Merged openstack/kolla: Fix RHEL builds https://review.openstack.org/241462 | 09:42 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 09:49 |
| openstackgerrit | Merged openstack/kolla: Adjust gate btrfs setup https://review.openstack.org/241465 | 09:49 |
| *** inc0_ has joined #kolla | 09:53 | |
| *** inc0 has quit IRC | 09:56 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Add additional ntp information around ceph https://review.openstack.org/236758 | 10:01 |
| *** inc0_ has quit IRC | 10:06 | |
| *** inc0 has joined #kolla | 10:07 | |
| *** inc0 has quit IRC | 10:17 | |
| *** inc0 has joined #kolla | 10:17 | |
| *** tfukushima has quit IRC | 10:23 | |
| *** egonzalez has quit IRC | 10:27 | |
| *** inc0 has quit IRC | 10:40 | |
| *** tzn has joined #kolla | 10:40 | |
| *** sdake has quit IRC | 10:41 | |
| *** dwalsh has joined #kolla | 10:41 | |
| openstackgerrit | Paul Bourke proposed openstack/kolla: Clean up keystone httpd pid files for RPM distros https://review.openstack.org/241980 | 10:41 |
| *** pbourke has quit IRC | 10:43 | |
| *** tzn has quit IRC | 10:46 | |
| *** inc0 has joined #kolla | 10:51 | |
| *** inc0 has quit IRC | 10:53 | |
| *** inc0 has joined #kolla | 10:54 | |
| SamYaple | yay! | 10:55 |
| SamYaple | the multinode gate found a bug! | 10:55 |
| SamYaple | a rather large one actually / | 10:55 |
| SamYaple | :/ | 10:55 |
| *** dwalsh has quit IRC | 10:59 | |
| *** inc0 has quit IRC | 11:00 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Add Docker in Docker container https://review.openstack.org/241444 | 11:02 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Allow disabling of sysctl values https://review.openstack.org/241856 | 11:02 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Restrucutre gate scripts https://review.openstack.org/241499 | 11:02 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 11:02 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: bootstrap takes place on the wrong node https://review.openstack.org/241989 | 11:02 |
| *** pbourke has joined #kolla | 11:15 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: bootstrap takes place on the wrong node https://review.openstack.org/241989 | 11:24 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 11:24 |
| *** dolpher has joined #kolla | 11:31 | |
| *** dolpher has quit IRC | 11:35 | |
| *** dwalsh has joined #kolla | 11:36 | |
| *** inc0 has joined #kolla | 11:44 | |
| *** inc0_ has joined #kolla | 11:46 | |
| *** dwalsh has quit IRC | 11:46 | |
| pbourke | SamYaple: so this is merged to ansible now? https://github.com/ansible/ansible-modules-core/pull/2258#event-435343008 | 11:47 |
| pbourke | SamYaple: can we bump our pin in kolla-ansible to remove our docker 1.8.3 restriction? | 11:47 |
| SamYaple | pbourke: nope | 11:48 |
| SamYaple | its merged into 2.0 ansible | 11:48 |
| inc0_ | when do we move to 2.0? | 11:48 |
| SamYaple | its also in stable/1.9 _however_ 1.9.5 will not be tagged | 11:48 |
| SamYaple | so 1.9.x will never be able to go past 1.8.2 | 11:48 |
| SamYaple | inc0_: when its ever released | 11:48 |
| *** inc0 has quit IRC | 11:49 | |
| *** ssurana has joined #kolla | 11:55 | |
| openstackgerrit | Merged openstack/kolla: Add Docker in Docker container https://review.openstack.org/241444 | 11:57 |
| *** achanda has joined #kolla | 12:03 | |
| *** achanda has quit IRC | 12:05 | |
| *** akwasnie has quit IRC | 12:10 | |
| *** sdake has joined #kolla | 12:10 | |
| sdake | morning | 12:10 |
| SamYaple | morning sdake | 12:11 |
| SamYaple | my multinode gate caught something! | 12:11 |
| sdake | damn sam you have been up for 2 days ;) | 12:11 |
| SamYaple | did you hear me?! | 12:11 |
| SamYaple | it caught somethign! | 12:11 |
| sdake | what ? | 12:11 |
| SamYaple | broken bootstrap process when compute is seperate from nova-api | 12:11 |
| sdake | and that is pretty cool ;) | 12:11 |
| SamYaple | _and_ the name of the compute host is lower in alphabetical order than that of the host running nova-api | 12:12 |
| SamYaple | you know sdake... we could techinically extend this multinode to be seperate services per each node like an ultra explodd out environment | 12:14 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Make the database json variable more readable https://review.openstack.org/242024 | 12:16 |
| openstackgerrit | Merged openstack/kolla: Restrucutre gate scripts https://review.openstack.org/241499 | 12:22 |
| *** dims_ has joined #kolla | 12:22 | |
| *** athomas has joined #kolla | 12:24 | |
| *** athomas has quit IRC | 12:25 | |
| *** athomas has joined #kolla | 12:25 | |
| *** dims has quit IRC | 12:25 | |
| *** athomas has quit IRC | 12:27 | |
| *** athomas has joined #kolla | 12:27 | |
| *** tzn has joined #kolla | 12:43 | |
| zigo | What've just read on http://docs.openstack.org/developer/kolla/quickstart.html horifies me... | 12:44 |
| zigo | curl -sSL https://get.docker.io | bash | 12:44 |
| SamYaple | that is the official docker way of doing it | 12:45 |
| SamYaple | dont look at us | 12:45 |
| zigo | Am I the only one believing this is unbelievably unsafe? | 12:45 |
| SamYaple | i never do it that way | 12:45 |
| zigo | Well, IMO, we shouldn't advise for that. | 12:45 |
| SamYaple | that is the official stated docker way | 12:45 |
| SamYaple | when docker changes things, thats how they get changed | 12:45 |
| zigo | "official" as in, from upstream? | 12:45 |
| SamYaple | yes | 12:45 |
| zigo | Upstream isn't always right... :P | 12:46 |
| zigo | Anyway, is there a minimum requirement for Docker's version? | 12:46 |
| zigo | Should Kolla depend on Docker by the way? | 12:46 |
| SamYaple | all the docker info should be in that guide | 12:46 |
| SamYaple | min and max | 12:47 |
| zigo | "Kolla release 1.0.0-liberty tested to work with docker 1.8.2" | 12:47 |
| zigo | That's all it says ... | 12:47 |
| *** tzn has quit IRC | 12:48 | |
| SamYaple | nope | 12:48 |
| zigo | (appart that 1.8.3 and later wont work) | 12:48 |
| zigo | In other words: I haven't seen a minimum required version. | 12:48 |
| SamYaple | it says more | 12:48 |
| zigo | Ok. | 12:48 |
| SamYaple | yea there is a table right there | 12:48 |
| SamYaple | 1.8.2 | 12:48 |
| zigo | Ah... so in fact, Kolla works *only* with 1.8.2? | 12:49 |
| SamYaple | that doesnt sound right, but thats what it says | 12:49 |
| SamYaple | i thought it worked with 1.6.0 and up to 1.8.2 | 12:49 |
| SamYaple | but whoever updated the docs may have newer info | 12:49 |
| zigo | Debian has only 1.5 | 12:50 |
| zigo | :( | 12:50 |
| SamYaple | i know | 12:50 |
| SamYaple | and as time moves on that version will only get higher and higher as a req for kolla | 12:50 |
| zigo | Ah no... | 12:50 |
| zigo | doker != docker.io ! :) | 12:50 |
| SamYaple | for instance when 1.10 comes out we will be using that | 12:50 |
| SamYaple | docker.io was ubuntus name | 12:50 |
| zigo | Jessie has 1.6.2, and Sid has 1.8.3. | 12:51 |
| SamYaple | are you sure its docker.io in debian? | 12:51 |
| SamYaple | ill trust you since those seem like legit versions. it should work with 1.6.2 | 12:51 |
| zigo | Well, it looks like docker.io in Ubuntu derives from the package in Debian. | 12:51 |
| SamYaple | sdake: comments ^ | 12:51 |
| zigo | Well, I can only upload to Sid, so I need kolla to work with 1.8.3, unfortunately. | 12:52 |
| zigo | What's the blocker? | 12:52 |
| SamYaple | ansible | 12:52 |
| zigo | Ah? | 12:52 |
| zigo | Could you expand? :) | 12:52 |
| SamYaple | for the record though zigo, when Mitaka comes out it will have a req for at _least_ 1.10 docker | 12:53 |
| SamYaple | ansible modules that are not in our control are broken | 12:53 |
| zigo | Who's working on these Ansible modules then? | 12:53 |
| sdake | ansible 1.9.z is abandoned | 12:53 |
| SamYaple | i already patched them | 12:53 |
| SamYaple | but they wont tag another 1.9 ansible | 12:53 |
| sdake | they are a 50 person company - can only focus on supporting 1 release not two | 12:54 |
| SamYaple | barely even one.... | 12:54 |
| SamYaple | hey sdake can you validate the version numbers in the quickstart guide? | 12:55 |
| SamYaple | they seem off | 12:55 |
| sdake | 1.8.2 is the max | 12:55 |
| sdake | not sure on the min | 12:55 |
| SamYaple | i dont think the minimum has changed from 1.6.0 | 12:56 |
| sdake | atlaewsat 1.6.0 | 12:56 |
| SamYaple | but i dont know | 12:56 |
| SamYaple | yea | 12:56 |
| sdake | 1.5.1 is broken | 12:56 |
| sdake | 1.5.0 is also broken | 12:56 |
| inc0_ | I don't suppose Debian will bump up version of docker when 1.10 releases | 12:57 |
| inc0_ | and we'll want to use it right away | 12:57 |
| inc0_ | (thin containers) | 12:57 |
| SamYaple | inc0_: thats only for mitaka (master) | 12:57 |
| SamYaple | not liberty | 12:57 |
| inc0_ | yeah | 12:58 |
| SamYaple | i dont think debian has the version of ansible needed either in packaged form | 12:58 |
| inc0_ | by the way, can you force version of docker using this curl install thingy? | 12:58 |
| inc0_ | if so, we should have it in our docs | 12:58 |
| SamYaple | this is why ubuntu has to maintain cloud-archive so they can support openstack | 12:58 |
| SamYaple | inc0_: you can package pin ahead of time | 12:58 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Make the database json variable more readable https://review.openstack.org/242024 | 12:59 |
| *** tobe has quit IRC | 13:03 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Make the database json variable more readable https://review.openstack.org/242024 | 13:03 |
| *** sdake_ has joined #kolla | 13:04 | |
| *** akwasnie has joined #kolla | 13:06 | |
| *** dwalsh has joined #kolla | 13:07 | |
| *** sdake has quit IRC | 13:08 | |
| *** tobe has joined #kolla | 13:10 | |
| *** tobe has quit IRC | 13:14 | |
| *** tobe has joined #kolla | 13:14 | |
| *** tobe has quit IRC | 13:19 | |
| *** tobe has joined #kolla | 13:20 | |
| *** dwalsh has quit IRC | 13:21 | |
| *** sdake_ has quit IRC | 13:21 | |
| *** tobe has quit IRC | 13:25 | |
| *** sdake has joined #kolla | 13:27 | |
| sdake | pbourke ping | 13:30 |
| openstackgerrit | Merged openstack/kolla: Clean up keystone httpd pid files for RPM distros https://review.openstack.org/241980 | 13:33 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Make the database json variable more readable https://review.openstack.org/242024 | 13:33 |
| *** diga has quit IRC | 13:33 | |
| *** diga has joined #kolla | 13:33 | |
| *** diga has quit IRC | 13:34 | |
| SamYaple | sdake: so for the gate, how do you want to play this | 13:35 |
| SamYaple | i dont think its reasonable to get 3 more gates for multinode | 13:35 |
| *** diga has joined #kolla | 13:35 | |
| sdake | i prefer that | 13:35 |
| SamYaple | i know | 13:35 |
| sdake | breaking stuff apart is easier to debug | 13:35 |
| SamYaple | thats what i made my coment | 13:35 |
| SamYaple | i dont think its reasonable to get 3 more gates for multinode | 13:36 |
| SamYaple | i would prefer more gates for other things | 13:36 |
| SamYaple | like excersiing different options | 13:36 |
| sdake | there are bo lmits on gates we can haved | 13:36 |
| sdake | bo/no | 13:36 |
| SamYaple | i am aware. my comment remains | 13:36 |
| SamYaple | i dont want to fight for gates | 13:37 |
| sdake | nova has like 60 gates | 13:37 |
| SamYaple | they do not | 13:37 |
| *** rhallisey has joined #kolla | 13:37 | |
| SamYaple | they have like 10 and then a bunch of thrid party ci | 13:37 |
| SamYaple | third party has that equipment | 13:37 |
| SamYaple | and we are not nova | 13:37 |
| sdake | i see | 13:37 |
| sdake | well infra doesn't mind if we have many gates, already had that discussion | 13:37 |
| sdake | they care about run length not numberof gates | 13:38 |
| sdake | what is the objection to mor egates? | 13:38 |
| sdake | the / your | 13:38 |
| SamYaple | ok then I want our AIO gates to have diferent options than our multinode gates | 13:38 |
| SamYaple | i want to test more options | 13:38 |
| SamYaple | and dont want 2000 gates | 13:38 |
| SamYaple | for every single permutation | 13:38 |
| *** inc0_ has quit IRC | 13:39 | |
| sdake | 2k gtates is probably not goin gto fly ;) | 13:39 |
| sdake | how many permutations do you want? | 13:39 |
| sdake | i'd be happy with basic stuff ;) | 13:39 |
| SamYaple | now we have spice vs novnc, linuxbridge vs openvswitch, etc. And that is just going to grow | 13:39 |
| SamYaple | i want the main ones | 13:41 |
| SamYaple | so if we can switch up AIO and multinode options I am ok with seperate gates | 13:41 |
| SamYaple | btw I am working on the logging situation | 13:42 |
| SamYaple | soon we will have a logs folder and not one massive console.html | 13:42 |
| SamYaple | so diving into the gate will be 1000% easier | 13:42 |
| SamYaple | including actual service log output | 13:42 |
| sdake | my motive for multiple gates isn't that its easy to debug its that the dashboard turns +1 / -1 eaisly | 13:43 |
| sdake | i expect debugging a failure will always be painful | 13:43 |
| SamYaple | thats why i said 'btw' | 13:43 |
| sdake | roger | 13:43 |
| *** tzn has joined #kolla | 13:44 | |
| sdake | i dont understand whaty ou mean by mixing up options | 13:44 |
| sdake | you mean having aio do spice and multinode do novnc? | 13:44 |
| SamYaple | AIO has spice and linuxbridge as options, multinode has novnc and ovs | 13:44 |
| SamYaple | something liek that | 13:44 |
| sdake | ya that wfm but i dont htink we can gate linuxbridge | 13:45 |
| sdake | i thikn we detemrined its not possible | 13:45 |
| SamYaple | whys that? | 13:45 |
| SamYaple | no it should be completely possible | 13:45 |
| sdake | not a networking rocket scientist | 13:45 |
| sdake | but you said long ago ovs was the only one we could get working | 13:45 |
| SamYaple | right before I was doing what im doing | 13:45 |
| SamYaple | should be good now | 13:45 |
| sdake | i tried long and hard to get linuxbridge to get one interface going | 13:45 |
| *** akwasnie has quit IRC | 13:45 | |
| sdake | and failed | 13:45 |
| SamYaple | yea we should be solid now | 13:46 |
| sdake | sepnt couple weeks on it atleast | 13:46 |
| sdake | well gotta say that is fantastic news ;) | 13:46 |
| SamYaple | i think im killing the node resources :( | 13:46 |
| SamYaple | it gets to horizon and breaks down | 13:46 |
| sdake | ya your ooming | 13:46 |
| sdake | this is why it hink dnd isn't viable | 13:46 |
| SamYaple | https://jenkins02.openstack.org/job/gate-kolla-deploy-images-ubuntu-source/32/console | 13:46 |
| sdake | we can't get more ram in teh vms unfortunately | 13:46 |
| SamYaple | its no worse than anything else | 13:46 |
| SamYaple | we can set the affinity for the processes | 13:47 |
| SamYaple | so only 2 copies of this, but 3 of that | 13:47 |
| SamYaple | etc | 13:47 |
| SamYaple | i thikn we can work it out | 13:47 |
| sdake | i'd like multinode to do multiple vms but ymmv :) | 13:47 |
| SamYaple | multiple 128mb ciros images... sure! | 13:48 |
| sdake | mainy to work arund the ram problem which is unsolveable | 13:48 |
| *** tzn has quit IRC | 13:48 | |
| SamYaple | well 'multinode' upstream is 2 vms right now | 13:48 |
| SamYaple | no more | 13:48 |
| SamYaple | thats not good enough | 13:48 |
| sdake | they said they woudl give us as many as we need (within reason) | 13:48 |
| sdake | after we got 2 owrking we could geet 3 after 3 working could get 4 etc | 13:49 |
| SamYaple | right. multinode upstream only works as 2 vms right now | 13:49 |
| SamYaple | unless something has changed... | 13:49 |
| sdake | infra folks confirmed they would give us many nodes if we incrementally prove our gates are reliable as we add nodes | 13:49 |
| sdake | we talked about 5 as a cap | 13:49 |
| sdake | i think we dcould get more like 10 or so | 13:49 |
| sdake | the reason they have 2 as a cap now is nobody can get 2 vms to successfully gate correctly | 13:50 |
| sdake | nova basicallyl gaveu pon multinode gating | 13:50 |
| SamYaple | well good luck with that | 13:50 |
| sdake | from my understanding | 13:50 |
| sdake | so infra hasn't added more nodes | 13:50 |
| SamYaple | we wont be able to gate 2-3-4 nodes due to the registry problem | 13:51 |
| SamYaple | a push in the gate even to a local registry will take ages | 13:51 |
| sdake | v1 registyr shouldn't be too bad | 13:51 |
| SamYaple | yes it will | 13:52 |
| sdake | ok well another option is to build on each node | 13:52 |
| sdake | ? | 13:52 |
| SamYaple | youre joking right? | 13:52 |
| sdake | just brainstorming | 13:52 |
| SamYaple | no were not doing that. thats teh exact oposite of kolla | 13:52 |
| SamYaple | i can trim the fat in the gate right now | 13:53 |
| sdake | maybe we can squeeeze out a 16gb vm from infra | 13:53 |
| SamYaple | if i can prove multinode on a single node... maybe? | 13:53 |
| sdake | yup | 13:53 |
| sdake | lets go ask and see if its possible | 13:53 |
| SamYaple | i mean it is from an L2 standpoint | 13:53 |
| SamYaple | completely | 13:53 |
| sdake | crickets probably too early | 13:54 |
| SamYaple | yea good luck with that | 13:55 |
| sdake | good luck with which? | 13:55 |
| SamYaple | im going to trim the gate fat right now and prove that multinode works | 13:55 |
| sdake | I have never asked for a 16gb vm | 13:55 |
| sdake | SamYaple join #openstack-infra | 13:56 |
| *** unicell1 has joined #kolla | 14:02 | |
| *** unicell has quit IRC | 14:03 | |
| *** cemmason1 has joined #kolla | 14:03 | |
| v1k0d3n | morning SamYaple (at least where i'm at) | 14:05 |
| *** cemmason has quit IRC | 14:05 | |
| v1k0d3n | i cleaned out my work from yesterdays issue...and i am still having problems. is there something i'm missing in the docs? | 14:05 |
| v1k0d3n | check login credentials (login_user, and login_password, which can be defined in ~/.my.cnf), check that mysql socket exists and mysql server is running | 14:05 |
| v1k0d3n | that's the error ^^ | 14:05 |
| v1k0d3n | i looked at /etc/my.cnf which includes /etc/my.cnf.d and points to mysql-clients and client.cnf | 14:07 |
| v1k0d3n | but not a lot in there...are there supposed to be changes or something added that i'm overlooking? | 14:08 |
| *** diogogmt has quit IRC | 14:13 | |
| *** ssurana1 has joined #kolla | 14:14 | |
| *** ssurana has quit IRC | 14:15 | |
| *** cemmason1 has quit IRC | 14:15 | |
| *** ssurana1 has quit IRC | 14:15 | |
| *** diogogmt has joined #kolla | 14:16 | |
| SamYaple | sdake: heat is the eater of ram | 14:21 |
| SamYaple | sdake: goodbye heat! | 14:21 |
| sdake | imo add swap ;) | 14:24 |
| sdake | i'd like to gate the big tent rather then copute kit | 14:24 |
| SamYaple | its just heat... | 14:24 |
| SamYaple | but well see | 14:24 |
| SamYaple | i want time for tests. with swap i have no time to actually test anything | 14:25 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 14:25 |
| *** absubram has joined #kolla | 14:26 | |
| SamYaple | sdake i think i can trim most services down to two nodes, only quorum ones requiring three nodes | 14:29 |
| SamYaple | drop horizon entirely from multinode | 14:29 |
| SamYaple | then well be better | 14:29 |
| sdake | recommend atleast testing more swap to see if that really has a bad performance ipact on testing | 14:30 |
| SamYaple | it does :) the system already has swap | 14:30 |
| SamYaple | the gate doesnt crash, it slows to a crawl and times out | 14:30 |
| SamYaple | its swapping | 14:30 |
| sdake | i see | 14:31 |
| sdake | well i guess big tent testing will be aio then | 14:31 |
| *** diogogmt has quit IRC | 14:31 | |
| SamYaple | maybe | 14:31 |
| SamYaple | we are probably at the limit there as well | 14:31 |
| SamYaple | dont know how much we can test | 14:31 |
| sdake | what is problem with building on each node with multivm? | 14:32 |
| SamYaple | they arent the same images and could have different packages and libraries? | 14:33 |
| SamYaple | aka everything Kolla is not | 14:33 |
| SamYaple | im -2 on that idea forever | 14:33 |
| sdake | why do you think v1 registry will be longer then 90 mins? | 14:34 |
| sdake | my v1 registry takes 15 mins on my gear | 14:34 |
| SamYaple | i dont. | 14:34 |
| SamYaple | the build will take 20-30 minutes. the push AT BEST takes 15 minutes | 14:34 |
| SamYaple | then we have to deploy | 14:34 |
| SamYaple | how can we test? | 14:34 |
| SamYaple | btw we only have 60 minute gates | 14:35 |
| SamYaple | but even at 90 we would time out | 14:35 |
| sdake | we can definatelly get 90 minute gates | 14:35 |
| SamYaple | i know | 14:35 |
| sdake | maybe able to get more time | 14:35 |
| sdake | 120 min seem to be the max people will tolerate in infra ;) | 14:36 |
| sdake | it sucks computing is all about constraints ;( | 14:37 |
| SamYaple | youre welcome to attempt multinode stuff | 14:37 |
| SamYaple | i will not be. this dind stuff is going to excrise more code paths than what you suggest | 14:38 |
| SamYaple | it already caught a problem multinode vms would not have because it can simulate many more nodes | 14:38 |
| *** dims has joined #kolla | 14:40 | |
| sdake | well gating compute kit is a good first start | 14:41 |
| sdake | better then 0 | 14:41 |
| *** guruMeditator has quit IRC | 14:41 | |
| SamYaple | not gating heat is not the same thigns as just compute kit | 14:41 |
| *** slotti has quit IRC | 14:42 | |
| *** sdake has quit IRC | 14:42 | |
| *** diga has quit IRC | 14:43 | |
| *** dims_ has quit IRC | 14:43 | |
| *** athomas has quit IRC | 14:43 | |
| *** egonzalez has joined #kolla | 14:44 | |
| *** kjelly has quit IRC | 14:44 | |
| *** athomas has joined #kolla | 14:45 | |
| *** tzn has joined #kolla | 14:45 | |
| *** sdake has joined #kolla | 14:47 | |
| *** tzn has quit IRC | 14:49 | |
| *** dims has quit IRC | 14:53 | |
| *** guruMeditator has joined #kolla | 14:56 | |
| guruMeditator | Greetings. I am attempting to deploy a Compute node with Neutron that has multiple provider networks. Each provider network is on a different interface. I see that the Neutron external interface is defined with "neutron_external_interface". Is it possible to accomplish this? Currently I have been manually changing the bridges to each neutron node | 14:56 |
| *** cemason has joined #kolla | 14:56 | |
| *** sdake has quit IRC | 14:57 | |
| *** cloudnautique has quit IRC | 15:07 | |
| *** athomas has quit IRC | 15:09 | |
| openstackgerrit | Paul Bourke proposed openstack/kolla: Clean up keystone httpd pid files for RPM distros https://review.openstack.org/242106 | 15:09 |
| rhallisey | SamYaple ^^ are you able to answer the above question? Because idk the answer :) | 15:11 |
| pbourke | guruMeditator: rhallisey: SamYaple: I wonder could you override neutron_external_interface per node in the inventory? | 15:12 |
| pbourke | would that solve it? | 15:12 |
| guruMeditator | That would allow a single neutron node to map to disparate interfaces, but, I don't think it would allow multiple provider networks to map to different interfaces | 15:14 |
| SamYaple | guruMeditator: not in an automated fashion with the current playbooks | 15:18 |
| SamYaple | its expected to have one provider interface | 15:18 |
| SamYaple | however you could override the appropriate config options and manually attach them to openvswitch | 15:18 |
| SamYaple | btw guys | 15:18 |
| SamYaple | multinode gate is a goooooo! | 15:18 |
| guruMeditator | Yes, that is what I have been doing (manually adding the bridges via OVS on all neutron nodes). Is this the best approach? Are there plans to add multiple provider networks mapping in the future? | 15:20 |
| SamYaple | guruMeditator: i would love to see it, but i will not have the cycles to add that support any time soon | 15:20 |
| SamYaple | that is my suggested approach for you | 15:21 |
| SamYaple | see if you can add the feature to kolla yourself or bug someone to do it for you :) | 15:21 |
| guruMeditator | Ok, thanks SamYaple. Appreciate the suggestions. | 15:21 |
| SamYaple | lol 230mb memory free in the gate after multinode with no heat or horizon | 15:23 |
| SamYaple | halarious | 15:23 |
| *** blahRus has joined #kolla | 15:27 | |
| v1k0d3n | are you guys aware of the broken ansible link in the main github README.md? | 15:28 |
| v1k0d3n | https://github.com/openstack/kolla/blob/master/doc/ansible-deployment.rst | 15:28 |
| SamYaple | v1k0d3n: probably not. feel free to correct it | 15:28 |
| SamYaple | does that dox still get published? | 15:28 |
| SamYaple | v1k0d3n: oh wait those links only work at docs.openstack.org | 15:29 |
| v1k0d3n | it looks like it's gone....or maybe that needs to point to: https://github.com/openstack/kolla/blob/master/doc/quickstart.rst? | 15:29 |
| SamYaple | they dont work in the get repo | 15:29 |
| v1k0d3n | ah! | 15:30 |
| v1k0d3n | so you're saying they should point to: http://docs.openstack.org/developer/kolla/dev-quickstart.html | 15:30 |
| *** sdake has joined #kolla | 15:30 | |
| *** sdake has quit IRC | 15:31 | |
| v1k0d3n | the title Quickstart to Deploying OpenStack using Ansible | 15:31 |
| SamYaple | v1k0d3n: no im saying that they will _work_ at docs.openstack.org | 15:31 |
| SamYaple | they dont work in teh git repo | 15:31 |
| SamYaple | i dont really understand the docs publishing or linkages to be honest | 15:31 |
| *** sdake has joined #kolla | 15:31 | |
| SamYaple | Kennan may be able to explain it better v1k0d3n | 15:31 |
| v1k0d3n | oh, i think that <https://github.com/openstack/kolla/blob/master/doc/ansible-deployment.rst> should probably read like [Title](https://www.likehere.com) for markdown. | 15:33 |
| v1k0d3n | i'm probably missing how the links are being used though. | 15:33 |
| SamYaple | its not markdown though | 15:33 |
| SamYaple | its for the sphinx docs | 15:33 |
| v1k0d3n | yeah, rst files i noticed. | 15:33 |
| *** sdake has quit IRC | 15:35 | |
| *** sdake has joined #kolla | 15:36 | |
| openstackgerrit | David Moreau Simard proposed openstack/kolla: Update kollaglue docker repository URL https://review.openstack.org/242118 | 15:36 |
| *** diogogmt has joined #kolla | 15:37 | |
| *** cloudnautique has joined #kolla | 15:38 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 15:38 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Remove vip for rabbitmq https://review.openstack.org/235777 | 15:43 |
| *** tzn has joined #kolla | 15:45 | |
| *** diogogmt has quit IRC | 15:49 | |
| *** diogogmt has joined #kolla | 15:49 | |
| *** tzn has quit IRC | 15:50 | |
| *** inc0 has joined #kolla | 15:58 | |
| pbourke | sdake: just saw ping | 16:01 |
| sdake | can you backport one of your recent changes for httpd.pid | 16:01 |
| pbourke | yup should be already done | 16:01 |
| pbourke | https://review.openstack.org/#/c/242106/ | 16:01 |
| diogogmt | i’m starting to get this errors when reloading the vagrant vm: http://pastebin.com/PAxgJAGv | 16:10 |
| sdake | thanks | 16:11 |
| diogogmt | i see a comment in the Vagrantfile pointing to this thread: http://askubuntu.com/questions/317338/how-can-i-increase-disk-size-on-a-vagrant-vm | 16:13 |
| diogogmt | what if instead we increased the size of the base image being used? | 16:13 |
| diogogmt | it looks like the base image being used is puppetlabs/centos-7.0-64-puppet, however, i don’t see any scripts to provision the vm with docker, ansible, python and all the other dependencies — how is the provision is done on the vagrant vm> | 16:14 |
| diogogmt | ? | 16:14 |
| diogogmt | can i use a different base image instead? | 16:15 |
| pbourke | diogogmt: it should have enough space out of the box | 16:16 |
| diogogmt | it only has a 20GB disk | 16:16 |
| pbourke | diogogmt: what steps have you taken so far | 16:16 |
| pbourke | diogogmt: yes but it maps /var/lib/docker to the host | 16:16 |
| pbourke | diogogmt: the provisioning is done using bootstrap.sh | 16:16 |
| pbourke | under kolla/dev/vagrant | 16:16 |
| diogogmt | pbourke: ok, i’ll take a look at the script, thanks | 16:17 |
| *** kproskurin has joined #kolla | 16:17 | |
| diogogmt | regarding the /var/lib/docker, i’m running on a mac, are you saying that all the docker images are being stored under /var/lib/docker ? | 16:17 |
| diogogmt | on my mac instead of the vm? | 16:17 |
| SamYaple | sdake: i had a successful deploy | 16:19 |
| SamYaple | no heat or horizon | 16:19 |
| SamYaple | 230MB left of ram | 16:19 |
| sdake | you can't run kolla on a mac | 16:20 |
| pbourke | sdake: no he's using a vm on his mac | 16:20 |
| sdake | unless you use vagrant | 16:20 |
| pbourke | yes | 16:20 |
| sdake | oh vms will work | 16:20 |
| sdake | pbourke nice summary of the sessions :) | 16:21 |
| pbourke | diogogmt: checking again, only the registry storage goes to the host | 16:21 |
| sdake | +2: ) | 16:21 |
| pbourke | sdake: cheers | 16:21 |
| pbourke | diogogmt: so maybe the disk is too small to build everything :/ | 16:21 |
| SamYaple | sdake: thought of how we can skirt the docker registry on multinode | 16:22 |
| SamYaple | true multinode | 16:22 |
| openstackgerrit | Merged openstack/kolla: Clean up keystone httpd pid files for RPM distros https://review.openstack.org/242106 | 16:22 |
| diogogmt | pbourke: i had the same issue before with some of my docker hosts running with vagrant, i had to manually increase the size of the box by converting the box back to vdi and then using lvm to increase the disk | 16:23 |
| pbourke | can anyone recommend dirt cheap public cloud that could work for kolla dev? | 16:23 |
| diogogmt | pbourke: do you see any issues if i use a different base image instead of puppetlabs/centos ? | 16:23 |
| sdake | rax woild work but it is fairly pricy i think | 16:23 |
| pbourke | diogogmt: yeah resizing can be a pain though :( I think another should work | 16:23 |
| pbourke | diogogmt: if you find a more suitable one let us know maybe we can update the vagrantfile | 16:24 |
| sdake | pbourke is a "top lad" a good thing? | 16:25 |
| pbourke | diogogmt: generally I dont build everything... you can use --profile default when building | 16:26 |
| pbourke | sdake: I think so?! | 16:26 |
| pbourke | diogogmt: that way you save resources but get everything you need for a working openstack | 16:26 |
| diogogmt | pbourke: i’ll check if the puppetlabs base image has lvm, if it does i might give it a try to resize it — do you have any concerns if we use ubuntu instead of centos for the vagrant base image? | 16:27 |
| diogogmt | pbourke: what do you mena by building everything? if we don’t build locally does kolla uses public images for some of the services? | 16:28 |
| *** daneyon has joined #kolla | 16:28 | |
| openstackgerrit | Merged openstack/kolla: Update kollaglue docker repository URL https://review.openstack.org/242118 | 16:29 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Allow disabling of sysctl values https://review.openstack.org/241856 | 16:30 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: bootstrap takes place on the wrong node https://review.openstack.org/241989 | 16:30 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Deploy multinode in gate https://review.openstack.org/241500 | 16:30 |
| SamYaple | thats two bugs for my multinode gate | 16:31 |
| *** daneyon_ has joined #kolla | 16:32 | |
| pbourke | diogogmt: oh sorry I thought you were building your own images | 16:33 |
| pbourke | diogogmt: so you're running out of space just from a deploy? | 16:33 |
| diogogmt | pbourke: no, i got the out of space problem while building the images | 16:34 |
| sdake | 20gb is not enough storage | 16:34 |
| sdake | 40gb | 16:34 |
| sdake | is what you need | 16:34 |
| diogogmt | i’ve been following the instructions from the wiki page | 16:34 |
| pbourke | yeah I thought the image we were using was 40 | 16:34 |
| sdake | which wiki page? | 16:34 |
| *** tzn has joined #kolla | 16:34 | |
| sdake | we dont have wiki docs that I know of | 16:34 |
| diogogmt | https://github.com/openstack/kolla/blob/master/doc/vagrant-dev-env.rst | 16:34 |
| sdake | docs.openstack.org/developer/kolla | 16:34 |
| diogogmt | so first i was building the images: kolla-build, and later planning to deploy them | 16:35 |
| diogogmt | but i hit the out of space error before i got to the deploy part | 16:35 |
| *** daneyon has quit IRC | 16:35 | |
| pbourke | If we mounted /var/lib/docker to the host that should fix things I think | 16:36 |
| diogogmt | sdake: pbourke: looks like the puppetlabs/cento7 vm is using lvm to manage the partions, i guess i’ll just increase the size of the base image manually then | 16:39 |
| diogogmt | later i can push the new resized box to https://atlas.hashicorp.com/scalefactory/boxes/ | 16:39 |
| diogogmt | do you think 40GB is enough? | 16:40 |
| diogogmt | or should i resize to 60GB to avoid future problems? | 16:40 |
| pbourke | go with 60GB sure | 16:42 |
| pbourke | Im wondering would it good if you supplied your instructions for resizing to the docs rather than pushing the box | 16:43 |
| pbourke | (though you're free to push the box also of course :)) | 16:44 |
| pbourke | people may be more comfortable using the puppetlabs maintained one and resizing themselves rather than a user submitted one | 16:44 |
| diogogmt | pbourke: makes sense, i’ll do it both ways | 16:48 |
| diogogmt | quick question regarding the branches, is master considered stable or is it better to go with liberty? | 16:49 |
| *** odyssey4me_ has joined #kolla | 16:51 | |
| pbourke | diogogmt: liberty would be more stable | 16:52 |
| *** Ti-mo- has joined #kolla | 16:53 | |
| pbourke | diogogmt: though master should work most of the time | 16:53 |
| diogogmt | pbourke: ok, i’ll start with liberty just to get a hang and move to master eventually | 16:55 |
| diogogmt | thanks for the pointers | 16:55 |
| pbourke | sounds good | 16:56 |
| *** sdake_ has joined #kolla | 16:56 | |
| *** Ti-mo- has quit IRC | 16:57 | |
| *** egonzalez has quit IRC | 16:59 | |
| *** barra204 has joined #kolla | 16:59 | |
| *** Ti-mo- has joined #kolla | 17:00 | |
| *** sdake has quit IRC | 17:04 | |
| *** v1k0d3n has quit IRC | 17:04 | |
| *** odyssey4me has quit IRC | 17:04 | |
| *** Ti-mo has quit IRC | 17:04 | |
| *** shakamunyi has quit IRC | 17:04 | |
| *** inc0 has quit IRC | 17:09 | |
| *** achanda has joined #kolla | 17:26 | |
| *** unicell1 has quit IRC | 17:33 | |
| *** kproskurin has quit IRC | 17:38 | |
| *** achanda has quit IRC | 17:43 | |
| *** ssurana has joined #kolla | 17:44 | |
| *** ssurana has quit IRC | 17:45 | |
| *** achanda has joined #kolla | 17:45 | |
| harmw | guys, are there reasons to _not_ venture into logging using the docker fluentd-driver? | 17:46 |
| harmw | instead of syslog | 17:46 |
| pbourke | only one I can think of is liberty is out in the wild using rsyslog so we'd need a migration path if we decided it was worth switching | 17:51 |
| *** ssurana has joined #kolla | 17:53 | |
| harmw | true, forgot that, offering fluentd as 2nd option sounds more friendly in that case :) | 17:53 |
| pbourke | options are always nice :) | 17:54 |
| harmw | but perhaps a full efk stack could be added | 17:54 |
| pbourke | im just reading about it now | 17:54 |
| harmw | in essence, it plain simple | 17:54 |
| pbourke | is it a fancy way of saying we'll store your logs in json? | 17:54 |
| harmw | :) | 17:54 |
| harmw | plus, it buffers | 17:55 |
| *** kfox1111_ is now known as kfox1111 | 17:55 | |
| harmw | btw, that Mesos discussion is awesome | 17:56 |
| harmw | I wonder where we'll be heading | 17:57 |
| sdake_ | we are leading with ansible | 17:57 |
| *** sdake_ is now known as sdake | 17:57 | |
| sdake | and a bare metal underlay | 17:57 |
| harmw | ofcourse we are | 17:57 |
| sdake | we are experimenting with mesos as the underlay and openstack as an overlay | 17:57 |
| sdake | atm, we have bare metal as an overlay | 17:57 |
| sdake | rather underlay | 17:58 |
| harmw | but using marathon to manage containers is cool :) | 17:58 |
| harmw | we're using it at $work | 17:58 |
| *** mbound has quit IRC | 17:58 | |
| *** unicell has joined #kolla | 18:09 | |
| *** ryansb has joined #kolla | 18:13 | |
| sdake | bmace you around | 18:13 |
| sdake | bmace__ around? | 18:13 |
| bmace__ | yup | 18:14 |
| *** achanda has quit IRC | 18:19 | |
| *** tzn has quit IRC | 18:20 | |
| *** tzn has joined #kolla | 18:24 | |
| *** tzn has quit IRC | 18:28 | |
| *** vilobhmm11 has joined #kolla | 18:34 | |
| *** jasonsb has quit IRC | 18:38 | |
| *** achanda has joined #kolla | 18:41 | |
| *** cemason has quit IRC | 18:45 | |
| *** cemason has joined #kolla | 18:46 | |
| *** gfidente has quit IRC | 18:48 | |
| openstackgerrit | David Moreau Simard proposed openstack/kolla: Fix documentation links in the README https://review.openstack.org/242188 | 18:50 |
| *** cemason has quit IRC | 18:51 | |
| dmsimard | btw <3 kolla, have high hopes for you guys. Good talks in Tokyo too :) | 18:52 |
| rhallisey | I <3 kolla | 18:53 |
| rhallisey | t-shirts sdake | 18:53 |
| rhallisey | dmsimard, thanks :) | 18:53 |
| sdake | dmsimard welcome to the party :) | 18:54 |
| sdake | glad you liked the talk and demo :) | 18:54 |
| sdake | rhallisey still working on stickers ;-) | 18:54 |
| rhallisey | selinux bug in docker 1.9 | 18:56 |
| dmsimard | I'm at Red Hat doing upstream RDO stuff and been toying with the idea of using Kolla for some stuff, please let me know if I can help in some way :) | 18:56 |
| rhallisey | :/ | 18:56 |
| *** mfalatic_ has joined #kolla | 18:56 | |
| sdake | dmsimard fyi for the future, docs changes do not require bug ids - just to speed thigns up | 18:56 |
| rhallisey | dmsimard, what stuff were you planning on using kolla for. Curious | 18:58 |
| *** mfalatic has quit IRC | 19:00 | |
| dmsimard | rhallisey: Nothing worth mentioning yet :) | 19:00 |
| sdake | dmsimard rhallisey is one of your coworkers :) | 19:00 |
| * dmsimard nods | 19:00 | |
| sdake | nothign worth mentioning - sounds dangerous :) | 19:00 |
| sdake | dmsimard there is all kinds of areas to help out in | 19:01 |
| dmsimard | Bah, I see potential in kolla for doing things that triple-o, packstack or rdo manager do right now but the hamster in my head is still at the spinning stage | 19:01 |
| rhallisey | dmsimard, do you happend to have +2 powers in rdo community? | 19:01 |
| rhallisey | dmsimard, neat | 19:01 |
| rhallisey | dmsimard, tripleo has come container support atm. Added it to the compute node | 19:02 |
| dmsimard | I assisted in some talks around magnum and kubernetes too - apparently there's some murano kubernetes cluster manager app or something too | 19:03 |
| dmsimard | rhallisey: +2 where? In packaging ? | 19:03 |
| rhallisey | dmsimard, ya | 19:03 |
| dmsimard | rhallisey: no but I work closely with people who do :) | 19:04 |
| rhallisey | dmsimard, can you ask someone to +2 https://review.gerrithub.io/#/c/251378/ | 19:04 |
| rhallisey | :) | 19:05 |
| rhallisey | no one responded when I asked earlier | 19:05 |
| *** tzn has joined #kolla | 19:05 | |
| *** JoseMello has joined #kolla | 19:05 | |
| sdake | dmsimard the cool part about kolla vs those other tools yoou mentioned is that it fits right in with rhel atomic | 19:08 |
| dmsimard | rhallisey: no added requirements to the package ? | 19:08 |
| sdake | those other tools require writeable /usr, where atomic has read only /usr | 19:08 |
| sdake | kolla designed to work with read/only /usr | 19:09 |
| sdake | dmsimard rhallisey can fill you in in more detail, I have limited exposure to this tech | 19:09 |
| dmsimard | sdake: my interest in kolla has a lot to do with the quick and easily reproducible deployments (think containers vs rdo manager that takes >1hr), just for the sake of CI this is *huge* | 19:10 |
| sdake | well with koll yuo still have to build the images | 19:10 |
| sdake | which on my high end gaer takes about 30 minutes | 19:10 |
| dmsimard | Once the images are built, though | 19:10 |
| sdake | roger that | 19:10 |
| dmsimard | I also like that you have the possibility to install either from source or from binaries | 19:11 |
| dmsimard | Lots of good design decisions :) | 19:11 |
| ryansb | has anyone gotten kolla going on a remote atomic host? | 19:11 |
| sdake | yo ryansb | 19:11 |
| ryansb | I'm trying to set the DOCKER_HOST env variable and it's not taking the address | 19:11 |
| sdake | i htink rhallisey had it running at one point but he could add mor edetail | 19:11 |
| rhallisey | ryansb, tripleo deploys onto atomic | 19:12 |
| ryansb | sdake: \o | 19:12 |
| ryansb | rhallisey: so if I want to build images against a remote docker host, how do I specify that? | 19:12 |
| sdake | ryansb what are you working on these days - heat still? | 19:12 |
| ryansb | yup, heat & some director | 19:12 |
| dmsimard | rhallisey: +2'd | 19:13 |
| rhallisey | dmsimard, thanks! | 19:13 |
| rhallisey | ryansb, I don't quite understand your setup. You want a remote host build images? | 19:14 |
| ryansb | rhallisey: so I've got kolla cloned locally, and my docker host is a separate box (baremetal atomic install) | 19:15 |
| ryansb | so what I want to do is have the build script use the equivalent of "docker -H atomic.local:4243 {subcommand}" | 19:16 |
| rhallisey | Ohh interesting I've never done that. I usually build locally | 19:16 |
| sdake | stackoverflow says | 19:16 |
| sdake | export DOCKER_HOST=tcp://192.168.59.103:2375 | 19:16 |
| sdake | is that what your doing ryansb ? | 19:16 |
| sdake | never tried DOCKER_HOST either | 19:16 |
| sdake | but one thing to keep in mind | 19:16 |
| sdake | I find it highly unlikekly our build.py processes DOCKER_HOST environment variable | 19:17 |
| sdake | yo uwill probably need to patch kolla to do such a thing | 19:17 |
| rhallisey | ya that would be a nice addition | 19:17 |
| ryansb | actually you do | 19:17 |
| ryansb | my problem was that I didn't have the "tcp://" prefix | 19:18 |
| ryansb | docker -H doesn't require the prefix | 19:18 |
| sdake | google ftw ;-) | 19:18 |
| rhallisey | nice | 19:18 |
| ryansb | you use kwargs_from_env() https://github.com/openstack/kolla/blob/master/kolla/cmd/build.py#L57 | 19:18 |
| ryansb | so you get it for free | 19:18 |
| ryansb | thanks for the tcp nudge, now we'll see if the images build properly | 19:19 |
| ryansb | rhallisey: yeah, my atomic baremetal has nicer hardware than my workstation, figured it might be faster | 19:20 |
| rhallisey | ryansb, cool. Using atomic for dev work can be a tough. No yum XD. But I'm sure you know your way around | 19:21 |
| ryansb | I actually rarely use the atomic tools currently, mostly just to remote docker host info | 19:22 |
| ryansb | err, "just to host remote docker containers" is what I meant | 19:22 |
| *** jasonsb has joined #kolla | 19:24 | |
| *** Hui has joined #kolla | 19:24 | |
| rhallisey | ryansb, cool | 19:24 |
| Hui | Hi, my latest neutron-server deployment fails from binary | 19:24 |
| Hui | Waiting for bootstrap container to exit > {"changed": true, "cmd": ["docker", "wait", "bootstrap_neutron"], "delta": "0:00:00.043556", "end": | 19:25 |
| Hui | does anyone have similar problem; it was working last few days | 19:25 |
| rhallisey | Hui, does the container show up in restarting? | 19:28 |
| Hui | this is a neutron bootstrap container | 19:29 |
| rhallisey | right, but you might still see it around | 19:29 |
| Hui | yes | 19:29 |
| Hui | in one of the compute node | 19:29 |
| Hui | because it is "runonce" on any node | 19:30 |
| Hui | do you know how to check the log? | 19:30 |
| rhallisey | well you can start the container by hand and shell into it | 19:30 |
| rhallisey | and run the commands manually | 19:30 |
| rhallisey | and see what they split back out at you | 19:31 |
| rhallisey | could be a db issue | 19:31 |
| Hui | but I see other service like glance and nova is booted fine | 19:31 |
| jasonsb | harmw: were you thinking of using fluentd on client side to forward the logs? | 19:32 |
| rhallisey | Hui, I'm not sure. Run the neutron-server container with '-it' flags and execute '/bin/bash' | 19:37 |
| rhallisey | then run entend-start.sh | 19:38 |
| *** blahRus1 has joined #kolla | 19:38 | |
| rhallisey | hopefully will reveal something | 19:38 |
| *** blahRus has quit IRC | 19:39 | |
| *** inc0 has joined #kolla | 19:51 | |
| inc0 | hey | 19:51 |
| *** mbound has joined #kolla | 19:53 | |
| sdake | yo inc0 | 20:22 |
| sdake | up late | 20:22 |
| *** ryansb has quit IRC | 20:26 | |
| *** ryansb has joined #kolla | 20:28 | |
| *** ryansb has quit IRC | 20:28 | |
| *** ryansb has joined #kolla | 20:28 | |
| *** achanda has quit IRC | 20:35 | |
| harmw | jasonsb: at $work we're using fluentd on each docker host (lets call those the clients) and they send data per interval to the fluentd containers that reside on the ElasticSearch nodes, from where they inject into localhost:9200 | 20:49 |
| harmw | and all services (containers) log their JSON to the local fluentd container, as is done using docker (stdout/stderr) | 20:50 |
| dmsimard | rhallisey: so, anyway, do you know about delorean ? There's probably a way to hook image builds in there :) | 21:04 |
| Hui | rhallisey, thanks. I found that somehow the neutron.conf is not pushed to all nodes | 21:05 |
| rhallisey | dmsimard, I do know about delorean, but not much | 21:05 |
| rhallisey | Hui, oh nice | 21:05 |
| Hui | so if that node is picked as a bootstrap node, the bootstrap of neutron will fail | 21:05 |
| rhallisey | dmsimard, how can we hook in images builds. That's sounds exactly what I'm looking for | 21:05 |
| rhallisey | Hui, see if you can file a bug around it | 21:06 |
| rhallisey | not sure how that happened | 21:06 |
| dmsimard | rhallisey: well, I don't have anything to show yet but delorean essentially watches upstream repos and then builds packages whenever there's a commit. It shouldn't be very hard to build an image instead of a package. Just have to figure out how to do it in a smart and efficient way. | 21:07 |
| rhallisey | dmsimard, oh neat. I was going to have a job that periodically builds the containers | 21:08 |
| rhallisey | but that sounds much better | 21:08 |
| Hui | rhallisey, sure, I will | 21:09 |
| rhallisey | Hui, cool thanks | 21:09 |
| dmsimard | rhallisey: well, this is for the RDO packaging side - I don't know what the other distro have/use | 21:11 |
| sdake | https://www.youtube.com/watch?v=BKYJuYsT4z4&feature=youtu.be | 21:13 |
| rhallisey | ^ check out 38 min. I'm on stage! :D | 21:15 |
| sdake | dmsimard you can regex the build | 21:15 |
| sdake | presentation was pretty solid | 21:16 |
| sdake | timed it right ;-) | 21:16 |
| dmsimard | You guys might want to A) pin that to liberty in the liberty branch or B) use the GA repo https://github.com/openstack/kolla/blob/stable/liberty/docker/base/Dockerfile.j2#L52-L53 | 21:19 |
| jasonsb | harmw: neat. any particular reason to use fluentd on source side? | 21:19 |
| dmsimard | Those links are the mitaka repos now (since liberty was cut) | 21:19 |
| jasonsb | harmw: is it to have same tooling to handle streamed syslog style and also picking up logs from files? | 21:19 |
| harmw | jasonsb: buffering :) | 21:20 |
| sdake | i thought we did that, but dms if you have links to the repo or want to wsubmit a patch itwould be appreciated | 21:20 |
| jasonsb | ah ok | 21:20 |
| harmw | the local container receives logs and pushes them per interval (10s) to fluentd container on ES hosts | 21:20 |
| jasonsb | is it qualitatively different than that provided by rsyslog? | 21:20 |
| dmsimard | I can submit a fix but what do you want ? Do you want to use the latest tested packages (closer to trunk liberty) or the stable packages (closer to release/milestone) | 21:20 |
| harmw | plus, we add some keys if the logs come from the crappy docker stdout/stderr | 21:20 |
| harmw | well, it's k/v JSON | 21:21 |
| jasonsb | rsyslog has queues i was thinking | 21:21 |
| rhallisey | dmsimard, agreed | 21:21 |
| harmw | but rsyslog logs in lines, not in JSON | 21:21 |
| harmw | right? | 21:22 |
| harmw | capturing stacktraces sucks bigtime that way | 21:22 |
| rhallisey | dmsimard, stable liberty | 21:23 |
| jasonsb | this is true. you have to parse the thing while making sure to keep the log lines in correct order | 21:23 |
| dmsimard | rhallisey: okay I'll submit a fix for that, straight in the liberty branch though since master we'll keep mitaka (master) | 21:23 |
| jasonsb | otoh you don't have to worry about local storage if you pay that piper | 21:24 |
| rhallisey | dmsimard, so into master and add 'backport: liberty' | 21:24 |
| dmsimard | rhallisey: You want to pin liberty in master ? | 21:24 |
| jasonsb | harmw: regardless of the tradeoffs, this is an interesting topic. have to look at it closer with operational POV | 21:25 |
| rhallisey | dmsimard, no sorry into liberty | 21:25 |
| rhallisey | I thought the whole time that branch was pinned to liberty | 21:26 |
| rhallisey | didn't notice | 21:26 |
| *** achanda has joined #kolla | 21:28 | |
| sdake | dmsimard for stable/liberty we want to use liberty release, for masster, we want to use delorean | 21:31 |
| inc0 | we have additional problem with rsyslog | 21:31 |
| inc0 | I can't get TRACEBACK to log to rsyslog:/ | 21:32 |
| inc0 | that's strange thing because everything else is logged | 21:32 |
| *** cemason has joined #kolla | 21:35 | |
| Hui | rhallisey, filed a bug at https://bugs.launchpad.net/kolla/+bug/1513598 | 21:35 |
| openstack | Launchpad bug 1513598 in kolla "neutron bootstrap container fails due to missing neutron.conf" [Undecided,New] - Assigned to Hui Kang (huikang27) | 21:35 |
| rhallisey | cool thanks Hui | 21:36 |
| harmw | jasonsb: we're using it in production, works like a breeze :) | 21:37 |
| harmw | let me know if you need more info | 21:37 |
| Hui | rhallisey my pleasure. If you have time, could you validate it? | 21:37 |
| jasonsb | harmw: will be in touch. i am interested in the cardinality of the source vs sink for fluentd | 21:37 |
| rhallisey | ya I'll try to get around to it | 21:38 |
| *** cemason has quit IRC | 21:39 | |
| *** _tzn has joined #kolla | 21:41 | |
| *** tzn has quit IRC | 21:41 | |
| *** inc0 has quit IRC | 21:43 | |
| *** absubram has quit IRC | 21:44 | |
| openstackgerrit | David Moreau Simard proposed openstack/kolla: Pin the liberty branch to the liberty RPM repositories https://review.openstack.org/242229 | 21:46 |
| *** JoseMello has quit IRC | 21:52 | |
| rhallisey | Hui, https://review.openstack.org/#/c/241989/3 | 21:53 |
| rhallisey | https://bugs.launchpad.net/kolla/+bug/1513439 | 21:53 |
| openstack | Launchpad bug 1513439 in kolla "bootstrap process can happen on the wrong node" [Critical,In progress] - Assigned to Sam Yaple (s8m) | 21:53 |
| rhallisey | Hui, ^^ that patch will fix it | 21:53 |
| openstackgerrit | Merged openstack/kolla: Fix documentation links in the README https://review.openstack.org/242188 | 21:54 |
| sdake | dmsimard if you plan to continue contributing to kolla, would you like to join the kolla-drivers team? | 21:57 |
| sdake | this team is responsible for bug triage of kolla (basically setting bug states on new bugs in kolla) | 21:57 |
| *** shardy has quit IRC | 21:58 | |
| dmsimard | I don't have any "official" time to spend on kolla right now :( | 21:59 |
| sdake | roger, open offer, keep in mind for future | 21:59 |
| * dmsimard nods | 21:59 | |
| *** rhallisey has quit IRC | 22:00 | |
| *** unicell1 has joined #kolla | 22:01 | |
| *** unicell has quit IRC | 22:02 | |
| *** unicell1 has quit IRC | 22:10 | |
| *** unicell has joined #kolla | 22:10 | |
| diogogmt | can somebody please clarify if there is a way on kolla to pull the images from a public repo instead of building them locally? i just want o build the image for a couple of services we want to add patches for, not all the openstack services images | 22:14 |
| *** achanda has quit IRC | 22:22 | |
| Hui | rhallisey, thanks | 22:25 |
| *** cloudnautique has quit IRC | 22:29 | |
| *** alisonh has quit IRC | 22:31 | |
| sdake | diogogmt atm the only tagged release of kolla is DOA becaue of a bug that was added after we tagged the repo to an upstream repository | 22:36 |
| sdake | diogogmt we fixed the problem in stable/liberty, however, we have not retagged a new 1.0.1 release | 22:36 |
| sdake | since we don't have a tag, we have not pushed a liberty releae to docker hub | 22:36 |
| sdake | so the short term answer is no | 22:37 |
| sdake | I guess we need to tag a 1.0.1 release but I've been super jet-lagged this week and my sleep is not good | 22:37 |
| sdake | basically I haven't had eneough energy to test and releae a 1.0.1 tag since summit | 22:37 |
| sdake | diogogmt if you already have imagess you can build specific imaes - build.py takes a regex as an argument that identifies the images to build | 22:40 |
| sdake | SamYaple are yo uabout | 22:42 |
| *** diogogmt has quit IRC | 22:44 | |
| *** daneyon_ has quit IRC | 22:46 | |
| *** alisonh has joined #kolla | 22:49 | |
| *** achanda has joined #kolla | 22:59 | |
| *** _tzn has quit IRC | 23:02 | |
| *** tzn has joined #kolla | 23:07 | |
| asalkeld | anyone running ubuntu 15.10 and kolla works? ubuntu wants to upgrade it's self | 23:08 |
| * asalkeld giving it a go | 23:11 | |
| *** sdake has quit IRC | 23:12 | |
| *** sdake has joined #kolla | 23:14 | |
| *** asalkeld has quit IRC | 23:15 | |
| *** tzn has quit IRC | 23:16 | |
| SamYaple | sdake: yea | 23:16 |
| *** tzn has joined #kolla | 23:16 | |
| sdake | SamYaple say can you test that build.py change on ubuntu wfm on centos | 23:18 |
| sdake | Hui around? | 23:19 |
| openstackgerrit | Merged openstack/kolla: Use kolla_internal_address in the example rc file https://review.openstack.org/239992 | 23:19 |
| *** Hui has quit IRC | 23:25 | |
| openstackgerrit | Steven Dake proposed openstack/kolla: Make two types of openrc example efiles https://review.openstack.org/242256 | 23:25 |
| sdake | SamYaple I inadvertnatly set w+1 to a patch and I think it caused the child to fail to merge, and recheck may not fix this type of oops | 23:27 |
| *** achanda has quit IRC | 23:31 | |
| *** achanda has joined #kolla | 23:32 | |
| *** Ti-mo- has quit IRC | 23:34 | |
| *** Ti-mo has joined #kolla | 23:36 | |
| *** asalkeld has joined #kolla | 23:40 | |
| sdake | samyaple do you tihnk you will be able to tackle the cap codebase, or willing to pass it to another cr? | 23:41 |
| sdake | for m1 (dec 4th) | 23:41 |
| asalkeld | dam, upgrading a desktop takes for ever ... | 23:43 |
| sdake | moring asalkeld | 23:44 |
| asalkeld | howdie, sorting a new spec rev. now | 23:44 |
| openstackgerrit | Angus Salkeld proposed openstack/kolla: Spec: Deploy Kolla images using Mesos https://review.openstack.org/241086 | 23:48 |
| openstackgerrit | Merged openstack/kolla: Remove vip for rabbitmq https://review.openstack.org/235777 | 23:49 |
| *** blahRus1 has quit IRC | 23:49 | |
| SamYaple | sdake: capping all of the versions? | 23:51 |
| SamYaple | i can do that | 23:51 |
| sdake | nah cap permissions | 23:51 |
| SamYaple | no | 23:51 |
| SamYaple | we need to drop user perms first | 23:51 |
| sdake | agree | 23:51 |
| SamYaple | cap permissions are just icing on the cake anyway | 23:52 |
| SamYaple | most people dont know what that is | 23:52 |
| sdake | i gueess i se it as one job but its probably two blueprints | 23:52 |
| SamYaple | obsolete/superceed | 23:52 |
| SamYaple | setup up the one blueprint with work items | 23:52 |
| sdake | if we are going to distribute the work i prefer etherpad for that | 23:54 |
| sdake | you ahd expressed interest in doing this work | 23:54 |
| sdake | wondering if you going to do it alone or want to distribute | 23:54 |
| SamYaple | id like to start using the work items rather than an etherpad | 23:56 |
| SamYaple | its why they are there | 23:56 |
| SamYaple | any cores around that can give me an ack for this so I can backport the patch afterward? https://review.openstack.org/#/c/241856/ | 23:57 |
| *** jasonsb_ has joined #kolla | 23:57 | |
| sdake | SamYaple back to hte original q do you want to do the job alone or distribute | 23:57 |
| SamYaple | i want to see the work items involved first :) | 23:58 |
| sdake | we can try it | 23:58 |
| sdake | open to giving it a go | 23:59 |
| sdake | if it doesn't work as well as etherpad we can stick with etherpad | 23:59 |
| sdake | if it is as effective we can switch permanetly | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!