| *** jtriley has quit IRC | 00:06 | |
| *** thetrav has quit IRC | 00:08 | |
| openstackgerrit | Martin André proposed stackforge/kolla: Add validation script for broken symlinks https://review.openstack.org/197465 | 00:10 |
|---|---|---|
| openstackgerrit | Martin André proposed stackforge/kolla: Add link validation to pep8 gate https://review.openstack.org/197466 | 00:10 |
| openstackgerrit | Martin André proposed stackforge/kolla: Add min env vars doc validation to pep8 gate https://review.openstack.org/197474 | 00:33 |
| openstackgerrit | Martin André proposed stackforge/kolla: Add validation script for min env vars documentation https://review.openstack.org/197473 | 00:33 |
| openstackgerrit | Martin André proposed stackforge/kolla: Sort variables and remove duplicates for min env doc https://review.openstack.org/197586 | 00:33 |
| *** vinkman has left #kolla | 00:37 | |
| *** thetrav has joined #kolla | 00:46 | |
| *** jtriley has joined #kolla | 00:57 | |
| *** dims_ has quit IRC | 01:19 | |
| *** jasonsb has quit IRC | 01:31 | |
| *** alisonh has joined #kolla | 01:31 | |
| openstackgerrit | Martin André proposed stackforge/kolla: Add min env vars doc validation to pep8 gate https://review.openstack.org/197474 | 01:38 |
| openstackgerrit | Martin André proposed stackforge/kolla: Sort variables and remove duplicates for min env doc https://review.openstack.org/197586 | 01:38 |
| openstackgerrit | Martin André proposed stackforge/kolla: Make tools/min-env.sh portable https://review.openstack.org/197806 | 01:38 |
| *** dims has joined #kolla | 01:39 | |
| *** erkules_ has joined #kolla | 01:51 | |
| *** erkules has quit IRC | 01:52 | |
| *** vinkman1 has joined #kolla | 02:04 | |
| *** vinkman1 has quit IRC | 02:04 | |
| *** jtriley has quit IRC | 02:24 | |
| *** jtriley has joined #kolla | 02:55 | |
| *** walters has quit IRC | 02:57 | |
| *** walters has joined #kolla | 03:02 | |
| *** dims has quit IRC | 03:22 | |
| *** jtriley has quit IRC | 03:24 | |
| *** coolsvap|away is now known as coolsvap | 03:45 | |
| *** alisonh has quit IRC | 03:51 | |
| *** jasonsb has joined #kolla | 04:00 | |
| *** bmace has joined #kolla | 04:00 | |
| *** alisonh has joined #kolla | 04:18 | |
| *** sdake has joined #kolla | 04:30 | |
| *** absubram has quit IRC | 04:35 | |
| *** alisonh has quit IRC | 04:44 | |
| *** alisonh has joined #kolla | 04:55 | |
| *** coolsvap is now known as coolsvap|away | 05:37 | |
| *** coolsvap|away is now known as coolsvap | 05:38 | |
| *** nihilifer has joined #kolla | 05:40 | |
| nihilifer | good morning | 05:40 |
| mandre | hi nihilifer | 05:41 |
| *** dasm|afk is now known as dasm | 05:52 | |
| *** coolsvap is now known as coolsvap|away | 05:54 | |
| vincent_vdk | hi | 05:55 |
| *** coolsvap|away is now known as coolsvap | 06:02 | |
| vincent_vdk | so how's the work on cinder going? | 06:05 |
| sdake | rhallisey was afk most of the day so unclear | 06:07 |
| sdake | hey mandre nih* | 06:07 |
| *** coolsvap is now known as coolsvap|away | 06:16 | |
| SamYaple | yo | 06:23 |
| SamYaple | sdake: you around | 06:23 |
| sdake | wound me | 06:23 |
| SamYaple | i fell asleep | 06:24 |
| sdake | that is good | 06:25 |
| sdake | sleep is necessary evil :) | 06:25 |
| vincent_vdk | so are people using kolla for production environments already? | 06:35 |
| SamYaple | vincent_vdk: i dont think | 06:35 |
| sdake | who knows such things | 06:35 |
| sdake | only the oracle | 06:35 |
| harmw | SamYaple: can you take a look at https://review.openstack.org/#/c/197761/? | 06:35 |
| SamYaple | harmw: im reviewing it now | 06:37 |
| harmw | ok, cool | 06:37 |
| SamYaple | im going to be a bit nitpicky so we can keep them the same across all projects, we can do back and forth on this and then use it as a template | 06:38 |
| harmw | I'm expecting you to do just that | 06:38 |
| harmw | this'll be the template for other containers indeed | 06:38 |
| vincent_vdk | just asking because we are testing several openstack "flavors" | 06:44 |
| vincent_vdk | to me, Kolla looks really interesting | 06:45 |
| vincent_vdk | so i'll be probably testing it in the near future | 06:45 |
| harmw | vincent_vdk: which other flavors are you trying out? | 06:45 |
| vincent_vdk | Helion which uses TripleO for deployment | 06:45 |
| vincent_vdk | I also briefly touched osad | 06:46 |
| harmw | ah ok | 06:46 |
| vincent_vdk | another colleague is testing Ubuntu currently | 06:46 |
| harmw | for true container experience, stick with kolla | 06:46 |
| vincent_vdk | i tend to follow you on that | 06:47 |
| vincent_vdk | also the over/undercloud architecture is something i don't really like | 06:47 |
| vincent_vdk | there is to much going on and to take into account when something fails | 06:48 |
| *** shardy has quit IRC | 06:48 | |
| vincent_vdk | but that is imho :) | 06:48 |
| harmw | ah yes, well, with kolla your laptop is your undercloudthingy :p running ansible and talking to the docker 'hypervisor' | 06:48 |
| *** shardy has joined #kolla | 06:48 | |
| harmw | triplo is basically just 2 clouds, right? | 06:48 |
| vincent_vdk | yes | 06:48 |
| thetrav | undercloud ftw | 06:48 |
| harmw | SamYaple: don't tell me I'm way off on Glance :p | 06:49 |
| thetrav | I thought compose was currently the kolla undercloud? | 06:49 |
| harmw | yea thetrav | 06:49 |
| * thetrav hasn't started using it yet | 06:49 | |
| harmw | we're working on ansible though | 06:49 |
| thetrav | I've been putting together a setup using ansible for my lab | 06:49 |
| vincent_vdk | compose just orchestrates your container deployment? | 06:49 |
| harmw | vincent_vdk: yes | 06:49 |
| vincent_vdk | thetrav: is that available somewhere | 06:50 |
| SamYaple | harmw: nope you are write on target functionally | 06:50 |
| thetrav | unfortunately not vincent | 06:50 |
| thetrav | it's also a bit of a hodge podge | 06:50 |
| vincent_vdk | just to help testing or eventually help fixing things | 06:50 |
| vincent_vdk | sure | 06:50 |
| thetrav | I've been trying to re-create: http://behindtheracks.com/2014/04/redundant-load-balancers-haproxy-and-keepalived/ using kolla containers where possible | 06:50 |
| harmw | vincent_vdk: thetrav , if you're interested we have a /ansible in our repo that will get a lot of love these next couple of weeks | 06:51 |
| vincent_vdk | I once started a deplyment framework to deploy docker containers with Ansible | 06:51 |
| vincent_vdk | ah, nice | 06:51 |
| thetrav | so while I've got you talking harmw it looks a bit like compose is handling the ansible like part... is there anything to handle the fleet/swarm/kubernetes part? | 06:52 |
| vincent_vdk | the fact you can deploy on thing like atomic is really interesting | 06:52 |
| thetrav | like... does compose actually deal with the multiple host thing? | 06:52 |
| vincent_vdk | i don't think so | 06:53 |
| sdake | vincent_vdk over last 3 weeks our project has made a conscience and unanimous decision to make our work deploy with ansible as our #1 priority | 06:53 |
| vincent_vdk | sdake: makes sense | 06:53 |
| sdake | our #2 priority is ha, #3 priority is from source installs | 06:53 |
| vincent_vdk | compose is nice for local dev imho | 06:54 |
| thetrav | have you hit the super annoying bug with ansible's docker module? | 06:54 |
| sdake | it took alot of work to get people on board with that, but now we are all in on that model | 06:54 |
| thetrav | need to have expose as well as port | 06:54 |
| sdake | so if you want to help, help implemenet ansible or test it :) | 06:54 |
| thetrav | took me about a week to figure out what was going on there | 06:54 |
| vincent_vdk | thetrav: which bug, it's been like 6 months ago i used Ansible+Docker | 06:54 |
| sdake | we use --pid=host rathe rthen expose and port | 06:54 |
| vincent_vdk | sdake: sure | 06:55 |
| thetrav | net: host in ansible | 06:55 |
| thetrav | I found I had to do that for some containers but not all | 06:55 |
| thetrav | it's one of those "avoid when possible" things | 06:55 |
| SamYaple | thetrav: in github.com/SamYaple/yaodu I do keepalived+haproxy and I will be copying that implmentation here | 06:55 |
| SamYaple | I even think I have a way we can rehup and load new configs without restarting haproxy | 06:56 |
| vincent_vdk | SamYaple: multi datacenter? | 06:56 |
| vincent_vdk | we plan on deploying OS spanning 2 datacenters | 06:56 |
| thetrav | I haven't managed to stick haproxy in a container | 06:56 |
| SamYaple | vincent_vdk: neither yaodu nor kolla supports regions or cells at the moment | 06:56 |
| sdake | vincent_vdk we plan to deploy openstack in one data center at present ;) | 06:57 |
| vincent_vdk | ok | 06:57 |
| sdake | if you want multi-datacenter not sure there is a container docker option for you | 06:57 |
| SamYaple | thetrav: sdake there will be | 06:57 |
| SamYaple | thetrav: sorry ignore | 06:57 |
| thetrav | sdake: what's the multi-datacenter option? | 06:57 |
| *** jmccarthy has joined #kolla | 06:57 | |
| sdake | will be and are are two different things | 06:57 |
| sdake | there is none I am aware of | 06:58 |
| SamYaple | thetrav: https://github.com/SamYaple/yaodu/tree/master/ansible/roles/docker_build/templates/ubuntu/haproxy that whole project shows my (working) implementation of haproxy in a container if you want some pointers | 06:58 |
| thetrav | ahh, I missed the "not sure" :P | 06:58 |
| SamYaple | sdake: I agree. but it will be easier to do than I initially though thanks to ansible variables | 06:59 |
| thetrav | thanks SamYaple that will be super useful | 06:59 |
| vincent_vdk | thing is we have a 10G link between both DC's | 06:59 |
| sdake | thetrav you ought to join us in kolla - if you want to implement that you can do so :) | 06:59 |
| vincent_vdk | but what would happen if it goes down | 06:59 |
| vincent_vdk | not sure keepalived handles this | 06:59 |
| thetrav | sdake, I very much want to do that | 06:59 |
| vincent_vdk | but we are not there yet | 06:59 |
| SamYaple | vincent_vdk: if the latency isn't to high, you can probably make that "one" cluster with different nova and availability zones | 07:00 |
| thetrav | might be able to swing it as an after hours activity | 07:00 |
| thetrav | haven't been able to convince my employers so far to contribute | 07:00 |
| vincent_vdk | SamYaple: indeed | 07:00 |
| SamYaple | vincent_vdk: you will have a much higher risk of splitbrain | 07:00 |
| vincent_vdk | SamYaple: that's what i'm afraid of | 07:00 |
| SamYaple | you would need to use pacemaker or similiar to build a quorum on top of that to ensure no splitbrain | 07:01 |
| SamYaple | we dont currently have that logic | 07:01 |
| vincent_vdk | uhu | 07:01 |
| sdake | thetrav roger, the model should be to use kolla as an upstream for your downstream products ;-) | 07:01 |
| sdake | after hours activity is good, sevearl of our best devs operate that way | 07:01 |
| thetrav | sdake that's the angle of attack I'm taking | 07:02 |
| thetrav | now that you guys have your own channel, and have given a bit of a preso at a summit I can probably make a stronger case | 07:03 |
| thetrav | "we wanted this thing to have this feature, we didn't want to build and maintain it ourselves, our vendors didn't want to do it, so we did it, pushed it upstream, now everyone supports it for us and our vendors are all fine with it" | 07:04 |
| thetrav | actually... | 07:05 |
| thetrav | did you guys do a preso? | 07:05 |
| thetrav | or was that a planned thing | 07:05 |
| thetrav | I never actually saw any video / write ups | 07:05 |
| vincent_vdk | there is one from the summit | 07:05 |
| vincent_vdk | which grabbed my attention for this project | 07:06 |
| sdake | there is a video on our wiki page | 07:06 |
| sdake | https://wiki.openstack.org/wiki/Kolla | 07:07 |
| sdake | along with our mission | 07:07 |
| vincent_vdk | you guys should submit something for FOSDEM | 07:07 |
| sdake | -ETOOMUCHTRAVEL | 07:09 |
| sdake | feel free to submit something and speak on our behalf :) | 07:10 |
| thetrav | found: https://www.openstack.org/summit/vancouver-2015/summit-videos/presentation/thanks-docker-the-pro-and-039s-and-con-and-039s-of-containerizing-your-openstack-services | 07:10 |
| SamYaple | sdake: what were your plans for labs/talks? I wanted to do a Kolla+Ansible lab | 07:10 |
| sdake | at summit? | 07:10 |
| sdake | not doing a lab | 07:10 |
| sdake | feel free ;) | 07:10 |
| SamYaple | cool | 07:10 |
| sdake | that would actually be pretty sweet | 07:10 |
| SamYaple | ill do an all-in-one grow to 3 node cluster | 07:11 |
| sdake | i think labs are 90 minute sessions | 07:11 |
| SamYaple | easy peasy | 07:11 |
| SamYaple | ill just have a local-private repo for download speed reasons | 07:11 |
| harmw | SamYaple: config-external just always does 1 thing, right? copy * files from A to B | 07:16 |
| SamYaple | it won't be * files | 07:16 |
| SamYaple | and the files wont all be going to the same place | 07:16 |
| harmw | ah | 07:16 |
| SamYaple | but it does copy from A-B yes | 07:16 |
| harmw | you got examples of where they don't go to the same place? | 07:17 |
| SamYaple | neutron | 07:17 |
| SamYaple | when we do support ceph, ceph has conf files in /etc/ceph | 07:17 |
| harmw | ok, with dnsmasq and neutron configs | 07:17 |
| SamYaple | there are a few scenarios | 07:17 |
| sdake | samyaple I dont know how the labs work - I've never done one | 07:17 |
| harmw | no no, what I mean is there a $source and a $destination and config-external should copy -all- files from $source to $destination | 07:18 |
| SamYaple | nope | 07:18 |
| harmw | there can me multiple $destination vars here? | 07:18 |
| harmw | eg. some files go in /etc/neutron, others to /etc/sysconfig? | 07:19 |
| SamYaple | yes | 07:19 |
| SamYaple | sysconfig is probably a bad example, but yea | 07:19 |
| sdake | diga you about | 07:19 |
| harmw | I know I know | 07:19 |
| sdake | ccolsvap|afk ^^ | 07:19 |
| harmw | I'd like to have just one config-external.sh though, instead of a c/p in everycontainer | 07:20 |
| *** coolsvap|away has quit IRC | 07:20 | |
| SamYaple | harmw: no thats never goign to work | 07:20 |
| SamYaple | and we are movign away from symlink speghetti | 07:20 |
| harmw | we could use a directory structure to point out where on / it should go? | 07:20 |
| SamYaple | yea thats really just over complicating it though | 07:20 |
| harmw | you think? | 07:21 |
| SamYaple | a straight forward uniquie bash script to handle unqie names and unique priviliges and permissiosn | 07:21 |
| harmw | I just hate seeing the same code in multiple places | 07:21 |
| harmw | powered by env vars | 07:21 |
| SamYaple | While I agree, this doesnt count as a DRY violation | 07:21 |
| SamYaple | no | 07:21 |
| SamYaple | ENV vars dont play a role here | 07:21 |
| harmw | (set through DEPLOYMENTTOOL) | 07:21 |
| SamYaple | nope | 07:21 |
| SamYaple | the container is the one place we have static control over things | 07:22 |
| SamYaple | we control the content of the container | 07:22 |
| SamYaple | give me a valid reason the deployer would need to adust the location of a file in a container | 07:22 |
| harmw | uhm, he shouldn't | 07:22 |
| harmw | never ever | 07:22 |
| SamYaple | I wont go _that_ far, but that would be the rare case he would need to, so no ENV vars needed here | 07:23 |
| *** coolsvap|away has joined #kolla | 07:23 | |
| harmw | I was merely thinking of using env vars to give power the copy-external script, setting a username that way to chown or setting the path wich holds the configs and settings a target path | 07:24 |
| harmw | that way the script is clean, and reusable | 07:24 |
| harmw | and ansible or whatever can set those vars | 07:24 |
| harmw | instead of a pure c/p of config-external to all containers | 07:24 |
| SamYaple | Yea I know what your getting at, but it seems like over complication to me | 07:25 |
| SamYaple | the only thing that _would_ work is a for loop on the outside of config-external called by start.sh so the copy and permission update code is only used once | 07:25 |
| harmw | lets note this down under future improvements | 07:26 |
| SamYaple | but that changes the logic from CONFIG_EXTERNAL_COPY_ONCE and CONFIG_EXTERNAL_COPY_ALWAYS | 07:26 |
| harmw | and go c/p for now, even though I hate it | 07:26 |
| SamYaple | harmw: agreed. it can be improved, lets just get it down for now | 07:26 |
| SamYaple | i know :/ | 07:26 |
| harmw | it's probably easier to figure out a good solution once everything is in place anyway, when we're all on-par with your ansible skills :P | 07:27 |
| SamYaple | I dont think ansible will come into play here | 07:30 |
| SamYaple | but once its all on paper we can see what we like and what we dont | 07:30 |
| *** mickt has joined #kolla | 07:33 | |
| harmw | SamYaple: rather then going through the BP I'll just ask you here (if I may), what should I do with KOLLA_BOOTSTRAP | 07:34 |
| SamYaple | You did it right in the glance review | 07:34 |
| SamYaple | well, except it shouldnt be on the glance-api at all | 07:34 |
| harmw | why not, thats what I want to understand :) | 07:35 |
| SamYaple | there is nothing to bootstrap | 07:35 |
| SamYaple | on the DB has to be bootstrapped and the glance-registry would be incharge of that | 07:35 |
| harmw | but where is glance-registry started? | 07:35 |
| SamYaple | (technically glance service container can do it) | 07:36 |
| SamYaple | ansible would start it | 07:36 |
| harmw | and why wouldn't it start glance-api? | 07:36 |
| SamYaple | It would. It just doesn't need to bootstrap glance-api | 07:36 |
| harmw | ok.. but start.sh will also start glance-api | 07:37 |
| SamYaple | glance-api and glance-registry are two different services | 07:37 |
| harmw | I know | 07:37 |
| SamYaple | then im confused | 07:37 |
| harmw | so am I | 07:37 |
| harmw | :) | 07:37 |
| SamYaple | glance-registry talks to and bootstraps the database | 07:37 |
| harmw | yep | 07:38 |
| SamYaple | then glance-api and glance-registry are started as new containers | 07:38 |
| harmw | ah! | 07:38 |
| harmw | missed that | 07:38 |
| SamYaple | external tool can use procedures and do things serially | 07:39 |
| SamYaple | much better :) | 07:39 |
| harmw | you init containers (bootstrap), have them restart and then they know they are not bootstrapping and can just start | 07:39 |
| SamYaple | ah yes hold ill show | 07:40 |
| SamYaple | https://github.com/stackforge/kolla/blob/master/ansible/roles/keystone/tasks/bootstrap.yml | 07:40 |
| SamYaple | those vars are plugged into https://github.com/stackforge/kolla/blob/master/ansible/roles/bootstrap.yml#L23 | 07:41 |
| SamYaple | then that whole container is cleaned up after the database is up to date https://github.com/stackforge/kolla/blob/master/ansible/roles/bootstrap.yml#L28 | 07:41 |
| SamYaple | after all of that _new_ containers are started normally for both services | 07:41 |
| harmw | wicked | 07:42 |
| *** thetrav has quit IRC | 07:45 | |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Glance containers https://review.openstack.org/197761 | 07:48 |
| *** loth has quit IRC | 07:49 | |
| sdake | bed time | 07:53 |
| sdake | night | 07:53 |
| SamYaple | night | 07:53 |
| sdake | samyaple use mailig list if ou have questions | 07:53 |
| SamYaple | for what? | 07:53 |
| *** loth has joined #kolla | 07:53 | |
| sdake | foks thre was a general qeury to use the mailing list | 07:53 |
| sdake | we idd't get to "for whta"just usjg irc better | 07:54 |
| harmw | SamYaple: could you please take another look on that review? :) | 07:54 |
| SamYaple | looking now | 07:54 |
| harmw | I wonder where that first bashate error comes from btw | 07:54 |
| SamYaple | harmw: i want the same comments inline from the keystone one | 07:55 |
| SamYaple | comments inline are not cludge and in the case of set_configs it was a super important comment | 07:55 |
| harmw | the same? I'm not allowed to expand on the topic? give a little more insights on the how/when/where/why? | 07:55 |
| SamYaple | you can expand, but you didnt copy any over at all | 07:56 |
| harmw | was there even a comment about set_configs? | 07:57 |
| harmw | let me check on that | 07:57 |
| harmw | I've skipped # loading common functions | 07:57 |
| harmw | for redundancy | 07:57 |
| SamYaple | if you aren't going to look at what i asked in my review why bother reviewing your change? | 07:57 |
| SamYaple | dont skip the comments just add it. there is no need to remove it | 07:58 |
| SamYaple | having a comment about what the source does, it just loads common function, rather than having it be code that is executed off to the side is important | 07:59 |
| SamYaple | sorry. youre right the comment is not in the keystone start.sh | 08:00 |
| SamYaple | https://review.openstack.org/#/c/196428/6/docker/common/galera/start.sh | 08:00 |
| harmw | SamYaple: aha! | 08:01 |
| harmw | ofcourse we need descriptive comments, I'm just saying we don't need the comment from L8 on that review you just pasted :) | 08:03 |
| harmw | Ill happily add the one on L11 though | 08:04 |
| SamYaple | I disagree | 08:04 |
| SamYaple | we do need it | 08:04 |
| SamYaple | `source` executes code | 08:04 |
| SamYaple | that comment says it only loads functions | 08:04 |
| harmw | because there just so happen to be only functions in kolla-common.sh | 08:04 |
| SamYaple | its not "just so happens" it is by design | 08:05 |
| SamYaple | hence the comment | 08:05 |
| harmw | indeed, why we don't need that comment | 08:05 |
| harmw | *wouldn't | 08:05 |
| harmw | but Ill add it back in, np | 08:05 |
| SamYaple | wrong. there is no way for the reader to know that is by design | 08:05 |
| *** erkules_ is now known as erkules | 08:06 | |
| *** erkules has quit IRC | 08:06 | |
| *** erkules has joined #kolla | 08:06 | |
| SamYaple | it could also say '# changing permissions and setting up users' if it was doing something, it is not doing something in this case. just loading functions | 08:06 |
| SamYaple | with the comments in place you can follow the entire flow of the program without leaving this file | 08:06 |
| SamYaple | without the comments you have to read through each sourced file | 08:06 |
| *** athomas has joined #kolla | 08:06 | |
| *** shardy has quit IRC | 08:09 | |
| harmw | ok, so, I've added the comments again | 08:09 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Glance containers https://review.openstack.org/197761 | 08:10 |
| harmw | well, typed 'em in | 08:10 |
| mickt | Quick Q regarding murano dashboard. Should it be part of horizon docker image or separate? I presume the latter. | 08:10 |
| SamYaple | mickt: is Murano a plugin to the horizon dash, or a standable web? | 08:15 |
| mickt | plugin | 08:18 |
| SamYaple | it would need to be in the Horizon container then | 08:18 |
| SamYaple | is it triggered to be active by something, or is it always active mickt ? | 08:19 |
| harmw | [ ! -f $cfg ] && continue, that shouldn't error out because there is not set -oerrorbla | 08:19 |
| SamYaple | harmw: it will because we are sourcing that file, not exec | 08:19 |
| harmw | ok | 08:20 |
| SamYaple | didn't we just agree to do that simple copy and come back around to this issue? | 08:20 |
| harmw | I thought this was simple | 08:20 |
| SamYaple | :/ | 08:20 |
| mickt | SY - mmmm, just started looking into it; I'll get back to you on that. I would presume the latter. | 08:21 |
| SamYaple | mickt: if it is always active then well need some wayto toggle it on/off | 08:21 |
| mickt | sorry former | 08:21 |
| SamYaple | oh if its triggered active then it should be fine to include | 08:21 |
| mickt | as I say, just started to look into it so I'll have a better idea later | 08:21 |
| SamYaple | sure thing! | 08:22 |
| SamYaple | mickt: did I talk to you about the Murano dashboard in #openstack a few weeks back? | 08:22 |
| mickt | no | 08:23 |
| SamYaple | k | 08:23 |
| SamYaple | carry on | 08:23 |
| *** gfidente has joined #kolla | 08:29 | |
| harmw | SamYaple: looks like your comments have evolved between keystone and galera as well, which is why I'm missing some bits | 08:33 |
| SamYaple | yea it was that damn rebasing 20 times | 08:33 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Glance containers https://review.openstack.org/197761 | 08:36 |
| SamYaple | harmw: "When bootstrapping is not required, start the service." that isn't correct since starting the service is not in the if block | 08:37 |
| vincent_vdk | kolla can be deployed on atomic hosts right? | 08:40 |
| harmw | should just work, I know CoreOS works vincent_vdk | 08:40 |
| vincent_vdk | really? | 08:41 |
| vincent_vdk | so how does it work with ovs? that lives on the host afaik | 08:41 |
| SamYaple | vincent_vdk: OVS doesn't work with the current deploymethods at this time | 08:41 |
| SamYaple | support still needs to be written for it | 08:42 |
| SamYaple | OVS + containers work _in_ _general_ because OVS is a kernel module, so the processes talk to the kernel module | 08:42 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Glance containers https://review.openstack.org/197761 | 08:44 |
| vincent_vdk | right, currently linux bridge is supported | 08:44 |
| harmw | yep | 08:44 |
| SamYaple | indeed | 08:44 |
| harmw | SamYaple: thanks for +2, finally! :) Ill use it as template tonight for other containers | 08:50 |
| SamYaple | sounds good. i need that glance patch to unblock ansible work | 08:52 |
| SamYaple | I need my other patches merged too | 08:52 |
| SamYaple | https://etherpad.openstack.org/p/ansibalising-containers | 08:52 |
| *** shardy has joined #kolla | 08:59 | |
| *** shardy_ has joined #kolla | 09:03 | |
| SamYaple | harmw: cluster that was running from ealier today, i nuked 1 of the three nodes completely | 09:05 |
| SamYaple | i then reran the playbooks and BAM its back | 09:05 |
| harmw | yea yea, you're godlike, we know | 09:06 |
| harmw | :p | 09:06 |
| harmw | nice man | 09:06 |
| SamYaple | Yaodu never worked this good, but i took a bunch of lessons from Kolla and Yaodu when building this | 09:06 |
| *** shardy_ has quit IRC | 09:13 | |
| *** openstackgerrit has quit IRC | 09:19 | |
| *** openstackgerrit has joined #kolla | 09:20 | |
| *** mstachow has joined #kolla | 09:20 | |
| *** dims has joined #kolla | 09:24 | |
| *** pbourke has quit IRC | 09:25 | |
| *** pbourke has joined #kolla | 09:25 | |
| *** bradjones has quit IRC | 09:33 | |
| *** bradjones has joined #kolla | 09:35 | |
| *** bradjones has joined #kolla | 09:35 | |
| *** mickt has quit IRC | 09:36 | |
| harmw | SamYaple: I know what *.*:USAGE does, just wondering why there is a need for that user in the first place since there is no database :) | 09:39 |
| harmw | a link where I can read it myself would suffice, np with that | 09:40 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Update centos-source-base from symlink to a real layer https://review.openstack.org/197897 | 09:46 |
| *** akwasnie has joined #kolla | 09:50 | |
| *** jmccarthy has quit IRC | 09:53 | |
| *** jmccarthy has joined #kolla | 09:54 | |
| openstackgerrit | Michal Rostecki proposed stackforge/kolla: Add --no-install-recommends option to Ubuntu image https://review.openstack.org/197481 | 09:54 |
| *** gfidente has quit IRC | 09:57 | |
| *** gfidente has joined #kolla | 09:58 | |
| mstachow | morning all | 10:09 |
| akwasnie | hi Michal :) | 10:10 |
| SamYaple | harmw: its for the future haproxy health check on the database. the user creation task is also being used to block until the database is active | 10:17 |
| harmw | ok | 10:28 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Update centos-source-base from symlink to a real layer https://review.openstack.org/197897 | 10:46 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add new script to fetch tarballs for source installation https://review.openstack.org/197919 | 10:54 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Update centos-source-base from symlink to a real layer https://review.openstack.org/197897 | 10:55 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Move declaration of TMPDIR to top of build-docker-image https://review.openstack.org/197027 | 10:55 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Keystone installation from source https://review.openstack.org/193166 | 10:55 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Keystone installation from source https://review.openstack.org/193166 | 10:56 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Keystone installation from source https://review.openstack.org/193166 | 10:56 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Keystone installation from source https://review.openstack.org/193166 | 10:57 |
| *** rhallisey has joined #kolla | 11:27 | |
| *** ogelbukh1 has quit IRC | 11:35 | |
| *** ogelbukh1 has joined #kolla | 11:35 | |
| SamYaple | so harmw are you happy with the wording for comments in glance start.sh? If so i will copy those for keystone since I have to update keystone | 11:48 |
| harmw | go ahead | 11:49 |
| SamYaple | roger roger | 11:50 |
| rhallisey | morning | 11:53 |
| rhallisey | sorry I missed the meeting yesterday | 11:53 |
| SamYaple | rhallisey: good morning! | 11:53 |
| mstachow | o/ rhallisey | 11:55 |
| *** jtriley has joined #kolla | 11:58 | |
| *** jtriley has quit IRC | 12:02 | |
| SamYaple | rhallisey: i poked at cinder. have to ideas im playing with to improve things. what are the biggest pain points do you see right now? | 12:05 |
| rhallisey | SamYaple, that is use to work :p | 12:05 |
| SamYaple | :) | 12:05 |
| rhallisey | if you do a nova volume-attach | 12:05 |
| rhallisey | you will see in cinder-volumes log and iscsi error | 12:06 |
| rhallisey | failed to create or something | 12:06 |
| SamYaple | yup | 12:06 |
| SamYaple | i think i know how to fix that | 12:06 |
| rhallisey | then in the volume-attach you will see another iscsi error : session not found | 12:06 |
| SamYaple | yea | 12:06 |
| rhallisey | when /dev was mounted this went away | 12:07 |
| rhallisey | with docker 1.7 | 12:07 |
| rhallisey | so I'm not sure what it is this time | 12:07 |
| SamYaple | right so i had to revert the thin neutron agents because docker reverted the patch in 1.7 that made them work, the same patch you were using | 12:07 |
| SamYaple | thats probably why it broke | 12:07 |
| rhallisey | that's what I was thinking too | 12:07 |
| SamYaple | https://github.com/docker/docker/commit/ff770d33cdee5ab4dc0bdeb06dcfea49aa72ab8e | 12:08 |
| SamYaple | between 1.7.0 rc2 and rc3 this commit happened | 12:08 |
| rhallisey | ya I saw you email to dwalsh | 12:08 |
| rhallisey | ok | 12:08 |
| SamYaple | if you test on rc2 and it works, then thats almost certainly whats happening | 12:08 |
| SamYaple | we will probably have to wait for 1.8 to get this fixed | 12:09 |
| SamYaple | thats 2 months out | 12:09 |
| rhallisey | I', using rc1 though | 12:09 |
| rhallisey | that's why I'm sure if it's docker | 12:10 |
| SamYaple | its broken with rc1? | 12:10 |
| rhallisey | it worked for a time | 12:10 |
| rhallisey | so I think it's something in cinder, but I'm not sure what since I solved this by moving to 1.7 | 12:10 |
| SamYaple | so we are not setting the keystone bind host or port anymore | 12:45 |
| SamYaple | it is binding to the default port on all ip addresses | 12:45 |
| *** jtriley has joined #kolla | 12:50 | |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Nova-controller containers https://review.openstack.org/197954 | 13:00 |
| *** alisonh has left #kolla | 13:00 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Update centos-source-base from symlink to a real layer https://review.openstack.org/197897 | 13:04 |
| *** shardy has quit IRC | 13:05 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add new script to fetch tarballs for source installation https://review.openstack.org/197919 | 13:06 |
| *** shardy has joined #kolla | 13:06 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Updated keystone for config-external https://review.openstack.org/197960 | 13:11 |
| *** prad has joined #kolla | 13:13 | |
| *** dasm is now known as dasm|afk | 13:15 | |
| SamYaple | harmw: recheck the glance and nova bootstrap sections | 13:25 |
| SamYaple | i just noticed the login shell in /etc/password wont allow executing commands so we need to override it | 13:25 |
| *** jtriley has quit IRC | 13:25 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Update centos-source-base from symlink to a real layer https://review.openstack.org/197897 | 13:26 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Move declaration of TMPDIR to top of build-docker-image https://review.openstack.org/197027 | 13:26 |
| *** thumpba has joined #kolla | 13:27 | |
| pbourke | sorry for the barage of patches folks | 13:29 |
| *** jtriley has joined #kolla | 13:29 | |
| pbourke | rebase hell | 13:29 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Keystone installation from source https://review.openstack.org/193166 | 13:29 |
| *** shardy has quit IRC | 13:38 | |
| *** shardy has joined #kolla | 13:44 | |
| *** nihilifer has quit IRC | 13:45 | |
| SamYaple | pbourke: cloning master is what we should be doing | 14:05 |
| SamYaple | stable/kilo is not the branch you are commiting to | 14:05 |
| SamYaple | you are commiting to master | 14:05 |
| pbourke | im not committing to anything :) | 14:06 |
| pbourke | i just thought people would want to build from kilo by default | 14:06 |
| pbourke | maybe im missing the point here | 14:06 |
| SamYaple | kolla master == stable/liberty | 14:06 |
| pbourke | ah | 14:06 |
| SamYaple | there is no stable/liberyty at the moment | 14:06 |
| pbourke | did you see my follow up comment though on that maybe we shouldnt do shallow clone | 14:07 |
| SamYaple | so shallow clones are what we need | 14:07 |
| SamYaple | this is why no buioldconfs in teh repo | 14:07 |
| SamYaple | thats a user config file | 14:07 |
| *** athomas has quit IRC | 14:08 | |
| pbourke | how will it work if someone wants a custom ref | 14:11 |
| *** akwasnie has quit IRC | 14:11 | |
| *** vbel has quit IRC | 14:12 | |
| pbourke | SamYaple: ^ | 14:13 |
| SamYaple | i responded in the review | 14:14 |
| pbourke | ah, thanks | 14:14 |
| pbourke | why worried about the tarball size though? we remove it once it's expanded | 14:15 |
| *** athomas has joined #kolla | 14:15 | |
| *** shardy has quit IRC | 14:18 | |
| SamYaple | you tar the whole git folder which means you add ALL of that into the container affecting final size | 14:18 |
| SamYaple | i dont care about the builders system at all | 14:18 |
| pbourke | ah, I see, even if I rm it within the container it gets added via the build context right? | 14:19 |
| *** dims has quit IRC | 14:19 | |
| SamYaple | ugh you switched back to COPY | 14:20 |
| SamYaple | we can use ADD | 14:21 |
| SamYaple | you just need to do i a bit different | 14:21 |
| SamYaple | also yes the tarball and the xtracted content will have layers and add to the size | 14:21 |
| pbourke | wrt copy - the tarballs from tarballs.openstack.org have a top level dir of keystone_2015.0.1/ | 14:22 |
| pbourke | if I use ADD I'll have to make sure to tar up the git repo with the same name | 14:22 |
| pbourke | we're bound to that naming for any other tarball method | 14:23 |
| SamYaple | ln -s keystone* /keystone | 14:23 |
| SamYaple | ^ | 14:23 |
| pbourke | kk | 14:24 |
| SamYaple | while in the real world that *might* be "bad practice" but in containers in a controlled environment it is safe and allows us to save that layer | 14:26 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add new script to fetch tarballs for source installation https://review.openstack.org/197919 | 14:37 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Updated keystone for wsgi in config-external https://review.openstack.org/197960 | 14:37 |
| *** dims has joined #kolla | 14:38 | |
| *** shardy has joined #kolla | 14:43 | |
| *** dims has quit IRC | 14:44 | |
| *** dims has joined #kolla | 14:49 | |
| *** mstachow_ has joined #kolla | 14:56 | |
| *** bradjones has quit IRC | 14:57 | |
| *** bradjones has joined #kolla | 14:58 | |
| *** bradjones has joined #kolla | 14:58 | |
| *** absubram has joined #kolla | 15:02 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Update centos-source-base from symlink to a real layer https://review.openstack.org/197897 | 15:05 |
| *** dims has quit IRC | 15:06 | |
| *** Guest42148 has joined #kolla | 15:07 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Keystone installation from source https://review.openstack.org/193166 | 15:08 |
| *** Guest42148 is now known as dims_ | 15:09 | |
| *** jtriley has quit IRC | 15:16 | |
| *** shardy_ has joined #kolla | 15:18 | |
| *** shardy has quit IRC | 15:21 | |
| *** jasonsb has quit IRC | 15:29 | |
| *** jtriley has joined #kolla | 15:30 | |
| *** jasonsb has joined #kolla | 15:30 | |
| *** jasonsb has quit IRC | 15:34 | |
| *** jtriley_ has joined #kolla | 15:41 | |
| *** jtriley has quit IRC | 15:44 | |
| *** jtriley_ has quit IRC | 15:46 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Keystone installation from source https://review.openstack.org/193166 | 15:48 |
| *** nihilifer has joined #kolla | 15:55 | |
| *** nihilifer has quit IRC | 16:05 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Keystone installation from source https://review.openstack.org/193166 | 16:07 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Keystone installation from source https://review.openstack.org/193166 | 16:10 |
| SamYaple | sdake: https://etherpad.openstack.org/p/ansibalising-containers | 16:12 |
| SamYaple | the two in "needs code review" should be good to go | 16:13 |
| SamYaple | once those merge you can test them out, follow the docs patch here https://review.openstack.org/#/c/197581/ | 16:13 |
| pbourke | SamYaple: while you're still awake - did you have anything in mind for how to default to a git url / tarball without specifying in buildconf? | 16:16 |
| diga | sdake: Hi | 16:22 |
| diga | I was travelling since morning today, so couldn't see your ping | 16:22 |
| diga | sdake: will be working over the weekend to finish up the things | 16:23 |
| *** jasonsb has joined #kolla | 16:26 | |
| *** mstachow has quit IRC | 16:29 | |
| *** mstachow_ has quit IRC | 16:30 | |
| pbourke | would anyone be opposed to using python for some of the config scripts rather than bash? | 16:31 |
| pbourke | specific case is where I need to specify ring info for swift and its going to be a bit ugly to parse a list of triplets in bash (ip,port,device_name,weight) | 16:32 |
| *** shardy_ has quit IRC | 16:37 | |
| *** shardy has joined #kolla | 16:42 | |
| *** vinkman has joined #kolla | 17:05 | |
| sdake | hey cats | 17:06 |
| sdake | pbourke python is fine imo | 17:06 |
| sdake | the shell scripts started at 10 lines | 17:06 |
| sdake | now they are hundreds | 17:06 |
| sdake | rhallisey cinder goes into reboot loops | 17:08 |
| sdake | rhallisey that should be an easy fix | 17:08 |
| rhallisey | roger that | 17:09 |
| sdake | there is a bug filed for that as well | 17:20 |
| *** absubram has quit IRC | 17:36 | |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Glance containers https://review.openstack.org/197761 | 17:39 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Nova-controller containers https://review.openstack.org/197954 | 17:39 |
| *** absubram has joined #kolla | 17:40 | |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Nova-compute containers https://review.openstack.org/198075 | 17:51 |
| *** absubram has quit IRC | 17:55 | |
| *** thumpba has quit IRC | 17:58 | |
| * sdake hates expense reporting day in our household | 18:07 | |
| *** jasonsb has quit IRC | 18:11 | |
| *** absubram has joined #kolla | 18:15 | |
| sdake | harmw where are you with the config internal exteranl | 18:21 |
| sdake | those two reviews gtg? | 18:21 |
| harmw | first 2 yea, nova-compute is new but probably good | 18:22 |
| harmw | working on neutron atm | 18:22 |
| harmw | if I continue at this rate it'll all be done this sunday | 18:22 |
| *** bmace has quit IRC | 18:29 | |
| sdake | harmw i have an update for your review | 18:31 |
| sdake | harmw try to save some work for diga ;) | 18:31 |
| sdake | harmw i'd like him to learn how the code works too | 18:31 |
| sdake | and coolsvap|afk if he contribute sin this area as well | 18:31 |
| *** bmace has joined #kolla | 18:32 | |
| harmw | hehe, I'm not taking over containers from other assignees sdake :) not without consulting them first, at the very least | 18:32 |
| sdake | thanks | 18:38 |
| sdake | they are actually volunteers, they weren't assigned iirc ;) | 18:39 |
| harmw | they assigned themselves, judging by the etherpad ;) | 18:39 |
| sdake | i assinged you some tho | 18:39 |
| harmw | yep | 18:40 |
| sdake | because i wanted them done fast and I now you are further upt o speed on the code base | 18:40 |
| sdake | now/now | 18:40 |
| sdake | know | 18:40 |
| sdake | if your attending midcycle in person pleae book hotel rooms asap - silicon valley rates are going up quickly! | 18:43 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Neutron containers https://review.openstack.org/198091 | 18:54 |
| *** athomas has quit IRC | 18:55 | |
| harmw | coolsvap|away: could you ping me when back? you opted to do Cinder so I'm keen on knowing if you've started on that or need help or anything :) re: https://etherpad.openstack.org/p/kolla-standard-start | 18:56 |
| *** sdake_ has joined #kolla | 18:58 | |
| harmw | sdake: I took the wrong spec? jeez | 18:58 |
| harmw | can I just change that on my end in the commitmsg, or is more work needed? | 19:00 |
| *** athomas has joined #kolla | 19:00 | |
| *** athomas has quit IRC | 19:00 | |
| *** sdake has quit IRC | 19:00 | |
| *** sdake has joined #kolla | 19:01 | |
| *** absubram has quit IRC | 19:01 | |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Nova-compute containers https://review.openstack.org/198075 | 19:03 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Glance containers https://review.openstack.org/197761 | 19:03 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Nova-controller containers https://review.openstack.org/197954 | 19:03 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Neutron containers https://review.openstack.org/198091 | 19:03 |
| *** sdake_ has quit IRC | 19:04 | |
| *** shardy has quit IRC | 19:16 | |
| *** shardy has joined #kolla | 19:17 | |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Heat containers https://review.openstack.org/198098 | 19:18 |
| harmw | now for swift | 19:19 |
| *** jasonsb has joined #kolla | 20:24 | |
| *** jasonsb has quit IRC | 20:24 | |
| *** jasonsb has joined #kolla | 20:25 | |
| sdake | i bought one of these last night http://www.theabsolutesound.com/articles/lehmannaudio-decade-phonostage/ | 20:30 |
| *** prad has quit IRC | 21:24 | |
| openstackgerrit | Merged stackforge/kolla: Remove duplicate PUBLIC_IP entry from tools/genenv https://review.openstack.org/197509 | 21:27 |
| *** gfidente has quit IRC | 21:33 | |
| *** shardy has quit IRC | 21:34 | |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add config-internal/config-external to Swift containers https://review.openstack.org/198140 | 21:39 |
| *** erkules_ has joined #kolla | 21:46 | |
| *** erkules has quit IRC | 21:47 | |
| sdake | harmw I assume your testing these changes with config internal config external in the config internal mode as you go? | 22:01 |
| *** sdake has quit IRC | 22:22 | |
| *** sdake has joined #kolla | 22:22 | |
| sdake | harmw I assume your testing these config internal external changes as you go | 22:23 |
| sdake | if fols dont' start filing techdebt bugs when they submit patches and I +2 them, I am going to start -1 patches until the techdebt bug is ifled or the bug is fixed ;) | 22:32 |
| sdake | just a FYI :) | 22:32 |
| *** sdake has quit IRC | 22:43 | |
| *** sdake has joined #kolla | 22:44 | |
| *** rhallisey has quit IRC | 22:45 | |
| *** sdake has quit IRC | 22:46 | |
| *** sdake has joined #kolla | 22:50 | |
| *** sdake has quit IRC | 22:51 | |
| *** sdake has joined #kolla | 22:52 | |
| *** sdake has quit IRC | 22:58 | |
| *** jasonsb has quit IRC | 23:07 | |
| *** jasonsb has joined #kolla | 23:08 | |
| *** jasonsb has quit IRC | 23:12 | |
| *** jasonsb has joined #kolla | 23:13 | |
| *** sdake has joined #kolla | 23:37 | |
| sdake | rhallisey you around | 23:40 |
| sdake | can you add me as an operator plz | 23:40 |
| sdake | apparently not around ;( | 23:40 |
| *** vinkman1 has joined #kolla | 23:50 | |
| *** vinkman has quit IRC | 23:51 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!