| *** jruano has joined #kolla | 00:00 | |
| *** achanda has quit IRC | 00:01 | |
| *** loth has left #kolla | 00:06 | |
| *** dims_ has joined #kolla | 00:08 | |
| *** inc0 has joined #kolla | 00:18 | |
| inc0 | good afternoon guys | 00:19 |
|---|---|---|
| *** shadower has quit IRC | 00:23 | |
| *** shadower has joined #kolla | 00:23 | |
| *** bmace has joined #kolla | 00:26 | |
| bmace | diga around? | 00:29 |
| *** zhiwei has joined #kolla | 00:57 | |
| *** inc0 has quit IRC | 01:05 | |
| SamYaple | hey guys! | 01:20 |
| SamYaple | guess who is not sleeping? | 01:20 |
| SamYaple | <--- this guy | 01:20 |
| *** jasonsb has quit IRC | 01:20 | |
| bmace | hi again SamYaple | 01:20 |
| SamYaple | hello | 01:22 |
| SamYaple | is jenkins all broked? | 01:22 |
| bmace | SamYaple sort of looks like it, don't see many V+1s lately | 01:25 |
| *** unicell has quit IRC | 01:34 | |
| *** erkules_ has joined #kolla | 01:59 | |
| *** erkules has quit IRC | 02:01 | |
| *** jtriley has joined #kolla | 02:09 | |
| *** jasonsb has joined #kolla | 02:25 | |
| *** vinkman has quit IRC | 02:26 | |
| *** openstackgerrit has quit IRC | 02:31 | |
| *** openstackgerrit has joined #kolla | 02:31 | |
| *** achanda has joined #kolla | 02:42 | |
| *** jtriley has quit IRC | 02:49 | |
| diga | bmace: Hi | 03:17 |
| SamYaple | hi | 03:17 |
| diga | SamYaple: Hi | 03:19 |
| SamYaple | o/ | 03:19 |
| diga | couldn't attend the Mid-cycle due to workload | 03:19 |
| bmace | hey diga, before i got to deep into it i wanted to ping you about the kolla-ansible command and see if you had done any work on it yet. | 03:20 |
| bmace | i got start working well in my environment and i'm looking at what it will take to do some of the other commands we talked about supporting at the mid-cycle | 03:21 |
| bmace | well, deploy | 03:21 |
| bmace | which does a start | 03:21 |
| SamYaple | yea bmace to do some of the stuff we talked about at the mid-cycle is going to require playbook tweaking | 03:21 |
| bmace | yup, i'm reading an ansible book :) looks like a decent amount of tweaking based on how it is implemented right now | 03:22 |
| diga | bmace: I think you can start with deploy https://blueprints.launchpad.net/kolla/+spec/kolla-ansible-script | 03:22 |
| bmace | diga ok, i wanted to make sure you didn't have any work in flight before i pushed something in or spent much time on it. | 03:23 |
| bmace | i'll probably get deploy in asap, and put some doc in the script related to the other commands we talked about supporting. teardown, deployconfig, start, stop <host>, etc. | 03:23 |
| diga | sure bmace | 03:23 |
| SamYaple | bmace: nah i dont think so | 03:24 |
| SamYaple | itll be simple | 03:24 |
| bmace | in case you are interested diga the list is from https://etherpad.openstack.org/p/liberty-kolla-midcycle and starts on line 109 | 03:24 |
| SamYaple | we can not use tags and call the config.yml directly | 03:25 |
| SamYaple | maybe | 03:25 |
| SamYaple | idk | 03:25 |
| SamYaple | i havent thought about it | 03:25 |
| bmace | yeah, the way we are doing things now makes it as non-obvious as possible, from an ansible point of view, as far as i can tell. everything seems made to go in one direction, up, at least on any host defined. you can of course mess with your inventory to make stuff go down, but i think we had not wanted that, at least on stop <host> for example. | 03:26 |
| bmace | hunting down the flow of things is, interesting, but maybe when i get better w/ ansible it won't be quite as bad. | 03:26 |
| SamYaple | i am doing non-standard ansible things to make dev easier. I want to use ansible, but we are Kolla and ansible isnt required | 03:27 |
| SamYaple | we start getting to fancy with ansible then kolla requires it | 03:27 |
| diga | let me go through the etherpad link bmace | 03:27 |
| SamYaple | i dont want that | 03:27 |
| *** echoingumesh has joined #kolla | 03:27 | |
| bmace | i am not sure those things need to be mutually exclusive. being able to deploy w/ ansible and being able to deploy independently, yet having reasonably understandable ansible flow. | 03:28 |
| *** achanda has quit IRC | 03:28 | |
| SamYaple | we have reasonable ansible flow | 03:28 |
| SamYaple | the _only_ wierd bit is the common playbooks in the upper directory | 03:29 |
| SamYaple | thats just to respect DRY | 03:29 |
| *** diogogmt has joined #kolla | 03:30 | |
| bmace | the main bits that you actually activate from ansible-playbook end up having pretty much no information about flow in them, which is fine, it just links via tags to other stuff. most of the other ansible playbooks i have looked at so far had far more obvious flow, but that was probably because they were simpler. | 03:31 |
| SamYaple | what do you mean? | 03:31 |
| *** dims_ has quit IRC | 03:33 | |
| bmace | i mean looking at what we have right now and knowing the flow of what files are called from where is pretty non-obvious. again, i am still reading up on ansible, but it looks to be possible to write playbooks in a way that it is more obvious what is going on. | 03:33 |
| bmace | i guess we end up getting into the main playbooks, from somewhere, which include all the config, start, etc.. and that causes all those to get kicked off, but as far as i can tell we don't document flow anywhere, probably because if you know a lot about ansible, maybe it is obvious. | 03:34 |
| SamYaple | i dont think you have gotten to the part about roles | 03:35 |
| SamYaple | roles are the future and this is very obvious to follow the structure | 03:35 |
| *** diogogmt has quit IRC | 03:37 | |
| SamYaple | bmace: http://docs.ansible.com/ansible/playbooks_roles.html | 03:40 |
| SamYaple | thats what you need to read | 03:40 |
| SamYaple | specifically the secont on roles | 03:40 |
| SamYaple | section | 03:40 |
| *** sdake_ has joined #kolla | 03:42 | |
| *** echoingumesh has quit IRC | 03:42 | |
| *** sdake has quit IRC | 03:43 | |
| *** sdake_ has quit IRC | 03:45 | |
| *** sdake has joined #kolla | 03:45 | |
| bmace | hrm, then to do some sort of conditional like a tear down or take down the containers on a specific host, you would prefer a different playbook w/ different roles, or to pass in conditionals on the ansible-playbook command line? | 03:45 |
| SamYaple | bmace: we can't have 'site.yml' for everything | 03:47 |
| SamYaple | I am perfectly fine with a destroy.yml that we can pass values | 03:47 |
| SamYaple | so it will be a seperate playbook | 03:47 |
| SamYaple | at that point its easy to make it do anything you want | 03:47 |
| bmace | sure, at least it would be the same inventory / host list, for the teardown. | 03:49 |
| *** coolsvap|away is now known as coolsvap | 03:49 | |
| SamYaple | yea thats not a problem at all to be honest | 03:51 |
| *** sdake_ has joined #kolla | 03:51 | |
| SamYaple | the real issue is the configs since we have the defaults/main.yml | 03:51 |
| SamYaple | i can make it work, but im not sure best practice | 03:51 |
| *** sdake has quit IRC | 03:52 | |
| bmace | that is the odd thing about ansible, where the goal is to reflect a final state, but for a config re-push and restart, we are really sort of in the end state already, which is having all the containers running on the hosts we want. doing certain programmatic flow sorts of things just feels, off, so far. | 03:56 |
| SamYaple | yea youll get over it bmace | 03:58 |
| SamYaple | its still better than puppet and chef | 03:58 |
| SamYaple | but i totally get whatyou mean | 03:59 |
| SamYaple | bmace: you do know how to do it though, right? | 04:01 |
| bmace | yeah, i love how light it is compared to puppet / chef | 04:01 |
| SamYaple | you can either check the configs match and use that true/false to rekick the container or you can remember that the configs changed and use that info | 04:02 |
| bmace | i like the idea of using the checksum to only re-start the containers that need it | 04:02 |
| SamYaple | bmace: well the containers dont restart now... | 04:05 |
| SamYaple | a config refresh wont restart _all_ containers | 04:05 |
| SamYaple | the current code only restarts "different than the desired end state containers" | 04:05 |
| bmace | right, so it would need to track, if we have the "change on config update" bit set, then also restart the container, since we want to be able to re-push configs without restarts, i assume | 04:09 |
| openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Fedora - Install from source for Neutron https://review.openstack.org/207709 | 04:15 |
| bmace | i'll play around with things. i set args for the inventory / playbook file on the kolla-ansible deploy command. maybe i'll update the docs and toss that in just so it is that much easier to get folks going w/ ansible. also, noticed the /etc/kolla stuff was hard coded in a bunch of files. any issues you know making that variable like we have so many other things? | 04:18 |
| openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Fedora - Install from source for Magnum https://review.openstack.org/207710 | 04:21 |
| *** sdake_ has quit IRC | 04:21 | |
| *** sdake has joined #kolla | 04:21 | |
| sdake | evening folks | 04:24 |
| bmace | hola | 04:25 |
| SamYaple | evening sdake | 04:27 |
| SamYaple | jenkins seems to not be +1 | 04:27 |
| *** sdake has quit IRC | 04:28 | |
| *** sdake_ has joined #kolla | 04:28 | |
| *** sdake has joined #kolla | 04:31 | |
| *** sdake_ has quit IRC | 04:31 | |
| *** sdake_ has joined #kolla | 04:35 | |
| *** sdake has quit IRC | 04:36 | |
| *** sdake_ has quit IRC | 04:41 | |
| *** sdake has joined #kolla | 04:42 | |
| *** sdake has quit IRC | 04:44 | |
| *** sdake has joined #kolla | 04:45 | |
| *** sdake has quit IRC | 04:47 | |
| *** sdake has joined #kolla | 04:47 | |
| sdake | only 6 purples left ;-) | 04:49 |
| bmace | not sure if you know, there is no kollaglue/centos-rdo-haproxy so it folks are just playing with kolla out of the box without doing their own local image build ansible won't work without some tweaking (like yanking haproxy) | 04:51 |
| sdake | what we reallly need is documentation that says to build your own image | 04:51 |
| sdake | andgive up on keeping the upstream dockerhub updated | 04:51 |
| sdake | because we re doing a poor job of it and its a manual process | 04:51 |
| bmace | yeah, we do fall behind pretty quickly, based on how quickly the code advances. not that we couldn't probably push up like once a day and overwrite latest, but we would need some resources and a script to do that. | 04:52 |
| bmace | depending on how onerous that is, it is sort of nifty to have them up there for people to pull. i think it easier to set up a system to run pre-built containers than to get container builds working. | 04:53 |
| SamYaple | bmace: i think the built contaienrs will be fine for "intro to kolla" and testing | 04:56 |
| SamYaple | but if youre really going to use Kolla, you should be building your own image | 04:56 |
| SamYaple | unless we can push after every commit | 04:56 |
| sdake | samyaple bmace's pont is there is no haproxy container | 04:56 |
| sdake | so kolla is unusable from the registry atm | 04:56 |
| SamYaple | sdake: there no lots of containers | 04:56 |
| SamYaple | it hasnt been usable for over 2 months | 04:57 |
| bmace | i don't think every commit is necessary. i think daily would be perfectly sufficient, even every other day. | 04:57 |
| SamYaple | bmace: but it would need to be every commit | 04:57 |
| SamYaple | otherwise it might break during a day | 04:57 |
| sdake | ideally it would be a gated version | 04:57 |
| sdake | which mean master, since master is gated | 04:57 |
| sdake | ;-) | 04:57 |
| SamYaple | until the gated version, people should be expected to build thier own | 04:58 |
| sdake | but our master gate isn't voting | 04:58 |
| bmace | i am just saying daily would be considerably better than what we have now, which is haphazard | 04:58 |
| SamYaple | bmace: no we need to push at the L1 L2 L3 stages | 04:58 |
| bmace | if the images we have up on dockerhub are unreliable, having them there at all is misleading | 04:58 |
| SamYaple | that way it works | 04:58 |
| SamYaple | you can checkout L2 and have the code work | 04:58 |
| SamYaple | you can expect master to work with old images | 04:58 |
| sdake | master wont work with l1 images samyaple | 04:59 |
| sdake | that only really applies after l2 | 04:59 |
| SamYaple | it would have had we pushed l1 | 04:59 |
| SamYaple | i mean git checkout L2 | 05:00 |
| SamYaple | even after we push L2, master cant be expected to work with the dockerhub images | 05:00 |
| SamYaple | but the L2 tag should | 05:00 |
| bmace | we might as well turn off the pull from docker hub by default, other tagged versions, if latest isn't going to ever work reliably in development. | 05:00 |
| bmace | the current behavior is a likely turn off to people trying it out for the first time. | 05:01 |
| SamYaple | bmace: but you would expect a private registry for your built images | 05:01 |
| sdake | samyaple bmace means peopel giving it a 10 minute spin | 05:01 |
| sdake | not people th are actually going to use it long term | 05:01 |
| SamYaple | i think this gets solved when we package kolla | 05:01 |
| SamYaple | we push at l2 and have pip install kolla==L2 | 05:02 |
| SamYaple | people who want to devel can git checkout | 05:02 |
| sdake | really i think what bmace wants is l2-latest vs l2 | 05:04 |
| bmace | i think the 10 minute spin actually functioning matters, even for people checking it out for something to contribute to. we can definitely say have a packaged kolla that works well, and that is what a certain set of people use, and in dev we always build images, and make sure we don't have stuff set up to by default pull from dockerhub | 05:04 |
| sdake | the 10 minut espin is important | 05:04 |
| SamYaple | if you git checkout you should be expected to build | 05:04 |
| sdake | agree we need to make that happen | 05:05 |
| SamYaple | 10 minute spinup means you use a package | 05:05 |
| SamYaple | or you can git checkout the L2 tag | 05:05 |
| sdake | samyaple it could happen vi a viarety ways not just packaging | 05:05 |
| SamYaple | like i said, the L2 tag | 05:05 |
| SamYaple | its going to break daily even if we push daily in master | 05:05 |
| sdake | it will braek until our gate is voting | 05:06 |
| sdake | an we can't make oru gat evoting until rdo has ha in their registry of rpms | 05:06 |
| sdake | rather their yum repos | 05:06 |
| *** vinkman has joined #kolla | 05:06 | |
| *** coolsvap is now known as coolsvap|away | 05:08 | |
| sdake | bmace if you make a script that canbe run from crontab that git pulls master, i'll lpony up the 50 bucks a month for a rackspace account to push from | 05:09 |
| sdake | git pulls master builds source builds binary and pushes and then cleans up | 05:09 |
| bmace | i'll have to take a look at the dev guide. for now maybe it just comes down to making sure people are always building their own images. code right now that defaults to doing pulls from dockerhub is problematic. | 05:09 |
| bmace | i'm torn. i don't disagree with sam that, even with daily pushes it is likely to be broken any given day, and for people playing we may be better off just having an l2 / l3 / etc.. and otherwise we just make sure we, as much as possible, document / force people to build the images and let them know that dockerhub images used against master are almost guaranteed to fail. | 05:13 |
| bmace | once we have better tests and better gating, being able to push something out that is known to work will be reasonable, but i don't think that is super close. | 05:14 |
| sdake | lets make an effort to get dockerhub rolling | 05:15 |
| sdake | but I need someone to write a build script | 05:15 |
| sdake | i am a bit busy atm and cant tackle it myself | 05:15 |
| sdake | but I can do the setup of the env | 05:15 |
| *** coolsvap|away is now known as coolsvap | 05:17 | |
| sdake | we will push l1, l2, l3 | 05:17 |
| sdake | but also l1-latest and l2-latest and l3-latest | 05:17 |
| sdake | l3-latest wil be master for example | 05:17 |
| sdake | in 2-4 weeks i will have gig-e at my house | 05:20 |
| sdake | and can push from there | 05:20 |
| sdake | it takes probably about an hour | 05:20 |
| SamYaple | sorry my internet connection tropped | 05:22 |
| sdake | uber refunded my trip yay :) | 05:23 |
| SamYaple | what my point was is if you push all the images. then someone changes start.sh. those images are broke until you push again | 05:23 |
| SamYaple | you can't expect master to be able to deploy built iamges | 05:24 |
| openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Fedora - Install from source for Ceilometer https://review.openstack.org/207723 | 05:24 |
| SamYaple | you _can_ expect tags to work | 05:24 |
| SamYaple | so you either pip install kolla, or git checkout --tag | 05:24 |
| SamYaple | those two should "just work" | 05:24 |
| sdake | yes i think we should want master to be tire-kickable as well tho | 05:25 |
| sdake | I get what your saying about tags, we should definatey do that | 05:25 |
| SamYaple | its not possible | 05:25 |
| SamYaple | no until the gate can push | 05:25 |
| sdake | its possible just not reliable | 05:25 |
| SamYaple | well then it sounds like we are currently doing it | 05:25 |
| SamYaple | if reliable isnt a requirement then were there | 05:26 |
| sdake | we are not doing it because nobody has written a script to do it daily | 05:26 |
| sdake | semi-reliable is a requirement ;) | 05:26 |
| sdake | as in we expect it to work some of the time | 05:26 |
| sdake | right nwo it doen't work any of the time | 05:26 |
| SamYaple | 100% reliable is tags and packages | 05:26 |
| SamYaple | we agreed no latest tags at the mid-cycle so what would these even be called | 05:26 |
| sdake | we can warn people in the documentation that th latest tag in dockerhub is not a reliable exeuction method but may or may not work ;) | 05:28 |
| SamYaple | but we agreed no latest tag unless it always works | 05:28 |
| *** jruano has quit IRC | 05:29 | |
| sdake | i dont recall that agreement but it was a busy two days :) | 05:29 |
| SamYaple | if there is no latest tag then its clear "ah i need to build the images" if shit starts downloading then breaks it makes Kolla seem broken | 05:29 |
| sdake | ya i tend to agree with you sam, a git pul from the tag and docker pull from the tag is the way to go - then it works 100% of the time | 05:30 |
| sdake | and if people want master to work they have ot build images | 05:30 |
| sdake | lets just document it | 05:30 |
| SamYaple | Kolla is super reliable by the design and using images. I dont want to mix in this "sometimes reliable" mess | 05:30 |
| sdake | because now nobody knows this information outside the dev team | 05:30 |
| SamYaple | sounds good | 05:30 |
| SamYaple | i totally get that | 05:30 |
| *** achanda has joined #kolla | 05:30 | |
| sdake | i'll do that doc fix | 05:31 |
| sdake | either tonight or tomorrow morning | 05:31 |
| SamYaple | so for L2 we need to change the default pull tag to liberty2 or something and push taht tag up | 05:31 |
| SamYaple | test that then tag the commit it came from | 05:31 |
| sdake | do we have pip registered for kolla? | 05:33 |
| *** sdake has quit IRC | 05:35 | |
| SamYaple | ha! as if | 05:35 |
| SamYaple | larsks: help us | 05:35 |
| SamYaple | ive personally never made a package for pip | 05:35 |
| SamYaple | larsks can probably hammer it out in a minute | 05:36 |
| SamYaple | I know it very basic, just never done it | 05:36 |
| *** sdake has joined #kolla | 05:41 | |
| sdake | ok i just registered kolla 0.0.1 with pypi | 05:47 |
| SamYaple | awesome! | 05:47 |
| sdake | now we need an install | 05:47 |
| sdake | ;-) | 05:47 |
| SamYaple | like packged and all? | 05:47 |
| SamYaple | oh | 05:47 |
| sdake | no not packaged | 05:47 |
| SamYaple | ok yea that shouldnt be hard | 05:47 |
| SamYaple | i just need some cycles for it | 05:48 |
| SamYaple | (or hopefully larsks saves the day, he isn't busy at all) | 05:48 |
| sdake | samyaple you ahve a bunch of things in needs code review ont he launchpad tracker for l2 | 05:49 |
| sdake | are any of those finished? | 05:49 |
| sdake | i was pretty sure aio worked for example | 05:49 |
| sdake | and the core services are done right? | 05:49 |
| SamYaple | in review | 05:49 |
| SamYaple | i need keepalived merged | 05:50 |
| SamYaple | were good past that | 05:50 |
| sdake | samyaple can yo update the states on your blueprints if they are implemented then | 05:51 |
| sdake | or are you waiting on keepalived? | 05:51 |
| SamYaple | waiting on keepalived | 05:51 |
| sdake | mandre you around? | 05:51 |
| coolsvap | pbourke, around? | 05:54 |
| *** coolsvap is now known as coolsvap|away | 06:05 | |
| *** coolsvap|away is now known as coolsvap | 06:13 | |
| *** unicell has joined #kolla | 06:13 | |
| *** unicell1 has joined #kolla | 06:24 | |
| *** unicell has quit IRC | 06:25 | |
| openstackgerrit | Merged stackforge/kolla: Cinder requires the automaton package to be installed https://review.openstack.org/206696 | 06:39 |
| *** vinkman has quit IRC | 06:51 | |
| *** coolsvap is now known as coolsvap|away | 07:02 | |
| *** coolsvap|away is now known as coolsvap | 07:10 | |
| *** achanda has quit IRC | 07:16 | |
| *** erkules_ is now known as erkules | 07:27 | |
| *** erkules has quit IRC | 07:27 | |
| *** erkules has joined #kolla | 07:27 | |
| *** gfidente has joined #kolla | 07:36 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add linuxbridge as option to ansible https://review.openstack.org/207760 | 07:50 |
| SamYaple | sdake: ^ that configuration wise should be working for linuxbridge | 07:50 |
| SamYaple | i cant get it to work though because i dont know how to setup linuxbridge | 07:50 |
| SamYaple | in ovs i would create br-ex and add the neutron_interface to br-ex | 07:51 |
| SamYaple | i dont know what to do with this | 07:51 |
| *** juggler_ is now known as juggler | 08:10 | |
| pbourke | coolsvap: morning | 08:24 |
| *** athomas has joined #kolla | 08:35 | |
| coolsvap | pbourke, did you run into issues for nova from source? | 08:36 |
| coolsvap | related to lvm | 08:36 |
| coolsvap | https://bugs.launchpad.net/kolla/+bug/1480077 | 08:36 |
| openstack | Launchpad bug 1480077 in kolla "nova install from source image fails for lvm" [High,Confirmed] - Assigned to Swapnil Kulkarni (coolsvap) | 08:36 |
| pbourke | coolsvap: not that I recall. Is then when trying to boot an image? | 08:40 |
| coolsvap | no while building the image from Dockerfile | 08:40 |
| coolsvap | i was testing it with fedora | 08:41 |
| coolsvap | it failed so i checked for centos as well | 08:41 |
| coolsvap | its failing there as well | 08:41 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Implement Swift Containers https://review.openstack.org/202171 | 08:41 |
| pbourke | strange no I didnt get that | 08:41 |
| coolsvap | pbourke, np I am trying to see whats wrong | 08:42 |
| pbourke | ok thanks | 08:42 |
| openstackgerrit | Merged stackforge/kolla: Add keepalived to ansible https://review.openstack.org/205570 | 08:50 |
| openstackgerrit | Merged stackforge/kolla: Forces metadata to bind to single interface https://review.openstack.org/205802 | 08:50 |
| openstackgerrit | Merged stackforge/kolla: Add a doc for Cinder https://review.openstack.org/207597 | 08:58 |
| openstackgerrit | Merged stackforge/kolla: Rename role name to service name https://review.openstack.org/205798 | 09:03 |
| openstackgerrit | Merged stackforge/kolla: Properly use groups throught Ansible https://review.openstack.org/205999 | 09:04 |
| *** dwalsh has joined #kolla | 09:07 | |
| *** dwalsh has quit IRC | 09:12 | |
| *** dwalsh has joined #kolla | 09:13 | |
| *** dwalsh has quit IRC | 09:52 | |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Add missing slash if docker registry is specified https://review.openstack.org/207463 | 10:03 |
| coolsvap | pbourke, got the reasons for failure, https://blueprints.launchpad.net/nova/+spec/consolidate-libvirt-fs-volume-drivers | 10:05 |
| coolsvap | with this bp the configuration options are chaging, but the imports are not changed | 10:05 |
| *** athomas has quit IRC | 10:14 | |
| openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Temporary change the nova tar url https://review.openstack.org/207825 | 10:37 |
| openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Fedora - Install from source for Nova https://review.openstack.org/207827 | 10:38 |
| *** athomas has joined #kolla | 10:40 | |
| openstackgerrit | Merged stackforge/kolla: Add build from source for gnocchi https://review.openstack.org/205952 | 10:50 |
| openstackgerrit | Merged stackforge/kolla: Fedora - Install from source for Neutron https://review.openstack.org/207709 | 10:53 |
| openstackgerrit | Merged stackforge/kolla: Fedora - Install from source for Magnum https://review.openstack.org/207710 | 10:54 |
| *** zhiwei has quit IRC | 10:57 | |
| *** dims_ has joined #kolla | 11:04 | |
| openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Fedora - Install from source for Horizon https://review.openstack.org/207838 | 11:07 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Vagrantfile and docs https://review.openstack.org/204238 | 11:14 |
| *** coolsvap is now known as coolsvap|away | 11:39 | |
| openstackgerrit | Merged stackforge/kolla: Fedora - Install from source for Ceilometer https://review.openstack.org/207723 | 11:40 |
| *** rhallisey has joined #kolla | 12:00 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fix some spelling and formatting errors in the Cinder doc https://review.openstack.org/207859 | 12:17 |
| *** thumpba has joined #kolla | 12:28 | |
| *** thumpba has quit IRC | 12:33 | |
| *** athomas has quit IRC | 12:35 | |
| *** coolsvap|away is now known as coolsvap | 12:36 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fix some spelling and formatting errors in the Cinder doc https://review.openstack.org/207859 | 12:38 |
| *** athomas has joined #kolla | 12:40 | |
| *** coolsvap is now known as coolsvap|away | 12:43 | |
| *** coolsvap|away is now known as coolsvap | 12:45 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Fix missing quotes in tools/genenv https://review.openstack.org/207888 | 13:08 |
| *** coolsvap is now known as coolsvap|away | 13:09 | |
| *** coolsvap|away is now known as coolsvap | 13:16 | |
| *** gfidente has quit IRC | 13:22 | |
| *** gfidente has joined #kolla | 13:26 | |
| *** gfidente has joined #kolla | 13:26 | |
| *** coolsvap is now known as coolsvap|away | 13:30 | |
| *** jtriley has joined #kolla | 13:34 | |
| *** diogogmt has joined #kolla | 13:37 | |
| pbourke | I think keystone may be broken | 13:45 |
| pbourke | doesn't seem to like use_stderr anymore | 13:46 |
| *** coolsvap|away is now known as coolsvap | 13:48 | |
| *** pbourke has quit IRC | 13:51 | |
| *** coolsvap is now known as coolsvap|away | 13:51 | |
| vbel | the log goes to apache log | 13:54 |
| vbel | Good morning/ evening, ubuntu server and oracle linux fail to run Kolla containers that use 'su' because of https://github.com/docker/docker/issues/5899 | 13:55 |
| *** pdb has joined #kolla | 13:56 | |
| vbel | At the same time su can be replaced with sudo and the workaround works quite well. Shall we replace it for all or just ubuntu/ol images? https://bugs.launchpad.net/kolla/+bug/1480209 | 13:56 |
| openstack | Launchpad bug 1480209 in kolla "containers fail to run commands with 'su'" [Undecided,New] - Assigned to Vladislav Belogrudov (vlad-belogrudov) | 13:56 |
| vbel | this was the reason ol-source-keystone did not populate its database - uses db sync via su | 13:57 |
| vbel | and later on it tried to load some federation code with non-existing xmldsig | 13:58 |
| rhallisey | oh interesting | 13:58 |
| rhallisey | ya put up a patch to fix that | 13:58 |
| vbel | will do, no objections for sudo? :) I tried this earlier in our containers because we met the same problem (--net=host containers cannot 'su') | 13:59 |
| rhallisey | I think it will be fine | 14:02 |
| *** diogogmt has quit IRC | 14:06 | |
| *** diogogmt has joined #kolla | 14:17 | |
| *** jruano has joined #kolla | 14:21 | |
| *** prad has joined #kolla | 14:26 | |
| openstackgerrit | Merged stackforge/kolla: Fedora - Install from source for Nova https://review.openstack.org/207827 | 14:34 |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Containers with host networking fail to 'su' on some kernels https://review.openstack.org/207925 | 14:38 |
| pdb | vbel: I seem to be getting this xmldsig problem now | 14:43 |
| pdb | vbel: using config-internal, tables seem to be populated fine | 14:44 |
| *** coolsvap|away is now known as coolsvap | 14:44 | |
| openstackgerrit | Merged stackforge/kolla: Temporary change the nova tar url https://review.openstack.org/207825 | 14:45 |
| vbel | pdb: strange, why could not we get it before? Because it was by itself (not in apache) | 14:46 |
| vbel | ? | 14:46 |
| pdb | vbel: no idea whats changed :( | 14:46 |
| vbel | I start disliking xmldsig because I could not pip install or compile it :( | 14:47 |
| pdb | things not going our way today | 14:47 |
| pdb | it seems to be in the ol image only, though can't confirm as centos is failing with another error around use_stderr | 14:48 |
| pdb | vbel: also on your sudo change, is the sudo command available in all images? | 14:48 |
| vbel | it should - will check. If not - will install it | 14:49 |
| pdb | well that was something I was always a little unsure of, as I know one or two images use sudo for some other things | 14:49 |
| pdb | but when I spin up a centos-rdo-base container for example I get sudo not found | 14:49 |
| vbel | I am adding sudo to bases | 14:55 |
| vbel | where necessary | 14:55 |
| SamYaple | morning | 14:55 |
| SamYaple | yay! thanks guys you merged my patchs | 14:56 |
| SamYaple | life is easier | 14:56 |
| vbel | morning | 14:57 |
| rhallisey | SamYaple, hey having some trouble with getting my allinone ansible going | 14:58 |
| rhallisey | ansible-playbook -i inventory/all-in-one -e ../etc/kolla/defaults.yml -e ../etc/kolla/globals.yml -e ../etc/kolla/passwords.yml site.yml | 14:58 |
| rhallisey | I'm not able to find my env-vars with that | 14:59 |
| SamYaple | rhallisey: i would recommend absolute paths, and I don't recommend using the etc directory in the repo since you really shouldnt be tweaking those files | 15:00 |
| SamYaple | but can you expand on what you mean by that? | 15:00 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Vagrantfile and docs https://review.openstack.org/204238 | 15:00 |
| rhallisey | I get: One or more undefined variables: 'node_config_directory' is undefined | 15:00 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add Ansible support for Heat https://review.openstack.org/205528 | 15:01 |
| vbel | I tried -e @../etc/kolla without luck too because it starts asking for /etc/kolla | 15:01 |
| SamYaple | like I said, it expects this stuff in /etc/kolla i dont know how well realitive paths are going to work | 15:01 |
| SamYaple | rhallisey: but you do need @ | 15:01 |
| SamYaple | ansible-playbook -e@/etc/kolla/defaults.yml -e@/etc/kolla/globals.yml -e@/etc/kolla/passwords.yml -i ~/kolla/ansible/inventory/all-in-one ~/kolla/ansible/site.yml | 15:01 |
| SamYaple | thats what I use | 15:02 |
| rhallisey | kk | 15:02 |
| bmace | but then keep in mind you still need all that .conf stuff still in /etc/kolla | 15:02 |
| SamYaple | bmace: thats what i mean | 15:02 |
| bmace | the "generated" ones end up in /opt/kolla on the machines running containers right? is someone messing with the variable and setting it to /opt/kolla your concern? otherwise the files i see it in certainly support variable substitution. | 15:04 |
| SamYaple | bmace: the destination host is all variable substituationed out | 15:04 |
| SamYaple | no issue there | 15:04 |
| SamYaple | the deploy host has some expectations | 15:04 |
| bmace | right, at this point all the expectations are defined in tasks, which would allow for substitution of /etc/kolla with something else, like in-place from your dev environment. | 15:06 |
| SamYaple | sure, but we probably should be testing with the defaults | 15:07 |
| SamYaple | i mean if you think of Kolla as an application /etc/kolla is the place you want to store confs | 15:07 |
| pdb | vbel: https://bugs.launchpad.net/keystone/+bug/1475796 | 15:07 |
| openstack | Launchpad bug 1475796 in Keystone kilo "using pysaml2 version 3.0.0 breaks keystone in kilo release 2015.1.0" [High,Fix committed] | 15:07 |
| SamYaple | pdb: o/ | 15:08 |
| *** pdb is now known as pbourke | 15:08 | |
| pbourke | SamYaple: morning | 15:08 |
| vbel | pdb: cool! Someone got it right ;) | 15:08 |
| SamYaple | does anyone here understand the setup for linuxbridge? | 15:08 |
| pbourke | vbel: yeah. now to decide which fix we want to take for kilo | 15:09 |
| SamYaple | for OVS i create br-ex and add the "neutron_interface" to that bridge. what do I do with linuxbridge? | 15:09 |
| harmw | SamYaple: nothing, linuxbridge doesn't require configuration, or only very little | 15:09 |
| harmw | Ill have a look tonight | 15:09 |
| harmw | but why bridge, if we have OVS? | 15:10 |
| pbourke | vbel: should we be using stable/kilo rather than the 2015.1.0 tag? | 15:10 |
| pbourke | seems the fix is already merged to stable/kilo | 15:11 |
| pbourke | but the tag isn't going to change | 15:11 |
| pbourke | harmw: I retested your vagrant work earlier, all good! | 15:11 |
| harmw | great! | 15:11 |
| pbourke | harmw: the only thing I dont really like is the bridge selection :/ | 15:12 |
| vbel | pbourke: yes, looks like stable/kilo is the way to go. Will we have 2015.2.0 ? | 15:12 |
| SamYaple | harmw: we support ALL the things | 15:12 |
| harmw | ok SamYaple | 15:12 |
| pbourke | vbel: yeah maybe they retag every so often | 15:12 |
| harmw | pbourke: you mean the specific interface or the fact that it's a public net? | 15:12 |
| harmw | the user may just change that ofcourse | 15:13 |
| pbourke | harmw: just the fact that when I type vagrant up it stops and asks to choose an interface | 15:13 |
| harmw | ah | 15:13 |
| pbourke | the confusion start to appear on my face | 15:13 |
| harmw | yes that sucks, and that's why there is a TODO listed :) | 15:13 |
| pbourke | :) | 15:13 |
| harmw | that realy, realy, realy sucks indeed | 15:13 |
| vbel | pbourke: http://tarballs.openstack.org/keystone/keystone-2015.1.1.tar.gz of July 28, wonder if it is there | 15:13 |
| pbourke | vbel: Id be surprised | 15:14 |
| SamYaple | harmw: so youre saying it magically warps L2 to the "neutron_interface"? it does not seem to be adding that interface to a bridge | 15:14 |
| SamYaple | btw did you guys see the big tent application? A comment reads 'Definitely "one of us"' | 15:15 |
| SamYaple | :D | 15:15 |
| pbourke | harmw: that's why I was kind of hinting could we get away with just another host only network for that adapter. but Im guessing its not so easy? | 15:15 |
| pbourke | SamYaple: :) | 15:15 |
| harmw | SamYaple: ill take a look tonight, ok :p | 15:16 |
| harmw | big tent, woohoo! | 15:16 |
| harmw | pbourke: yea, well my thing was I wanted a public net to get by with testing float stuff and just be 'fully connected' on that end | 15:16 |
| pbourke | I think we need 10 votes? | 15:16 |
| SamYaple | pbourke: only 7 | 15:17 |
| harmw | you could always edit the Vagrant file btw | 15:17 |
| *** dims_ has quit IRC | 15:17 | |
| harmw | sub-optimal, but still | 15:17 |
| vbel | pbourke: https://review.openstack.org/#/c/184598/ included in 2015.1.1 | 15:17 |
| *** dims_ has joined #kolla | 15:17 | |
| vbel | so we should rebase :) | 15:17 |
| pbourke | vbel: cool. just bump the tag in our buildconf no? | 15:18 |
| *** blahRus has joined #kolla | 15:18 | |
| vbel | yes, we can try | 15:18 |
| pbourke | does this tag exist for every project? | 15:19 |
| vbel | looks like that | 15:19 |
| vbel | 2015.1.1 tarballs of July 28 for several projects | 15:20 |
| SamYaple | you guys keep in mind | 15:20 |
| SamYaple | in Liberty everyone is independant tagging | 15:20 |
| SamYaple | that wont match for Liberty | 15:20 |
| SamYaple | make it configurable per project so we can but per project | 15:20 |
| *** dims_ is now known as dimsum__ | 15:21 | |
| vbel | ok for me | 15:21 |
| *** pbourke has quit IRC | 15:30 | |
| *** pdb has joined #kolla | 15:31 | |
| *** pdb is now known as pbourke | 15:32 | |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Add missing slash if docker registry is specified https://review.openstack.org/207463 | 15:33 |
| *** daneyon has joined #kolla | 15:36 | |
| *** jruano has quit IRC | 15:37 | |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Add missing slash if docker registry is specified https://review.openstack.org/207463 | 15:38 |
| pbourke | quick question on the above | 15:41 |
| SamYaple | shoot | 15:42 |
| pbourke | could we fix it in globals by having: | 15:42 |
| pbourke | docker_registry: | 15:42 |
| pbourke | docker_registry: "{{ docker_registry ~ '/' if docker_registry else '' }}" | 15:42 |
| SamYaple | but... thats where we are setting docker_registry | 15:42 |
| SamYaple | i want to keep that stuff hidden from view | 15:43 |
| pbourke | fair enough | 15:43 |
| pbourke | just checking :) | 15:43 |
| vbel | one thing on styling - do we use spaces instead of tabs everywhere? | 15:47 |
| rhallisey | yikes the dependencies are killing me :( | 15:47 |
| rhallisey | from oslo_i18n import _lazy ImportError: cannot import name _lazy | 15:47 |
| rhallisey | having trouble with this one ^ | 15:47 |
| vbel | e.g. in Dockerfiles I see tabs .. | 15:48 |
| SamYaple | vbel: tabs are baaaad | 15:48 |
| rhallisey | vbel, spaces | 15:48 |
| vbel | ok, then will correct RHEL one :) | 15:49 |
| SamYaple | vbel: spaces gaurantee formatting remains the same no matter your system | 15:49 |
| SamYaple | assuming you use a fixed-width font | 15:49 |
| *** Slower has joined #kolla | 15:49 | |
| vbel | my vim is 4 space tab replace mode too | 15:49 |
| vbel | thanks! | 15:49 |
| SamYaple | mine too | 15:49 |
| rhallisey | SamYaple, is it unsanitary to do the allinone ansible on a single baremetal host? | 15:50 |
| *** Slower has quit IRC | 15:50 | |
| SamYaple | rhallisey: define "unsanitary" | 15:50 |
| SamYaple | its mostly how i dev | 15:50 |
| SamYaple | then i test against both that and multihost | 15:50 |
| rhallisey | I'm just trying to get AIO going on my machine | 15:51 |
| pbourke | does that strike you as a little ironic given its one of the reasons we use docker? | 15:51 |
| rhallisey | I don't need a vm or anything right? | 15:51 |
| rhallisey | just set the ips to the hosts | 15:51 |
| rhallisey | pbourke, it does | 15:52 |
| pbourke | I guess config-internal gives you the fully isolated approach | 15:52 |
| pbourke | I use VMs for all dev but not required | 15:52 |
| pbourke | rhallisey: when you come to try multinode check out harmw Vagrantfile work | 15:53 |
| *** Slower has joined #kolla | 15:53 | |
| *** jtriley has quit IRC | 15:53 | |
| rhallisey | pbourke, kk. I gota get a simple AIO going first | 15:53 |
| pbourke | makes sense | 15:53 |
| rhallisey | stuck dependency limbo | 15:54 |
| pbourke | actually, you could still use the vagrantfile | 15:54 |
| pbourke | and just do AIO on the operator node | 15:54 |
| SamYaple | rhallisey: correct anything works no vm or anything | 15:54 |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Containers with host networking fail to 'su' on some kernels https://review.openstack.org/207925 | 15:54 |
| SamYaple | the docs need updates now that we have keeplaived and haproxy landing | 15:54 |
| rhallisey | SamYaple, have you seen from oslo_i18n import _lazy ImportError: cannot import name _lazy | 15:54 |
| rhallisey | no idea how to get around that | 15:54 |
| SamYaple | rhallisey: old pip? | 15:55 |
| SamYaple | i havent seen it but i suggest pip and wheels | 15:55 |
| SamYaple | pip install -U pip | 15:55 |
| SamYaple | pip install -U wheel | 15:55 |
| SamYaple | seems to be the root of alot of fedora problems | 15:56 |
| rhallisey | SamYaple, are you using fedora or ubuntu? | 15:57 |
| SamYaple | ubuntu | 15:57 |
| SamYaple | with the centos images | 15:57 |
| SamYaple | but i have tested on fedora | 15:57 |
| SamYaple | well F21 | 15:57 |
| pbourke | centos works well as a host also | 15:58 |
| *** coolsvap is now known as coolsvap|away | 16:02 | |
| sdake | morning | 16:10 |
| SamYaple | morning sdake | 16:11 |
| SamYaple | my stuff merged. huge load off | 16:11 |
| *** ashuk has joined #kolla | 16:13 | |
| sdake | xwe got 2 mor erollcall votes ;) | 16:19 |
| sdake | yay victory :) | 16:19 |
| SamYaple | sdake: i posted it in the channler earlier | 16:20 |
| SamYaple | 15:15 < SamYaple> btw did you guys see the big tent application? A comment reads 'Definitely "one of us"' | 16:20 |
| *** unicell has joined #kolla | 16:22 | |
| *** unicell1 has quit IRC | 16:23 | |
| *** ashuk has quit IRC | 16:25 | |
| *** sdake has quit IRC | 16:26 | |
| *** sdake has joined #kolla | 16:37 | |
| *** vinkman has joined #kolla | 16:37 | |
| pbourke | long weekend here in Ireland, gone till Tuesday | 16:39 |
| SamYaple | pbourke: sounds like youll be working on kolla till monday | 16:41 |
| SamYaple | eh | 16:41 |
| SamYaple | eh | 16:41 |
| sdake | beer-o-holiday ;) | 16:41 |
| pbourke | sdake's version might be more accurate | 16:41 |
| SamYaple | so normal day in ireland? | 16:42 |
| rhallisey | is it Arthur day? I celebrate that :) | 16:42 |
| pbourke | ha no they had to cancel that | 16:42 |
| rhallisey | since I love Guinness | 16:42 |
| rhallisey | they canceled that O.o | 16:42 |
| pbourke | it was turning into patricks day 2.0 | 16:42 |
| SamYaple | vbel have you tested https://review.openstack.org/#/c/207925/ across the project? | 16:43 |
| vinkman | pbourke: so what's wrong with that??? Isn't 2.0 better than 1.0 :D | 16:43 |
| bmace | maybe wait for 2.1? :) | 16:44 |
| sdake | hey i heard they run guiness in the streets in pipes - is it true | 16:44 |
| pbourke | lol dont think so | 16:45 |
| bmace | they had it in every pub i went to in england, but i don't think i ever saw anyone order it. | 16:45 |
| sdake | i mean under the streets | 16:45 |
| vinkman | so do I just bring my own tap ? | 16:45 |
| pbourke | we have better beer here now :) | 16:45 |
| sdake | nothing like an irish guiness | 16:46 |
| rhallisey | my favorite | 16:46 |
| vinkman | pbourke: As long as you don't say Bud ;) | 16:46 |
| sdake | even though its a "generic" drink in ireland, there i something about it there | 16:46 |
| sdake | fresher or soething | 16:46 |
| pbourke | trade you guys some guinness for west coast ipa anytime | 16:46 |
| sdake | the us has alot of fine beers no dobut about it | 16:46 |
| *** jasonsb has quit IRC | 16:48 | |
| *** jasonsb has joined #kolla | 16:48 | |
| *** vbel has quit IRC | 16:48 | |
| *** vbel has joined #kolla | 16:49 | |
| *** athomas has quit IRC | 16:53 | |
| *** jasonsb has quit IRC | 16:53 | |
| *** kevsi has quit IRC | 16:57 | |
| sdake | for our core members that may not get budget for meeting in Tokyo summit | 17:06 |
| sdake | there is a travel assistance program that the foundation funds your hotel+airfare | 17:06 |
| sdake | your on your own for food/beer | 17:07 |
| sdake | but its better then not goign :) | 17:07 |
| sdake | https://wiki.openstack.org/wiki/Travel_Support_Program | 17:08 |
| *** achanda has joined #kolla | 17:09 | |
| harmw | sdake: now thats nice! | 17:17 |
| harmw | though Tokyo is no option, it's baby time by then :) more or less | 17:18 |
| sdake | harmw snce your independent i suspect you would be a non brainer for the foundation to support travel wise | 17:18 |
| sdake | well congratulatons! | 17:19 |
| sdake | boy or girl? | 17:19 |
| harmw | boy, okt. 15 is due date | 17:19 |
| sdake | ya hard to abandon the wife during that one ;) | 17:19 |
| harmw | yep | 17:19 |
| harmw | where'll it be held in April, is that known yet? | 17:19 |
| sdake | I had heard barcelona | 17:20 |
| sdake | but not certain | 17:20 |
| sdake | the international travel is putting me in an earlygrave :( | 17:20 |
| harmw | damn, I'd almost say that's to close :p | 17:20 |
| SamYaple | wasnt it supposed to be in Austin again sdake | 17:24 |
| sdake | oh right austin then barcelona | 17:25 |
| harmw | thats like, TX right? | 17:25 |
| SamYaple | yes | 17:25 |
| harmw | well, my English is better than Spanish :p | 17:26 |
| SamYaple | si | 17:26 |
| sdake | we got 6 epics that need to turn into green amchines | 17:27 |
| sdake | (epics = purple in launchpad) | 17:27 |
| sdake | can folks update them now that the erview queue is straightened up | 17:28 |
| daneyon | SamYaple ping re: https://review.openstack.org/#/c/207760/ | 17:28 |
| SamYaple | daneyon: reviewing now | 17:28 |
| SamYaple | daneyon: Does network_agent_name make sense for nova-net | 17:29 |
| daneyon | sdake said the code is not completely working | 17:29 |
| SamYaple | since you can specify openvswitch linubridge or nova-net | 17:29 |
| daneyon | the feedback i provided is general | 17:29 |
| daneyon | none of the agents make sense for nova-net | 17:29 |
| daneyon | i would go with neutron_network_agent_name | 17:30 |
| SamYaple | right but that is the ansible paramater where you would specfic nova-net is my point | 17:30 |
| daneyon | something like network_provider should specify whether to use nova-net or neutron | 17:30 |
| daneyon | or network_manager | 17:30 |
| SamYaple | i cant really do that for the structure of the playbooks | 17:31 |
| sdake | i like network manager since nova networking is a type of manager | 17:31 |
| daneyon | can u bring me up to speed on what's not working with the patch? | 17:31 |
| SamYaple | daneyon: it doesnt work at all for passing traffic. i dont know linuxbridge | 17:31 |
| SamYaple | for OVS i would create br-ex and add the "neutron_interface" | 17:31 |
| SamYaple | for linuxbridge i dont know what that looks like | 17:31 |
| daneyon | passing traffic between vm's on the same host, external traffic or over a vxlan tunnel? | 17:32 |
| SamYaple | daneyon: they dont even get dhcp | 17:32 |
| sdake | pbourke around still? | 17:33 |
| daneyon | ok, if you cerate a neutron net/subnet/router, you should see the dhcp/l3 agents spawn a router/dhcp instance. they should be able to ping one another. Is that the case? | 17:34 |
| SamYaple | daneyon: nope. i see dhcp traffic come out the tap, but i never see it in the namespace | 17:34 |
| SamYaple | dhcp* namespace | 17:34 |
| SamYaple | daneyon: is there no bridge setup requirement like with OVS? | 17:35 |
| SamYaple | ovs-vsctl add-br br-ex; ovs-vsctl add-port br-ex <interface_name> | 17:36 |
| SamYaple | nothing that needs to be done for linuxberidge/ | 17:36 |
| SamYaple | linuxbridge?* | 17:36 |
| daneyon | SamYaple we do need to map the phy int to the LB bridge, but that should not cause the problems you're seeing. | 17:38 |
| daneyon | here is the interface mapping that you will need to perform: https://github.com/stackforge/kolla/blob/stable/kilo/docker/neutron/neutron-agents/config-scripts/config-linuxbridge-agent.sh#L24-L27 | 17:38 |
| daneyon | SamYaple ^ for the LB agent | 17:38 |
| daneyon | It almost sounds like something is going on with the vif plugging | 17:39 |
| *** sdake has quit IRC | 17:39 | |
| SamYaple | daneyon: yea i do that in the conf | 17:40 |
| SamYaple | daneyon: does linuxbridge need a special vif | 17:40 |
| daneyon | i would enable debug on n-compute and neutron server/agents and see if you're getting any ERRORs releated to the vif plugging | 17:40 |
| SamYaple | i am not | 17:40 |
| *** sdake has joined #kolla | 17:40 | |
| daneyon | nope | 17:40 |
| SamYaple | the port comes online all is good | 17:41 |
| SamYaple | i mean like i said, i can tcpdump and see the dhcp traffic on the tap | 17:41 |
| SamYaple | linuxnet_interface_driver = nova.network.linux_net.BridgeInterfaceDriver | 17:42 |
| SamYaple | thats correct, right? | 17:42 |
| daneyon | when u enable debug, can you look at the logs for the details of the conf file that gets loaded. make sure the config that gets loaded matches what is being set by ansible | 17:42 |
| SamYaple | daneyon: i did double check that | 17:42 |
| SamYaple | i had to make some changes, but it sees it all now | 17:42 |
| daneyon | linuxnet_interface_driver = nova.network.linux_net.BridgeInterfaceDriver is correct | 17:43 |
| daneyon | and compute should be using the noop fw driver | 17:43 |
| SamYaple | yup | 17:43 |
| SamYaple | just like OVS | 17:43 |
| daneyon | You may want to try reverting this: https://review.openstack.org/#/c/207760/1/docker/common/neutron/neutron-linuxbridge-agent/start.sh | 17:44 |
| *** sdake has quit IRC | 17:44 | |
| SamYaple | i had to change that to get the options to recognize | 17:45 |
| *** sdake has joined #kolla | 17:45 | |
| SamYaple | it didnt work the other way either | 17:45 |
| daneyon | i thought i saw a bug at some point where neutron did not respect the --config-file /etc/neutron/plugins/ml2/ml2_conf.ini flag | 17:45 |
| sdake | samyaple can I make this one completed https://blueprints.launchpad.net/kolla/+spec/build-script | 17:45 |
| daneyon | ok | 17:45 |
| sdake | you canfile bugs for the additional thingsy ou want to tackle | 17:45 |
| SamYaple | sdake: no i need push | 17:45 |
| sdake | the code is in the code base and afaik nothing else needs reviews | 17:45 |
| sdake | is that going to happe ntoday? | 17:46 |
| sdake | today is deadline for all the things ;) | 17:46 |
| SamYaple | sdake: well then that blueprint isnt done and needs to be pushed back | 17:46 |
| sdake | ok bouncing to l3 then? | 17:46 |
| SamYaple | yea | 17:46 |
| sdake | sounds good | 17:46 |
| *** gfidente has quit IRC | 17:46 | |
| daneyon | SamYaple not sure what else I can say to help you out... sorry. | 17:47 |
| SamYaple | daneyon: no its been a good sanity check | 17:47 |
| SamYaple | its probably a bad config option for me | 17:47 |
| SamYaple | ill hammer on it | 17:47 |
| daneyon | if i were in your shoes, i would just set neutron/nova to deb and trace every step in the path and look for an error or some other anomaly. | 17:48 |
| SamYaple | does the "neutron_interface" ever get pulled into a bridge daneyon? | 17:48 |
| SamYaple | if so that is not happening | 17:49 |
| daneyon | SamYaple neutron_interface as it stands gets pulled into the external LB/OVS bridge. | 17:49 |
| sdake | harmw ping re vagrant | 17:49 |
| SamYaple | daneyon: does the neutron agent take care of that for linuxbridge? because the OVS agent does not handle that | 17:49 |
| harmw | sdake: shoot | 17:49 |
| sdake | the patch is in review | 17:49 |
| sdake | what is its status | 17:49 |
| sdake | is it done and just needs approval? | 17:49 |
| SamYaple | sdake: ITS GOT TABS | 17:50 |
| sdake | keep in mind it doesns't need to be perfect | 17:50 |
| SamYaple | DO NOT MERGE | 17:50 |
| harmw | wait wut | 17:50 |
| sdake | lol its got tabs | 17:50 |
| SamYaple | lol past that the patch looks good | 17:50 |
| daneyon | again, the interface should not cause the problems you are seeing. You can never map a phy interface for ovs or lb and vm's on the same host should get an ip from the dhcp agent and communicate with one another over the internal ovs/lb bridge | 17:50 |
| sdake | july 31 (today) is our deadline for liberty 2 | 17:50 |
| SamYaple | daneyon: thats a good point. | 17:50 |
| harmw | SamYaple: vagrant got tabs? | 17:50 |
| sdake | i am trying to clean up the state of the launchpad tracker | 17:50 |
| SamYaple | harmw: i left a review | 17:50 |
| harmw | jeez | 17:51 |
| harmw | ill take a look | 17:51 |
| harmw | thanks | 17:51 |
| sdake | harmw besides the tabs, is it goo to go? | 17:51 |
| SamYaple | daneyon: ok well let me think about this | 17:51 |
| sdake | harmw please get an untabbed review in the equeue, I'll find a couple cores to review it assuming you think its ready to go | 17:51 |
| daneyon | ok | 17:51 |
| harmw | sdake: it's usefull already | 17:51 |
| sdake | nice | 17:51 |
| sdake | if there are techdebt bugs that need filign please file those for librty-3 | 17:51 |
| SamYaple | daneyon: I appreciate the help! | 17:52 |
| harmw | uhm openstackgerrit ... | 17:53 |
| harmw | wake up please | 17:53 |
| harmw | sdake: SamYaple new Vagrant review is up | 17:54 |
| sdake | harmw are there techdebt bugs that need filing? | 17:54 |
| *** jasonsb has joined #kolla | 17:54 | |
| harmw | there are some TODO's but I'd rather see ppl use it first, Ill make sure anything usefull gets in the LP bugtracker | 17:54 |
| sdake | harmw are the nics created automatically by vagrant? | 17:55 |
| SamYaple | TABS I TELL YOU | 17:55 |
| sdake | (reading the documentation now) | 17:55 |
| harmw | again tabs? | 17:55 |
| SamYaple | tab | 17:55 |
| SamYaple | but yea | 17:55 |
| sdake | harmw you said 3 nics get created, are they created by vagrant? | 17:56 |
| harmw | yep | 17:56 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Vagrantfile and docs https://review.openstack.org/204238 | 17:56 |
| harmw | SamYaple: god I missed that | 17:56 |
| daneyon | SamYaple anytime | 17:56 |
| SamYaple | i found a use case for running both linuxbridge and openvswitch in the same cloud | 17:58 |
| jasonsb | sdake: i spoke with some folks about contributing elk for openstack logs | 17:58 |
| SamYaple | it _appears_ that I may be able to do a migration from linuxbridge-openvswitch live! | 17:58 |
| jasonsb | sdake: got a warm response | 17:58 |
| sdake | jasonsb nice :) | 17:59 |
| SamYaple | jasonsb: thats awesome | 17:59 |
| sdake | harmw there is a -1 on that vagrant review | 17:59 |
| jasonsb | sdake: if we conttributed it it, when would the review have to go up? | 17:59 |
| harmw | sdake: did you refresh? | 17:59 |
| harmw | ah | 17:59 |
| sdake | harmw it is a requet to improve the docs | 17:59 |
| harmw | yea looking at it right away | 17:59 |
| SamYaple | sdake: you just -1 no comment | 17:59 |
| sdake | then we will merge it | 17:59 |
| sdake | I left a comment in review #4 | 17:59 |
| jasonsb | SamYaple: i was happy. i'm sure you have something which would be a nice contribution too | 18:00 |
| sdake | rather review #10 sorry | 18:00 |
| SamYaple | jasonsb: openstack logs done right with no configuration? that alone is enough | 18:00 |
| sdake | harmw check out my review comment in #10 | 18:00 |
| jasonsb | sdake: (IE; when would review have to go up to have a chance of merging for this cycle) | 18:01 |
| sdake | this cycle ends today | 18:01 |
| sdake | then we have l3 | 18:01 |
| sdake | which is august 30 iir | 18:01 |
| sdake | jasonsb one moment | 18:01 |
| harmw | sdake: I would say that is something specific to Vagrant, since Vagrant just creates all the virtual machines to my liking and with the ammounts of RAM and NICS I tell it | 18:02 |
| sdake | jasonsb: https://wiki.openstack.org/wiki/Liberty_Release_Schedule | 18:02 |
| jasonsb | sdake: i think you are going to say review in two weeks | 18:02 |
| jasonsb | to give two weeks to get it through review process | 18:02 |
| sdake | harmw got it, what I want in the docs is for people to know they dont need 3 nics, but vagrant virtualizes that | 18:02 |
| sdake | jasonsb first step - create a blueprint | 18:02 |
| sdake | i'll place in discussion phase | 18:02 |
| sdake | once we have a week of discussion or so, i'll approve blueprint (or any other core reviewer can approve it) | 18:03 |
| sdake | see the schedule above | 18:03 |
| jasonsb | sdake: ok | 18:03 |
| harmw | sdake: ofcourse they don't, thats my point :) Vagrant takes care of that by instructing the desired virtualisation tool to create specific vms and such, I'm not sure if I should put that in the docs since Vagrant in general covers that aspect | 18:03 |
| sdake | the feature freeze is september 4th | 18:03 |
| sdake | that means no new blueprints after september 4th | 18:03 |
| jasonsb | sdake: i dont have a commit yet from my people, but i think its likely to see good support from all parties. | 18:03 |
| sdake | jasonsb cool start with a blueprint | 18:04 |
| jasonsb | sdake: the trick will be in getting it through lots of hands quickly to get approval | 18:04 |
| jasonsb | sdake: you got it | 18:04 |
| sdake | een if you dont do it , someone else may implement it | 18:04 |
| jasonsb | sdake: i have two i owe you | 18:04 |
| harmw | openstack + ELK sounds nice jasonsb | 18:04 |
| sdake | new blueprints after september 4th will be pushed into m | 18:05 |
| sdake | so jasonsb I htink we dont really need elastic search | 18:05 |
| sdake | but we definately want logstash | 18:05 |
| sdake | so this is the type of discussion that would happen in the blueprint during the discussion phase | 18:06 |
| jasonsb | hmm.. we will have to discuss in bp a bit. ours is a little bit variation | 18:06 |
| jasonsb | it uses fluentd | 18:06 |
| sdake | jasonsb if you get er in the blueprint we can discuss live during our next wednesday meeting | 18:07 |
| jasonsb | sdake: you got it | 18:08 |
| sdake | we had talked about using a data container to store all the logs for hte node | 18:08 |
| sdake | i like that idea alot | 18:08 |
| harmw | why not use the docker log driver and a rsyslog container to store it all? | 18:08 |
| SamYaple | sdake: yea the elkstacke container can just do that | 18:08 |
| SamYaple | I have a patch in progress for the logging | 18:08 |
| sdake | we definately need diag+audit sorted out by l3 | 18:08 |
| jasonsb | SamYaple: and have the logs within easy reach of user | 18:09 |
| SamYaple | harmw: we discussed this at the mid-cycle and rejected it | 18:09 |
| SamYaple | I can go over the reasons if youd like | 18:09 |
| SamYaple | jasonsb: yup :) | 18:09 |
| sdake | samyaple might as well store that up for our wednesday meeting ;0 | 18:09 |
| harmw | SamYaple: please do, I'm interested :) | 18:09 |
| sdake | ;) I mean | 18:09 |
| SamYaple | harmw: metting it is! | 18:09 |
| sdake | well if harmw wants to know now feel free to tell him now | 18:09 |
| harmw | sure thing | 18:09 |
| sdake | but you will probably have to repat yourself | 18:09 |
| SamYaple | i should probably build an etherpad for quick consumption | 18:09 |
| sdake | harmw can you add somethign to the docs plz | 18:10 |
| harmw | sdake: working on it | 18:10 |
| sdake | saying that vagrant sets up the eth tuff | 18:10 |
| sdake | cool thanks :) I wasn't sure if you were going to do that or not | 18:10 |
| sdake | jasonsb do you have questions about the above schedule | 18:10 |
| jasonsb | harmw: thank you for kind words. its hard to right. crazy long tail problem | 18:10 |
| sdake | samyaple ya put together an etherpad when your not busy (i.e. tomorrow) | 18:11 |
| sdake | and we can link that in the blueprint that jasonsb will file | 18:11 |
| SamYaple | sdake: lol not busy | 18:11 |
| jasonsb | sdake: yes kind of. how much time would you like to give to review process. | 18:11 |
| jasonsb | sdake: is two weeks enough? | 18:11 |
| sdake | jasonsb who knows - depends on how good the code is | 18:11 |
| sdake | the review process is meant to fix bugs in the code not to bikeshed the design | 18:12 |
| SamYaple | jasonsb: or how small you make the patchsets... | 18:12 |
| sdake | we bikeshed the design during the blueprint process | 18:12 |
| sdake | ya avoiding reviewer fatigue = win :) | 18:12 |
| jasonsb | ok, one step at a time. lets see if you like the ideas first | 18:12 |
| sdake | lets get a solid design - and discuss at next week's meeting | 18:12 |
| sdake | its not if "i" ike the design | 18:13 |
| sdake | its if the team collectively can live with it :) | 18:13 |
| jasonsb | (collective you) | 18:13 |
| jasonsb | sorry for sloppy language | 18:13 |
| SamYaple | jasonsb: no sdake didnt want the configs the way they are | 18:13 |
| SamYaple | community overruled | 18:13 |
| SamYaple | now hes happy | 18:13 |
| sdake | configs the way they are in which way samyaple | 18:14 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Vagrantfile and docs https://review.openstack.org/204238 | 18:14 |
| sdake | you mean bindmount? | 18:14 |
| SamYaple | sdake: you wanted that environment variable this | 18:14 |
| SamYaple | yea | 18:14 |
| sdake | that is because kevinfox foudn a solution that gives us immutability | 18:14 |
| SamYaple | yea i know | 18:14 |
| sdake | I dont care about env vs non env, I cared about immutability | 18:14 |
| SamYaple | its not an attack against you | 18:14 |
| SamYaple | to be fair, i promised that with teh configs... | 18:14 |
| sdake | yup my point is, have to understand the reason - the community didnt overrule me, they came up with a solution to my problem ;) | 18:15 |
| harmw | sdake: there, sattisfied now/ :p | 18:15 |
| sdake | harmw reviewing moment | 18:15 |
| SamYaple | well the commnity overruled and _then_ came up with a solution | 18:16 |
| sdake | harmw maybe I'm dense but I don't see the change | 18:16 |
| sdake | I wrote the spec based upon kevin fox's input - pretty sure I recall the exact ordering :) | 18:16 |
| harmw | don't you compare specific patchsets sdake ? | 18:17 |
| sdake | harmw with my eyeballs | 18:17 |
| sdake | is there another way? | 18:17 |
| sdake | is there a compare button? | 18:17 |
| sdake | if so do teach me sometihing new pls :) | 18:17 |
| SamYaple | sdake: yes | 18:17 |
| SamYaple | you can compare patchsets | 18:17 |
| sdake | how | 18:17 |
| harmw | https://review.openstack.org/#/c/204238/11..12/docs/vagrant.md | 18:17 |
| harmw | there | 18:17 |
| SamYaple | sdake: on a per file basis | 18:17 |
| harmw | thats how I've been going tthrough subsequent reviews since before time :p | 18:17 |
| SamYaple | you can select the patchset at the top of each file | 18:18 |
| sdake | ok one of you will have to walk me through that when I am less busy | 18:18 |
| harmw | sure thing old man :P | 18:18 |
| sdake | but looks like a fantastic feature I need to know about | 18:18 |
| harmw | oh it is :) | 18:18 |
| SamYaple | man how have you survived! and kept up reviews?! | 18:18 |
| SamYaple | im very happy we can share this with you | 18:18 |
| harmw | but the above link'll show you exactly what I added in the latest patchset to the docs | 18:18 |
| *** unicell has quit IRC | 18:19 | |
| sdake | samyaple an you review harm's work pls | 18:19 |
| sdake | I spend couple hours a day doing reviews | 18:19 |
| sdake | and manually inspect each request vs change | 18:19 |
| SamYaple | :O | 18:19 |
| sdake | clearly this seems like a better way | 18:19 |
| sdake | this being what harm just showed me | 18:19 |
| SamYaple | yea thats what were talking about | 18:19 |
| sdake | ok can you review harm's review samyaple plz | 18:20 |
| SamYaple | on it | 18:20 |
| sdake | ta | 18:20 |
| sdake | ok so we need to make all our tags liberty2 | 18:23 |
| sdake | Iguess I'll do that | 18:23 |
| * sdake groans | 18:23 | |
| sdake | first commit after tag will change it to liberty3 | 18:23 |
| sdake | samyaple does that match up with what we discussed yesterday? | 18:23 |
| daneyon | SamYaple sdake I am seeing the following error when starting nova containers: ImportError: No module named cryptography | 18:23 |
| daneyon | are you aware of a patch that fixes this? | 18:23 |
| daneyon | or maybe it's unique to me. | 18:24 |
| SamYaple | daneyon: no but i havent rebuilt today | 18:24 |
| sdake | daneyon did you just rebuild your containers? | 18:24 |
| daneyon | my images were built within the last day | 18:24 |
| sdake | the base image needs that dependency added for rdo | 18:24 |
| daneyon | looks like my nova images were built 46 hours ago | 18:24 |
| sdake | that is a essential bug, can you file it daneyon | 18:24 |
| daneyon | this is my tip: 4d6eaa2 Merge "Tells rabbitmq to listen on single ip" | 18:25 |
| sdake | daneyon it is because delorean added a dependency but it wasn't added to the rpm | 18:25 |
| daneyon | sdake i built the nova images 46 hours ago, from tip: 4d6eaa2 Merge "Tells rabbitmq to listen on single ip" | 18:25 |
| sdake | daneyon yes i just read that :) | 18:25 |
| daneyon | so do a pip install cryptography to the nova-base images, correct? | 18:26 |
| daneyon | s/images/image | 18:26 |
| sdake | i'd put it in centos base | 18:26 |
| sdake | with the othe pip installs | 18:26 |
| sdake | some other package will probably use cryptography | 18:26 |
| sdake | we are int eh process in openstack of deprecating all our crypto code an using some new library | 18:27 |
| sdake | I think cryptography is it | 18:27 |
| daneyon | sdake weird the centos base already has pip installing crypto: ../kolla/docker/centos/binary/nova/nova-base/Dockerfile:RUN pip install cryptography --upgrade | 18:28 |
| sdake | daneyon moment let me look at repo | 18:29 |
| sdake | daneyon pulling nova repo to see what crypto dep is in requirements.txt still downloading | 18:31 |
| sdake | daneyon this change just went in in last couple of days https://github.com/openstack/nova/blob/master/requirements.txt#L15 | 18:34 |
| sdake | they replaced openssl with cryptography library | 18:34 |
| *** jmccarthy has quit IRC | 18:36 | |
| *** jmccarthy has joined #kolla | 18:36 | |
| sdake | jmccarthy ping | 18:38 |
| *** sdake has quit IRC | 18:42 | |
| harmw | SamYaple: | 18:42 |
| harmw | name: Writing hostname of host with existing cluster files to temp file | 18:42 |
| harmw | thats also trying to chown root | 18:42 |
| harmw | which fails in multihost, just like the other one last week | 18:42 |
| SamYaple | harmw: yea but your patch addressed that right? | 18:43 |
| SamYaple | it just hasnt landed | 18:44 |
| harmw | no, mine didn't look at this piece | 18:44 |
| harmw | it's essentially the same thing | 18:44 |
| SamYaple | oh | 18:44 |
| SamYaple | yea its just another copy | 18:44 |
| *** sdake has joined #kolla | 18:44 | |
| harmw | indeed, I should look through all files perhaps for chowns to root | 18:44 |
| SamYaple | harmw: thats the only one | 18:44 |
| harmw | oh ok | 18:44 |
| harmw | hehe | 18:45 |
| sdake | daneyon https://github.com/openstack/nova/blob/master/requirements.txt#L15 | 18:45 |
| SamYaple | thats the only localaction thats used thoughtout the playbooks | 18:45 |
| sdake | daneyon cryptography was just added recently | 18:45 |
| *** jruano has joined #kolla | 18:46 | |
| daneyon | sdake adding the pip install to base | 18:47 |
| daneyon | will test and let u know | 18:47 |
| sdake | harmw can you ack https://review.openstack.org/#/c/205528/ | 18:47 |
| sdake | daneyon nice if youcn file a bug i'd appreciat it as wel (and link inchannel so I can set properties appropriately) | 18:48 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Remove the explicit chown root https://review.openstack.org/208018 | 18:48 |
| harmw | sdake: done | 18:48 |
| sdake | samyaple so is the deal that heat is busted atm even once that eview goes in? | 18:49 |
| sdake | because of domains? | 18:49 |
| SamYaple | sdake: not sure | 18:49 |
| SamYaple | but its not complete | 18:49 |
| SamYaple | it bootstraps and stack lists | 18:49 |
| sdake | can you create a stack? | 18:49 |
| sdake | not you specifically but can it be done | 18:50 |
| sdake | rhallisey ping | 18:50 |
| SamYaple | sdake: idk, test it out | 18:51 |
| sdake | samyaple ack | 18:51 |
| SamYaple | i cant build at the moment because of the breakage | 18:51 |
| sdake | which breakage | 18:51 |
| sdake | the nova one? | 18:51 |
| SamYaple | the base image braeks for me | 18:51 |
| openstackgerrit | Merged stackforge/kolla: Fedora - Install from source for Horizon https://review.openstack.org/207838 | 18:52 |
| sdake | we may have to push our deadline a couple of days as we are entering the weekend | 18:52 |
| sdake | and it sounds like we have several critical shipping issues | 18:52 |
| SamYaple | we do | 18:52 |
| SamYaple | the mariadb bootstrap patch is done but needs testing | 18:52 |
| harmw | feels like last deadline all over again :p | 18:52 |
| SamYaple | cant build | 18:52 |
| SamYaple | harmw: blame non-source builds | 18:52 |
| harmw | yep | 18:53 |
| sdake | I think oru last deadline we were pretty on target ;) | 18:53 |
| openstackgerrit | Merged stackforge/kolla: Fix some spelling and formatting errors in the Cinder doc https://review.openstack.org/207859 | 18:53 |
| sdake | mandre pbourke rhallisey jpeeler around? | 18:54 |
| sdake | need a second review on https://review.openstack.org/#/c/205528/ | 18:54 |
| SamYaple | found breakage | 18:56 |
| SamYaple | gcc not in base | 18:56 |
| sdake | we have 3 purps left | 18:56 |
| SamYaple | Running setup.py install for posix-ipc | 18:56 |
| SamYaple | Complete output from command /usr/bin/python -c "import setuptools, tokenize;__file__='/tmp/pip-build-VPsikn/posix-ipc/setup.py';exec(compile(getattr(tokenize, 'open', open)(__file__).read().replace('\r\n', '\n'), __file__, 'exec'))" install --record /tmp/pip-bViZnu-record/install-record.txt --single-version-externally-managed --compile: | 18:56 |
| *** SamYaple has left #kolla | 18:56 | |
| *** SamYaple has quit IRC | 18:56 | |
| sdake | rofl | 18:56 |
| *** SamYaple has joined #kolla | 18:56 | |
| harmw | omfg :P | 18:57 |
| harmw | awesome | 18:57 |
| SamYaple | sorry | 18:57 |
| SamYaple | i quit for a reason | 18:57 |
| SamYaple | paste.openstack.org/?reply_to=406622 | 18:58 |
| sdake | is this rdo? | 18:59 |
| SamYaple | yea | 18:59 |
| SamYaple | this is centos-binary | 18:59 |
| SamYaple | we cant be pip-upgrading all packages like this | 18:59 |
| daneyon | sdake i cant get crypto to install https://gist.github.com/danehans/fc4ea4f48f7a3b22cd2b | 18:59 |
| sdake | the correect solution is not to use pip in the binary packaging | 19:00 |
| sdake | and use copr to build the packages we need | 19:00 |
| SamYaple | ep | 19:00 |
| sdake | samyaple which package did you add for pip building there that generated taht problem? | 19:01 |
| SamYaple | none | 19:01 |
| sdake | daneyon yoru have the same problem, missing dependencies | 19:01 |
| SamYaple | i just did a fresh build | 19:01 |
| SamYaple | looks like oslo upstream updated versions | 19:02 |
| SamYaple | and thats pulling in stuff | 19:02 |
| sdake | ok lets have a time out and figure out what is needed for the remaining 3 blueprints besides this charlie foxtrot that just presented itself on tag day | 19:03 |
| SamYaple | which three blueprints | 19:03 |
| sdake | https://launchpad.net/kolla/+milestone/liberty-2 | 19:03 |
| sdake | https://blueprints.launchpad.net/kolla/+spec/ansible-service | 19:04 |
| sdake | https://blueprints.launchpad.net/kolla/+spec/ansible-multi | 19:04 |
| sdake | https://blueprints.launchpad.net/kolla/+spec/vagrant-devenv | 19:04 |
| SamYaple | does heat count as core here? | 19:04 |
| sdake | the last one is making it through zuul atm | 19:04 |
| SamYaple | if not thats good to go | 19:04 |
| sdake | no heat does not count as core | 19:05 |
| sdake | we may change that definition later ;) | 19:05 |
| SamYaple | down to vagrant then | 19:05 |
| sdake | core is what we originally defined | 19:05 |
| sdake | so the other two are done? | 19:05 |
| SamYaple | everything i have is merged | 19:05 |
| sdake | can you mark them as such | 19:05 |
| SamYaple | well except bugs | 19:05 |
| SamYaple | already done | 19:05 |
| sdake | yes we can tackle bugs separately | 19:05 |
| sdake | ok bugs next | 19:06 |
| sdake | we have a mountain of bugs | 19:06 |
| sdake | about 20 that I see | 19:06 |
| sdake | lets figure out which ones we absolutely must fix for l3 | 19:06 |
| sdake | rather for l2 | 19:06 |
| sdake | https://bugs.launchpad.net/kolla/+bug/1468779 | 19:07 |
| openstack | Launchpad bug 1468779 in kolla "Need way to select between linuxbridge and OVS" [High,Confirmed] | 19:07 |
| sdake | i'm assigning this to you sam | 19:07 |
| SamYaple | sdake: yup | 19:07 |
| sdake | linuxbridge is not madnatory for l2 | 19:07 |
| sdake | https://bugs.launchpad.net/kolla/+bug/1473062 | 19:08 |
| openstack | Launchpad bug 1473062 in kolla "consoleauth is broken" [Critical,Triaged] | 19:08 |
| sdake | what is consoleauth | 19:08 |
| SamYaple | sdake: provides token auth for accessing console | 19:09 |
| SamYaple | required for console access | 19:09 |
| sdake | via horizon? | 19:09 |
| SamYaple | via web | 19:09 |
| SamYaple | doesnt have to be horizon | 19:09 |
| SamYaple | doesnt require horizon | 19:09 |
| jasonsb | SamYaple: where is your rsyslog configs for all of the services? | 19:09 |
| sdake | ok we should try to fix that - if its even still a problem | 19:09 |
| jasonsb | SamYaple: is it merged yet? | 19:09 |
| SamYaple | jasonsb: it is not, logging is L3 so i have been focusing on closing L2 stuff | 19:10 |
| SamYaple | sdake: i dont know that it is, but i cant confirm atm | 19:10 |
| sdake | I think all other bugs besides 1473062 can go in l3 | 19:10 |
| SamYaple | at one point i got a token | 19:10 |
| jasonsb | SamYaple: is it in git-review? | 19:10 |
| SamYaple | jasonsb: nope | 19:10 |
| jasonsb | SamYaple: okedokey thnx | 19:10 |
| sdake | jasonsb we are totally focused for the next days on tagging l2 | 19:10 |
| sdake | but feel free to file blueprints for l3 ;) | 19:10 |
| SamYaple | but im not using rsyslog, im creating a docker volume and using regular file logging for the openstack services | 19:11 |
| SamYaple | jasonsb: since not all services support rsyslog, but all can log to file thats the most consistent way to go | 19:11 |
| sdake | if we can't get good consensus from the blueprint process on logging we may hav to bit the bullet and do a spec | 19:11 |
| sdake | but i'd prefer not to do so ;) | 19:11 |
| sdake | ok so besides the CF that is the pip install in the binary images, kolla is ready to go? | 19:12 |
| SamYaple | sdake: https://bugs.launchpad.net/kolla/+bug/1479970 | 19:12 |
| openstack | Launchpad bug 1479970 in kolla "Galera bootstrap is broken when starting on multinode" [Critical,Triaged] - Assigned to Sam Yaple (s8m) | 19:12 |
| SamYaple | that is required l2 | 19:12 |
| SamYaple | its basically broken for multinode without that | 19:12 |
| sdake | ok we need that fixed then | 19:12 |
| sdake | consoleauth is lower priority then our pip problem or our galera bootstrap | 19:13 |
| SamYaple | i have the fix, trying to build and test | 19:13 |
| SamYaple | i removed what we were pip installing | 19:13 |
| SamYaple | they may have fixed the packages upstream so we dont need to | 19:13 |
| SamYaple | testing now | 19:13 |
| sdake | if we still have to pip install stuff, I can make copr repos for them | 19:13 |
| sdake | if youc an get me a list samyaple | 19:13 |
| sdake | i'll do the rpm packaging and knock it out today | 19:14 |
| SamYaple | will let oyu know if i find anything | 19:14 |
| SamYaple | the only other important thing we pip install is in barbican | 19:15 |
| SamYaple | and thats known broken | 19:15 |
| sdake | i'm wililng to ship with a busted barbican | 19:15 |
| sdake | ;) | 19:15 |
| SamYaple | for shame! | 19:15 |
| sdake | its just a milestone not a stable relese | 19:15 |
| SamYaple | i really really want trove for Liberty | 19:15 |
| SamYaple | but :/ | 19:15 |
| SamYaple | well everything built | 19:18 |
| SamYaple | time for a test | 19:18 |
| sdake | oslo.service is now available as an rpm | 19:19 |
| SamYaple | i know | 19:19 |
| sdake | oslo.reports is now available as an RPM | 19:19 |
| SamYaple | thats why i remved it | 19:19 |
| sdake | cachetools is available as an rpm | 19:20 |
| SamYaple | looksl ike futurist is also pulledi n properly | 19:20 |
| sdake | furusist | 19:20 |
| sdake | ok well all the shit we install via pip is in the delorean repos now | 19:20 |
| sdake | except for cryptography | 19:20 |
| sdake | so I'll do a copr of cryptography | 19:20 |
| sdake | and then we will be set | 19:20 |
| sdake | sam can you file a bug to remove all that pip stuff that is there atm | 19:21 |
| SamYaple | yup | 19:21 |
| SamYaple | i have the patch | 19:21 |
| sdake | and target it for l2 | 19:21 |
| sdake | nice | 19:21 |
| rhallisey | I'll have to see about autmaton | 19:21 |
| SamYaple | just confirming correctness and booting | 19:21 |
| SamYaple | rhallisey: are you pip installign that? | 19:21 |
| sdake | automaton is in the centos repos | 19:21 |
| SamYaple | yea | 19:21 |
| sdake | rather then delorean repos | 19:21 |
| sdake | so remove automaton too samyaple in your patch plz | 19:22 |
| SamYaple | the only other thing we pip install is supervisord | 19:22 |
| sdake | the only thing we are missing is cryptography | 19:22 |
| SamYaple | i dont know why we do, but im not touching it | 19:22 |
| sdake | oh supervisord | 19:22 |
| SamYaple | we can ignore it for now | 19:22 |
| sdake | we can probably find an rpm of that :) | 19:22 |
| SamYaple | probably | 19:22 |
| sdake | ya if it works with pip lets leae it alone for now | 19:22 |
| sdake | i'll go build an rpm of cryptography | 19:23 |
| sdake | I may drop off for a bit - my network is wierd atm | 19:23 |
| sdake | samyaple is there anything else we pip install besides supervisord (and the need for cryptography) | 19:24 |
| SamYaple | ugh ansible inventory bug | 19:24 |
| SamYaple | not a problem. simple fix | 19:24 |
| SamYaple | sdake: nope | 19:24 |
| SamYaple | not in centos-banry at least | 19:24 |
| sdake | can you grep the codebaes for pip | 19:25 |
| SamYaple | well lots for the source install | 19:25 |
| sdake | I dont have acess atm to my dev system | 19:25 |
| SamYaple | so thats a bad idea | 19:25 |
| sdake | cd to the rdo directory ;) | 19:25 |
| SamYaple | like i said, centos-binary is clean | 19:25 |
| SamYaple | except what we know of | 19:25 |
| sdake | sound sgood | 19:25 |
| sdake | just got another rv+1 | 19:25 |
| sdake | we are allmost there folks :) | 19:25 |
| sdake | an zero complaints | 19:26 |
| sdake | minus the bugs in the review which were fixed | 19:26 |
| SamYaple | https://bugs.launchpad.net/kolla/+bug/1480438 | 19:28 |
| openstack | Launchpad bug 1480438 in kolla "remove unneeded pip installs" [Critical,Triaged] - Assigned to Sam Yaple (s8m) | 19:28 |
| SamYaple | https://bugs.launchpad.net/kolla/+bug/1480439 | 19:28 |
| openstack | Launchpad bug 1480439 in kolla "ansible inventory is missing glance entries" [Critical,Triaged] - Assigned to Sam Yaple (s8m) | 19:28 |
| sdake | ok i am clearing the bugs that hae not been fixed and reassigning to l3 | 19:29 |
| SamYaple | those two i have patches for | 19:29 |
| SamYaple | and are needed for l2 | 19:29 |
| sdake | lets get the patches up sam | 19:29 |
| sdake | and I'll review/ rhallisey can review | 19:29 |
| SamYaple | testing now | 19:29 |
| sdake | ack | 19:29 |
| sdake | nova is busted btw ;) | 19:29 |
| sdake | so you wont totally be able to test | 19:29 |
| sdake | because of the cryptography depepdnecy which is unpackaged | 19:29 |
| SamYaple | works fine for me | 19:29 |
| sdake | it will take 1-2 hours for me to copr it | 19:30 |
| rhallisey | ya I'm having that issue | 19:30 |
| sdake | are you doing --no-cache? | 19:30 |
| SamYaple | sdake: im even rebuilding the base remember | 19:30 |
| SamYaple | let me check | 19:30 |
| sdake | well its deffinately in the code base | 19:30 |
| sdake | because both rhallisey and daneyon can't get nova to run | 19:30 |
| *** bmace has quit IRC | 19:30 | |
| SamYaple | but they had the pip installs in the base | 19:30 |
| sdake | and building cyrptogrpahy via pip requires gcc, so it needs to be rpm'ified | 19:30 |
| SamYaple | but ill confirm | 19:30 |
| rhallisey | sdake, sorry I miss typed. I'm *not* having that issue | 19:31 |
| sdake | cryptography package is in the nova requirements.txt but not in the delorean repo | 19:31 |
| SamYaple | sdake: nope all good | 19:31 |
| sdake | ok before I go do that work | 19:31 |
| sdake | lets get th epatches up and let daneyon rebuild | 19:31 |
| sdake | daneyon how are you building? | 19:31 |
| sdake | build.py? | 19:31 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Removed unneeded pip installs https://review.openstack.org/208036 | 19:32 |
| SamYaple | instances start and can access networking | 19:32 |
| SamYaple | cryptography seems unneeded at this time | 19:32 |
| *** openstack has joined #kolla | 19:35 | |
| -cameron.freenode.net- [freenode-info] if you're at a conference and other people are having trouble connecting, please mention it to staff: http://freenode.net/faq.shtml#gettinghelp | 19:35 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Update inventory examples for Ansible https://review.openstack.org/208039 | 19:36 |
| *** sdake has quit IRC | 19:38 | |
| *** sdake has joined #kolla | 19:39 | |
| sdake | samyaple cn you be pro and pick this up for l2 https://bugs.launchpad.net/kolla/+bug/1476294 | 19:40 |
| openstack | Launchpad bug 1476294 in kolla "openrc needs to be made compatible with ansible deploy" [Critical,Confirmed] - Assigned to Sam Yaple (s8m) | 19:40 |
| sdake | should be pretty straight forward | 19:40 |
| sdake | so I was thinking, daneyon is probbly using compose rather then ansible to launch with | 19:40 |
| sdake | maye compose requires cryptography | 19:40 |
| daneyon | sdake i am using compose | 19:42 |
| sdake | samyaple give compose a spin | 19:42 |
| SamYaple | sdake: ive only used compose a few times | 19:43 |
| SamYaple | i wont be the quickest to test compose | 19:43 |
| sdake | daneyon another thing is you might try rebuilding your image since they are 46 hours old | 19:43 |
| sdake | samypale rebuilt *just now* | 19:43 |
| *** bmace has joined #kolla | 19:43 | |
| sdake | make sure to rebuidl with no-cache option | 19:44 |
| sdake | use build.py its faster | 19:44 |
| SamYaple | ./build.py -T<number_of_parallel_build> --no-cache | 19:44 |
| SamYaple | daneyon: ^ | 19:44 |
| daneyon | OK | 19:44 |
| SamYaple | oh youre using compose | 19:44 |
| SamYaple | ./build.py -T<number_of_parallel_build> --no-cache --type rdo | 19:44 |
| SamYaple | youll need to do that | 19:45 |
| sdake | ok console auth | 19:45 |
| *** echoingumesh has joined #kolla | 19:45 | |
| SamYaple | sdake: dont know testing now | 19:45 |
| sdake | can youtest that samyaple | 19:45 |
| SamYaple | BAM | 19:45 |
| SamYaple | from __future__ import answersdake | 19:45 |
| SamYaple | sdake: well have L3 deadline and ansible 2.0 issues | 19:47 |
| sdake | lots of green -> https://launchpad.net/kolla/+milestone/liberty-2 | 19:47 |
| SamYaple | but we can discuss them after L2 | 19:47 |
| sdake | we can backport bug fixes | 19:47 |
| sdake | into our stable branch | 19:47 |
| SamYaple | yea but itll be major changes | 19:47 |
| sdake | if its required to keep something maintained it is what it is | 19:48 |
| SamYaple | ok | 19:48 |
| SamYaple | fair enough | 19:48 |
| sdake | the only hard rule is no db migrations | 19:48 |
| sdake | since we dont have a db atm that doesn't affect us | 19:48 |
| SamYaple | well that cant stay a requirement | 19:48 |
| SamYaple | since its release indepentandt inside a stabel branch can have db upgrades now | 19:49 |
| SamYaple | not our choice, requirement from upstream project | 19:49 |
| sdake | rhallisey or harmw we need reviews on: https://review.openstack.org/#/c/208036/ | 19:50 |
| bmace | yeah, had the same question harm.. | 19:51 |
| bmace | very odd to just comment out vs. remove unless it is temporary | 19:51 |
| SamYaple | sdake: to handle the openrc i can build a quick playbook that generates an openrc for the localhost | 19:51 |
| sdake | plase review https://review.openstack.org/#/c/208039/ | 19:52 |
| SamYaple | it needs to be discussed if we dump the openrc on each deploy host | 19:52 |
| sdake | bmace I found it odd too | 19:52 |
| sdake | but its not rally critical either way | 19:52 |
| sdake | samyaple what I'd like is for genenv to just generate an openrc.ansible version | 19:52 |
| harmw | hence the +2, I figured we'd want to revert this somewhere in the near future | 19:52 |
| sdake | and we can tackle something better later | 19:52 |
| SamYaple | sdake: genenv has all kinds of crap that ansible will never use | 19:53 |
| SamYaple | genenv-ansible is ok | 19:53 |
| sdake | wfm | 19:53 |
| SamYaple | but genenv is for compose | 19:53 |
| sdake | just do somethign simple for now | 19:53 |
| sdake | so I can cut the release plz :) | 19:53 |
| SamYaple | ok | 19:53 |
| bmace | i can add a link into it from some kolla-ansible command too if you want, and if you want a kolla-ansible that just does start i could have that in for l2 (today) | 19:54 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Removed unneeded pip installs https://review.openstack.org/208036 | 19:54 |
| sdake | harmw yo uwill have to re-review that | 19:55 |
| sdake | bmace that woudl be ok but olike next 1 hr or so? | 19:55 |
| sdake | its getting late core reviewers will start falling off for the weekend soon :) | 19:56 |
| sdake | we can always do that for l3 | 19:56 |
| harmw | it already weekend here :p | 19:56 |
| bmace | as far as start, which works well, i could toss it up now. | 19:56 |
| sdake | cool lets gt what you got in the repo then | 19:56 |
| sdake | you can pop up extra work if you want later | 19:57 |
| sdake | maybe sam could add the generate env to your tool | 19:57 |
| bmace | the ansible stuff does or does not still use ./openrc? i left the check to see if it was there from the kolla command. | 19:57 |
| SamYaple | sdake: i can do a genopenrc script | 19:57 |
| sdake | ansible uses a different openrc format bmace | 19:58 |
| SamYaple | it has to be processed by ansible | 19:58 |
| SamYaple | sdake: no thats not it | 19:59 |
| *** sdake_ has joined #kolla | 20:01 | |
| *** sdake has quit IRC | 20:05 | |
| openstackgerrit | BorneMace proposed stackforge/kolla: Initial kolla-ansible command - deploy support only https://review.openstack.org/208047 | 20:05 |
| bmace | it is a pretty simple call into the current behavior described in the ansible doc, with optional variables for inventory and playbook | 20:06 |
| bmace | a small step up from needing to type it all in by hand. | 20:07 |
| sdake_ | bmace since the job on kolla-ansible likley not to be done by l2, i'll go ahead and leae it in for l3 | 20:07 |
| SamYaple | bmace: any reason not to write this properly in python? | 20:07 |
| SamYaple | this script is going to grow alot | 20:08 |
| SamYaple | if we dont write it in python now we will just rewrite it | 20:08 |
| sdake_ | ya it will be in python at some point | 20:08 |
| sdake_ | howeever i am good with it in shell for this releaes | 20:09 |
| SamYaple | for what we need it to do, it needs to be in python for L3 | 20:09 |
| bmace | i guess it depends on what the plan is for this script. this isn't what i have been working on to manipulate the inventory file, etc. this is just calling into ansible-playbook to perform some operations. the full cli is done in python, of course, w/ cliff, babel, etc. | 20:09 |
| sdake_ | with an immediat eplan to refactor it into python | 20:09 |
| SamYaple | sdake_: thats the wrong call. its going to have to be rewritten before L3 is over | 20:09 |
| sdake_ | what do the arguments do | 20:09 |
| bmace | one arg sets the inventory, the other the playbook, if not passed in it uses the defaults, all-in-one and site.yml | 20:10 |
| sdake_ | yes so this code already exists | 20:10 |
| sdake_ | no reason to rewrite it in thenext few hours | 20:10 |
| *** Hui has joined #kolla | 20:10 | |
| sdake_ | we can do that next week ;) | 20:10 |
| sdake_ | i dont htink bmace has objections to rewriting this code in python | 20:10 |
| sdake_ | i just have objection to rewriting it at last minute in python | 20:10 |
| SamYaple | sdake_: i set the milestone on this for L3 like a week ago though | 20:11 |
| SamYaple | no reason to merge code we wont use | 20:11 |
| sdake_ | yes it is is still l3 | 20:11 |
| bmace | i'm not re-writing this in python right now. if we don't want it, dump it. if we want something so people don't need to call ansible-playbook for l2, keep it | 20:11 |
| sdake_ | i think people will use it | 20:11 |
| SamYaple | dont agree with that decision | 20:12 |
| sdake_ | i see no harm in merging now and rewriting in python later | 20:12 |
| SamYaple | i think we all understand each other, just a difference of opinion | 20:12 |
| sdake_ | bmace are you going to get angry if soemone comes along and rewrites this script in python later? | 20:13 |
| bmace | not in the least :) | 20:13 |
| sdake_ | that would be the only harm I could see happen | 20:13 |
| SamYaple | like i said, we are all on the same page, same information. Just different opinion | 20:13 |
| SamYaple | no problem. we are all adults :) | 20:13 |
| bmace | i'm not attached to the stuff i write. depending on what it is to become do it in Go or Java for all I care :) | 20:13 |
| sdake_ | so samyaple can you live with it then :) | 20:13 |
| SamYaple | yep | 20:13 |
| sdake_ | cool thansk :) | 20:14 |
| bmace | i hate shell scripts, so i like it to be something else, but for how simple it needed to be at present, if it was just calling into ansible-playbook to do things anyway, basing it on the kolla command was the fastest and simplest path. | 20:14 |
| sdake_ | cool I hate typing the million commands to nasible anyway | 20:15 |
| sdake_ | bmace i -1'ed the review, couple things need addressing | 20:15 |
| bmace | ah, yup, sorry, forgot to hit it w/ bashate locally :/ | 20:15 |
| sdake_ | and the argument list | 20:16 |
| sdake_ | so if we can somehow get the environment gen in this script that would be good :) | 20:16 |
| bmace | actually, tox -e bashate didn't complain | 20:17 |
| bmace | and that was the same spacing from the kolla command | 20:17 |
| bmace | maybe it is different for case statement? | 20:17 |
| sdake_ | wfm then as long as it passes the voting gate i'm good :) | 20:18 |
| sdake_ | zuul needs more minerals | 20:20 |
| SamYaple | damn we really need to do the core/contrib thing | 20:21 |
| rhallisey | bmace, O.o I was going to write that nice | 20:22 |
| Hui | @Sam, to run the ansible kolla script, is it mandatory to run it on the target node? | 20:23 |
| SamYaple | Hui: ansible can be run from anywhere that has ssh access to the destination nodes as defined in the inventory | 20:23 |
| SamYaple | so with the right inventory, you dont have to run it on the target node | 20:24 |
| Hui | ansible-playbook -i inventory/tygris2-multi -e @../etc/kolla/defaults.yml -e @../etc/kolla/globals-tygris2.yml -e @../etc/kolla/passwords.yml site-tygris2.yml --tags "mariadb" | 20:24 |
| Hui | this is the command I used to start only the mariadb | 20:24 |
| bmace | anything special you want for arg usage output sdake? something like: Usage: $0 deploy <inventory> <playbook> ok? | 20:24 |
| sdake_ | wfm bmce | 20:25 |
| bmace | kk | 20:25 |
| SamYaple | Hui: right now starting mutli node mariadb is broke. I have a fix in place waiting to test it | 20:25 |
| sdake_ | might put the defaults in there | 20:25 |
| SamYaple | shoudl be merged by end of day Hui | 20:25 |
| sdake_ | eg inventory - defults to all-in-one | 20:25 |
| sdake_ | playbook defaults - to site.yml | 20:25 |
| sdake_ | or something | 20:25 |
| Hui | I changed to the all in one file | 20:26 |
| SamYaple | Hui: for the record, it expects all the files to be present in /etc/kolla/ | 20:26 |
| openstackgerrit | BorneMace proposed stackforge/kolla: Initial kolla-ansible command - deploy support only https://review.openstack.org/208047 | 20:27 |
| sdake_ | need another core review on https://review.openstack.org/#/c/208047/ | 20:28 |
| Hui | Sam: do you mean I can not use relative path? | 20:28 |
| sdake_ | hui it must be in /etc/kolla | 20:29 |
| sdake_ | so right no relative path | 20:29 |
| Hui | let me try again | 20:29 |
| SamYaple | Hui: you can use a relative path for the configs, but there are other configs while running the playbooks that require /etc/kolla/config* | 20:29 |
| SamYaple | so for all intents and purposes, you can't use realative paths | 20:30 |
| daneyon | SamYaple do you see this error when running the build script: https://gist.github.com/danehans/4ed37a196392a524eab6 | 20:36 |
| SamYaple | daneyon: yea gnocchi doesnt build | 20:38 |
| SamYaple | you should get a summary at the end | 20:39 |
| sdake_ | daneyon those are like warnings "xyz didn't build" | 20:39 |
| sdake_ | the way it comes out looks like the world ended | 20:39 |
| sdake_ | we need to tidy up that error message :) | 20:39 |
| daneyon | ok, just wanted to make sure the script will continue on | 20:39 |
| SamYaple | yea jpeeler is working on improving the output daneyon but thats safe to ignore | 20:39 |
| SamYaple | yea it should be "Processing: neutron-linuxbridge-agent" right now | 20:39 |
| SamYaple | i dont log errors, i just spit out the traceback in that initial script | 20:40 |
| SamYaple | jpeeler is doing proper logging | 20:40 |
| sdake_ | actually is doing the marriage thing atm :) | 20:40 |
| SamYaple | i hope he is properly logging | 20:41 |
| *** rhallisey has quit IRC | 20:44 | |
| sdake_ | do we hve any outstanding patches that need reviews? | 20:44 |
| sdake_ | or are we ggood to go once the queue finishes merging | 20:45 |
| SamYaple | sdake_: got patches i havent pushed due to lack of testing | 20:45 |
| SamYaple | one required for l2 | 20:45 |
| sdake_ | this is the db problem? | 20:45 |
| SamYaple | yea | 20:45 |
| sdake_ | do you need any help or just time? | 20:46 |
| SamYaple | time unfortunately | 20:46 |
| sdake_ | how much | 20:46 |
| SamYaple | idk an hour maybe | 20:46 |
| sdake_ | cool | 20:46 |
| SamYaple | https://review.openstack.org/#/c/205528/ | 20:46 |
| *** sdake_ is now known as sdake | 20:46 | |
| SamYaple | thats heat, if you want it in | 20:46 |
| sdake | yes but Ican't review it | 20:46 |
| SamYaple | i know | 20:47 |
| sdake | I'd like it merged | 20:47 |
| sdake | rhallisey is the only core that can merge it atm | 20:47 |
| sdake | and he just dropped off irc | 20:47 |
| SamYaple | pbourke: | 20:47 |
| sdake | pbourke is gone for the day iirc :) | 20:47 |
| sdake | its drink-o-clock in ireland atm | 20:47 |
| SamYaple | anyone have his password? | 20:47 |
| SamYaple | I should really take a day off one of these years | 20:47 |
| *** achanda has quit IRC | 21:08 | |
| *** prad has quit IRC | 21:14 | |
| openstackgerrit | Merged stackforge/kolla: Vagrantfile and docs https://review.openstack.org/204238 | 21:16 |
| daneyon | SamYaple sdake you don't see this error when starting the keystone container: https://gist.github.com/danehans/6df342656bd5f89499ef | 21:16 |
| sdake | victory | 21:17 |
| sdake | daneyon I havn't used compose for a week or two | 21:17 |
| daneyon | ok | 21:17 |
| sdake | so it may be a compose only problem | 21:17 |
| openstackgerrit | Merged stackforge/kolla: Remove the explicit chown root https://review.openstack.org/208018 | 21:17 |
| openstackgerrit | Merged stackforge/kolla: Update inventory examples for Ansible https://review.openstack.org/208039 | 21:17 |
| harmw | I'm out guys, Ill check back in tomorrow to ack stuff :) | 21:17 |
| sdake | i've been busy focusing on ansible | 21:17 |
| sdake | harmw sounds good bro thanks for styaing up :) | 21:18 |
| harmw | yoo | 21:18 |
| sdake | green is good : https://launchpad.net/kolla/+milestone/liberty-2 | 21:18 |
| sdake | 4 more to go | 21:18 |
| SamYaple | daneyon: no but we dont use pki for config-external | 21:19 |
| SamYaple | daneyon: for the record i just rebuilt from scratch on new node and deployed too | 21:19 |
| sdake | samyaple we use pki in compose | 21:19 |
| SamYaple | sdake: i know | 21:19 |
| SamYaple | but thats why i dont see the error, config-external does not | 21:19 |
| sdake | nah the error is abou tconfig_stderr or something | 21:20 |
| sdake | it has nothing to do with pki | 21:20 |
| SamYaple | sdake: the words at the top are Setting up PKI | 21:20 |
| SamYaple | its in the set of pki stage, is it not? | 21:20 |
| SamYaple | oh nevermind | 21:20 |
| SamYaple | i just read the trace | 21:20 |
| SamYaple | probably deperacted/removed option | 21:21 |
| sdake | likely the case | 21:21 |
| sdake | i want our compose code to work when we releae l2 so daneyon if youcould file a bug I'll fix it up | 21:21 |
| *** echoingumesh has quit IRC | 21:30 | |
| *** echoingumesh has joined #kolla | 21:30 | |
| *** superflyy has joined #kolla | 21:31 | |
| *** echoingumesh has quit IRC | 21:32 | |
| *** echoingumesh has joined #kolla | 21:33 | |
| openstackgerrit | BorneMace proposed stackforge/kolla: Initial kolla-ansible command - deploy support only https://review.openstack.org/208047 | 21:40 |
| bmace | this should be far more "normal bash" via use of getopt, arg shifting, etc. | 21:40 |
| sdake | samyaple can you take a look at bmace's patch | 21:43 |
| SamYaple | not atm but i will | 21:44 |
| bmace | guess zuul is getting hammered by lots of people trying to get stuff in for the end of l2? :) | 21:50 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Use sleep during database bootstrap https://review.openstack.org/208066 | 21:51 |
| SamYaple | sorry sdake. got one more L2 bug | 21:53 |
| SamYaple | https://bugs.launchpad.net/kolla/+bug/1480468 | 21:54 |
| openstack | Launchpad bug 1480468 in kolla "race condition when creating services and users in ansible" [Critical,Triaged] - Assigned to Sam Yaple (s8m) | 21:54 |
| *** sdake has quit IRC | 21:54 | |
| *** sdake has joined #kolla | 21:54 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fixes race condition in ansible playbooks https://review.openstack.org/208067 | 21:56 |
| openstackgerrit | Merged stackforge/kolla: Removed unneeded pip installs https://review.openstack.org/208036 | 21:57 |
| *** sdake has quit IRC | 22:02 | |
| SamYaple | alright sdake with all the patches in the queue I can successfully start AIO, multinode, and AIO scaled to multinode | 22:02 |
| echoingumesh | https://blueprints.launchpad.net/kolla/+spec/central-logging-service | 22:04 |
| jasonsb | SamYaple: sdake: bp for elk service ^^ | 22:05 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Use sleep during database bootstrap https://review.openstack.org/208066 | 22:05 |
| *** Hui has quit IRC | 22:06 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Use sleep during database bootstrap https://review.openstack.org/208066 | 22:06 |
| SamYaple | bmace: got the correct bug this time :/ | 22:07 |
| *** Hui has joined #kolla | 22:07 | |
| bmace | nice, other than that the fix looked reasonable :) | 22:07 |
| SamYaple | jasonsb: running into those timing issues for rabbitmq and galera. its fine i can fix them but you called that. I thought I had structured around that issue | 22:17 |
| *** sdake has joined #kolla | 22:25 | |
| ccrouch | SamYaple: can we not do something other than sleep there? | 22:25 |
| ccrouch | https://github.com/stackforge/kolla/blob/master/docker/common/base/kolla-common.sh#L42 | 22:25 |
| SamYaple | ccrouch: we don't use wait_for in config-external | 22:26 |
| SamYaple | i mean not that function, ansible has a wait_for | 22:26 |
| SamYaple | what is your issue with sleep there ccrouch ? | 22:27 |
| sdake | storming badly | 22:27 |
| sdake | must have angered the rain gods ;0 | 22:28 |
| SamYaple | zuul soundsl ike a water god | 22:28 |
| SamYaple | makes sense | 22:28 |
| ccrouch | SamYaple: just a pet-peeve, i don't like using sleep if there is an alternative of polling for the actual desired state to occur | 22:28 |
| ccrouch | that may not be possible here | 22:29 |
| bmace | zuul is the gatekeeper of gozer | 22:29 |
| ccrouch | but it is possible in a lot of places | 22:29 |
| ccrouch | i rewatched ghostbusters a few weeks ago. even better than before :-) | 22:29 |
| SamYaple | ccrouch: yea i am with you on that. i think that function is very dynamic, it will 'wait_for' whatever command you give it to return true | 22:29 |
| jasonsb | SamYaple: what was the timing issue? sorry forgot! | 22:30 |
| bmace | ccrouch definitely the best of the ghostbuster movies. the all female version should be.. interesting. if it is like most remakes, disappointing. star trek as an exception. | 22:30 |
| ccrouch | SamYaple: +1 its very useful, i borrowed it from TripleO where it was a standalone script | 22:30 |
| SamYaple | jasonsb: fresh rabbitmq cluster; start two or more nodes at the same time and they break | 22:30 |
| SamYaple | jasonsb: one at a time (same config) all good | 22:30 |
| ccrouch | bmace: fingers crossed! | 22:30 |
| ccrouch | +1 on startrek reboot, both excellent | 22:31 |
| SamYaple | bmace: its an exception in the fact that it never happened? | 22:31 |
| SamYaple | sure wish they would make another star trek movie... | 22:31 |
| SamYaple | so far weve gotten two star wars movies without lightsabers | 22:31 |
| sdake | another star trek moie is coming but different director | 22:32 |
| SamYaple | simon pegg is writting it so it might be good | 22:32 |
| bmace | SamYaple it is an exception in that it was actually a well made reboot. I think the second was excellent also. Love Cumberbatch. | 22:32 |
| sdake | he does verey well in sherlock - worth watching if your into nerc shows :) | 22:32 |
| SamYaple | bmace: yea im with you. it wasn't bad. but it wasn't star trek imo | 22:32 |
| SamYaple | but hey im a TNG guy | 22:32 |
| bmace | we are all caught up on his sherlock. they need to make more of those :/... 4 eps every 2 or 3 years is painful. | 22:33 |
| *** echoingumesh has quit IRC | 22:33 | |
| SamYaple | 3 eps | 22:33 |
| SamYaple | every "wherever they feel like it" | 22:33 |
| sdake | the best one is where he bets him he cn get him to kill himself ;) | 22:33 |
| bmace | even worse.... it is like "new season, yay!!!" then one afternoon of watching and it is done | 22:33 |
| sdake | with the poison or whatever | 22:33 |
| *** echoingumesh has joined #kolla | 22:33 | |
| SamYaple | sdake: consoleauth looks good to me im closing that bug | 22:34 |
| sdake | one of my stereo channels isnt working | 22:34 |
| sdake | me growls | 22:34 |
| sdake | ya mark it invalid | 22:34 |
| sdake | and untag from release | 22:34 |
| sdake | err mielstone | 22:34 |
| *** echoingu_ has joined #kolla | 22:35 | |
| bmace | SamYaple: i think TNG is optimal star trek as well, but i think the sort of pacing you can do with a series like that just doesn't work for a movie. would love another trek tv series, though everything they have done since tng has sort of sucked.. voyager / ds9 / enterprise. | 22:35 |
| *** echoingumesh has quit IRC | 22:35 | |
| sdake | tng rocks - best episode is where picard gets tortured and says "I saw 5 lights" | 22:35 |
| sdake | been there ;) | 22:36 |
| SamYaple | voyager i liked, ds9 not so much BUT it was pretty cool for a few reasons | 22:36 |
| SamYaple | enterprise, well they started messign with time travel in the first episode thats lazy writing | 22:36 |
| sdake | lol sam | 22:36 |
| sdake | so matter of factly ;) | 22:36 |
| SamYaple | "lets take a story thats already been written, wrap time travel around it and tell it again" - star trek for the past 2 decades | 22:37 |
| jasonsb | SamYaple: oh thats right. do you think its worth looking at the rabbit patch? | 22:37 |
| SamYaple | jasonsb: yea do you still ahve that link? | 22:37 |
| jasonsb | SamYaple: on one hand its non-standard rabbit, on the other i think the user experience is probably better | 22:37 |
| bmace | really, you don't pull out time travel or prime directive breaking in ep1.. just start of mellow, getting to know the characters.. maybe go to an alien planet, pick up some of the locals, sniff some flowers and go crazy. | 22:38 |
| SamYaple | prime directive was always up for debate too. loved that | 22:39 |
| SamYaple | the episode with the proto-vulcan race and picard brings the woman up to the ship and she thinks they are gods | 22:39 |
| SamYaple | great stuff | 22:39 |
| *** jruano has quit IRC | 22:40 | |
| jasonsb | SamYaple: https://github.com/rabbitmq/rabbitmq-clusterer | 22:40 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fixes race condition in ansible playbooks https://review.openstack.org/208067 | 22:40 |
| SamYaple | jasonsb: i wonder if we should just push to rabbit 3.5.4 | 22:41 |
| SamYaple | then no other changes are needed | 22:41 |
| jasonsb | looks like it | 22:42 |
| SamYaple | well it wont happen in L2 | 22:43 |
| jasonsb | this is an interesting thing. on one hand, kolla can deploy (in best practices way) many different versions of openstack (theoretically) | 22:43 |
| SamYaple | well just have to live with rabbit crashing | 22:43 |
| SamYaple | jasonsb: yes! | 22:43 |
| jasonsb | in this rabbit case, hand in hand with best practices is an opinionated install of rabbit | 22:43 |
| SamYaple | if the playbooks dont change, it can deploy everything | 22:43 |
| SamYaple | well jasonsb you can use 3.5.4 with juno or kilo | 22:43 |
| jasonsb | yes, i think you can kill both birds if done right | 22:43 |
| jasonsb | opinionated install of rabbit would be 3.5.4 and clusterer | 22:43 |
| jasonsb | and operators can opt in if this resonates with them | 22:44 |
| SamYaple | since we generate the config file we dont even need clusterer jasonsb | 22:44 |
| jasonsb | if it doesn't then leave as-is | 22:44 |
| sdake | almost done: https://launchpad.net/kolla/+milestone/liberty-2 | 22:44 |
| sdake | one mor eptach | 22:44 |
| daneyon | sdake SamYaple what does your openrc file look like? | 22:44 |
| SamYaple | export OS_PROJECT_DOMAIN_ID=default | 22:44 |
| SamYaple | export OS_USER_DOMAIN_ID=default | 22:44 |
| SamYaple | export OS_PROJECT_NAME=admin | 22:44 |
| SamYaple | export OS_USERNAME=admin | 22:45 |
| SamYaple | export OS_PASSWORD=password | 22:45 |
| SamYaple | export OS_AUTH_URL=http://192.168.31.10:35357 | 22:45 |
| SamYaple | that should be enough daneyon | 22:45 |
| SamYaple | but thats for the config-external. the compose stuff works different i think | 22:45 |
| jasonsb | SamYaple: i think still need clusterer? export RABBITMQ_BOOT_MODULE=rabbit_clusterer | 22:45 |
| jasonsb | SamYaple: and place compiled plugin into . | 22:45 |
| jasonsb | SamYaple: not sure | 22:45 |
| SamYaple | hmm ill test a new rabbit version and get back to you jasonsb | 22:46 |
| jasonsb | SamYaple: neat. i'm curious what you think of it | 22:48 |
| jasonsb | SamYaple: btw, the rabbit guy who suggested is Michael Klishin | 22:48 |
| sdake | daneyon the openrc for ansible is different then for compose | 22:48 |
| jasonsb | SamYaple: he said that rabbit people are pledging to support openstack | 22:48 |
| sdake | samyaple pasted the ansible version | 22:48 |
| jasonsb | SamYaple: so you have a resource if you want | 22:48 |
| daneyon | i am having issues with crus config'ing the keystone user/tenant/ep Trying to figure out the root cause | 22:49 |
| jasonsb | SamYaple: i think they hired a fte solely for this | 22:49 |
| daneyon | I am wondering if crux supports the keystone v3 api | 22:49 |
| SamYaple | jasonsb: awesome! anyone I can reach out too? | 22:49 |
| SamYaple | daneyon: it doesnt | 22:49 |
| daneyon | then i don;t think compose will work | 22:49 |
| daneyon | sdake ^ | 22:49 |
| sdake | why wont ompose work daneyon | 22:51 |
| sdake | it worked a couple weeks ago lst i tested it | 22:51 |
| *** superflyy has quit IRC | 22:52 | |
| SamYaple | jasonsb: oh wow. looking at clusterer more that is fantastic! we can use it much like haproxy where we reload the config on the fly i think | 22:53 |
| daneyon | sdake here is the gist of the error: https://gist.github.com/danehans/3b364989d07458245845 | 22:53 |
| sdake | daneyon you have to use th openrc created by genenv | 22:54 |
| sdake | not the one pasted by samyaple | 22:54 |
| sdake | openrc from genev = config internal | 22:54 |
| SamYaple | sdake: thats inside the keystone container i looks like | 22:54 |
| sdake | openrc that is stored on my harddisk = config external | 22:54 |
| sdake | while v2 keystone is deprecated, it isn't "deleted" yet | 22:55 |
| sdake | so config-internal can still make use of it | 22:55 |
| jasonsb | SamYaple: i've been itching to try clusterer too. I guess pivotal got permission from emc just couple weeks back | 22:55 |
| jasonsb | SamYaple: this is pretty fresh stuff. Michael Klishin said that they have been using internally for a while for the exact case of | 22:55 |
| jasonsb | SamYaple: having a rabbit which you don't have to babysit as much | 22:55 |
| jasonsb | SamYaple: IE: an appliance for instance | 22:56 |
| SamYaple | jasonsb: i just hesistate to move away from this rabbit packaged by the repo because of updates | 22:56 |
| SamYaple | most rabbit updates require a full cluster stop which == downtime | 22:56 |
| jasonsb | SamYaple: this is a good question to pose to rabbit people | 22:56 |
| jasonsb | I would say reach out to Michael Klishin <mklishin@pivotal.io> | 22:56 |
| SamYaple | very good | 22:57 |
| jasonsb | and ask him if rabbit people can hang out on kolla | 22:57 |
| daneyon | maybe it's /usr/sbin/httpd -DFOREGROUND keystone is not starting | 22:57 |
| SamYaple | wgsi might have some deps on V3.... | 22:57 |
| SamYaple | i dont know | 22:57 |
| jasonsb | michael is also active on rabbit-users maillist | 22:58 |
| sdake | daneyon what is the problem exactly yo urun into | 22:59 |
| sdake | keystone worked last time I fixed it | 22:59 |
| daneyon | the keystone container is not starting properly. The container starts but keystone is not running | 23:01 |
| daneyon | logs show keystone_1 | waiting for Keystone @ http://172.29.74.86:35357/v3 | 23:01 |
| jasonsb | SamYaple: i reached out to Michael. give a little bit and see if he replies? | 23:01 |
| SamYaple | jasonsb: ah sure | 23:02 |
| SamYaple | ill save the draft! | 23:02 |
| jasonsb | SamYaple: I asked him if there were rabbit people who could help kolla people | 23:02 |
| daneyon | the keystone db is successfully created and init'd | 23:02 |
| openstackgerrit | Merged stackforge/kolla: Initial kolla-ansible command - deploy support only https://review.openstack.org/208047 | 23:03 |
| daneyon | sdake SamYaple according to the OS docs, we need to configure some apache files: http://docs.openstack.org/kilo/install-guide/install/yum/content/keystone-install.html | 23:04 |
| daneyon | is the ansible deploy config'ing these files? | 23:04 |
| SamYaple | daneyon: yes, but so is config-internal | 23:06 |
| daneyon | SamYaple I just reviewed the /etc/httpd/conf/httpd.conf and I don't see any of the req config | 23:09 |
| daneyon | SamYaple can u point me to where config-internal is configuring the apache files? | 23:11 |
| daneyon | SamYaple... nm I see where it's being config'd in the Dockerfile | 23:13 |
| jasonsb | daneyon: do you have a <VirtualHost *:35357> section in your apache? | 23:15 |
| jasonsb | daneyon: should have a bunch of WSGIXXXX stuff in there | 23:16 |
| daneyon | jasonsb i do | 23:18 |
| daneyon | SamYaple and jsonb, As it relates to the Apache config it appears the keystone dockerfile does not match up completely to the OS docs | 23:20 |
| *** jasonsb_ has joined #kolla | 23:23 | |
| *** jasonsb has quit IRC | 23:23 | |
| SamYaple | daneyon: i dont know about the docs | 23:23 |
| daneyon | jasonsb sdake or SamYaple does keystone start for you w/o having ServerName <IP_ADDRESS> in /etc/httpd/conf/httpd.conf | 23:23 |
| SamYaple | i do know that it worked for sdake at one point | 23:23 |
| SamYaple | daneyon: yes | 23:24 |
| SamYaple | thats justa warning | 23:24 |
| SamYaple | sdake sets that to * i believe | 23:24 |
| daneyon | And keystone starts for you w/o step 4 in the apache section Copy the WSGI components from the upstream repository into this dir | 23:25 |
| daneyon | curl http://git.openstack.org/cgit/openstack/keystone/plain/httpd/keystone.py?h=stable/kilo \ | tee /var/www/cgi-bin/keystone/main /var/www/cgi-bin/keystone/admin | 23:25 |
| SamYaple | daneyon: we do that in teh dockerfile | 23:25 |
| daneyon | SamYaple I don;t see the curl comand in the Dockerfile | 23:26 |
| daneyon | Can you point it out for me? | 23:26 |
| SamYaple | i dont think we curl, we copy it from /usr/share | 23:26 |
| SamYaple | finding line | 23:27 |
| SamYaple | https://github.com/stackforge/kolla/blob/master/docker/centos/binary/keystone/Dockerfile#L15 | 23:27 |
| SamYaple | daneyon: ^ | 23:27 |
| daneyon | i see it | 23:27 |
| bmace | hey SamYaple. right now we have a mix of values being in the /etc/kolla/defaults.yml /etc/kolla/globals.yml and others being in the roles/xxservice/defaults/main.yml and there doesn't seem to be a flow where values in globals / defaults from /etc/kolla override those service defaults? | 23:30 |
| *** daneyon has quit IRC | 23:30 | |
| bmace | like i would sort of expect all default values for each service to be in their role defaults/main.yml and they just override them as necessary in /etc/kolla/globals.yml ? it is sort of a strange mix right now | 23:30 |
| *** daneyon has joined #kolla | 23:31 | |
| *** achanda has joined #kolla | 23:31 | |
| SamYaple | bmace: defaults in the role are the lowlever | 23:34 |
| SamYaple | everything overrides them | 23:34 |
| SamYaple | commandline options (the default/global/password) are the high level | 23:35 |
| SamYaple | everything overrides | 23:35 |
| SamYaple | err rather they override everything else | 23:35 |
| bmace | sure, but for some services some values which are in default / global, for example, are not in service/defaults.main.yml | 23:35 |
| SamYaple | bmace: like what? | 23:35 |
| bmace | nova_api_port for example | 23:36 |
| SamYaple | because defaults.yml is required | 23:36 |
| SamYaple | and we need to reference nova_api_port in several roles | 23:36 |
| SamYaple | same with all other ports | 23:36 |
| bmace | it is still odd to have some service defaults in the global defaults, and some in service specific default files. | 23:36 |
| SamYaple | those arent service defaults | 23:36 |
| SamYaple | ports are global | 23:36 |
| daneyon | seems like i need to specify ServerName controller <IP> in /etc/httpd/conf/httpd.conf for apache to start. | 23:37 |
| SamYaple | nova_api_port is needed by haproxy, neutron, and nova | 23:37 |
| bmace | it is the the nova_api. i appreciate other services need to know it | 23:37 |
| SamYaple | you can inherit variables up bmace | 23:37 |
| SamYaple | cant* | 23:37 |
| *** blahRus has quit IRC | 23:37 | |
| SamYaple | but to be fair, the defaults yaml may be moving to the group_vars/all.yml | 23:38 |
| SamYaple | i just need to be sure inheritance still works | 23:38 |
| bmace | i guess the programmatic term would be scope, so no service can see data scoped to roles other than their own, so only data from globals / defaults can be seen across services | 23:38 |
| SamYaple | daneyon: strange, is that an httpd config option you can tweak? | 23:38 |
| SamYaple | bmace: this is scoped. they are referenced at the lowest scope they are needed | 23:39 |
| SamYaple | in this case thats toplevel | 23:39 |
| daneyon | yes, by default it's commented-out | 23:39 |
| SamYaple | could that default have changed in a package version daneyon ? | 23:39 |
| bmace | i am good with breaking the vars out of a single massive env file like we had it for the docker stuff, but having them in a format that can be more easily processed separate from the jinja stuff in the yml files and in common area would be excellent, imho | 23:40 |
| bmace | err.. docker stuff meaning compose | 23:40 |
| SamYaple | bmace: well if you can find a away to do it without duplicating the variable in ansible you go for it | 23:41 |
| bmace | the way it is right now, allowing programmatic control over those values is a little obnoxious. | 23:41 |
| SamYaple | bmace: anything in globals.yml will override everything else | 23:41 |
| SamYaple | perioud | 23:41 |
| sdake | daneyon yes i read thoe docs an applied them when I did the keystone conversion to httpd | 23:42 |
| sdake | samyaple I think the only thing remaining is https://bugs.launchpad.net/kolla/+bug/1476294 - any chance you will wrap up soon on that? | 23:45 |
| openstack | Launchpad bug 1476294 in kolla "openrc needs to be made compatible with ansible deploy" [Critical,Confirmed] - Assigned to Sam Yaple (s8m) | 23:45 |
| *** daneyon has quit IRC | 23:46 | |
| SamYaple | sdake: i said before i cant just write a script for that | 23:46 |
| SamYaple | it has to go through ansible | 23:46 |
| SamYaple | since ansible has all the info | 23:46 |
| sdake | so ansible is going to write out the openrc then? | 23:46 |
| SamYaple | yea, but to where | 23:46 |
| sdake | cwd I guess | 23:47 |
| SamYaple | doesnt work like that | 23:47 |
| sdake | pass in path? | 23:47 |
| SamYaple | where is the teplate going to go? | 23:47 |
| SamYaple | template* | 23:48 |
| sdake | hel if I know make something up ;) | 23:48 |
| SamYaple | i dont want to throw this in. this needs to be thoughtout | 23:48 |
| SamYaple | this is your passwords | 23:49 |
| sdake | ok then put the defualt one in that we create in the home dir | 23:49 |
| sdake | i'd be good with that | 23:49 |
| sdake | with some instructions that it needs customization | 23:49 |
| SamYaple | what do yo mean? | 23:49 |
| sdake | i mean the one youpasted the one that is on my harddisk somewhere just put that in the repo | 23:49 |
| sdake | some kind of documentation or file or something that people can get rolling with | 23:50 |
| SamYaple | ok | 23:50 |
| sdake | this is l2, doen't hafe to be perfect just hsa to work ;) | 23:50 |
| SamYaple | like tools/openrc-example ? | 23:50 |
| sdake | we have an entire 2 months to gold place this part | 23:50 |
| sdake | wfm | 23:50 |
| SamYaple | sounds good | 23:50 |
| jasonsb_ | sdake: bp is up for logging service (if you would like to put on agenda for weekly) | 23:56 |
| *** sdake has quit IRC | 23:57 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Adds an example openrc file https://review.openstack.org/208097 | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!