| *** sdake_ has joined #kolla | 00:07 | |
| *** sdake has quit IRC | 00:11 | |
| *** jruano has joined #kolla | 00:14 | |
| *** sdake has joined #kolla | 00:18 | |
| *** sdake_ has quit IRC | 00:21 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: WIP: Start tgtd in cinder-volume container https://review.openstack.org/200139 | 00:22 |
|---|---|---|
| *** mfalatic has quit IRC | 00:26 | |
| *** diga has quit IRC | 00:37 | |
| *** dims has quit IRC | 00:40 | |
| *** jasonsb has quit IRC | 00:48 | |
| *** dolpher has joined #kolla | 01:04 | |
| *** dolpher1 has joined #kolla | 01:18 | |
| *** dolpher has quit IRC | 01:21 | |
| *** erkules_ has joined #kolla | 01:22 | |
| *** dims has joined #kolla | 01:24 | |
| *** erkules has quit IRC | 01:25 | |
| *** dims has quit IRC | 01:25 | |
| *** dims has joined #kolla | 01:26 | |
| *** diogogmt has joined #kolla | 01:26 | |
| *** sdake_ has joined #kolla | 01:29 | |
| *** sdake has quit IRC | 01:32 | |
| *** dims has quit IRC | 01:36 | |
| *** dims has joined #kolla | 01:37 | |
| *** sdake has joined #kolla | 01:38 | |
| *** tobe has joined #kolla | 01:39 | |
| *** sdake_ has quit IRC | 01:41 | |
| *** dims has quit IRC | 01:42 | |
| sdake | anyone around that has launched a vm with nova lately? | 02:03 |
| sdake | getting 2015-07-10 01:43:46.234 1 ERROR nova.compute.manager [instance: ce6bfd6c-a160-4302-bb99-70c9b8e2e317] libvirtError: Failed to create controller cpu for group: No such file or directory | 02:03 |
| *** vinkman has joined #kolla | 02:09 | |
| *** jasonsb has joined #kolla | 02:20 | |
| *** achanda has joined #kolla | 02:23 | |
| sdake | would appreciate if someone could confirm by either reproducing from a build of master or indicate whether their version works... Make sure to do a clean build. | 02:26 |
| sdake | https://bugs.launchpad.net/kolla/+bug/1473270 | 02:26 |
| openstack | Launchpad bug 1473270 in kolla "compute container prints exception libvirtError: Failed to create controller cpu for group: No such file or directory" [Critical,Triaged] - Assigned to Steven Dake (sdake) | 02:26 |
| *** dims has joined #kolla | 02:38 | |
| *** jruano has quit IRC | 02:39 | |
| *** achanda has quit IRC | 02:42 | |
| *** dims has quit IRC | 02:43 | |
| sdake | samyaple ping me when you arrive pls ;) | 02:55 |
| *** diogogmt has quit IRC | 03:08 | |
| *** jruano has joined #kolla | 03:30 | |
| *** gfidente has quit IRC | 04:01 | |
| *** jruano has quit IRC | 04:07 | |
| *** achanda has joined #kolla | 04:20 | |
| *** tobe has quit IRC | 04:44 | |
| *** sdake_ has joined #kolla | 04:58 | |
| *** sdake has quit IRC | 05:00 | |
| *** sdake has joined #kolla | 05:00 | |
| *** dims has joined #kolla | 05:03 | |
| *** sdake_ has quit IRC | 05:03 | |
| *** dims_ has joined #kolla | 05:04 | |
| *** dims has quit IRC | 05:08 | |
| *** dims_ has quit IRC | 05:08 | |
| *** vinkman has quit IRC | 05:19 | |
| *** Slower has joined #kolla | 05:38 | |
| harmw | it sucks $office is requiring a proxy | 05:48 |
| harmw | makes building containers impossible | 05:48 |
| *** dims has joined #kolla | 06:01 | |
| *** dims has quit IRC | 06:09 | |
| *** achanda has quit IRC | 06:11 | |
| *** dims has joined #kolla | 06:13 | |
| *** dims_ has joined #kolla | 06:14 | |
| *** erkules_ is now known as erkules | 06:15 | |
| *** erkules has joined #kolla | 06:15 | |
| harmw | pbourke: gen-source-tar.sh is missing +x | 06:16 |
| *** dims has quit IRC | 06:17 | |
| *** dims_ has quit IRC | 06:21 | |
| *** dims has joined #kolla | 06:22 | |
| *** dims has quit IRC | 06:27 | |
| *** inc0 has joined #kolla | 06:29 | |
| inc0 | good morning | 06:30 |
| harmw | https://github.com/docker/docker/pull/9176#issuecomment-118222682 there, building now behind a proxy :) | 06:32 |
| inc0 | yes please | 06:34 |
| harmw | building all containers now, looks fine so far :) | 06:34 |
| *** tobe has joined #kolla | 06:45 | |
| *** shardy has joined #kolla | 07:09 | |
| harmw | done | 07:14 |
| harmw | failed: magnum-conductor, barbican, swift | 07:16 |
| harmw | The command '/bin/sh -c cd /tmp && curl -L https://github.com/GoogleCloudPlatform/kubernetes/releases/download/v0.15.0/kubernetes.tar.gz -o /tmp/kubernetes.tar.gz && /usr/bin/tar -xzvf /tmp/kubernetes.tar.gz && cp -a /tmp/kubernetes/platforms/linux/amd64/kubectl /usr/bin/kubectl && rm -rf /tmp/kubernetes' returned a non-zero code: 35 | 07:20 |
| harmw | why is there even a -o there :/ | 07:20 |
| vincent_vdk | morning | 07:22 |
| *** jmccarthy has joined #kolla | 07:23 | |
| *** SamYaple has quit IRC | 07:29 | |
| harmw | building magnum now, https_proxy wasn't set properly | 07:29 |
| harmw | morning vincent_vdk | 07:29 |
| *** shardy_ has joined #kolla | 07:33 | |
| harmw | and it's done, great | 07:34 |
| *** shardy has quit IRC | 07:35 | |
| harmw | uhm, why is tools/kolla requiring root again? | 07:37 |
| *** shardy_ has quit IRC | 07:38 | |
| *** shardy has joined #kolla | 07:39 | |
| vincent_vdk | so i added python-oslo-reports to the cinder-base yum install line and rebuild | 07:41 |
| vincent_vdk | but still geting ImportError: No module named oslo_reports | 07:42 |
| harmw | isn't that in the base-base? | 07:46 |
| harmw | and perhaps --no-cache is needed | 07:46 |
| vincent_vdk | i'll add it to base-base and see what happens | 07:51 |
| harmw | it's not litteraly base-base | 07:51 |
| harmw | sry | 07:51 |
| harmw | it's centos-base or something similar | 07:52 |
| vincent_vdk | i have docker/centos/rdo/base | 07:52 |
| harmw | just check whatever cinder-base has for FROM | 07:52 |
| harmw | yep, ok | 07:52 |
| vincent_vdk | i added it to cinder-base | 07:52 |
| vincent_vdk | i'll add it to base-base now | 07:52 |
| vincent_vdk | hmm, oslo.service comes from pip | 07:53 |
| vincent_vdk | all the others from rpm | 07:53 |
| harmw | it got rebranded to oslo_service recently | 07:53 |
| harmw | *everything oslo | 07:53 |
| vincent_vdk | so base should have everything then | 07:53 |
| vincent_vdk | ah, No package python-oslo-reports available | 07:56 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fixes an issue with AIO and galera clusters https://review.openstack.org/199406 | 08:02 |
| harmw | heat looks broken: ERROR: openstack 'NoneType' object does not support item assignment | 08:02 |
| harmw | not sure why yet | 08:02 |
| openstackgerrit | Merged stackforge/kolla: Add source keystone image for oraclelinux https://review.openstack.org/191833 | 08:05 |
| *** jmccarthy1 has joined #kolla | 08:08 | |
| *** jmccarthy has quit IRC | 08:08 | |
| *** jmccarthy1 has quit IRC | 08:08 | |
| *** jmccarthy has joined #kolla | 08:08 | |
| harmw | sdake: nova is stuck in SPAWNING | 08:25 |
| *** shadower has joined #kolla | 08:25 | |
| *** coolsvap|afk is now known as coolsvap|away | 08:26 | |
| *** coolsvap|away is now known as coolsvap | 08:28 | |
| harmw | "Build of instance 26fdc82a-5ca1-4b51-a781-b69af030d54f aborted: Failed to allocate the network(s), not rescheduling." | 08:28 |
| *** gfidente has joined #kolla | 08:37 | |
| *** SamYaple has joined #kolla | 08:39 | |
| SamYaple | hola | 08:40 |
| *** dims has joined #kolla | 08:40 | |
| jmccarthy | o/ | 08:41 |
| *** dims has quit IRC | 08:45 | |
| inc0 | SamYaple, hello, tell me plz, where do you register endpoint in glance | 08:49 |
| inc0 | still in container? | 08:49 |
| inc0 | or it'll be in register.yml? | 08:49 |
| SamYaple | GARRAGAGARAR | 08:50 |
| SamYaple | i dont yet. | 08:50 |
| SamYaple | working on it | 08:50 |
| SamYaple | https://review.openstack.org/#/c/199463/ | 08:50 |
| inc0 | no worries, I'm just looking for a place for VIPs | 08:51 |
| inc0 | ahh, you're making an module! | 08:51 |
| SamYaple | unfortunately :(* | 08:52 |
| inc0 | yeah, I know, I'll look for haproxy module then | 08:52 |
| inc0 | maybe there is something like that already | 08:52 |
| inc0 | http://docs.ansible.com/haproxy_module.html there is and it's in main tree! | 08:53 |
| SamYaple | wait. why should we need an haproxy moule | 08:53 |
| inc0 | I was thinking on changing our haproxy vips dynamically | 08:53 |
| SamYaple | you cant do that with haproxy | 08:54 |
| SamYaple | requires a config reload | 08:54 |
| SamYaple | the socket is super limited | 08:54 |
| inc0 | yes, I know | 08:54 |
| SamYaple | the haproxy module only talks to the socket though | 08:54 |
| inc0 | but we don't want to make haproxy "fire and forget" | 08:54 |
| SamYaple | didnt we have this conversation yesterday? | 08:55 |
| inc0 | ah so you can't add backend dynamically? | 08:55 |
| SamYaple | right | 08:55 |
| inc0 | forget about that then, gonna play with config | 08:55 |
| SamYaple | here is my plan | 08:55 |
| inc0 | go on | 08:55 |
| SamYaple | drop confs in conf.d per service. docker exec haproxy to trigger a "config update" script which will pull in the new configs. HUP the process to reread the configs (without killing existing connections) | 08:56 |
| SamYaple | not sure we can HUP the pid 1 in docker without crashing the contaienr, so we may need supervisord or something | 08:56 |
| inc0 | I'll figure it out | 08:57 |
| vincent_vdk | you can use consul and registrator to dynamically change the HA config | 09:04 |
| SamYaple | no to consul | 09:04 |
| vincent_vdk | why not? | 09:05 |
| SamYaple | it wont do exactly what we want, but more importantly, no external dependacies or additional services like that | 09:05 |
| vincent_vdk | what are no external dependacies or additional services | 09:06 |
| SamYaple | let me change the question vincent_vdk, where would consul run? | 09:07 |
| vincent_vdk | in a docker container | 09:07 |
| vincent_vdk | i used it for a project | 09:08 |
| SamYaple | that is an additional service unrelated to deploying openstack | 09:08 |
| vincent_vdk | HAproxy is one too then. or not? | 09:08 |
| SamYaple | no, because that is _required_ to get it work HA | 09:09 |
| SamYaple | consul is not | 09:09 |
| SamYaple | we have gone over this in the community before, and the community shot down consul. You can propose it again if you would like | 09:09 |
| *** jmccarthy has quit IRC | 09:10 | |
| *** jmccarthy has joined #kolla | 09:10 | |
| *** athomas has joined #kolla | 09:12 | |
| vincent_vdk | it would solve your issue with static HA config though | 09:15 |
| vincent_vdk | but nevermind | 09:16 |
| harmw | he, lets deploy on top of mesos while at it! :> | 09:16 |
| SamYaple | vincent_vdk: what issue? | 09:16 |
| vincent_vdk | SamYaple: nevermind | 09:17 |
| vincent_vdk | i'm just strugeling with an HA idea for our own use here | 09:17 |
| *** dolpher1 has quit IRC | 09:25 | |
| vincent_vdk | but to give you an idea http://blog.xebia.com/2015/03/24/a-high-available-docker-container-platform-using-coreos-and-consul/ | 09:33 |
| harmw | that stuff is more microservice-oriented, not neccesarily something like Openstack | 09:34 |
| SamYaple | vincent_vdk: oh i know how consul works. I happen to not like it | 09:34 |
| harmw | would be fun to make up something using consul and the likes though | 09:35 |
| SamYaple | i mean, the containers can be used standalone/with a different deploy method | 09:36 |
| SamYaple | thats kinda the point of this whole project | 09:36 |
| vincent_vdk | well, one could take kolla and build consul around it for its own usage | 09:39 |
| vincent_vdk | so the current idea is to deploy a set with ansible and add these to the HAproxy config | 09:41 |
| SamYaple | vincent_vdk: its not one large haproxy config file, its a config file per service, but yes | 09:41 |
| vincent_vdk | ok, that matches the current TripleO deployment more | 09:43 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Fixes an issue with AIO and galera clusters https://review.openstack.org/199406 | 09:53 |
| harmw | is there a bug targetting magnum not having +x on those config- scripts? | 09:59 |
| SamYaple | harmw: only start.sh and config-internal.sh need +x | 10:07 |
| harmw | fine, but they currently lack that :) | 10:07 |
| harmw | just like Horizon, but there's a merge pending on that | 10:08 |
| SamYaple | no idea then | 10:08 |
| harmw | ERROR (InstanceInErrorState): Build of instance f343e2a6-e3c0-445b-949b-6f52b5f5a254 aborted: Failed to allocate the network(s), not rescheduling. | 10:08 |
| harmw | that sucks however | 10:08 |
| harmw | booting a new instance on a fresh kolla cloud | 10:08 |
| *** jmccarthy has quit IRC | 10:11 | |
| *** jmccarthy has joined #kolla | 10:11 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add temporary Ansible keystone modules https://review.openstack.org/199463 | 10:16 |
| SamYaple | harmw: would you mind https://review.openstack.org/#/c/199406/ | 10:18 |
| SamYaple | had to rebase | 10:18 |
| SamYaple | wait pbourke why did you commit to my patchset | 10:19 |
| SamYaple | pbourke: please dont do commit to other peoples patchsets unless you have discussed it with them | 10:21 |
| pbourke | SamYaple: I just used the rebase button | 10:21 |
| pbourke | SamYaple: didn't push any code | 10:21 |
| SamYaple | yea but i had already rebased | 10:21 |
| SamYaple | and it screws with stats | 10:22 |
| pbourke | sorry | 10:22 |
| SamYaple | its cool, just dont mess with patchsets that arent yours without tlaking to the owner | 10:23 |
| *** dims has joined #kolla | 10:28 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fixes an issue with AIO and galera clusters https://review.openstack.org/199406 | 10:29 |
| *** dims_ has joined #kolla | 10:29 | |
| SamYaple | pbourke: that ended up being helpful. made me catch a mistake | 10:30 |
| pbourke | SamYaple: that was my intention all along ;) | 10:30 |
| SamYaple | mwahaha | 10:31 |
| *** dims has quit IRC | 10:33 | |
| *** dims_ has quit IRC | 10:34 | |
| *** coolsvap is now known as coolsvap|away | 10:43 | |
| harmw | interesting, docker-compose -f compose/horizon.yml logs | 11:08 |
| harmw | is showing me all services | 11:08 |
| harmw | instead of only horizon | 11:08 |
| *** jmccarthy has quit IRC | 11:08 | |
| *** jmccarthy has joined #kolla | 11:08 | |
| *** rhallisey has joined #kolla | 11:14 | |
| *** tobe has quit IRC | 11:15 | |
| *** tobe has joined #kolla | 11:16 | |
| harmw | pbourke: you remember what was up with this error? NovaException: Unexpected vif_type=binding_failed | 11:16 |
| pbourke | harmw: yeah that one is a pain | 11:16 |
| pbourke | harmw: I cant really remember the specifics though :( I think there's a number of reasons that can cause it, anything to do with neutron not been able to succesfully create the virtual interfaces | 11:18 |
| harmw | all I remember is you caught it once and i hinted at a solution :p just can't remember the solution anymore | 11:19 |
| pbourke | I have something in my notes around nova-server not registering mech drivers | 11:20 |
| harmw | ah | 11:20 |
| harmw | EUTRON_FLAT_NETWORK_INTERFACE=eth1 | 11:20 |
| harmw | that's a problem | 11:20 |
| pbourke | yeah if that doesn't work make sure nova is loading the l2 mechanism drivers correctly | 11:21 |
| *** gfidente has quit IRC | 11:25 | |
| *** tobe has quit IRC | 11:45 | |
| *** slagle has joined #kolla | 11:45 | |
| SamYaple | i know that error harmw! i have done lots of answers for that | 11:46 |
| harmw | the solution is up there | 11:47 |
| harmw | :) | 11:47 |
| harmw | (eth1 doesn't exist) | 11:47 |
| harmw | genenv asumed it did, or I didn't bother to change it :p | 11:47 |
| SamYaple | yea eth1 not existing is the solution to fix the environment | 11:48 |
| SamYaple | but that error is alot more complicated than that | 11:48 |
| harmw | oh yes | 11:48 |
| harmw | I've seen it numerous times | 11:48 |
| harmw | with different causes | 11:48 |
| SamYaple | anyway its the source of lots of networking problems because neutron uses subprocess to wrap commands rather than useful python libraries | 11:48 |
| *** prad has joined #kolla | 11:55 | |
| SamYaple | so crux only supports keystone v2.0 which explains why i wasn't able to use it in the container | 12:04 |
| harmw | wicked, instance is running again sdake | 12:06 |
| harmw | Ill file a bug+fix tonight | 12:07 |
| SamYaple | VICTORY | 12:08 |
| *** absubram has joined #kolla | 12:08 | |
| SamYaple | database rabbit keystone glance cominug up and registering themselves | 12:08 |
| SamYaple | idempotently as well | 12:09 |
| harmw | cool stuff Sam | 12:09 |
| *** absubram has quit IRC | 12:10 | |
| *** absubram has joined #kolla | 12:10 | |
| *** dims has joined #kolla | 12:15 | |
| *** dims has quit IRC | 12:20 | |
| *** dwalsh has joined #kolla | 12:27 | |
| *** jruano has joined #kolla | 12:29 | |
| *** ccrouch has joined #kolla | 12:37 | |
| *** diogogmt has joined #kolla | 12:39 | |
| *** dims has joined #kolla | 12:41 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add temporary Ansible keystone modules https://review.openstack.org/199463 | 12:42 |
| *** absubram has quit IRC | 12:43 | |
| *** coolsvap|away has quit IRC | 12:47 | |
| *** shadower has quit IRC | 12:48 | |
| *** sdake_ has joined #kolla | 12:50 | |
| *** coolsvap|away has joined #kolla | 12:51 | |
| *** coolsvap|away is now known as coolsvap | 12:53 | |
| *** sdake has quit IRC | 12:54 | |
| *** diogogmt has quit IRC | 13:01 | |
| *** diogogmt has joined #kolla | 13:04 | |
| *** gfidente has joined #kolla | 13:05 | |
| *** gfidente has joined #kolla | 13:05 | |
| *** inc0 has quit IRC | 13:07 | |
| *** thumpba has joined #kolla | 13:09 | |
| *** jmccarthy has quit IRC | 13:10 | |
| *** jmccarthy has joined #kolla | 13:11 | |
| *** athomas has quit IRC | 13:13 | |
| *** athomas has joined #kolla | 13:20 | |
| *** rhallisey has quit IRC | 13:20 | |
| *** rhallisey has joined #kolla | 13:23 | |
| *** shardy_ has joined #kolla | 13:37 | |
| *** shardy has quit IRC | 13:39 | |
| *** dolpher has joined #kolla | 13:39 | |
| *** shardy_ has quit IRC | 13:42 | |
| *** shardy has joined #kolla | 13:43 | |
| *** sdake_ has quit IRC | 13:51 | |
| *** dolpher1 has joined #kolla | 14:04 | |
| *** dolpher has quit IRC | 14:06 | |
| *** dolpher has joined #kolla | 14:09 | |
| *** jmccarthy has quit IRC | 14:10 | |
| *** jmccarthy has joined #kolla | 14:10 | |
| *** dolpher1 has quit IRC | 14:10 | |
| *** pbourke has quit IRC | 14:20 | |
| *** pbourke has joined #kolla | 14:21 | |
| *** thumpba has quit IRC | 14:39 | |
| *** dims has quit IRC | 14:40 | |
| *** jtriley has joined #kolla | 14:43 | |
| *** jtriley has quit IRC | 14:43 | |
| *** jtriley has joined #kolla | 14:43 | |
| *** jtriley has quit IRC | 14:43 | |
| *** jtriley has joined #kolla | 14:43 | |
| *** coolsvap is now known as coolsvap|away | 14:50 | |
| *** britthouser has quit IRC | 14:53 | |
| *** inc0 has joined #kolla | 15:06 | |
| *** dims has joined #kolla | 15:07 | |
| *** dims_ has joined #kolla | 15:07 | |
| *** dims has quit IRC | 15:11 | |
| inc0 | guys, especially SamYaple, you think keystone should put its haproxy to haproxy | 15:14 |
| inc0 | haproxy's conf.d? or haproxy should do that? | 15:14 |
| inc0 | if former, how do we make decision if to put it there alltogether, and if latter, how to provide full list of services to be configured? | 15:15 |
| inc0 | I'm talking about ansible. | 15:15 |
| inc0 | thoughts? | 15:16 |
| *** sdake has joined #kolla | 15:20 | |
| sdake | harmw which bug did you find | 15:21 |
| SamYaple | inc0: what are you asking again? | 15:22 |
| SamYaple | should keystone have a seperate conf.d file, or should it existi nthe haproxy conf? | 15:22 |
| inc0 | no, there will be haproxy conf.d | 15:22 |
| sdake | rhallisey ping | 15:23 |
| rhallisey | sdake, hey | 15:23 |
| inc0 | and if we run ha-setup, keystone on deployment will add its set of endpoints to this folder (/opt/kolla/haproxy/conf.d) | 15:23 |
| sdake | is anyone able to get instanscas into the running state | 15:23 |
| inc0 | and call hup on haproxy container | 15:23 |
| sdake | rhallisey why does cinder-api require python-eystone | 15:23 |
| SamYaple | sdake: for authentication? | 15:24 |
| rhallisey | let me find the bug.. | 15:24 |
| inc0 | that's my idea | 15:24 |
| *** shardy_ has joined #kolla | 15:24 | |
| SamYaple | oh keystone, not keystoneclient | 15:24 |
| sdake | samyaple in the same container... | 15:24 |
| SamYaple | inc0: yea thats what i was proposing to you, except it is going to have to work a bit different than that | 15:24 |
| inc0 | other idea would be to haproxy somehow know which services it's supposed to configure, and add these configs by itself | 15:24 |
| rhallisey | https://bugs.launchpad.net/kolla/+bug/1469182 | 15:25 |
| openstack | Launchpad bug 1469182 in kolla "Cinder API requires python-keystone package" [High,Fix committed] - Assigned to Ryan Hallisey (rthall14) | 15:25 |
| SamYaple | inc0: that model works poorly, i used that for yaodu | 15:25 |
| inc0 | I'm in favor of first one as well | 15:25 |
| SamYaple | inc0: so basically the keystone role _itself_ cannot add the conf to haproxy | 15:25 |
| SamYaple | well have to do it a bit different | 15:25 |
| inc0 | why? | 15:25 |
| sdake | rhallisey we odnt need all of eystone for that, we only need the eystonemiddleware package | 15:25 |
| *** shardy has quit IRC | 15:25 | |
| SamYaple | because haproxy is not gauranteed to run on the same nodes as keystone | 15:25 |
| inc0 | hmm...good point | 15:26 |
| SamYaple | inc0: i know the model that will work, just get the container functional | 15:26 |
| SamYaple | ill work with you on how the model should look | 15:26 |
| inc0 | care to share it with me? | 15:26 |
| SamYaple | test the HUP thing esspecially, because that determines if we need supervisord | 15:27 |
| SamYaple | inc0: yea but i have to run out | 15:27 |
| inc0 | ok, I'll ping you on monday | 15:27 |
| SamYaple | cool, im always around | 15:27 |
| SamYaple | yea you get the contaienr solid and ill help you with the ansible bits | 15:27 |
| inc0 | I'm ok with ansible bits, I want to learn it anyway | 15:28 |
| inc0 | just let me know what you have in mind;) | 15:28 |
| rhallisey | sdake, ya fine with me, I wasn't sure where to get it by itself | 15:28 |
| *** absubram has joined #kolla | 15:28 | |
| sdake | rhallisey no big deal just saw it in the ode and was like "wtf" :) | 15:29 |
| inc0 | SamYaple, a callback plugin maybe | 15:29 |
| sdake | my vms never enter a running state | 15:29 |
| inc0 | ? | 15:29 |
| sdake | so i am totallly jammed up | 15:29 |
| *** shardy_ has quit IRC | 15:29 | |
| SamYaple | inc0: the issue is the role itself cannot address hosts in a seperate group than the ones specified initially | 15:30 |
| *** shardy has joined #kolla | 15:30 | |
| inc0 | but can it call some sort of callback on ansible itself, and ansible in turn will do stuff? | 15:30 |
| rhallisey | sdake, :) | 15:31 |
| SamYaple | yea but thats far from best practice :) | 15:31 |
| SamYaple | inc0: in fact, if you want to just on #ansible i should really ping them about the best way to handle this | 15:31 |
| inc0 | I'll do that | 15:31 |
| inc0 | and we'll see if this matches your idea | 15:31 |
| inc0 | we'll get more problems like this one as we go on I guess | 15:32 |
| SamYaple | inc0: i do need to go, so ive asked the question, feel free to move the discussion along | 15:33 |
| inc0 | kk | 15:33 |
| inc0 | cya later | 15:33 |
| SamYaple | well i see where you came up with callbacks | 15:34 |
| inc0 | not exactly this one, but yeah, that would work I guess | 15:35 |
| SamYaple | rhallisey: hey. crux. it only supports v2.0 keystone | 15:37 |
| rhallisey | oh really | 15:38 |
| rhallisey | I think we can remove it at some point | 15:38 |
| inc0 | hmm...ha playbook that will be called on register might work | 15:38 |
| rhallisey | just do the keystone calls | 15:38 |
| sdake | harmw did you fix some problem? | 15:40 |
| sdake | can ppl tell me if they have got nova booting in kolla lately? | 15:40 |
| sdake | as in the last 2 weeks | 15:40 |
| SamYaple | as a whole we should talk about v3 keystone | 15:41 |
| inc0 | SamYaple, how about this set_facts? | 15:44 |
| SamYaple | nope | 15:44 |
| inc0 | why? if we'd set them in ks and configure haproxy afterwards | 15:44 |
| inc0 | based on them | 15:44 |
| SamYaple | one, thats horribly ugly, two it still wouldnt do what we need because that means the haproxy role lays down ALL the configs | 15:45 |
| inc0 | not too bad tho | 15:45 |
| inc0 | one template per config | 15:45 |
| inc0 | I mean one template, and facts determine how its filled | 15:46 |
| SamYaple | inc0: its pretty bad actually | 15:46 |
| inc0 | this will keep all haproxy login in haproxy | 15:46 |
| inc0 | which is good. | 15:46 |
| SamYaple | well that would happen anyway | 15:46 |
| inc0 | and DRY-happy | 15:46 |
| inc0 | we add new service to playbook, rerun playbook and at the end of haproxy play do a hup | 15:47 |
| SamYaple | what im trying to avoid is including the haproxy hosts in each role and using a when conditional for all tasks | 15:47 |
| inc0 | but you wont | 15:48 |
| inc0 | you'll just fill facts | 15:48 |
| inc0 | in each role | 15:48 |
| inc0 | which isn't that bad anyway considering things like keystone port | 15:48 |
| SamYaple | facts dont cross roles | 15:49 |
| inc0 | ah... | 15:49 |
| SamYaple | also this means the haproxy role is either run after EACH service, or it is run at the end of everything | 15:49 |
| inc0 | is there anything that does? | 15:49 |
| SamYaple | both bad | 15:49 |
| SamYaple | across roles? i dont think so | 15:49 |
| inc0 | end of everything isn't that bad tho | 15:49 |
| SamYaple | inc0: it is because all the services have to hit haproxy to talk to keystone and eachother :) | 15:50 |
| SamYaple | and the database | 15:50 |
| inc0 | hmf | 15:50 |
| inc0 | ok, I won't keep you Sam, thanks | 15:50 |
| inc0 | I'll read a bit more about features we could potentially use | 15:51 |
| SamYaple | i think the best way might be to call the haproxy role initially and have it lay down all the configs that will be used | 15:51 |
| SamYaple | well have all the facts to do so and it will just online serice as they check in | 15:51 |
| inc0 | but we dont know which ones will | 15:51 |
| SamYaple | yes we do | 15:51 |
| inc0 | where and how? | 15:52 |
| SamYaple | we have access to all the same variables "use_cinder: True" then for the config file "when: use_cinder" | 15:52 |
| inc0 | I wanted to avoid having list of services hardcoded in config... | 15:52 |
| SamYaple | inc0: well we have to do that.... | 15:53 |
| inc0 | we'd need to dig into vars from roles anyway | 15:53 |
| SamYaple | how so | 15:53 |
| inc0 | because keystone_port might be something else than 5000 | 15:53 |
| SamYaple | that stuff exists in global for a few reasons | 15:53 |
| SamYaple | ports and groups are all we need and we have it | 15:53 |
| *** achanda has joined #kolla | 15:54 | |
| *** mfalatic has joined #kolla | 15:55 | |
| SamYaple | k gotta go for realzz | 15:55 |
| inc0 | ok SamYaple, go to sleep, cya in Monday | 15:55 |
| SamYaple | bai | 15:55 |
| *** mfalatic_ has joined #kolla | 15:56 | |
| *** dolpher1 has joined #kolla | 15:56 | |
| *** dolpher has quit IRC | 15:57 | |
| *** mfalatic has quit IRC | 16:00 | |
| sdake | later inc0 | 16:07 |
| inc0 | hehe I was only now writting that I'm going off | 16:07 |
| inc0 | you're fast sdake! | 16:07 |
| inc0 | sooo...cya all next week:) | 16:08 |
| *** inc0 has quit IRC | 16:08 | |
| *** jmccarthy has quit IRC | 16:10 | |
| *** jmccarthy has joined #kolla | 16:10 | |
| *** vinkman has joined #kolla | 16:12 | |
| sdake | samyaple did we decide to use our own module or the osad module | 16:13 |
| sdake | samyaple and by we I mean you :) | 16:14 |
| *** sdake_ has joined #kolla | 16:19 | |
| openstackgerrit | BorneMace proposed stackforge/kolla: Fixed all remaining pep8 failures in the codebase https://review.openstack.org/200612 | 16:21 |
| *** sdake has quit IRC | 16:22 | |
| harmw | sdake_: yes, gimme a sec | 16:34 |
| *** jtriley has quit IRC | 16:35 | |
| harmw | sdake_: https://bugs.launchpad.net/kolla/+bug/1473500 | 16:35 |
| openstack | Launchpad bug 1473500 in kolla "neutron-linuxbridge fails because of missing ebtables binary" [Undecided,New] | 16:35 |
| *** jtriley has joined #kolla | 16:40 | |
| *** sdake_ has quit IRC | 16:43 | |
| *** jmccarthy has quit IRC | 16:43 | |
| *** dolpher1 has quit IRC | 16:44 | |
| *** sdake has joined #kolla | 16:45 | |
| *** dims_ has quit IRC | 16:46 | |
| harmw | any tried to build a base image lately? | 16:46 |
| harmw | like, last couple of hours? | 16:46 |
| *** dims has joined #kolla | 16:46 | |
| *** dims has quit IRC | 16:47 | |
| *** dims has joined #kolla | 16:47 | |
| SamYaple | sdake: once the OSAD one merges ill look at it | 16:48 |
| SamYaple | sdake: its probably a week from merge or so. by then we may just keep using mine if it works like we want | 16:48 |
| harmw | looks like there is an issue with delorean: curl http://trunk.rdoproject.org/centos70/current/delorean.repo | 16:49 |
| SamYaple | sdake: i rewrote deguerri modules as well and put an explicit asl2.0 license so there is no confusion | 16:49 |
| *** mfalatic_ is now known as mfalatic | 16:51 | |
| harmw | sdake: who should I bug regarding a 404 on above url | 16:51 |
| harmw | just rdo ML? | 16:51 |
| sdake | harmw hang tight a moment | 16:54 |
| sdake | harmw join #rdo | 16:55 |
| *** bitblt has joined #kolla | 17:01 | |
| sdake | harmw try http://trunk.rdoproject.org/centos7/current/delorean.repo | 17:10 |
| sdake | hey lets change things for no good reason! | 17:10 |
| harmw | and not include a redirect | 17:11 |
| harmw | I want a redirect | 17:11 |
| harmw | lamers | 17:11 |
| sdake | i'll file a bug | 17:11 |
| sdake | getting redirects on fedora infra is alot of work | 17:11 |
| *** britthouser has joined #kolla | 17:13 | |
| harmw | right, and changing some folder is not | 17:13 |
| harmw | baseurl=http://209.132.178.14/centos7/1a/4c/1a4c958c1c98dc27bd8f8ecd2779e8adad4303d4_9c7271d9 | 17:14 |
| harmw | wtf | 17:14 |
| harmw | why is there an IP in there | 17:14 |
| sdake | https://bugs.launchpad.net/kolla/+bug/1473513 | 17:16 |
| openstack | Launchpad bug 1473513 in kolla "delorean repo location changed" [Critical,Confirmed] - Assigned to Harm Weites (harmw) | 17:16 |
| sdake | harmw can you apply a fix | 17:17 |
| harmw | noes! | 17:17 |
| harmw | I'm building a new base | 17:17 |
| sdake | nah anyon ecan control folder destination | 17:17 |
| harmw | so gimme a sec | 17:17 |
| sdake | its just a mv operation | 17:17 |
| harmw | thats insane | 17:17 |
| *** gfidente has quit IRC | 17:17 | |
| sdake | a redirect requires admin privs | 17:17 |
| harmw | ppl should *think* before mv | 17:18 |
| harmw | like mv -i | 17:18 |
| *** sdake_ has joined #kolla | 17:19 | |
| sdake_ | harmw do you have an ebltables patch? | 17:21 |
| sdake_ | or know what to do to fix? | 17:21 |
| harmw | yep | 17:21 |
| harmw | just add the package | 17:22 |
| sdake_ | wich package to which docker file | 17:22 |
| harmw | I wanted to test the fix, fully | 17:22 |
| sdake_ | i just want to see if that is why nova wouldn't boot | 17:22 |
| harmw | ebtables, to neutron-linuxbridge | 17:22 |
| harmw | I got the fix right here | 17:22 |
| *** sdake has quit IRC | 17:23 | |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Fix the delorean repo url https://review.openstack.org/200637 | 17:24 |
| bmace | any other core around able to take a quick look at https://review.openstack.org/#/c/200612 ? i would love to get the pep8 test actually working correctly today :) | 17:29 |
| sdake_ | harmw if you c an plz look above | 17:31 |
| harmw | bmace: https://review.openstack.org/#/c/200612/1/docker/centos/binary/neutron/neutron-base/ip_wrapper.py L41 | 17:33 |
| harmw | does that work? | 17:33 |
| sdake_ | syntax looks ok to me | 17:33 |
| harmw | I simply don't know, could be some kind of appender is needed | 17:34 |
| harmw | SamYaple: can you ack that repo fix? | 17:35 |
| bmace | L41? the comment line? | 17:35 |
| harmw | comment? no that print() | 17:35 |
| harmw | print( | 17:35 |
| harmw | "bla" | 17:35 |
| harmw | "blabla" | 17:35 |
| harmw | ) | 17:35 |
| bmace | ah, i looked it up. i had tossed a \ in there to split lines but flake8 told me to yank it | 17:35 |
| harmw | that just works? or should there be a + after the first "bla"? | 17:36 |
| harmw | ok, sounds like it's ok | 17:36 |
| harmw | sdake_: you confirmed the ebtables fix? | 17:36 |
| bmace | and i ran it through the python interpreter like that and it works :) | 17:36 |
| harmw | cool bmace :) | 17:36 |
| sdake_ | have not | 17:37 |
| harmw | will do? | 17:37 |
| sdake_ | still building | 17:37 |
| harmw | or should I? | 17:37 |
| harmw | ok | 17:37 |
| sdake_ | did you think that owuld fix the nova instances not starting? | 17:37 |
| sdake_ | i am using neutron in my startup (demos/heat/launch) | 17:37 |
| harmw | it did earlier today | 17:37 |
| harmw | I got an error regarding neutron, and in logs it bailed because of not finding ebtables | 17:37 |
| openstackgerrit | Merged stackforge/kolla: Fixed all remaining pep8 failures in the codebase https://review.openstack.org/200612 | 17:38 |
| sdake_ | harmw is there an ebtables bug filed and a patch up? | 17:41 |
| harmw | bug yes, patch not yet | 17:42 |
| harmw | I'm almost ready to test | 17:42 |
| harmw | and submit | 17:42 |
| sdake_ | link to bug plz | 17:42 |
| sdake_ | so i can set it appropriately | 17:42 |
| harmw | https://bugs.launchpad.net/kolla/+bug/1473500 | 17:42 |
| openstack | Launchpad bug 1473500 in kolla "neutron-linuxbridge fails because of missing ebtables binary" [Critical,Confirmed] - Assigned to Harm Weites (harmw) | 17:42 |
| sdake_ | or you can | 17:42 |
| sdake_ | looks like its set | 17:42 |
| sdake_ | all goo then | 17:42 |
| *** prad has quit IRC | 17:53 | |
| *** sdake has joined #kolla | 17:55 | |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Install ebtables in neutron-linuxbridge https://review.openstack.org/200642 | 17:56 |
| sdake | jpeeler or rhallisey can you ack those last two changes | 17:57 |
| *** shardy has quit IRC | 17:58 | |
| *** sdake_ has quit IRC | 17:59 | |
| bmace | interesting, apparently the "hacking" version of flake8 used via most other ostack projects is more strict than normal flake8, so there are still some issues to fix :/ | 18:01 |
| sdake | the rules are determined by the pep8 file | 18:03 |
| sdake | rather by the tox.ini file | 18:03 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add +x to Magnum's config-internal.sh scripts https://review.openstack.org/200645 | 18:03 |
| bmace | that isn't the only thing. if you use just vanilla flake8 that you grab via pip you will get different output than if you use the flake8 you get from the hacking package ostack projects usually use. | 18:04 |
| sdake | rux tox -e pep8 and it will apply the hacking rules | 18:04 |
| bmace | it adds stuff like complains about missing apache 2 headers, etc. | 18:04 |
| sdake | but wasn't aware of that - although differnett rules aer applied | 18:04 |
| bmace | based on what we had before, it wouldn't. we didn't even have "hacking" as a test requirement | 18:04 |
| openstackgerrit | Merged stackforge/kolla: Install ebtables in neutron-linuxbridge https://review.openstack.org/200642 | 18:10 |
| *** sdake_ has joined #kolla | 18:13 | |
| *** coolsvap|away is now known as coolsvap | 18:13 | |
| *** sdake has quit IRC | 18:16 | |
| sdake_ | harmw nice work - confirmed those two changes get me active instances | 18:19 |
| openstackgerrit | Merged stackforge/kolla: Fix the delorean repo url https://review.openstack.org/200637 | 18:20 |
| sdake_ | rhallisey how do you attach a cinder volume to a nova instance | 18:21 |
| rhallisey | nova volume-attach <instance_name> <cinder_id> | 18:21 |
| rhallisey | then check nova-compute for volume attach successful | 18:21 |
| rhallisey | or something like that | 18:21 |
| sdake_ | seems to attach | 18:22 |
| sdake_ | i can't actually login to the vm to verify unfortunately | 18:25 |
| * sdake_ groans | 18:25 | |
| *** sdake_ is now known as sdake | 18:25 | |
| sdake | one of these days my network will work consistently | 18:25 |
| sdake | harmw can you test out cinder | 18:27 |
| sdake | harmw sincey our can get int oyour vms apparentl.y with one nic magically :) | 18:28 |
| harmw | I should verify that once | 18:28 |
| sdake | oh you havenever sshed in? | 18:28 |
| harmw | nova console-log is just cheating | 18:28 |
| harmw | I have, just not recently | 18:28 |
| sdake | oh let me try console log | 18:28 |
| harmw | console-log is just showing the console.log from your instance | 18:29 |
| *** dims has quit IRC | 18:29 | |
| harmw | so you can atleast check if it finished booting | 18:29 |
| harmw | without issues | 18:29 |
| sdake | it doesn't appear to show the drive attached | 18:29 |
| harmw | I've got one more bugfix thats annoying the shit out of me for ages :p | 18:30 |
| harmw | https://bugs.launchpad.net/kolla/+bug/1471331 | 18:30 |
| openstack | Launchpad bug 1471331 in kolla "tools/kolla concluding to soon deployment has succeeded" [Undecided,New] - Assigned to Harm Weites (harmw) | 18:30 |
| rhallisey | sdake, if you're not seeing any complaints on nova compute it should be working | 18:30 |
| sdake | rhallisey i get this after a cinder attach http://ur1.ca/n35dm -> http://paste.fedoraproject.org/242773/14365530 | 18:31 |
| sdake | http://ur1.ca/n35e2 -> http://paste.fedoraproject.org/242774/65531421 | 18:32 |
| sdake | is this the error you got rhallisey? | 18:32 |
| sdake | cinder volume missing config options | 18:34 |
| sdake | http://ur1.ca/n35ee -> http://paste.fedoraproject.org/242776/53242143 | 18:34 |
| rhallisey | oh | 18:34 |
| rhallisey | getting closer nice | 18:35 |
| rhallisey | where's my lock_path patch.. | 18:35 |
| *** diogogmt has quit IRC | 18:37 | |
| rhallisey | sdake, just set lock_path via crudini | 18:37 |
| rhallisey | I think I have a missing branch for my cinder work I need to find.. | 18:38 |
| harmw | rhallisey: you had that, once | 18:38 |
| sdake | set to what? | 18:38 |
| harmw | I remember we uncovered that | 18:38 |
| sdake | should it be shared between nova as well? | 18:38 |
| sdake | nova has a lack path for instances | 18:38 |
| rhallisey | harmw, ya I remember | 18:38 |
| rhallisey | /tmp/cinder_lock | 18:39 |
| rhallisey | or something | 18:39 |
| sdake | taht should be persistent - so /var/run/cinder | 18:40 |
| sdake | /var/lib/cinder/lock that is | 18:40 |
| *** diogogmt has joined #kolla | 18:40 | |
| sdake | but my question is is the same lock path used in nova | 18:40 |
| rhallisey | ok that works | 18:40 |
| sdake | or do you know/not know? | 18:40 |
| rhallisey | I don't know | 18:40 |
| rhallisey | lock_path showed up about a few days before liberty | 18:40 |
| sdake | what other gems are hiding in your lost repo :) | 18:43 |
| rhallisey | not sure, a lot of this stuff showed up right before the release | 18:44 |
| rhallisey | it's odd cause I remember talking with harmw about lock_path | 18:44 |
| harmw | yep, we did | 18:47 |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Properly wait for Openstack services to be ready https://review.openstack.org/200655 | 18:47 |
| harmw | ^^ woohoo to that | 18:48 |
| sdake | does it go in the default section? | 18:49 |
| sdake | I am going to set disable_process_locking to true instead | 18:49 |
| sdake | but not sure which section it goe in | 18:49 |
| *** athomas has quit IRC | 18:51 | |
| rhallisey | sdake, ya default | 18:52 |
| sdake | default is deprecated | 18:54 |
| sdake | its sactually oslo_concurrency | 18:54 |
| *** athomas has joined #kolla | 18:56 | |
| bmace | hrm.. am i still connected? put up a review but no message in the channel from gerrit | 19:08 |
| *** bmace has quit IRC | 19:09 | |
| *** bmace has joined #kolla | 19:09 | |
| *** dims has joined #kolla | 19:12 | |
| harmw | sdake: rhallisey if you got the time, could you take a look at that review up above? | 19:14 |
| harmw | sdake: I can't bridge my wifi, so I'll give singlenic another go within my virtualboxed CoreOS | 19:15 |
| harmw | see what happens | 19:15 |
| *** achanda has quit IRC | 19:16 | |
| *** vinkman has left #kolla | 19:16 | |
| *** dims has quit IRC | 19:17 | |
| *** dims has joined #kolla | 19:17 | |
| harmw | bah, building containers is probably going to take a while | 19:31 |
| openstackgerrit | BorneMace proposed stackforge/kolla: Fixes the last of the hacking based pep8 errors https://review.openstack.org/200690 | 19:39 |
| bmace | and, after these two reviews are done we can get pep8 actually turned on! :) | 19:40 |
| sdake | rhallisey looks like there are errors here: | 19:47 |
| sdake | https://bugs.launchpad.net/kolla/+bug/1469323 | 19:47 |
| openstack | Launchpad bug 1469323 in kolla "cinder nova volume attach not working" [Critical,Triaged] - Assigned to Ryan Hallisey (rthall14) | 19:47 |
| sdake | I added to the bug log | 19:47 |
| sdake | rhallisey please make it a priority to get this resolved ;) | 19:47 |
| sdake | we are half way thorugh liberty 2 and this bug remains open | 19:48 |
| rhallisey | ya blew up my env today will get to fixing that.. | 19:48 |
| harmw | sdake: https://blueprints.launchpad.net/kolla/+spec/standard-start | 19:49 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Add oslo.reports as a global dependency https://review.openstack.org/200298 | 19:49 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Change permissions on cinder config-internal/external https://review.openstack.org/200055 | 19:49 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: WIP: Start tgtd in cinder-volume container https://review.openstack.org/200139 | 19:49 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Turn off oslo_conccurrency locking in volume driver https://review.openstack.org/200693 | 19:49 |
| bmace | harmw: about your comment. that comment needs to either have a summary added to it, or be squashed enough to be a single line comment. right now it is a broken multi-line comment :/ | 19:49 |
| bmace | harmw: and a decent summary didn't immediately come to mind. | 19:50 |
| harmw | bmace: uhm, we can't have multi-line comments? | 19:50 |
| sdake | you can just put end in " and put in " | 19:50 |
| bmace | harmw: not of that structure, no, multi-line comments require a summary on the """ start line | 19:50 |
| sdake | the next line | 19:50 |
| sdake | hae to wrap at 79 characters | 19:50 |
| sdake | oh thattype of comment | 19:50 |
| harmw | i know sdake , but bmace why not do just that? | 19:51 |
| sdake | ya bmace is correct | 19:51 |
| harmw | change the commenttype? | 19:51 |
| harmw | summary: organize bla | 19:51 |
| harmw | and BAM, you're done :) | 19:51 |
| harmw | (and can insert the multiline) | 19:52 |
| bmace | yeah, i'll put in a very brief summary. | 19:52 |
| *** sdake_ has joined #kolla | 19:52 | |
| *** sdake has quit IRC | 19:56 | |
| openstackgerrit | BorneMace proposed stackforge/kolla: Fixes the last of the hacking based pep8 errors https://review.openstack.org/200690 | 19:57 |
| harmw | sdake_: you agree on closing the standard-start spec which I just did? | 19:57 |
| sdake_ | if its done | 19:57 |
| sdake_ | it looks done to me | 19:57 |
| harmw | I've updated the spec's whiteboard | 19:58 |
| harmw | to reflect the missing bits | 19:58 |
| sdake_ | what is missing | 19:58 |
| harmw | eg. mariadb, mongodb, hautoproxy | 19:58 |
| sdake_ | ok | 19:58 |
| harmw | and how they will be dealt with | 19:58 |
| sdake_ | we should probably open a new blueprint for those 3 | 19:58 |
| sdake_ | especially mongodb - should be done in this blueprint if possible | 19:59 |
| harmw | it needs additional work, since it's using entrypoint crap | 19:59 |
| harmw | may be combined, but still, plus I don't know if the damn thing even works :p | 20:00 |
| harmw | which is why I put it aside | 20:00 |
| harmw | since it's a dependecny for ceilometer, there could very well be just some get-ceilometer-to-operate spec | 20:01 |
| sdake_ | there already is | 20:01 |
| harmw | cool | 20:01 |
| sdake_ | but it doesnt include dependencies | 20:01 |
| * sdake_ would like to see cinder working asap ;-) | 20:01 | |
| harmw | hehe | 20:01 |
| harmw | I'm building all containers atm | 20:01 |
| harmw | waiting for that to finish | 20:02 |
| openstackgerrit | Merged stackforge/kolla: Fixes the last of the hacking based pep8 errors https://review.openstack.org/200690 | 20:02 |
| harmw | sdake_: you keep an eye on tagging all those for L2? | 20:02 |
| harmw | or can I do that as well? | 20:03 |
| *** britthou_ has joined #kolla | 20:04 | |
| *** britthouser has quit IRC | 20:08 | |
| *** jruano has quit IRC | 20:13 | |
| *** bitblt has quit IRC | 20:14 | |
| bmace | one more review is up now and when it is merged i'll post the tox / test-requirement changes and we will be 100% pep8 compliant! :) | 20:16 |
| *** prad has joined #kolla | 20:17 | |
| openstackgerrit | Swapnil Kulkarni proposed stackforge/kolla: Remove requirement for root user in scripts https://review.openstack.org/200702 | 20:17 |
| *** prad has quit IRC | 20:19 | |
| coolsvap | bmace, do we need requirement change for docker-py>=1.2.0 <=1.3.0? | 20:20 |
| coolsvap | i am not able to run kolla start unless i do this change | 20:21 |
| coolsvap | anyone else facing the similar issue | 20:21 |
| harmw | coolsvap: I had something similar earlier today in my venv | 20:21 |
| harmw | after fixing that I had to upgrade websockets (?) as well | 20:21 |
| coolsvap | harmw, yes | 20:22 |
| bmace | coolsvap not sure. my requirement change is just adding the hacking package to test-requirements. | 20:22 |
| bmace | and it is test-requirements only | 20:22 |
| coolsvap | i think it will be a change | 20:23 |
| coolsvap | in requirements.txt | 20:23 |
| bmace | for docker.py it would be, yes | 20:23 |
| coolsvap | hmm | 20:24 |
| coolsvap | harmw, do you have the log handy for the failure? | 20:27 |
| harmw | no, was at $work | 20:28 |
| harmw | python errors | 20:28 |
| coolsvap | hmm, i wanted to create bug, will do that later sometime | 20:29 |
| harmw | I'm not sure if it is, realy | 20:29 |
| harmw | upgrading websockets doesn't feel like something we should fix | 20:29 |
| openstackgerrit | Swapnil Kulkarni proposed stackforge/kolla: WIP: Update docker-py in requirement.txt https://review.openstack.org/200705 | 20:30 |
| coolsvap | harmw, but i think we will need to update docker-py | 20:30 |
| *** britthou_ has quit IRC | 20:30 | |
| coolsvap | correct? | 20:30 |
| harmw | don't know, realy, since I didn't check what the actual issue was (apart from having to apply an update all of a sudden) since I didn't trust my python env either :) | 20:32 |
| bmace | if it does need to be updated wouldn't it be >1.2.0,<=1.3.0 if your assumption is that 1.2.0 won't work | 20:33 |
| sdake_ | coolsvap don't hard cap requirements pls | 20:34 |
| openstackgerrit | Swapnil Kulkarni proposed stackforge/kolla: Remove requirement for root user in scripts https://review.openstack.org/200702 | 20:34 |
| coolsvap | sdake_, no i am not | 20:35 |
| harmw | coolsvap: I think you missed my first comment on that bug :) | 20:35 |
| sdake_ | [13:20:38] <coolsvap>bmace, do we need requirement change for docker-py>=1.2.0 <=1.3.0? | 20:35 |
| sdake_ | 1.3.0 is a hard cap in this instanc | 20:35 |
| sdake_ | hard caps are evil for a bunch of reasons | 20:35 |
| coolsvap | harmw, no, i did check the scripts with same user | 20:35 |
| sdake_ | i am fine with exclusionis | 20:36 |
| sdake_ | (when we know an upstream version is busted) | 20:36 |
| sdake_ | did that cleanup-containers wor as nonroot | 20:36 |
| coolsvap | sdake_, yes | 20:36 |
| sdake_ | i was curious about the release of the netns stuff | 20:36 |
| sdake_ | release/clenaup | 20:36 |
| harmw | coolsvap: then you misunderstood my comment, my question was to add in some logic to do (in a clever way) what the current check achieves | 20:37 |
| harmw | "since root can use docker, lets check if YOU are root!" | 20:37 |
| harmw | (and bail if you aren't) | 20:37 |
| coolsvap | harmw, sure i will add, i will put it wip | 20:37 |
| harmw | put workflow to -1 and when you're done put a new review up :) | 20:38 |
| harmw | (atleast thats what I would do) | 20:38 |
| sdake_ | i think it is fine as is | 20:38 |
| sdake_ | see comments in review | 20:38 |
| sdake_ | we justneed to document how to add user to docerrott group | 20:38 |
| sdake_ | or point at upstream docs on that | 20:38 |
| sdake_ | having to run sudo for some commands and not others is annoying | 20:39 |
| sdake_ | and the checking is different per distro and packaging | 20:39 |
| *** coolsvap is now known as coolsvap|away | 20:39 | |
| sdake_ | some distros use dockerroot, some distro use the docker group | 20:39 |
| harmw | right, so mayhem will occur if $USER didn't read the docs... | 20:39 |
| sdake_ | mayhem already occurs if user doesn't read the docs ;) | 20:40 |
| sdake_ | trust me I always start out with "did you read the quickstart guide" :) | 20:40 |
| harmw | I didn't ask to check if $USER was in a specific group, just to check if docker was usable :) | 20:40 |
| sdake_ | harmw that actually sounds good | 20:40 |
| sdake_ | i'm not sure how you would do that docker ps perhaps | 20:40 |
| harmw | or docker info | 20:41 |
| harmw | I don't care, realy | 20:41 |
| sdake_ | but lets do that as a seperate change if possible | 20:41 |
| sdake_ | that way we can move on past the requireing orot | 20:41 |
| harmw | isn't that perfect for this changeset? | 20:41 |
| sdake_ | harmw request a techdebt bug be filed in the bug | 20:41 |
| sdake_ | and dont approve until it is :) | 20:41 |
| harmw | I'll file it this instant, way easier | 20:41 |
| sdake_ | coolsvap has other higher priority things to work on I think | 20:41 |
| sdake_ | like ansibllizing code base | 20:42 |
| sdake_ | I asked him to fix this small probelm up in the meantime as I sort out our next work streams | 20:42 |
| sdake_ | if coolsvap is good with doing the job now i'm ok with either approach ;) | 20:42 |
| harmw | fair enough | 20:42 |
| harmw | oh, he's away | 20:42 |
| sdake_ | ya unfortunately he is afk | 20:42 |
| sdake_ | so leave a comment in the bug as to how to proceed pls :) | 20:43 |
| sdake_ | rather in the review | 20:43 |
| harmw | yea | 20:43 |
| harmw | yup | 20:43 |
| sdake_ | thanks fine sir | 20:43 |
| *** achanda has joined #kolla | 20:43 | |
| sdake_ | and i'll sort out oour next couple of weeks of work | 20:43 |
| harmw | ack the init-runonce review pls :) | 20:43 |
| harmw | rhallisey: you to | 20:43 |
| sdake_ | link harmw | 20:43 |
| sdake_ | i usually do reviews in the morning unless they are blocking the code from working | 20:44 |
| rhallisey | harmw, kk | 20:44 |
| sdake_ | then ask for out of band acks | 20:44 |
| harmw | https://review.openstack.org/#/c/200655/ | 20:44 |
| harmw | ok sdake_ :) | 20:44 |
| sdake_ | nice job harmw ! | 20:45 |
| sdake_ | I wanted to do that for like oh 3 months :) | 20:45 |
| sdake_ | and never got around to it | 20:45 |
| harmw | same here :) | 20:46 |
| harmw | until now | 20:46 |
| *** sdake_ is now known as sdake | 20:46 | |
| sdake | wish my client auto nick changed | 20:46 |
| sdake | i'll hav eto see if textual can do that | 20:46 |
| rhallisey | so many things, so little tine O.o | 20:46 |
| rhallisey | nice harmw | 20:46 |
| openstackgerrit | Merged stackforge/kolla: Properly wait for Openstack services to be ready https://review.openstack.org/200655 | 20:47 |
| harmw | cool! | 20:47 |
| harmw | thanks guys | 20:47 |
| sdake | i got some reviews which need attention too - the stream on cinder | 20:47 |
| sdake | if folks could ack those or neg them | 20:47 |
| sdake | so I can fix them | 20:47 |
| sdake | would appreciate it | 20:47 |
| bmace | i have one more pep8 review, then i can do the commit for the actual pep8 enablement, if you want it today, if not i can finish it up monday. https://review.openstack.org/#/c/200679/ | 20:47 |
| sdake | dont ack the wip: tgt one yet | 20:47 |
| harmw | I never touch WIP's :) | 20:48 |
| harmw | unless explicitly asked | 20:48 |
| sdake | bmace will look - this is an example of a blocking reviwew harmw :) | 20:48 |
| harmw | I'm still building containers, damn | 20:48 |
| harmw | either my wifi sucks, or virtualbox does | 20:49 |
| bmace | i thought i saw something yesterday about the build time being super short now? | 20:49 |
| sdake | are you using build.py -t rdo -T 1000? | 20:49 |
| sdake | it taes me about 5 minutes to build there | 20:49 |
| sdake | bmace not yet | 20:49 |
| harmw | crap, I should've done that sdake | 20:49 |
| sdake | but there is a much faster builder tool | 20:49 |
| bmace | you sure you want the T 1000? isn't that the liquid Terminator? :) | 20:49 |
| sdake | and it seems reliable | 20:49 |
| sdake | bmace termination ftw ;) | 20:49 |
| sdake | samyaple did a really nice job with the patch | 20:50 |
| sdake | or the code rather | 20:50 |
| sdake | since it wasn't really a patch :) | 20:50 |
| harmw | yea, I just totally forgot that script | 20:50 |
| sdake | the plan is to remove the existing build once we get the new one into good woring order and can trust its results | 20:50 |
| sdake | atm it seems like the command line around caching doesn't work for example | 20:50 |
| sdake | harmw can you review the link in the scrollback from bmace | 20:53 |
| sdake | we need this because the pep8 gate is soon to break ;) | 20:53 |
| harmw | I was saving that for sunday | 20:53 |
| harmw | actually :p | 20:53 |
| bmace | putting in more pep8 breaking code harmw? ;) | 20:55 |
| sdake | if someone could review this stream: | 20:56 |
| sdake | https://review.openstack.org/#/c/200693/1 | 20:56 |
| sdake | jpeeler or rhallisey - harmw has already reviewed | 20:56 |
| harmw | lol@https://review.openstack.org/#/c/200679/1/docker/centos/binary/neutron/neutron-base/ip_wrapper.py | 20:58 |
| harmw | great stuff :P | 20:58 |
| bmace | yeah. the hacking stuff added to pep8 gets pretty picky | 20:59 |
| harmw | I apparently like picky :> | 20:59 |
| bmace | yeah, i really don't mind needing them to be alphabetized. | 20:59 |
| bmace | especially with larger import lists | 21:00 |
| harmw | it looks clean | 21:00 |
| harmw | less mess | 21:00 |
| openstackgerrit | Merged stackforge/kolla: Fixed additional OpenStack pep8 issues https://review.openstack.org/200679 | 21:01 |
| harmw | there you go | 21:01 |
| -openstackstatus- NOTICE: Gerrit will be unavailable from 22:00 to 22:30 UTC for project renames | 21:01 | |
| bmace | awesome, now i'll put up the review to make it all official :) | 21:02 |
| bmace | sdake, still want the pep8 to run all those validate-contents tests until infra finished that other review? | 21:04 |
| *** jtriley has quit IRC | 21:06 | |
| sdake | bmace might as well | 21:09 |
| sdake | whatever gets it done fastest without blociking the gate | 21:09 |
| sdake | I dont want to do any emergency reverts because it annoys the infra tream :) | 21:09 |
| bmace | yup, the only downside of checking in the final version of what we want to do would be that we wouldn't actually haver the validate-contents coverage until infra was done with their review | 21:10 |
| sdake | well we do that in our repo right? | 21:10 |
| bmace | if i commit both tests covering everything at some point we just need to yank the extra tests from our pep8 test | 21:10 |
| bmace | yup | 21:10 |
| sdake | if so, lets just do all the tests | 21:10 |
| bmace | kk | 21:11 |
| sdake | changes in our repo are fast | 21:11 |
| sdake | I can harass ppl to review and get em done in like 5-10 minutes during the work week typically | 21:11 |
| sdake | I am more concerned about having long review cycles on the infra side | 21:11 |
| sdake | bmace btw the reason infra has not reviewed your patch is because this is rename week | 21:11 |
| sdake | when projects go from stackforge to openstack | 21:11 |
| bmace | ah, extra chaos | 21:12 |
| sdake | its a super super busy time for them | 21:12 |
| *** dwalsh has quit IRC | 21:12 | |
| bmace | ok, review up shortly | 21:12 |
| sdake | it requires all their attention to not fuck up the system ;) | 21:12 |
| bmace | btw, found out that PyYAML wasn't needed as a dependency for those validate tests | 21:12 |
| sdake | ok so we need to revert a change then? | 21:12 |
| sdake | or new patch? | 21:13 |
| bmace | nope.. again, just in our tree, it is part of my review going up soon | 21:13 |
| sdake | sounds good | 21:13 |
| *** achanda has quit IRC | 21:14 | |
| openstackgerrit | BorneMace proposed stackforge/kolla: Enable pep8 tests https://review.openstack.org/200729 | 21:15 |
| sdake | are ou sure about that line 38 change | 21:16 |
| sdake | could you explain why you did that? | 21:16 |
| bmace | the yank of test-requirements from images? | 21:16 |
| sdake | yes | 21:16 |
| bmace | it was added in base testenv | 21:16 |
| sdake | ok | 21:16 |
| bmace | so it is available for all the tests | 21:16 |
| sdake | got it | 21:16 |
| bmace | so once that infra stuff is done we just yank the blah/tools/*.sh from testenv:pep8 and we are done. | 21:18 |
| bmace | but those tests are fast anyway. isn't super painful to have them in both places | 21:18 |
| sdake | what is the deal with this? | 21:19 |
| sdake | http://logs.openstack.org/98/200298/3/check/check-kolla-functional-f21/812e876/console.html | 21:19 |
| sdake | bmace ya just makes tracking down the problem harder | 21:19 |
| sdake | each of those should really be a separate gatei mo :) | 21:20 |
| *** Slower has quit IRC | 21:21 | |
| bmace | sdake hrm.. not sure. i didn't get that in my environment | 21:21 |
| * sdake really wants check-kolla-functioanl-f21 to start being more reliable not less reliable ;( | 21:22 | |
| bmace | let me see if i can find out what is going on | 21:22 |
| *** sdake_ has joined #kolla | 21:24 | |
| bmace | looks like something mock is using, but i don't think i did anything that should have effected that. i only added the hacking to test-requirement and if it was some sort of issue at that level i would have expected to see it on my system too :/ | 21:24 |
| openstackgerrit | Merged stackforge/kolla: Change permissions on cinder config-internal/external https://review.openstack.org/200055 | 21:24 |
| openstackgerrit | Merged stackforge/kolla: Add oslo.reports as a global dependency https://review.openstack.org/200298 | 21:24 |
| sdake_ | bmace yes it wasn'tdirected at you | 21:24 |
| sdake_ | just a general question for people to help sort out :) | 21:24 |
| sdake_ | bmace i'm pretty ssure something broke in the gate around that code | 21:25 |
| sdake_ | jpeeler you should take a look too if you can | 21:25 |
| bmace | sdake_ i gotcha. i know that the requirements can be really touchy about ordering | 21:25 |
| sdake_ | there shouldn't be touchiness around ordering | 21:25 |
| sdake_ | i personally haven't experienced that | 21:25 |
| bmace | a lot of other projects specifically mention in their requirements / test-requirements that the order should be left as is since issues can happen if it isn't :/ | 21:26 |
| sdake_ | what good is an installation system (pip) if its not deterministic | 21:26 |
| *** achanda has joined #kolla | 21:26 | |
| sdake_ | bmace i still haven't experienced that | 21:26 |
| bmace | i don't disagree :) | 21:26 |
| sdake_ | I know it says it across many projects, but i ordered them asci-betically and it had no impact | 21:27 |
| sdake_ | other then being properly sorted ;) | 21:27 |
| *** sdake has quit IRC | 21:27 | |
| harmw | oh my, /me just created a linked profile :/ | 21:29 |
| sdake_ | your famous now ! :) | 21:34 |
| sdake_ | btw harmw how old areoyu if i can ask :) | 21:34 |
| harmw | 29 :) | 21:34 |
| sdake_ | nice i remembe r29 | 21:34 |
| harmw | hehe | 21:34 |
| sdake_ | i felt so much better back then | 21:34 |
| sdake_ | 41 here | 21:34 |
| rhallisey | woo still the youngest | 21:34 |
| harmw | 29 and kid #3 on his way :p | 21:34 |
| sdake_ | you now they make operations and pills to prevent that harmw :) | 21:35 |
| harmw | yep | 21:35 |
| sdake_ | both of our chidren came on birth control | 21:35 |
| sdake_ | yay for birth control | 21:35 |
| harmw | interesting | 21:35 |
| sdake_ | not that my kids dont rock | 21:35 |
| sdake_ | i'm glad we had em when we did | 21:36 |
| sdake_ | or we may never have | 21:36 |
| bmace | they would not be stopped! :) | 21:36 |
| harmw | :) | 21:36 |
| *** rhallisey has quit IRC | 21:37 | |
| bmace | last review request of the day if someone other than sdake could look at https://review.openstack.org/#/c/200729/ otherwise if someone sneaks in pep8 breaking changes before it goes in the gate might start failing :/ | 21:37 |
| jpeeler | sdake_: adding funcsigs to test-requirements.txt should fix it, i don't understand why pip isn't doing the job | 21:39 |
| jpeeler | https://github.com/testing-cabal/mock/blob/master/requirements.txt | 21:40 |
| jpeeler | it's listed there | 21:40 |
| jpeeler | this comment does not inspire confidence: https://github.com/testing-cabal/mock/blob/master/mock/mock.py#L85 | 21:41 |
| harmw | right, so, after filling in countless crappy forms I can finally invite $world to my linkedin profile | 21:50 |
| harmw | jeez | 21:50 |
| harmw | the hassle | 21:50 |
| *** achanda has quit IRC | 21:50 | |
| sdake_ | jpeeler loll | 21:52 |
| sdake_ | thats a good one ;) | 21:52 |
| sdake_ | "import funcsigs" ?Has this been fixed? | 21:53 |
| *** achanda has joined #kolla | 21:54 | |
| -openstackstatus- NOTICE: Gerrit is unavailable from approximately 22:00 to 22:30 UTC for project renames | 21:56 | |
| *** ChanServ changes topic to "Gerrit is unavailable from approximately 22:00 to 22:30 UTC for project renames" | 21:56 | |
| jpeeler | bmace: is it not the python way to keep separate dependency lists? | 21:57 |
| *** vinkman has joined #kolla | 21:58 | |
| jpeeler | i guess it's easier to maintain, but installing a bunch of clients in the pep8 virtual env seems weird to me | 21:58 |
| bmace | most of the other projects i had looked at for sort of "standards" guidance had the requirements / test-requirements right in testenv like that | 21:59 |
| jpeeler | yeah, both nova and glance do it your way (but neutron does it my way) | 22:03 |
| jpeeler | won't make a big deal out of it though | 22:03 |
| jpeeler | well gerrit just went down, so i can't approve it | 22:03 |
| jpeeler | should have translated 22:00 UTC before i started looking around | 22:04 |
| bmace | jpeeler btw, i added your name into a couple of TODOs when doing my pep8 fixes earlier, based on the git blame line saying you were the last one to touch it, hopefully that was true :) | 22:07 |
| jpeeler | git is lying! | 22:07 |
| bmace | lol | 22:08 |
| *** Slower has joined #kolla | 22:10 | |
| *** sdake has joined #kolla | 22:11 | |
| *** sdake_ has quit IRC | 22:14 | |
| *** britthouser has joined #kolla | 22:19 | |
| bmace | seems like it is alive again? | 22:29 |
| *** ChanServ changes topic to "Please register for the the Kolla-Palooza midcycle July 28th, July 29th @ https://www.eventbrite.com/e/kolla-palooza-midcycle-event-tickets-17628669826 - Channel is Logged @ http://eavesdrop.openstack.org/irclogs/%23kolla/" | 22:29 | |
| bmace | and ChanServ agrees ;) | 22:31 |
| *** diogogmt has quit IRC | 23:00 | |
| *** diogogmt has joined #kolla | 23:01 | |
| *** diogogmt has quit IRC | 23:02 | |
| *** jruano has joined #kolla | 23:04 | |
| *** britthouser has quit IRC | 23:08 | |
| vinkman | So, for the config-external bits, I think we are missing a hook in there for people to setup pieces that would be required by their specific config, for example if someone wants to have oslo lock dir somewhere sane, they would need to be able to do the mkdir -p and chown etc. | 23:11 |
| *** dims has quit IRC | 23:13 | |
| *** dims has joined #kolla | 23:17 | |
| *** openstackgerrit has quit IRC | 23:39 | |
| *** openstackgerrit has joined #kolla | 23:39 | |
| *** jruano has quit IRC | 23:42 | |
| bmace | gotta run guys, try to get that pep8 review in or else we might end up with gate issues if other breaking changes get in first :/ | 23:52 |
| bmace | thanks much jpeeler :) | 23:55 |
| openstackgerrit | Merged stackforge/kolla: Enable pep8 tests https://review.openstack.org/200729 | 23:55 |
| jpeeler | np | 23:55 |
| bmace | well, now that is done, weekend time! have a great one all | 23:55 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!