| *** masber has joined #openstack-ansible | 00:07 | |
| *** vnogin has quit IRC | 00:08 | |
| *** gouthamr has quit IRC | 00:09 | |
| SamYaple | cloudnull: templates for nginx+uwsgi https://gist.github.com/SamYaple/e1d99a41fe35cf465e99fc38ac0e7147 https://gist.github.com/SamYaple/208ef3537ec9bb5605a40294e459fb7e | 00:23 |
|---|---|---|
| SamYaple | i run all services behind those templates | 00:23 |
| SamYaple | (full ssl) | 00:24 |
| SamYaple | and an example of the keystone wsgi (which OSA alwready has) https://github.com/SamYaple/home-salt/blob/master/salt/files/openstack/keystone/keystone-wsgi.py | 00:25 |
| SamYaple | with horizon being a bit trickier https://github.com/SamYaple/home-salt/blob/master/salt/files/openstack/horizon/horizon-wsgi.py | 00:26 |
| *** savvas has joined #openstack-ansible | 00:36 | |
| *** savvas has joined #openstack-ansible | 00:36 | |
| *** exodusftw has quit IRC | 00:36 | |
| *** yifei has joined #openstack-ansible | 00:40 | |
| *** exodusftw has joined #openstack-ansible | 00:45 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_glance master: Update glance NFS for systemd https://review.openstack.org/526930 | 00:59 |
| *** gouthamr has joined #openstack-ansible | 01:08 | |
| *** savvas_ has joined #openstack-ansible | 01:26 | |
| *** savvas has quit IRC | 01:28 | |
| *** adreznec has joined #openstack-ansible | 01:43 | |
| *** asettle_ is now known as asettle | 01:45 | |
| *** markvoelker has quit IRC | 02:13 | |
| *** dave-mccowan has joined #openstack-ansible | 02:14 | |
| *** dave-mcc_ has joined #openstack-ansible | 02:15 | |
| *** dave-mccowan has quit IRC | 02:19 | |
| *** oanson has quit IRC | 02:19 | |
| *** oanson has joined #openstack-ansible | 02:21 | |
| *** vnogin has joined #openstack-ansible | 02:21 | |
| *** japestinho has joined #openstack-ansible | 02:23 | |
| *** vnogin has quit IRC | 02:27 | |
| *** dave-mccowan has joined #openstack-ansible | 02:33 | |
| *** dave-mcc_ has quit IRC | 02:35 | |
| *** hybridpollo has quit IRC | 02:35 | |
| *** hybridpollo has joined #openstack-ansible | 03:03 | |
| *** vnogin has joined #openstack-ansible | 03:23 | |
| *** hw_wutianwei has joined #openstack-ansible | 03:25 | |
| *** nshetty has joined #openstack-ansible | 03:26 | |
| *** vnogin has quit IRC | 03:27 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_glance master: Update glance NFS for systemd https://review.openstack.org/526930 | 03:40 |
| *** dave-mccowan has quit IRC | 03:41 | |
| *** gouthamr has quit IRC | 03:45 | |
| *** woodard has joined #openstack-ansible | 03:49 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_glance master: Update glance NFS for systemd https://review.openstack.org/526930 | 03:54 |
| *** markvoelker has joined #openstack-ansible | 04:13 | |
| *** vnogin has joined #openstack-ansible | 04:23 | |
| *** udesale has joined #openstack-ansible | 04:27 | |
| *** vnogin has quit IRC | 04:28 | |
| *** masuberu has joined #openstack-ansible | 04:30 | |
| *** masber has quit IRC | 04:33 | |
| *** masuberu has quit IRC | 04:36 | |
| *** nshetty has quit IRC | 04:43 | |
| *** markvoelker has quit IRC | 04:48 | |
| *** nshetty has joined #openstack-ansible | 04:55 | |
| *** hybridpollo has quit IRC | 05:27 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-galera_server master: Further simplify the upgrade check https://review.openstack.org/524086 | 05:31 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Change the galera health check for better cluster health https://review.openstack.org/520673 | 05:31 |
| *** udesale__ has joined #openstack-ansible | 05:34 | |
| *** udesale has quit IRC | 05:37 | |
| *** masber has joined #openstack-ansible | 05:42 | |
| *** markvoelker has joined #openstack-ansible | 05:45 | |
| *** cshen has joined #openstack-ansible | 06:03 | |
| *** indistylo has joined #openstack-ansible | 06:07 | |
| *** ianychoi_ is now known as ianychoi | 06:12 | |
| *** indistylo has quit IRC | 06:13 | |
| *** Prateek has joined #openstack-ansible | 06:14 | |
| Prateek | How i can make public endpoint URL on http instead of https?? | 06:14 |
| *** indistylo has joined #openstack-ansible | 06:14 | |
| *** threestrands has quit IRC | 06:17 | |
| *** markvoelker has quit IRC | 06:19 | |
| *** wspthr has quit IRC | 06:20 | |
| *** vnogin has joined #openstack-ansible | 06:24 | |
| prometheanfire | Prateek: https://github.com/openstack/openstack-ansible/blob/master/group_vars/all/all.yml#L60 | 06:27 |
| *** vnogin has quit IRC | 06:29 | |
| *** cshen has quit IRC | 06:29 | |
| Prateek | Thanks prometheanfire: i will try this | 06:31 |
| *** udesale has joined #openstack-ansible | 06:46 | |
| *** udesale__ has quit IRC | 06:48 | |
| *** jbadiapa has joined #openstack-ansible | 06:51 | |
| *** jafeha__ is now known as jafeha | 07:08 | |
| jafeha | good morning everyone | 07:08 |
| evrardjp | Prateek: it's by default | 07:12 |
| evrardjp | oh http instead of https | 07:12 |
| evrardjp | yeah | 07:12 |
| evrardjp | jafeha: good morning | 07:13 |
| jafeha | good morning evrardjp | 07:13 |
| evrardjp | could people vote on https://review.openstack.org/#/c/522778/ ? | 07:13 |
| *** markvoelker has joined #openstack-ansible | 07:16 | |
| *** jvidal has joined #openstack-ansible | 07:16 | |
| *** cshen has joined #openstack-ansible | 07:16 | |
| *** jafeha has quit IRC | 07:18 | |
| *** hamza21 has joined #openstack-ansible | 07:20 | |
| *** jafeha has joined #openstack-ansible | 07:21 | |
| prometheanfire | moin | 07:22 |
| *** vnogin has joined #openstack-ansible | 07:25 | |
| *** vnogin has quit IRC | 07:30 | |
| ThomasS | hi | 07:33 |
| ThomasS | good morning | 07:33 |
| ThomasS | i have a question | 07:33 |
| ThomasS | how do i update packages inside a container the correct way | 07:33 |
| ThomasS | for example the ceph mon containers i want to update to luminous 12.2.2 from 12.2.1 | 07:36 |
| ThomasS | the osds is simple because they are on metal | 07:36 |
| ThomasS | maybe some has a hint | 07:36 |
| ThomasS | the packages i guess come from the repo container | 07:36 |
| *** nshetty is now known as nshetty|lunch | 07:38 | |
| evrardjp | distro packages are following the regular package management process | 07:44 |
| evrardjp | we just compile openstack related things from source when we need | 07:45 |
| evrardjp | I don't know for ceph, we are using ceph-ansible, could you check there? | 07:45 |
| evrardjp | ThomasS: see the 3 lines above ^ | 07:45 |
| *** markvoelker has quit IRC | 07:49 | |
| ThomasS | hi thanks for the input | 07:54 |
| ThomasS | ok | 07:54 |
| ThomasS | so how does this work for distor packages? | 07:55 |
| ThomasS | for compute everything is metal | 07:55 |
| ThomasS | so just updateing as normal | 07:55 |
| ThomasS | but how can i update the containers? | 07:55 |
| ThomasS | maybe i need a dummies guide to lxc :-) | 07:56 |
| ThomasS | for example bash inside the container | 07:57 |
| *** mardim_PTO is now known as mardim | 07:57 | |
| evrardjp | ThomasS: these are machines | 07:59 |
| evrardjp | you can ssh to them | 07:59 |
| evrardjp | or you can attach from the host | 08:00 |
| evrardjp | but all the packages needing updating should already be taken care by the playbooks | 08:00 |
| evrardjp | so you don't have to | 08:00 |
| evrardjp | if you're saying we don't run the latest version of luminous maybe there is a variable we can pass to ceph ansible to do the update | 08:01 |
| andymccr | SamYaple: nice on the uwsgi/nginx stuff - had a question around neutron. Have they changed the way it reads in the plugin .ini files? it used to not read them in so it'd just fail to work properly and you can't tell it to in uwsgi (I see you haven't either and the neutron-wsgi.py doesnt look like it does it) | 08:03 |
| *** ivve has quit IRC | 08:06 | |
| *** nshetty|lunch is now known as nshetty | 08:16 | |
| *** ivve has joined #openstack-ansible | 08:24 | |
| *** ppetit has joined #openstack-ansible | 08:26 | |
| *** vnogin has joined #openstack-ansible | 08:26 | |
| *** gkadam has joined #openstack-ansible | 08:27 | |
| *** ppetit has left #openstack-ansible | 08:28 | |
| openstackgerrit | Lihi Wishnitzer proposed openstack/openstack-ansible-os_neutron master: Update Dragonflow configuration https://review.openstack.org/523870 | 08:30 |
| *** vnogin has quit IRC | 08:31 | |
| *** cshen_ has joined #openstack-ansible | 08:38 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_congress master: Introduce os_congress role in gerrit https://review.openstack.org/522491 | 08:42 |
| evrardjp | andymccr: I am close to John Snow on this topic, but I have seen changes in neutron on the uwsgi side: https://review.openstack.org/#/c/526307/1/templates/api-paste.ini.j2 | 08:44 |
| andymccr | evrardjp: hmm must be some changes then - i had a look to see if their devstack gate was using it, it's not. | 08:45 |
| andymccr | only nova/keystone/glance have apache conf setup for wsgi on the neutron gate | 08:45 |
| andymccr | and i think the initial patch to do it was abandoned | 08:45 |
| evrardjp | yeah that's what jmccrory said | 08:46 |
| evrardjp | but I think jmccrory got some results recently | 08:46 |
| *** markvoelker has joined #openstack-ansible | 08:46 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Change the galera health check for better cluster health https://review.openstack.org/520673 | 08:46 |
| evrardjp | Wow it's heavily snowing here | 08:47 |
| prometheanfire | nice | 08:47 |
| prometheanfire | still in .be? | 08:47 |
| evrardjp | yeah my garden looks nice :) | 08:47 |
| evrardjp | yeah | 08:47 |
| prometheanfire | it snowed here as well, think it was friday | 08:48 |
| evrardjp | last week of work | 08:48 |
| prometheanfire | nice | 08:48 |
| evrardjp | so I heard | 08:48 |
| prometheanfire | back up to the 20s this week | 08:48 |
| evrardjp | how was it? | 08:48 |
| evrardjp | not too many problems on the road? | 08:48 |
| prometheanfire | it was funny, no one knows how to drive in it | 08:48 |
| evrardjp | haha | 08:48 |
| prometheanfire | but not too bad | 08:48 |
| evrardjp | noobs! | 08:48 |
| prometheanfire | being from the north helps :D | 08:48 |
| prometheanfire | n00bs indeed | 08:48 |
| evrardjp | well, we are terrible too tbh | 08:49 |
| evrardjp | we have no mandatory thing that forces ppl to use winter tyres | 08:49 |
| evrardjp | or tires | 08:49 |
| evrardjp | I don't know | 08:49 |
| evrardjp | tyres looks weird | 08:50 |
| evrardjp | thanks wikipedia for your help: A tire or tyre is a ring-shaped component that surrounds a wheel's rim to transfer a vehicle's load from the axle through the wheel to the ground and to provide traction on the surface traveled over | 08:50 |
| prometheanfire | ya, that could hurt | 08:50 |
| evrardjp | yeah so basically people with BMW are terrible on the roads without winter tires | 08:51 |
| evrardjp | so it gets fun | 08:51 |
| evrardjp | New car here, no winter tires, and its a BMW, so I guess I won't take the road in the future. | 08:51 |
| evrardjp | :p | 08:51 |
| evrardjp | near future* | 08:51 |
| prometheanfire | :D | 08:51 |
| prometheanfire | nice | 08:51 |
| prometheanfire | ya, shannon had winter tires still | 08:51 |
| evrardjp | it works well :) | 08:52 |
| evrardjp | amazed by the difference | 08:52 |
| prometheanfire | I didn't but I wasn't going over 50, so lower speeds | 08:52 |
| *** shardy has joined #openstack-ansible | 08:53 | |
| evrardjp | prometheanfire: you don't get a lot of snow in texas do you? | 08:55 |
| prometheanfire | nope, first in years | 08:57 |
| prometheanfire | like I said, it was funny | 08:57 |
| *** nshetty_ has joined #openstack-ansible | 08:59 | |
| *** nshetty has quit IRC | 08:59 | |
| *** nshetty_ is now known as nshetty | 09:03 | |
| evrardjp | :) | 09:05 |
| *** Jack_Iv has joined #openstack-ansible | 09:07 | |
| *** epalper has joined #openstack-ansible | 09:11 | |
| *** DanyC has joined #openstack-ansible | 09:14 | |
| SamYaple | andymccr: so neutron is a *bit* tricky | 09:17 |
| SamYaple | uwsgi is whats reading the uwsgi.ini of course (or you can set all that on the uwsgi cli) | 09:18 |
| SamYaple | the neutron-api is the wsgi file you want to reference | 09:18 |
| SamYaple | im going to gist what i use pre-pike (which isbasically upstream) | 09:18 |
| SamYaple | the tricky part is neutron-server used to be neutron-api and neutron-rpc-server runnign together | 09:18 |
| SamYaple | well, neutron-rpc-server doesnt really work standalone | 09:19 |
| SamYaple | so you have to run nginx+uwsgi refercing the neutron-api wsgi file *and* run neutron-server (not neutron-rpc-server) | 09:19 |
| SamYaple | ijust bind neutron-server to 127.0.0.1:9686 to get it out of the way, its not used. | 09:19 |
| *** markvoelker has quit IRC | 09:20 | |
| SamYaple | its just there to make neutron-rpc-server work right | 09:20 |
| SamYaple | https://bugs.launchpad.net/neutron/+bug/1708389 | 09:20 |
| openstack | Launchpad bug 1708389 in neutron "wsgi and rpc-server has issues with dhcp when creating subnet" [Undecided,New] | 09:20 |
| SamYaple | andymccr: https://gist.github.com/SamYaple/18145b2a452770de832708056be606ef thats the pre-pike neutron-api-wsgi file i use | 09:21 |
| SamYaple | with that wsgi file yo ucan load up any config files you want | 09:23 |
| SamYaple | for unrelated reasons i put everything from all the conf files into neutron.conf, so i dont have to do that | 09:23 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_congress master: Introduce os_congress role in gerrit https://review.openstack.org/522491 | 09:23 |
| SamYaple | im prepping for a etcd type config pull down in containers is the main reason | 09:23 |
| evrardjp | SamYaple: we would be interested, we are thinking about using etcd more, now that it's a base service. | 09:25 |
| SamYaple | evrardjp: etcd3 is needed for alot of the fancier things... you should look at pulling down and running/packaging that binary | 09:25 |
| evrardjp | it's already done | 09:25 |
| SamYaple | oh sweet! | 09:26 |
| SamYaple | fantastic. ican port over my galera stuff now if cloudnull hasnt already figured out a much better way todo it :) | 09:26 |
| evrardjp | etcd3 is ok under centos and suse (so I heard), so the only bit that was needed was ubuntu. We have a ppa thanks to lihi | 09:26 |
| evrardjp | I will deal with the politics to include this under UCA | 09:26 |
| SamYaple | andymccr: sorry for wall o text. hit me up on any implementation details and ill tell you how i got everything under nginx+uwsgi | 09:26 |
| *** vnogin has joined #openstack-ansible | 09:26 | |
| evrardjp | can someone help me on some ansible? | 09:27 |
| evrardjp | I am stuck for a few days. | 09:27 |
| SamYaple | evrardjp: `pip uninstall ansible; apt install salt` | 09:28 |
| SamYaple | huehuehue | 09:28 |
| evrardjp | :) | 09:28 |
| evrardjp | https://i.imgur.com/6arEaCj.gif?noredirect | 09:28 |
| SamYaple | where did you get a video of me! | 09:29 |
| evrardjp | hue hue hue | 09:29 |
| evrardjp | You're piling them up SamYaple ! | 09:29 |
| evrardjp | anyway have a look at this wonderful gate mystery | 09:30 |
| evrardjp | http://logs.openstack.org/24/504224/27/check/openstack-ansible-deploy-all_metal-ubuntu-xenial/1920546/job-output.txt.gz#_2017-12-10_22_37_50_294178 | 09:30 |
| evrardjp | the issue is : /openstack/venvs/keystone-17.0.0.0b2/bin/keystone-manage --debug fernet_setup --keystone-user keystone --keystone-group keystone no such file or directory | 09:30 |
| SamYaple | as it would happen.... i recently hit that | 09:30 |
| evrardjp | but as you can see a few lines above: /openstack/venvs/keystone-17.0.0.0b2/bin/keystone-manage executable: true. | 09:30 |
| evrardjp | from a stat. | 09:31 |
| SamYaple | mine was cause by two things | 09:31 |
| evrardjp | I thought it was fernet folder not existing | 09:31 |
| SamYaple | one, /etc/keystone didnt exist. two, when it did exist, it was read-only | 09:31 |
| evrardjp | but it's not the case | 09:31 |
| *** vnogin has quit IRC | 09:31 | |
| SamYaple | it will autocreate the fernet folder | 09:31 |
| SamYaple | it wont autocreate /etc/keystone though | 09:31 |
| evrardjp | let me double check that then | 09:31 |
| SamYaple | not saying thats your issue, but thats where i found it. /etc/keystone missing and /etc/keystone read only | 09:32 |
| evrardjp | http://logs.openstack.org/24/504224/27/check/openstack-ansible-deploy-all_metal-ubuntu-xenial/1920546/job-output.txt.gz#_2017-12-10_22_36_53_194422 | 09:32 |
| evrardjp | http://logs.openstack.org/24/504224/27/check/openstack-ansible-deploy-all_metal-ubuntu-xenial/1920546/job-output.txt.gz#_2017-12-10_22_36_53_950721 | 09:32 |
| evrardjp | we are creating the damn thing. | 09:32 |
| evrardjp | also, it works on my machine. | 09:32 |
| evrardjp | (yes it's still the same issue than last week and the week before that) | 09:32 |
| SamYaple | hmmm you definetly dont need to create fernet-keys, and you have the permissions wrong from when keystone *does*create it | 09:33 |
| SamYaple | i might suggest trying to 755 the keystone folder and not create the fernet-keys folder as a test | 09:33 |
| evrardjp | sure thing | 09:34 |
| *** zkynet has joined #openstack-ansible | 09:34 | |
| evrardjp | thanks SamYaple | 09:35 |
| *** hamza21 has quit IRC | 09:37 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_keystone master: [DNM] debug why on metal fails on gates and not on my machine. https://review.openstack.org/524954 | 09:39 |
| openstackgerrit | Arno Kellermann proposed openstack/openstack-ansible-os_gnocchi master: fix gnocchi ceph name variable https://review.openstack.org/527037 | 09:40 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Build everything on metal https://review.openstack.org/504224 | 09:40 |
| openstackgerrit | Arno Kellermann proposed openstack/openstack-ansible-os_gnocchi master: fix gnocchi ceph name variable https://review.openstack.org/527037 | 09:53 |
| evrardjp | andymccr: if you want you can vote on the future: https://review.openstack.org/#/c/522778/ | 09:56 |
| evrardjp | and there: https://review.openstack.org/#/c/504224/ | 09:58 |
| *** vnogin has joined #openstack-ansible | 09:58 | |
| *** mrch has joined #openstack-ansible | 09:59 | |
| evrardjp | second is maybe a little presumptuous :p | 09:59 |
| evrardjp | but it's a heads up | 09:59 |
| *** wspthr has joined #openstack-ansible | 10:00 | |
| odyssey4me | o/ I'm unfortunately still ill, but well enough to use a keyboard now. | 10:09 |
| hwoarang | good morning. odyssey4me sorry to hear that. Get well soon! | 10:11 |
| *** cshen_ has quit IRC | 10:13 | |
| *** electrofelix has joined #openstack-ansible | 10:19 | |
| *** mbuil has joined #openstack-ansible | 10:28 | |
| *** nshetty_ has joined #openstack-ansible | 10:29 | |
| *** nshetty has quit IRC | 10:29 | |
| tnogisto | anybody has a working openstack-ansible deployment that has openvswitch self-service networks? | 10:31 |
| *** nshetty_ is now known as nshetty|afk | 10:34 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts stable/pike: Remove LXC yum repo from containers https://review.openstack.org/526768 | 10:39 |
| *** stuartgr has joined #openstack-ansible | 10:39 | |
| odyssey4me | hurrah, the deploy guide for newton/ocata/pike is now updated :) | 10:42 |
| *** armaan has joined #openstack-ansible | 10:46 | |
| evrardjp | thanks for your work odyssey4me | 10:49 |
| *** vnogin has quit IRC | 10:54 | |
| *** nshetty|afk is now known as nshetty | 11:08 | |
| *** Jack_Iv has quit IRC | 11:08 | |
| *** vnogin has joined #openstack-ansible | 11:15 | |
| *** markvoelker has joined #openstack-ansible | 11:17 | |
| evrardjp | 2.4 is ready to ship for the integrated repo https://review.openstack.org/#/c/522778/ | 11:17 |
| *** Jack_Iv has joined #openstack-ansible | 11:18 | |
| *** udesale has quit IRC | 11:19 | |
| *** Jack_Iv has quit IRC | 11:23 | |
| odyssey4me | except for the merge conflict? | 11:23 |
| odyssey4me | do we have 2.4 testing for the roles? have we ironed out any resulting lint issues from that? | 11:23 |
| evrardjp | most of the issues | 11:24 |
| evrardjp | oh that's something you can help with: | 11:24 |
| evrardjp | so | 11:24 |
| odyssey4me | I see https://github.com/openstack/openstack-ansible-tests/commit/6713b5e9f7b334fef88173a939608a1af6fe5d9e | 11:24 |
| evrardjp | the tests repo was bumped to 2.4 | 11:24 |
| evrardjp | This blocks integrated gates. We need to fix the path under tests for ansible lint before merging this again. | 11:25 |
| odyssey4me | I remember seeing you describe a resulting issue, perhaps that the lint test wasn't doing the right thing? | 11:25 |
| evrardjp | so let me show you the bug | 11:25 |
| evrardjp | https://bugs.launchpad.net/openstack-ansible/+bug/1737310 | 11:26 |
| openstack | Launchpad bug 1737310 in openstack-ansible "integrated linting test is busted" [Undecided,New] | 11:26 |
| evrardjp | because ansible lint is using tests repo it makes sense that roles will be checked out in /root/.ansible/roles | 11:26 |
| evrardjp | it doesn't make sense why it's using this folder though | 11:26 |
| odyssey4me | why? in https://github.com/openstack/openstack-ansible/commit/f3dccff4ea40b498c09954ad6b9c2a6c40f95bea I switched up the lint tests to use the tests repo with the same method as all the roles so that we get a more consistent result and it's easier to manage | 11:27 |
| evrardjp | I think think we can merge the 2.4 patch, as there is only a handful of tasks that are broken into the integrated repo (I will fix them in a separate commit), and most of the other failures are from external roles (like etcd, resolvconf), due to the changes of systemd module | 11:28 |
| odyssey4me | I did that because I started hitting all sorts of weird lint failures and discovered that our previous implementation was missing all sorts of things we'd fixed in the roles. | 11:28 |
| odyssey4me | evrardjp you can't merge it - gerrit is howing a merge conflict | 11:28 |
| odyssey4me | *showing | 11:28 |
| evrardjp | is it? It's not showing on my page | 11:29 |
| evrardjp | let me refresh | 11:29 |
| evrardjp | darn | 11:29 |
| odyssey4me | Look at 'Strategy' | 11:29 |
| odyssey4me | It says 'Cannot Merge' | 11:29 |
| odyssey4me | I think some recent merges to group_vars have caused that. | 11:29 |
| evrardjp | yeah | 11:29 |
| evrardjp | I will rebase again | 11:29 |
| evrardjp | anyway | 11:30 |
| odyssey4me | if you want this merged urgently, you'll likely have to block other master merges until it's in. | 11:30 |
| evrardjp | it's not urgent, I just want to get these two done before I am on holiday | 11:31 |
| evrardjp | two-> baremetal + ansible 2.4 | 11:31 |
| evrardjp | but to come back to tests | 11:31 |
| odyssey4me | heh, merge it and go on holiday - let everyone else suffer the consequences ;) | 11:31 |
| evrardjp | I am not like that :) | 11:31 |
| evrardjp | so... I don't mind using the tests repo for linting | 11:32 |
| evrardjp | the problem is not that | 11:32 |
| evrardjp | the problem is that it doesn't work | 11:32 |
| odyssey4me | it looks like you're assuming the lint test works like the integrated build - it won't, and shouldn't | 11:32 |
| odyssey4me | where the roles are stored is immaterial for the lint test | 11:32 |
| odyssey4me | the lint test not working is something we can figure out and fix | 11:33 |
| evrardjp | the playbook that get tested is important, or the folder it resides in | 11:33 |
| evrardjp | the integrated repo linting should test linting of the integrated repo, not of everything else around it. | 11:33 |
| odyssey4me | what do you mean 'not of everything else around it' ? | 11:34 |
| evrardjp | Because it's already tested (for our roles) or we shouldn't care (for external roles) | 11:34 |
| odyssey4me | sure, but then we need to make sure that we apply the right things | 11:34 |
| evrardjp | well it's testing all the roles that are checked out into /root/.ansible/roles (which is every role checked out from zuul + even more) | 11:34 |
| evrardjp | odyssey4me: yeah that's my point in the bug | 11:34 |
| evrardjp | we should do the appropriate overrides | 11:34 |
| evrardjp | we have almost everything we need | 11:35 |
| evrardjp | like the ANSIBLE_ROLE_PATH | 11:35 |
| evrardjp | and the playbooks to test | 11:35 |
| odyssey4me | I don't understand what ANSIBLE_ROLE_PATH has to do with anything | 11:35 |
| evrardjp | but it seems that the playbooks to test is not even working | 11:35 |
| evrardjp | well if we set it to /etc/ansible/roles for the integrated, the folder that we tests for lint tests would be empty, so roles linting would be skipped | 11:36 |
| odyssey4me | yes, but we're not setting it to that | 11:36 |
| evrardjp | ROLE_DIR | 11:36 |
| odyssey4me | that var isn't used by the lint test either | 11:36 |
| evrardjp | yeah but we could that's what I meant | 11:36 |
| evrardjp | sorry I meant ANSIBLE_ROLE_DIR | 11:37 |
| odyssey4me | but ansible-lint doesn't use that var | 11:37 |
| evrardjp | https://github.com/openstack/openstack-ansible-tests/blob/master/test-ansible-env-prep.sh#L156-L160 | 11:37 |
| odyssey4me | we're telling it to test the playbooks, and by doing that it will also test the roles used byt he playbooks | 11:37 |
| *** ianychoi has quit IRC | 11:37 | |
| evrardjp | ok | 11:37 |
| evrardjp | so that doesn't matter then | 11:38 |
| odyssey4me | the roles need to be there to test the playbooks, so yes - we download the roles | 11:38 |
| odyssey4me | but we don't tell ansible-lint to test the roles, it just does that because the role is used by the playbook | 11:38 |
| evrardjp | well then there is something wrong, for lint test we download in a place that's not consistent with the integrated repo | 11:38 |
| odyssey4me | if we want to skip tests for a role, we have to tag it appropriately to be skipped | 11:38 |
| evrardjp | odyssey4me: then on my second part of my bug: see what's happening on the second playbook | 11:39 |
| odyssey4me | yes, we do that on purpose - it's a lint test, not an integrated build | 11:39 |
| evrardjp | well, long story short: I don't think we should lint all the roles again | 11:39 |
| evrardjp | and certainly not the external roles | 11:39 |
| odyssey4me | the way that ansible-lint works, we don't really have that much of a choice - but yes, I'll take a look at the bug and see what we can do there | 11:40 |
| odyssey4me | can you tell me what actually raised this as an issue, because it seems to me that linting is passing currently | 11:40 |
| evrardjp | so the linting in ansible 2.3 works | 11:40 |
| odyssey4me | can you perhaps gimme some sort of config or change I can use to replicate it | 11:40 |
| evrardjp | the linting in ansible 2.4 doesn't because we have to bump the ansible-lint version | 11:41 |
| *** askb has quit IRC | 11:41 | |
| evrardjp | when bumping ansible-lint you get more checks | 11:41 |
| evrardjp | one of it is systemd | 11:41 |
| *** vnogin has quit IRC | 11:41 | |
| evrardjp | systemd can now take daemon_reload without a service name, bare. | 11:41 |
| odyssey4me | yes, good | 11:41 |
| evrardjp | So all the places where we have systemctl daemon_reload can be replaced | 11:41 |
| odyssey4me | ok, so the trick to replicate it is to bump the ansible-lint version | 11:42 |
| evrardjp | but not for ansible 2.3 | 11:42 |
| evrardjp | yeah, for the integrated | 11:42 |
| odyssey4me | oh, and I see why you'd like to decouple the role tests from the integrated build - it allows us to fix things piece by piece | 11:42 |
| evrardjp | I mean you bump in tests for linting | 11:42 |
| odyssey4me | ok, let me see what I can figure out | 11:42 |
| evrardjp | odyssey4me: yes! | 11:43 |
| evrardjp | got it :) | 11:43 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Use Ansible 2.4 https://review.openstack.org/522778 | 11:46 |
| *** vnogin has joined #openstack-ansible | 11:49 | |
| *** ianychoi has joined #openstack-ansible | 11:50 | |
| *** markvoelker has quit IRC | 11:50 | |
| *** rodolof has joined #openstack-ansible | 11:54 | |
| *** savvas_ has quit IRC | 11:59 | |
| *** savvas has joined #openstack-ansible | 11:59 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/pike: Fix tempest plugins installation https://review.openstack.org/527068 | 12:05 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/ocata: Fix tempest plugins installation https://review.openstack.org/527069 | 12:05 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/newton: Fix tempest plugins installation https://review.openstack.org/527070 | 12:05 |
| *** zkynet has quit IRC | 12:05 | |
| *** dave-mccowan has joined #openstack-ansible | 12:06 | |
| *** Jack_Iv has joined #openstack-ansible | 12:08 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Update existing container_networks https://review.openstack.org/494515 | 12:08 |
| *** dave-mcc_ has joined #openstack-ansible | 12:13 | |
| *** dave-mccowan has quit IRC | 12:16 | |
| *** zkynet has joined #openstack-ansible | 12:19 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Build everything on metal https://review.openstack.org/504224 | 12:20 |
| *** Jack_Iv_ has joined #openstack-ansible | 12:23 | |
| *** Jack_Iv has quit IRC | 12:27 | |
| *** ppetit has joined #openstack-ansible | 12:38 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Use correct name for networking-odl git group https://review.openstack.org/521260 | 12:39 |
| *** askb has joined #openstack-ansible | 12:40 | |
| *** smatzek has joined #openstack-ansible | 12:45 | |
| *** hw_wutianwei has quit IRC | 12:45 | |
| *** markvoelker has joined #openstack-ansible | 12:47 | |
| *** savvas has quit IRC | 12:49 | |
| *** savvas has joined #openstack-ansible | 12:49 | |
| *** taseer1 has joined #openstack-ansible | 12:50 | |
| *** Taseer has quit IRC | 12:51 | |
| *** taseer2 has joined #openstack-ansible | 12:51 | |
| *** yifei has quit IRC | 12:51 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_glance stable/pike: Fix systemd init template for program override https://review.openstack.org/527080 | 12:54 |
| *** taseer1 has quit IRC | 12:55 | |
| *** ivve has quit IRC | 12:57 | |
| *** vnogin has quit IRC | 13:00 | |
| *** rodolof has quit IRC | 13:02 | |
| *** rodolof has joined #openstack-ansible | 13:02 | |
| *** udesale has joined #openstack-ansible | 13:03 | |
| *** ChanServ has quit IRC | 13:08 | |
| *** ivve has joined #openstack-ansible | 13:09 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Remove is_metal from playbooks https://review.openstack.org/527081 | 13:10 |
| *** askb has quit IRC | 13:17 | |
| *** ChanServ has joined #openstack-ansible | 13:18 | |
| *** barjavel.freenode.net sets mode: +o ChanServ | 13:18 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Remove is_metal from playbooks https://review.openstack.org/527081 | 13:20 |
| *** markvoelker has quit IRC | 13:21 | |
| *** zkynet has quit IRC | 13:23 | |
| mgariepy | good morning | 13:24 |
| tnogisto | openstack ansible is deprecating containers? | 13:26 |
| *** zkynet has joined #openstack-ansible | 13:28 | |
| odyssey4me | tnogisto nope, just making it easier to deploy without them | 13:30 |
| tnogisto | whee | 13:30 |
| ThomasS | hi | 13:32 |
| *** yifei has joined #openstack-ansible | 13:32 | |
| ThomasS | wanted to a an extra service to haproxy | 13:32 |
| ThomasS | where can i find an example for haproxy_backend_nodes: ? | 13:32 |
| ThomasS | cannot find the haproxy_config.yml | 13:32 |
| odyssey4me | ThomasS you can use the var haproxy_extra_services for that. | 13:33 |
| odyssey4me | see https://docs.openstack.org/openstack-ansible-haproxy_server/latest/configure-haproxy.html#configuring-additional-services | 13:34 |
| ThomasS | yes | 13:34 |
| ThomasS | thanks | 13:34 |
| ThomasS | how is the syntax for haproxy_backend_nodes: | 13:35 |
| ThomasS | ? | 13:35 |
| odyssey4me | it needs to contain a list of the hosts for the backend - an example is provided there | 13:35 |
| odyssey4me | the defaults are in https://github.com/openstack/openstack-ansible/blob/master/group_vars/haproxy_all/haproxy.yml#L40 for master, and probably pike too | 13:36 |
| ThomasS | ok | 13:36 |
| *** markvoelker has joined #openstack-ansible | 13:38 | |
| *** savvas has quit IRC | 13:38 | |
| ThomasS | still not so clear to me where are the vars defined | 13:38 |
| ThomasS | for example | 13:38 |
| *** savvas has joined #openstack-ansible | 13:38 | |
| ThomasS | "{{ groups['glance_registry'] | default([]) }}" | 13:38 |
| ThomasS | so where is the list is coming from? | 13:39 |
| *** ianychoi has quit IRC | 13:40 | |
| odyssey4me | it comes from the inventory | 13:40 |
| odyssey4me | it takes the list of hosts from the group named 'glance_registry' | 13:40 |
| odyssey4me | if that group doesn't exist, it provides an empty list | 13:41 |
| ThomasS | ok so if i would only like to add one ip in this case | 13:41 |
| ThomasS | how would i do it best? | 13:41 |
| ThomasS | or better easiest :-) | 13:41 |
| odyssey4me | if the ip is static, then just add that IP | 13:41 |
| odyssey4me | so it would be: "[ '<your IP address here>' ]" | 13:42 |
| ThomasS | and port with : i guess? | 13:42 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-galera_server stable/pike: Implement a proper WSREP check for galera https://review.openstack.org/527087 | 13:43 |
| odyssey4me | the port is in a different var, for example: https://github.com/openstack/openstack-ansible/blob/master/group_vars/haproxy_all/haproxy.yml#L46 | 13:43 |
| *** nshetty is now known as nshetty|afk | 13:43 | |
| ThomasS | ok | 13:45 |
| ThomasS | is the default part required? | 13:45 |
| odyssey4me | what do you mean by 'the default part' ? | 13:46 |
| ThomasS | | default([]) }}" | 13:46 |
| ThomasS | this | 13:46 |
| odyssey4me | not if you're providing a static value, no | 13:46 |
| ThomasS | ok | 13:46 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-galera_server stable/pike: Implement a proper WSREP check for galera https://review.openstack.org/527087 | 13:46 |
| *** vnogin has joined #openstack-ansible | 13:46 | |
| ThomasS | is the haproxy port then same as the target port? | 13:46 |
| odyssey4me | yes | 13:47 |
| ThomasS | so if i provide 6800 haproxy will use 6800 also to connect ? | 13:47 |
| ThomasS | ok | 13:47 |
| ThomasS | great | 13:47 |
| ThomasS | i will try now with --tags haproxy-config | 13:47 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-haproxy_server stable/pike: Add option to set the check port https://review.openstack.org/527089 | 13:49 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-haproxy_server stable/ocata: Add option to set the check port https://review.openstack.org/527090 | 13:49 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-haproxy_server stable/newton: Add option to set the check port https://review.openstack.org/527091 | 13:49 |
| evrardjp | ThomasS: run all the tags not only config :) | 13:52 |
| ThomasS | ok | 13:53 |
| ThomasS | removed --tags and only run haproxy-install.yml | 13:53 |
| ThomasS | ok now it fails | 13:55 |
| ThomasS | error looks like this | 13:56 |
| ThomasS | https://pastebin.com/vDhjE2na | 13:56 |
| ThomasS | user_vars like this: https://pastebin.com/DjHa8c7X | 13:57 |
| odyssey4me | ah, that IP is not defined as a host in the inventory, and it sems that we have something that assumes that it's a host | 13:57 |
| ThomasS | ok how can i fix this? | 13:58 |
| ThomasS | without breaking anything :-) | 13:58 |
| odyssey4me | ja, looks like this is where it's falling over: http://git.openstack.org/cgit/openstack/openstack-ansible-haproxy_server/tree/templates/service.j2#n105 | 13:58 |
| odyssey4me | so either that host will have to be in the inventory, or we'll need to patch the role up to allow a host outside of the inventory to be used as a back-end | 14:00 |
| ThomasS | can i add it to the inventory without breaking anything? | 14:00 |
| *** woodard has quit IRC | 14:00 | |
| odyssey4me | off-hand I'm not sure which is best, but I expect adding it to the inventory is easiest right now... I'm just not sure how one would do that exactly | 14:00 |
| *** woodard has joined #openstack-ansible | 14:00 | |
| odyssey4me | it'd need to be in the inventory, but not part of any of the normal groups | 14:01 |
| odyssey4me | maybe some of the clever folks in here can advise how to do that | 14:01 |
| ThomasS | ok so not that easy? | 14:01 |
| *** ppetit has quit IRC | 14:02 | |
| ThomasS | i guess editing openstack_inventory.json is not the best idea | 14:02 |
| *** udesale has quit IRC | 14:05 | |
| odyssey4me | oh no, that's absolutely the wrong idea | 14:06 |
| ThomasS | :-) | 14:06 |
| odyssey4me | is this a test or prod environment? | 14:07 |
| odyssey4me | what I mean is - can you afford to make mistakes in it | 14:08 |
| odyssey4me | (which may require a rebuild) | 14:08 |
| *** vnogin has quit IRC | 14:09 | |
| *** esberglu has joined #openstack-ansible | 14:12 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Change the galera health check for better cluster health https://review.openstack.org/527099 | 14:13 |
| *** ziitel023 has joined #openstack-ansible | 14:15 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_tempest master: Define cache_timeout by default https://review.openstack.org/527100 | 14:15 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-haproxy_server stable/pike: Resolve python lint failures https://review.openstack.org/527101 | 14:16 |
| *** taseer2 is now known as Taseer | 14:18 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-haproxy_server stable/pike: Add option to set the check port https://review.openstack.org/527089 | 14:18 |
| ziitel023 | sorry guys, i have got a problem with OVS after successful deploy all services. I tested my openstack then i found that after i reboot a host infra1 neutron plugin cant find datapathID of openvswitch. | 14:19 |
| *** nshetty|afk is now known as nshetty | 14:20 | |
| ziitel023 | Any idea this is error on infra1-neutron-agent neutron.plugins.ml2.drivers.openvswitch.agent.ovs_neutron_agent [-] No datapath_id on bridge br-int Agent terminated!: RuntimeError: No datapath_id on bridge br-int | 14:20 |
| *** shardy has quit IRC | 14:23 | |
| ziitel023 | If i restart only lxc neutron still work, but with host restart ovs-neutron-agent on infra1-neutron-agent is die. | 14:24 |
| ziitel023 | i follow https://medium.com/@travistruman/configuring-openstack-ansible-for-open-vswitch-b7e70e26009d this link for using OVS with openstack ansible | 14:26 |
| *** gouthamr has joined #openstack-ansible | 14:26 | |
| *** neith has quit IRC | 14:28 | |
| ThomasS | it is a poc | 14:28 |
| *** neith has joined #openstack-ansible | 14:28 | |
| ThomasS | sorry for the late reply | 14:28 |
| hwoarang | evrardjp: when is the ETA for m2? | 14:28 |
| *** savvas_ has joined #openstack-ansible | 14:28 | |
| evrardjp | hwoarang: it's on its way why? | 14:29 |
| *** smatzek has quit IRC | 14:30 | |
| *** savvas has quit IRC | 14:30 | |
| evrardjp | we can still release other milestones :) | 14:30 |
| *** smatzek has joined #openstack-ansible | 14:30 | |
| ThomasS | odyssey4me: so it could break as long as it can be reverted | 14:30 |
| *** smatzek has quit IRC | 14:30 | |
| ThomasS | so what can i try? | 14:31 |
| hwoarang | evrardjp: just wondering about the -2 https://review.openstack.org/#/c/525264/ | 14:31 |
| *** shardy has joined #openstack-ansible | 14:31 | |
| odyssey4me | ThomasS then I'd suggest getting an AIO up to try something in isolation. | 14:31 |
| evrardjp | hwoarang: yeah the last thing merged recently, yesterday evening or this morning, let me remove that vote | 14:31 |
| hwoarang | ok great thanks! | 14:32 |
| Mr_Smurf | when will you abandon newton? | 14:32 |
| evrardjp | odyssey4me: you're moving the content of opendaylight file right? | 14:32 |
| odyssey4me | in theory you should be able to add an arbitrary group into openstack_user_config with your host in it. That should add it to the inventory with no repurcussions... but I'm not certain of that and doing it in your PoC may cause inventory breakage which will involve some rather non-fun surgery to fix. | 14:32 |
| odyssey4me | evrardjp yep | 14:32 |
| odyssey4me | https://review.openstack.org/521260 | 14:33 |
| Mr_Smurf | will 14.2.13 be the last update? | 14:33 |
| ThomasS | the host i need is in the inventory | 14:33 |
| ThomasS | i will try to add it | 14:33 |
| tnogisto | if I need to recreate a container with different network interfaces, any hints? | 14:34 |
| odyssey4me | Mr_Smurf probably not - we're waiting for ironic and nova to EOL their kilo branches | 14:34 |
| odyssey4me | once that's done, we'll do a final update | 14:34 |
| tnogisto | wait, I has it | 14:34 |
| odyssey4me | ThomasS if it *is* in the inventory, then instead of the IP in your extra services - use the host name as ansible knows it | 14:35 |
| evrardjp | odyssey4me: thanks | 14:35 |
| ThomasS | already done | 14:35 |
| ThomasS | installing now :-) | 14:35 |
| *** savvas_ has quit IRC | 14:37 | |
| tnogisto | oki, still broken | 14:37 |
| odyssey4me | cores, I'd appreciate prompt reviews for these backports: https://review.openstack.org/#/q/I81c924464aa4b19c2a62f37b5bf26c3c0453786a | 14:37 |
| *** savvas has joined #openstack-ansible | 14:37 | |
| odyssey4me | also https://review.openstack.org/527087 | 14:37 |
| tnogisto | so, for some reason the container only gets generated with one interface, any ideas? | 14:38 |
| ThomasS | ok still fails | 14:38 |
| ThomasS | https://pastebin.com/iexsVGtz | 14:38 |
| tnogisto | also, playbook gives missing "Gather hardware addresses to be used as facts" error for only one, but should give it for 3 interfaces | 14:38 |
| odyssey4me | ThomasS ok, so you're actually trying to setup a LB for the 'ceph-mon' group, right? | 14:40 |
| odyssey4me | I take it this is for monitoring or something? | 14:40 |
| odyssey4me | it's likely best to use an active-standby config | 14:40 |
| ThomasS | no ceph-mgr | 14:40 |
| ThomasS | but only the one on infra2 | 14:41 |
| odyssey4me | ah ok | 14:41 |
| ThomasS | because this is the only one which is active | 14:41 |
| ThomasS | :-) | 14:41 |
| ThomasS | so adding all mons will not work | 14:41 |
| ThomasS | ;-) | 14:41 |
| odyssey4me | well, the name you gave seems to be something ansible can't find hostvars for | 14:41 |
| ThomasS | it is part of ceph-mon | 14:41 |
| *** yifei has quit IRC | 14:42 | |
| odyssey4me | yeah, I think it may be a formatting issue with the var not sure though | 14:42 |
| ThomasS | mhh | 14:42 |
| ThomasS | haproxy_backend_nodes: "[ 'infra2_ceph-mon_container-3c83f7fb' ]" | 14:43 |
| odyssey4me | if you do something along the lines of this does it work... | 14:43 |
| ThomasS | ? | 14:43 |
| odyssey4me | ansible -m debug -a "var=hostvars['infra2_ceph-mon_container-3c83f7fb']['ansible_host']" localhost | 14:43 |
| ThomasS | yepp success | 14:44 |
| ThomasS | "hostvars['infra2_ceph-mon_container-3c83f7fb']['ansible_host']": "10.0.100.226" | 14:44 |
| odyssey4me | can you pastebin the user_vars entry? | 14:44 |
| ThomasS | sure | 14:45 |
| odyssey4me | oh, I think I know what the issue is actually | 14:45 |
| ThomasS | https://pastebin.com/ekbrtvbL | 14:45 |
| ThomasS | ? | 14:45 |
| odyssey4me | take the quotes away - it's taking the list as a string instead of a list | 14:45 |
| ThomasS | " " | 14:46 |
| ThomasS | = | 14:46 |
| odyssey4me | so: haproxy_backend_nodes: [ 'infra2_ceph-mon_container-3c83f7fb' ] | 14:46 |
| ThomasS | ok | 14:46 |
| ThomasS | let's try | 14:46 |
| odyssey4me | of course as you suggested this should also work: haproxy_backend_nodes: "{{ groups['ceph-mon'] }}" | 14:47 |
| ThomasS | not sure | 14:47 |
| ThomasS | only one is active | 14:47 |
| odyssey4me | yeah, haproxy will configure them as back-ends, but never use them because they're always down | 14:48 |
| odyssey4me | but meh, same-same | 14:48 |
| *** weezS has joined #openstack-ansible | 14:49 | |
| ThomasS | maybe i will try this later | 14:49 |
| ThomasS | now the config is writen | 14:49 |
| ThomasS | thank you very much! | 14:49 |
| ThomasS | ok bad gateway | 14:50 |
| ThomasS | will debug the haproxy issue | 14:50 |
| ThomasS | thanks! | 14:50 |
| ThomasS | hatop show backend up | 14:53 |
| ThomasS | strange | 14:53 |
| *** nshetty has quit IRC | 14:55 | |
| *** woodard has quit IRC | 14:56 | |
| *** ziitel023 has quit IRC | 14:58 | |
| *** weezS_ has joined #openstack-ansible | 15:00 | |
| *** weezS has quit IRC | 15:03 | |
| *** weezS_ is now known as weezS | 15:03 | |
| *** rromans has quit IRC | 15:04 | |
| *** smatzek has joined #openstack-ansible | 15:06 | |
| *** indistylo has quit IRC | 15:06 | |
| *** armaan has quit IRC | 15:11 | |
| *** armaan has joined #openstack-ansible | 15:12 | |
| *** epalper has quit IRC | 15:12 | |
| *** armaan has quit IRC | 15:12 | |
| *** epalper has joined #openstack-ansible | 15:13 | |
| *** armaan has joined #openstack-ansible | 15:13 | |
| *** rromans has joined #openstack-ansible | 15:22 | |
| evrardjp | oh interesting I just managed to have a good run with neutron master. Found the problem: We need to use the plugins :p | 15:23 |
| evrardjp | glad we are fixing it on time :) | 15:23 |
| savvas | Just checking in again to see if there's anyone here familiar with cluster timeout errors on OA Magnum | 15:25 |
| odyssey4me | savvas what do you mean by 'cluster'? which cluster? there are many ;) | 15:26 |
| savvas | I get it on all types | 15:26 |
| *** rodolof has quit IRC | 15:26 | |
| *** epalper has quit IRC | 15:27 | |
| *** rodolof has joined #openstack-ansible | 15:27 | |
| savvas | docker, kubernetes etc | 15:27 |
| *** epalper has joined #openstack-ansible | 15:27 | |
| *** hw_wutianwei has joined #openstack-ansible | 15:27 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_neutron master: Include neutron plugins for neutron testing https://review.openstack.org/527127 | 15:27 |
| *** armaan has quit IRC | 15:28 | |
| savvas | it runs for a while, seems to create the VMs but then at some point I lose connectivity towards the VMs and the cluster creation fails with a timeout error odyssey4me | 15:28 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_neutron master: Include neutron plugins for neutron testing https://review.openstack.org/527127 | 15:29 |
| openstackgerrit | Merged openstack/openstack-ansible master: Avoid using 'roles' reserved variable https://review.openstack.org/524660 | 15:29 |
| *** rodolof has quit IRC | 15:29 | |
| evrardjp | savvas: I think it's probably worth asking to magnum guys | 15:29 |
| *** rodolof has joined #openstack-ansible | 15:30 | |
| *** gloksinia has joined #openstack-ansible | 15:30 | |
| evrardjp | but there are ppl here that are using magnum, so it was worth a try, not sure who is available though. | 15:30 |
| *** phalmos has joined #openstack-ansible | 15:35 | |
| *** vnogin has joined #openstack-ansible | 15:36 | |
| cloudnull | mornings | 15:38 |
| cloudnull | evrardjp: odyssey4me: I know you guys were working on the pip install things -- https://review.openstack.org/#/q/topic:pypi-index+(status:open+OR+status:merged) -- If you have a moment I think these two changes are worth getting in, assuming they're not conflicting with the larger plan. | 15:41 |
| savvas | evrardjp: I thought so but I found a particular bug report on launchpad | 15:41 |
| savvas | for using Magnum on OpenStack Ansible with the same issue | 15:41 |
| savvas | and I suspect the issue is networking related, somewhere between the Magnum containers and the VMs | 15:41 |
| savvas | Magnum logs show no errors | 15:42 |
| savvas | so it leads me back here | 15:42 |
| *** armaan has joined #openstack-ansible | 15:44 | |
| *** zkynet has quit IRC | 15:45 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Converge neutron agents onto Baremetal https://review.openstack.org/454450 | 15:47 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add nspawn container driver https://review.openstack.org/477017 | 15:47 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Update test tooling for manual testing https://review.openstack.org/527137 | 15:47 |
| odyssey4me | cloudnull hmm, those are interfering a bit -but let me look more closely | 15:49 |
| odyssey4me | right now that pip_install change will nullify the use of the repo build results - the nest step is to change the repo build so that the pip links aren't needed any more | 15:50 |
| cloudnull | it should still use the repo-build results. the pypi-server is just serving the pool and the package install is limited by the set constraints file. | 15:51 |
| cloudnull | unless I'm onfused. | 15:51 |
| cloudnull | which is possible | 15:51 |
| odyssey4me | pypiserver is not currently serving the pool at all | 15:52 |
| cloudnull | that was this change | 15:52 |
| odyssey4me | currently it serves nothing, because nothing is being put into the package folder it serves | 15:52 |
| cloudnull | https://review.openstack.org/#/c/526886/ | 15:52 |
| cloudnull | so with those two changes it works really well. | 15:52 |
| odyssey4me | oh, hmm - interesting | 15:53 |
| *** zkynet has joined #openstack-ansible | 15:53 | |
| odyssey4me | I'll give those a try, but if you don't mind I'd like to ask if you -w them for now. | 15:53 |
| cloudnull | sure | 15:53 |
| *** vnogin has quit IRC | 15:54 | |
| cloudnull | done | 15:54 |
| odyssey4me | That's a different approach, so thanks for doing it - but I'd like to test and evaluate how this affects what I was thinking of as next steps. | 15:54 |
| odyssey4me | This might be a nice interim change though to buy me some time. | 15:54 |
| cloudnull | ++ | 15:54 |
| *** openstacking_123 has joined #openstack-ansible | 15:54 | |
| odyssey4me | I'm surprised that it works with the pools folder, actually - I thought it only allowed the use of a flat folder. | 15:54 |
| cloudnull | yea it does a recursive scan which is lovely | 15:55 |
| cloudnull | especially when dealing with mixed arch | 15:55 |
| cloudnull | or distro | 15:55 |
| odyssey4me | well, that's brilliant :) | 15:55 |
| cloudnull | I mean , i've only got a 5 node cluster but its working for me :) | 15:55 |
| odyssey4me | that really is helpful to help us re-use previously build wheels - it'll still allow us to ditch all the symlinks which makes me happy | 15:55 |
| *** jvidal has quit IRC | 15:55 | |
| cloudnull | ++ | 15:56 |
| cloudnull | i was thinking the next change would be to remove the symlink tasks in repo-build | 15:56 |
| *** phalmos_ has joined #openstack-ansible | 15:56 | |
| *** epalper has quit IRC | 15:56 | |
| cloudnull | but i stopped wanting to sync with you as I knew you were working on it. | 15:56 |
| *** epalper has joined #openstack-ansible | 15:56 | |
| odyssey4me | ja, the general idea I'm thinking is to simplify the repo build in a few ways | 15:57 |
| odyssey4me | 1 - the symlinks are no longer required | 15:57 |
| evrardjp | hey | 15:57 |
| evrardjp | if you add a link to the docs showing that it works recursively I am game. Else we don't have evidence of that in our commit message. | 15:58 |
| evrardjp | let me check the other patch just to be 100% sure | 15:58 |
| evrardjp | odyssey4me: what would be good is that you share what you think your next steps are. | 15:58 |
| evrardjp | because what we have all in mind could be different | 15:58 |
| cloudnull | evrardjp: its in the server docs | 15:59 |
| odyssey4me | 2 - we don't have to pre-build wheels, then use them to build venvs... rather only pre-build wheels in the _required_distro_packages and _optional_pip_packages lists... then just build venvs. In most cases we don't use the wheels we build from git except in the venvs, so why bother building the wheels at all... just the venvs will do. | 15:59 |
| *** phalmos has quit IRC | 15:59 | |
| odyssey4me | evrardjp I would be, if I thought anyone else was working on it with me. Up until now it's just been me, and I have a notebook with what I know so far. | 15:59 |
| cloudnull | evrardjp: "pypiserver scans this directory recursively for packages. It skips packages and directories starting with a dot. Multiple package directories can be specified." https://pypi.python.org/pypi/pypiserver | 16:00 |
| cloudnull | evrardjp: are you saying i should just put a link to the package docs in my commit message? | 16:01 |
| *** gouthamr has quit IRC | 16:01 | |
| odyssey4me | cloudnull evrardjp if it's of interest, I can put up an etherpad to share my notes as I go | 16:02 |
| evrardjp | odyssey4me: I am not aware of anyone touching this except you, except now that we are 3 SME on that topic apparently :) | 16:02 |
| evrardjp | Still worth syncing. | 16:03 |
| odyssey4me | I guess I could also submit reviews to the spec repo to add work items | 16:03 |
| odyssey4me | most of this is being done in personal time, and largely by feel as I go :p | 16:03 |
| cloudnull | ++ | 16:04 |
| evrardjp | cloudnull: yes, links to docs, some tangible evidence we can do it and deserve the move for someone external. | 16:04 |
| evrardjp | odyssey4me: it's fine for me, but if there is a risk of conflicts like cloudnull 's patch, we should all be aware of it | 16:04 |
| evrardjp | also, we can help you | 16:04 |
| evrardjp | "It's dangerous to go alone" | 16:04 |
| cloudnull | it was going to implement something similar on my home cloud, but then found odyssey4me's work with the pypi-server which is briliant. | 16:04 |
| evrardjp | yes I like it too. | 16:05 |
| evrardjp | I will like it better when this will be able to move to deploy node by default, and wire haproxy to it for example until the deployment repo servers are seeded, because then it simplifies a lot of things too, and further helps the offline story. | 16:06 |
| evrardjp | but that's probably not the same goals as everyone else, so I'll patiently follow that :D | 16:06 |
| odyssey4me | I'm doing it piece by piece and trying not to break all the other work going on at the same time. | 16:07 |
| evrardjp | odyssey4me: perfect! | 16:08 |
| odyssey4me | So yeah, cloudnull I'll try not to hold it back too long - will fire up a test env this evening to test it out and validate that it's not doing anything that I think will break anything. | 16:08 |
| cloudnull | the one thing i was not able to get to work was `pip search`. | 16:09 |
| cloudnull | it works with the pypi-server | 16:09 |
| cloudnull | but not with the reverse proxy | 16:09 |
| odyssey4me | yes | 16:09 |
| odyssey4me | but that's because pypiserver will respond every time, it won't return a 404 | 16:09 |
| odyssey4me | so meh, small price to pay | 16:10 |
| cloudnull | I'm sure we can get it to work | 16:10 |
| cloudnull | i mean its no loss | 16:10 |
| cloudnull | search presently doesn't work | 16:10 |
| *** dave-mcc_ is now known as dave-mccowan | 16:11 | |
| cloudnull | so we're a lot closer than we were :) | 16:11 |
| *** DanyC has quit IRC | 16:12 | |
| evrardjp | cloudnull: so I am confused a little | 16:12 |
| evrardjp | I understand you want to have a pool into the pypiserver | 16:13 |
| evrardjp | of all our wheels | 16:13 |
| cloudnull | yup | 16:13 |
| evrardjp | but how can the search work with the reverse proxy indeed? | 16:13 |
| evrardjp | would pypi itself be the reverse proxy? | 16:13 |
| evrardjp | or is that pip that is smart enough, depending on its configuration? | 16:13 |
| evrardjp | Maybe I should learn more about pypiserver | 16:14 |
| evrardjp | what it can do | 16:14 |
| cloudnull | pypi-server is on localhost and nginx is in front | 16:14 |
| evrardjp | odyssey4me: I am also confused about what you said about what to not pre-build, and only do venv, would you care to explain a little more? | 16:14 |
| evrardjp | Because I saw optimizations of the repo build process that could be completely useless if we skip things. | 16:15 |
| odyssey4me | evrardjp currently the repo build does a build of all wheels, then uses those wheels to build venvs | 16:15 |
| evrardjp | cloudnull: that's something I understand, so if 404, redirect to upstream | 16:15 |
| evrardjp | cloudnull: that's the part I can understand easily. Now the other parts is basically the features of this pypiserver, and if it can be made smarter | 16:16 |
| odyssey4me | evrardjp what do you mean by 'smarter' ? | 16:16 |
| evrardjp | odyssey4me: please continue? | 16:16 |
| odyssey4me | evrardjp so, you understand that part - about the wheels, then venvs? | 16:16 |
| evrardjp | I know the current process you just explained yes. | 16:17 |
| * hwoarang is paying attention to the class | 16:17 | |
| evrardjp | I don't know how you arrive to skipping the wheel building | 16:17 |
| odyssey4me | ok, most of those wheels are never used again | 16:17 |
| evrardjp | and just do the venv | 16:17 |
| odyssey4me | they're used for the venvs, then they're just data | 16:17 |
| odyssey4me | nothing ever installs them | 16:17 |
| odyssey4me | the only times we ever install wheels on their own is when we do one of two things | 16:17 |
| evrardjp | developer mode which doesn't use repo | 16:18 |
| odyssey4me | 1 - we install them on hosts because they're in the _required_pip_packages list for something | 16:18 |
| evrardjp | we are only using venvs indeed, but we are ensuring the installation of pip packages during the installation process, right? | 16:18 |
| *** chyka has joined #openstack-ansible | 16:18 | |
| odyssey4me | ideally I'd like to replace those instances with putting a venv down, or using a distro package | 16:18 |
| *** weezS has quit IRC | 16:19 | |
| evrardjp | ok so you're talking about pip itself, setuptools and wheel | 16:19 |
| odyssey4me | evrardjp leave dev mode out of it for the moment - I'm speaking specifically of where a repo build is used | 16:19 |
| evrardjp | and maybe a few more | 16:19 |
| evrardjp | ok | 16:19 |
| evrardjp | still need to understand this | 16:19 |
| odyssey4me | no, in most cases it's python-keystoneclient, to facilitate ansible registering the service | 16:19 |
| evrardjp | yeah | 16:20 |
| evrardjp | ok | 16:20 |
| odyssey4me | we can replace that easily with a venv that ansible uses on all hosts instead of putting stuff down into system packages | 16:20 |
| odyssey4me | sometimes it's some other stuff for ansible, so again we either use distro packages or we use a venv | 16:20 |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: test-log-collect.sh: Collect all environment variables https://review.openstack.org/525649 | 16:20 |
| odyssey4me | but going into that is a distraction | 16:20 |
| odyssey4me | the point is that one instance of where we actually use the wheels is for the items in the _required_pip_packages lists | 16:21 |
| evrardjp | ok so you mean that if we change all the roles to USE existing venvs instead of installing required packages, we could trim down the list of wheels to build separately | 16:21 |
| odyssey4me | the other instance is where the wheel is in a _optional_pip_packages list | 16:21 |
| odyssey4me | this is where we don't install something by default, but if you enable it we add it into the venv | 16:22 |
| *** chyka has quit IRC | 16:22 | |
| odyssey4me | for example neutron plugins, horizon dashboards, etc | 16:22 |
| evrardjp | yeah, but there is one last big item if I am not mistaken | 16:22 |
| evrardjp | every other thing? | 16:22 |
| *** chyka has joined #openstack-ansible | 16:22 | |
| odyssey4me | those two sets are actually a very small set of wheels... if we only build those, instead of all of them, we can probably save a bunch of time | 16:22 |
| evrardjp | ok | 16:23 |
| evrardjp | gotcha | 16:23 |
| odyssey4me | in fact, I don't think we *need* to build them, we can just cache them unless a deployer *wants* to build them | 16:23 |
| odyssey4me | ie for offline installs | 16:23 |
| *** ppetit has joined #openstack-ansible | 16:23 | |
| evrardjp | I am still not sure | 16:23 |
| evrardjp | let me ask the question differently | 16:24 |
| odyssey4me | that will mean that instead of building, for example, nova - we use 'pip install file://nova-git-repo' directly into the venv and totally skip the wheel build entirely | 16:24 |
| evrardjp | yes, got that. | 16:24 |
| odyssey4me | this is all theory - I'm not entirely sure either - but this is what I hope will work | 16:24 |
| evrardjp | well | 16:24 |
| evrardjp | for me there is a problem with that | 16:24 |
| odyssey4me | using this method will basically kill around 50% of the repo-build role | 16:25 |
| evrardjp | if you don't actually build the wheels | 16:25 |
| evrardjp | let's take an example: https://github.com/openstack/openstack-ansible-os_rally/blob/master/defaults/main.yml#L54-L58 | 16:25 |
| evrardjp | rally needs those for example | 16:25 |
| odyssey4me | yes | 16:25 |
| evrardjp | let's talk about pymysql | 16:25 |
| evrardjp | https://github.com/openstack/openstack-ansible-os_rally/blob/master/tasks/rally_install.yml#L36 | 16:25 |
| odyssey4me | all of those except 'rally' are available via pypi... so we can install through the cache | 16:25 |
| evrardjp | ok so you mean that ALL of this would be cached, and never fetched on the repo server, which basically DECREASES the quality of offline installs | 16:26 |
| evrardjp | if we go that route, we need to make it optional | 16:26 |
| odyssey4me | yes, but we still then retain the option to build all the wheels if you want to - it just becomes optional, rather than required | 16:26 |
| evrardjp | ok | 16:27 |
| evrardjp | agreed then. | 16:27 |
| evrardjp | it will also be slower | 16:27 |
| odyssey4me | for 'rally' we just switch it up to work like dev mode does | 16:27 |
| odyssey4me | ie we do 'pip install rally -c constraints.txt' and the constraints has the appropriate git sha fixed | 16:27 |
| evrardjp | because when the cache will be invalidated, all the pip installs will go through the reverse proxy then the upstream, where the other, we have the file handy | 16:28 |
| evrardjp | but that's fine | 16:28 |
| odyssey4me | so instead of doing our complicated repo build resolution of all wheels down to one thing, we can set constraints per project (falling back to the global constraints) | 16:28 |
| evrardjp | odyssey4me: we can chain the -c now, right? | 16:28 |
| odyssey4me | so, if you have a repo, you'll still have all the venvs, and whatever wheels are required | 16:28 |
| odyssey4me | the build will just be faster, and we'll allow the switches to turn on building everything for anyone who wants to | 16:29 |
| odyssey4me | yes if you chain -c (like we do for the role tests), pip takes the first applicable constraint | 16:29 |
| evrardjp | not sure what you mean by that? We still need to build our own constraint file | 16:30 |
| odyssey4me | yes we will, but it will be way simpler | 16:30 |
| evrardjp | have I understood you correctly? | 16:30 |
| evrardjp | ok. | 16:30 |
| odyssey4me | somehting like suggested here: https://github.com/openstack/openstack-ansible-specs/blob/master/specs/queens/python-build-install-simplification.rst#L214-L224 | 16:30 |
| evrardjp | yeah I remember that part | 16:31 |
| odyssey4me | so we don't actually *build* one in our own tooling like we do today with all that complicated jinja | 16:31 |
| evrardjp | that's why I was saying chaining them | 16:31 |
| evrardjp | not understood what you meant by "first applicable constraint" | 16:31 |
| odyssey4me | we just take the vars specified, and the upstream bits, and put them into a set of files | 16:31 |
| evrardjp | yeah we just consume them | 16:31 |
| odyssey4me | so if you specify pip==9.0.1 in the first constraint file, and pip==8.1.2 in the second one - the resulting constraint will be 9.0.1 | 16:31 |
| evrardjp | ok | 16:32 |
| evrardjp | first match ruleset | 16:32 |
| odyssey4me | so we'll have a very simple override mechanism | 16:32 |
| evrardjp | ok | 16:32 |
| evrardjp | question | 16:33 |
| evrardjp | if the repo server becomes a lightweight pip thing, why don't we run this on all physical nodes? | 16:33 |
| evrardjp | or at least all infrastructure nodes | 16:34 |
| openstackgerrit | Periyasamy Palanisamy proposed openstack/openstack-ansible master: [WIP] install quagga for configuring ODL for BGP https://review.openstack.org/523907 | 16:34 |
| Tahvok | Hey guys! Can anyone take a look at this 2? https://review.openstack.org/#/c/510650 and https://review.openstack.org/#/c/510654 | 16:34 |
| cloudnull | evrardjp: its a web service we'd want to insulate the host from that, IMO | 16:34 |
| cloudnull | I dislike the pushing things to the host unless there's a reason to do it, like access to hardware devices, drives, etc. | 16:35 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-tests stable/pike: test-log-collect.sh: Collect all environment variables https://review.openstack.org/527155 | 16:36 |
| odyssey4me | evrardjp well, given that pypiserver is happy to have a flat folder of wheels to serve, there's no reason why the deploy host couldn't facilitate getting the wheels/venvs there... and if there's no repo server, then it can push to the hosts | 16:37 |
| openstackgerrit | Periyasamy Palanisamy proposed openstack/openstack-ansible master: [WIP] install quagga for configuring ODL for BGP https://review.openstack.org/523907 | 16:37 |
| cloudnull | Tahvok: https://review.openstack.org/#/c/510654 lgtm | 16:37 |
| odyssey4me | I think it may be possible down the line to have the first host of each kind do the build, then pull it back to the deploy host/repo server... but that's an idea for later down the line | 16:38 |
| evrardjp | odyssey4me: I meant that if it's just an app that runs under uwsgi, we can just follow the same process as every other app running uwsgi | 16:38 |
| cloudnull | Tahvok: I have the same question as evrardjp on https://review.openstack.org/#/c/510650/6/group_vars/all/designate.yml | 16:38 |
| odyssey4me | evrardjp yes, maybe - but with all the other apps we're still running them in containers where applicable | 16:39 |
| evrardjp | but I am going too far. | 16:39 |
| cloudnull | ++ | 16:39 |
| odyssey4me | and yes, it can run on the host if there are no containers | 16:39 |
| cloudnull | ++ | 16:39 |
| evrardjp | not if we are using fastrouter | 16:39 |
| evrardjp | then the location doesn't matter | 16:39 |
| evrardjp | you'll just get wired | 16:39 |
| cloudnull | the present implementation of pypi-server is within a unit file. seems to work really well. | 16:39 |
| evrardjp | you have a free cache available | 16:39 |
| odyssey4me | ultimately in time we'll see how we can evolve it to be even better, but for now this is a step in the right direction | 16:39 |
| cloudnull | and fronted with nginx is working great | 16:40 |
| odyssey4me | yeah, I figured putting it behind uwsgi was just another layer for no reason | 16:40 |
| cloudnull | ++ | 16:40 |
| odyssey4me | and using a systemd unit was simple | 16:40 |
| cloudnull | the current implementation is doing what the docs say is best | 16:40 |
| evrardjp | I am not critizing the code or the software itself, I am trying to understand what you are achieving, and how much it helps | 16:40 |
| odyssey4me | my only concern would be for scale... but nginx does caching so I think we're ok there | 16:40 |
| evrardjp | my concern is simple | 16:41 |
| cloudnull | also adding the pypi-server[cache] package gives it even better update performance | 16:41 |
| Tahvok | cloudnull, evrardjp: not sure what do you mean. The designate_service_adminurl is what's needed for the neutron template to work with neutron_designate_enabled: https://review.openstack.org/#/c/510654/5/templates/neutron.conf.j2 | 16:41 |
| evrardjp | the more apps we run, the harder it is. The more CACHING we have, the harder it will be to troubleshoot | 16:41 |
| cloudnull | evrardjp: odyssey4me: https://pypi.python.org/pypi/pypiserver#serving-thousands-of-packages | 16:42 |
| cloudnull | odyssey4me: if we modify the proxy setup to statically service /packages it should also improve performance | 16:42 |
| cloudnull | but not required | 16:43 |
| odyssey4me | cloudnull yeah, nice find that | 16:43 |
| cloudnull | something to add once we actually have the pypi-server serving packages. | 16:43 |
| odyssey4me | cloudnull if we did that we'd have to make a more complex setup - I don't think it's worth it | 16:43 |
| odyssey4me | the aim here is to try and make that whole repo thing simple | 16:44 |
| cloudnull | I figured, that's why I didn't change it in my other pr. | 16:44 |
| odyssey4me | fyi I actually tested just using nginx as a reverse proxy to pypi - with a bit of pypi emulation through serving from a local folder first... it sucked a bit more and was less reliable | 16:45 |
| odyssey4me | we would have had to have both the index (to the rev proxy) and extra links (like we do now) to have it work, and that would then have pip query both before installing | 16:45 |
| odyssey4me | this is way faster because it'll serve a local file if it has it build or cached | 16:46 |
| odyssey4me | if not, it'll fetch the remote file and cache it for next time | 16:46 |
| odyssey4me | it also allows a form of lock down through building a local set | 16:46 |
| odyssey4me | evrardjp to answer your question - you're right, this may not be easy to troubleshoot... but it should be easier to understand and therefore troubleshoot than the current repo build process | 16:47 |
| Tahvok | cloudnull: evrardjp: I think I understand you now. You mean I should put this variables in group_vars/neutron_all.yml ? | 16:47 |
| odyssey4me | the plan here is to remove moving parts and hard to understand scripts | 16:47 |
| mgariepy | re | 16:51 |
| evrardjp | let me summarize what I think: I like the fact we're gonna simplify each role with the constraints, and directly hit a pip server. | 16:53 |
| evrardjp | pypiserver* | 16:53 |
| evrardjp | for offline installs, we can simply tell: just roll your own. | 16:54 |
| *** Jack_Iv_ has quit IRC | 16:54 | |
| evrardjp | for online installs, we just ignore the repo server | 16:54 |
| evrardjp | that's for me the simplest ever. | 16:54 |
| odyssey4me | I disagree with that - offline installs will happen, but they will require you to pre-build your venvs and hold them somewhere. | 16:54 |
| *** weezS has joined #openstack-ansible | 16:55 | |
| *** Jack_Iv has joined #openstack-ansible | 16:55 | |
| odyssey4me | we will tool it | 16:55 |
| odyssey4me | to achieve it there's a lot more than just the venvs/wheels though | 16:55 |
| evrardjp | odyssey4me: can't we just load this pypi server up with a list of packages to have? | 16:55 |
| *** Smeared_Beard has joined #openstack-ansible | 16:55 | |
| odyssey4me | we also have to facilitate repo keys, repositories, etc | 16:55 |
| evrardjp | so for example, we produce an artifact on each tag that list all the packages we have installed into each env ? | 16:55 |
| odyssey4me | yes, that's exactly what will happen for an offline build - you have CI to build everything, then you either use that directly (online) or you package those and push them to the repo server when you build it | 16:56 |
| odyssey4me | yes, I expect we'll end up doing something like that | 16:57 |
| odyssey4me | we already do it - we just don't publish it yet | 16:57 |
| *** weezS has quit IRC | 16:57 | |
| odyssey4me | I expect we'll be able to do stuff like this in Rocky. | 16:57 |
| odyssey4me | For now, though, we have to prep the execution to work with something like that. | 16:57 |
| *** SmearedBeard has quit IRC | 16:57 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_neutron master: Handle Pike changes for Neutron https://review.openstack.org/525499 | 16:59 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Avoid using 'roles' reserved variable https://review.openstack.org/527166 | 17:01 |
| *** electrofelix has quit IRC | 17:02 | |
| *** KMT has joined #openstack-ansible | 17:03 | |
| KMT | hello | 17:04 |
| *** electrofelix has joined #openstack-ansible | 17:04 | |
| KMT | how can I integrate openstack service with in which openstack ansible does not does not support like storlets freezer designate octavia tricircle senlin | 17:05 |
| *** zkynet has quit IRC | 17:05 | |
| KMT | I want to use openstack ansible but want to use these services also | 17:06 |
| *** Jack_Iv_ has joined #openstack-ansible | 17:06 | |
| openstackgerrit | Albert Mikaelyan proposed openstack/openstack-ansible master: designate variables for neutron integration https://review.openstack.org/510650 | 17:06 |
| johnsom | KMT https://docs.openstack.org/openstack-ansible-os_octavia/latest/ and https://github.com/openstack/openstack-ansible-os_octavia | 17:06 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: [DNM] Build everything on metal https://review.openstack.org/504224 | 17:07 |
| evrardjp | stupid gates I am tired of those | 17:07 |
| evrardjp | woops | 17:07 |
| evrardjp | sorry | 17:07 |
| *** woodard has joined #openstack-ansible | 17:07 | |
| openstackgerrit | Albert Mikaelyan proposed openstack/openstack-ansible master: designate variables for neutron integration https://review.openstack.org/510650 | 17:08 |
| evrardjp | KMT octavia and designate are in. | 17:08 |
| KMT | johnsom this is for octavia what about other service I just named few | 17:08 |
| evrardjp | the rest are up to everyone, including you :) | 17:08 |
| *** Jack_Iv has quit IRC | 17:09 | |
| KMT | ask the idea how I integrage manualy | 17:09 |
| KMT | asking | 17:09 |
| evrardjp | well if you're doing the effort, why not making it automated, and sharing your work? | 17:10 |
| evrardjp | we can help you there | 17:10 |
| * mhayden opens up cloudnull's nspawn patch | 17:10 | |
| * mhayden gulps | 17:10 | |
| cloudnull | :) | 17:11 |
| cloudnull | its a lot of little changes. | 17:11 |
| KMT | :) nice but unfortunatily I am not much expert in this | 17:11 |
| mhayden | cloudnull: so you said macvlan with existing bridges, eh? | 17:11 |
| odyssey4me | KMT if you understand the service, then we can help you get the deployment automated | 17:11 |
| odyssey4me | we have enough OSA and ansible experts around, what we don't have enough of are people who understand how these peripheral services work | 17:12 |
| odyssey4me | that's the reason why no-one has built roles for them | 17:12 |
| Tahvok | evrardjp: cloudnull: please check now: https://review.openstack.org/#/c/510650 | 17:12 |
| *** epalper has quit IRC | 17:12 | |
| cloudnull | mhayden: yup | 17:12 |
| cloudnull | keeps current compat | 17:12 |
| cloudnull | I implemented https://cloudnull.io/2017/06/nspawning-openstack-ansible/ | 17:13 |
| openstackgerrit | Merged openstack/openstack-ansible-galera_server stable/pike: Implement a proper WSREP check for galera https://review.openstack.org/527087 | 17:13 |
| cloudnull | yay! | 17:13 |
| * mhayden reads | 17:14 | |
| cloudnull | Tahvok: looking now | 17:15 |
| *** armaan has quit IRC | 17:17 | |
| odyssey4me | jmccrory logan- any objections to porting https://review.openstack.org/#/q/I81c924464aa4b19c2a62f37b5bf26c3c0453786a back to ocata/newton ? | 17:18 |
| *** armaan has joined #openstack-ansible | 17:18 | |
| evrardjp | mgariepy: too I guess ^ | 17:21 |
| odyssey4me | yep, good idea - thanks | 17:22 |
| evrardjp | logan-: probably doesn't care with his own haproxy :) | 17:22 |
| evrardjp | well I don't know. | 17:22 |
| *** SimAloo has joined #openstack-ansible | 17:26 | |
| *** cshen_ has joined #openstack-ansible | 17:29 | |
| *** DanyC has joined #openstack-ansible | 17:29 | |
| *** KMT has quit IRC | 17:29 | |
| jmccrory | odyssey4me i don't mind, looks like it won't change anything unless someone overrides a service config to make use of it. was there a release note to let people know about the feature though? | 17:30 |
| odyssey4me | jmccrory nope, one wasn't added in master | 17:30 |
| *** gkadam has quit IRC | 17:30 | |
| jmccrory | oh ok | 17:30 |
| *** DanyC_ has joined #openstack-ansible | 17:30 | |
| openstackgerrit | Merged openstack/openstack-ansible-haproxy_server stable/pike: Resolve python lint failures https://review.openstack.org/527101 | 17:32 |
| openstackgerrit | Merged openstack/openstack-ansible-haproxy_server stable/pike: Add option to set the check port https://review.openstack.org/527089 | 17:32 |
| *** DanyC has quit IRC | 17:34 | |
| *** indistylo has joined #openstack-ansible | 17:36 | |
| *** eumel8 has joined #openstack-ansible | 17:37 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-openstack_openrc master: Add defaults for keystone insecure variables https://review.openstack.org/527194 | 17:38 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-galera_server master: Further simplify the upgrade check https://review.openstack.org/524086 | 17:39 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests master: Allow additional parameters for ansible-lint https://review.openstack.org/527195 | 17:41 |
| openstackgerrit | Merged openstack/openstack-ansible-os_tempest stable/pike: Fix tempest plugins installation https://review.openstack.org/527068 | 17:42 |
| *** vnogin has joined #openstack-ansible | 17:44 | |
| *** ppetit has quit IRC | 17:44 | |
| *** cshen_ has quit IRC | 17:46 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-lxc_hosts master: Create and use a local LXC RPM repo https://review.openstack.org/526739 | 17:46 |
| mhayden | hwoarang: ^^ hoping that's clearer | 17:46 |
| odyssey4me | mhayden why not just add a reverse proxy in openstack-infra? | 17:48 |
| *** cshen_ has joined #openstack-ansible | 17:48 | |
| mhayden | that helps for CI, but not for prod | 17:49 |
| mhayden | well, it marginally helps for CI | 17:49 |
| eumel8 | hi ho ... andymccr, are you around? | 17:50 |
| SamYaple | never when you need in :'( | 17:51 |
| odyssey4me | I think andymccr is on holiday, spelunking in Oz or something. | 17:51 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-lxc_hosts stable/ocata: Remove LXC yum repo from containers https://review.openstack.org/527196 | 17:51 |
| eumel8 | hahaha, is he still in Downunder? | 17:51 |
| odyssey4me | Or perhaps he'll be working from there, but has yet to arrive. It is rather a long way away. | 17:51 |
| mhayden | doesn't traveling to australia involve time travel, too ? | 17:52 |
| odyssey4me | no idea personally, but I've seen several asks for him over the day and no response... so I'm guessing that's the case | 17:52 |
| eumel8 | lucky guy, will send an email, thx | 17:52 |
| *** vnogin has quit IRC | 17:55 | |
| *** jwitko_ has joined #openstack-ansible | 17:56 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Skip ansible-lint test against roles https://review.openstack.org/527198 | 17:57 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests master: Revert "Revert "Use Ansible 2.4"" https://review.openstack.org/527200 | 18:05 |
| odyssey4me | evrardjp ^ right-o, I think that series of patches resolves the issue we discussed earlier :) | 18:06 |
| *** armaan has quit IRC | 18:06 | |
| openstackgerrit | Hugo Nicodemos proposed openstack/openstack-ansible master: Fix roles variable name https://review.openstack.org/527201 | 18:09 |
| odyssey4me | mhayden ^ you broke it again ;) | 18:09 |
| mhayden | oh crud | 18:10 |
| *** cshen_ has quit IRC | 18:12 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Avoid using 'roles' reserved variable https://review.openstack.org/527166 | 18:12 |
| eumel8 | You tried ansible 2.4 on OSA? Funny thing ;) | 18:12 |
| odyssey4me | mhayden no worries, I think evrardjp might be beating you at the moment given that he broke the integrated build for 2 weeks ;) | 18:13 |
| odyssey4me | I suppose if you aint braking stuff, you aint doin' nuthin' | 18:13 |
| odyssey4me | eumel8 it's coming: https://review.openstack.org/522778 | 18:14 |
| *** cshen_ has joined #openstack-ansible | 18:17 | |
| eumel8 | odyssey4me: hold my beer? ;) | 18:17 |
| openstackgerrit | Merged openstack/openstack-ansible-haproxy_server stable/ocata: Add option to set the check port https://review.openstack.org/527090 | 18:19 |
| openstackgerrit | Merged openstack/openstack-ansible-haproxy_server stable/newton: Add option to set the check port https://review.openstack.org/527091 | 18:19 |
| odyssey4me | eumel8 yup, we typically pin to a series for every openstack series - it helps us reduce the amount of bugs and unexpected things going on | 18:19 |
| odyssey4me | for queens we're pinning to ansible 2.4 | 18:19 |
| odyssey4me | pike is 2.3, ocata was 2.2, and newton 2.1 | 18:19 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Change the galera health check for better cluster health https://review.openstack.org/527099 | 18:21 |
| odyssey4me | ooh, zuul has some fancy new dadhboardness | 18:21 |
| odyssey4me | *dashboardness | 18:21 |
| eumel8 | odyssey4me: yeah, I saw this in the past in your code and it's the best way to protect against surprises. | 18:23 |
| odyssey4me | alright, I'm out for the night - cheers all! time for me to !computer | 18:23 |
| eumel8 | cu | 18:24 |
| spotz | night odyssey4me and welcome back:) | 18:25 |
| *** Jack_Iv_ has quit IRC | 18:28 | |
| *** jafeha has quit IRC | 18:30 | |
| *** jafeha__ has joined #openstack-ansible | 18:30 | |
| evrardjp | lol no I didn't break it for two weeks. | 18:32 |
| *** mbuil has quit IRC | 18:34 | |
| *** jafeha has joined #openstack-ansible | 18:36 | |
| *** jafeha__ has quit IRC | 18:36 | |
| *** cshen_ has quit IRC | 18:37 | |
| *** Jack_Iv has joined #openstack-ansible | 18:38 | |
| *** cshen_ has joined #openstack-ansible | 18:38 | |
| *** shardy has quit IRC | 18:40 | |
| *** hamza21 has joined #openstack-ansible | 18:41 | |
| *** Jack_Iv has quit IRC | 18:43 | |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: test-log-collect.sh: Collect package and repo information from containers https://review.openstack.org/526741 | 18:44 |
| *** cshen_ has quit IRC | 18:46 | |
| *** cshen_ has joined #openstack-ansible | 18:51 | |
| *** vnogin has joined #openstack-ansible | 18:54 | |
| *** hamza21 has quit IRC | 18:54 | |
| *** hamza21 has joined #openstack-ansible | 18:54 | |
| openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible-os_congress master: Introduce os_congress role in gerrit https://review.openstack.org/522491 | 18:54 |
| *** mrch has quit IRC | 18:55 | |
| *** MasterOfBugs has joined #openstack-ansible | 18:57 | |
| *** pramodrj07 has joined #openstack-ansible | 18:57 | |
| *** hamza21 has quit IRC | 18:58 | |
| *** poopcat has joined #openstack-ansible | 18:59 | |
| *** hamza21 has joined #openstack-ansible | 18:59 | |
| *** gkadam has joined #openstack-ansible | 18:59 | |
| *** gkadam has quit IRC | 19:02 | |
| *** gkadam has joined #openstack-ansible | 19:02 | |
| *** poopcat1 has joined #openstack-ansible | 19:03 | |
| *** armaan has joined #openstack-ansible | 19:03 | |
| *** Jack_Iv has joined #openstack-ansible | 19:06 | |
| *** poopcat has quit IRC | 19:06 | |
| *** hamza21 has quit IRC | 19:06 | |
| *** cshen_ has quit IRC | 19:07 | |
| *** hamza21 has joined #openstack-ansible | 19:07 | |
| *** MasterOfBugs has quit IRC | 19:07 | |
| *** pramodrj07 has quit IRC | 19:07 | |
| *** cshen_ has joined #openstack-ansible | 19:09 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-repo_server master: Set pypi-server to cache and use known built wheels https://review.openstack.org/526886 | 19:15 |
| *** chyka has quit IRC | 19:23 | |
| *** stuartgr has quit IRC | 19:31 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_tempest master: Define cache_timeout by default https://review.openstack.org/527100 | 19:33 |
| *** pramodrj07 has joined #openstack-ansible | 19:35 | |
| *** MasterOfBugs has joined #openstack-ansible | 19:35 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Allow disabling container restarts https://review.openstack.org/524735 | 19:44 |
| *** MasterOfBugs has quit IRC | 19:46 | |
| *** pramodrj07 has quit IRC | 19:46 | |
| *** MasterOfBugs has joined #openstack-ansible | 19:46 | |
| *** cshen_ has quit IRC | 19:46 | |
| *** pramodrj07 has joined #openstack-ansible | 19:46 | |
| *** cshen_ has joined #openstack-ansible | 19:48 | |
| *** openstacking_123 has quit IRC | 19:50 | |
| *** vnogin has quit IRC | 19:51 | |
| *** Jack_Iv has quit IRC | 19:59 | |
| *** eumel8 has left #openstack-ansible | 20:03 | |
| *** hybridpollo has joined #openstack-ansible | 20:06 | |
| *** hamza21 has quit IRC | 20:06 | |
| *** hamza21 has joined #openstack-ansible | 20:07 | |
| *** vnogin has joined #openstack-ansible | 20:07 | |
| *** vnogin has quit IRC | 20:09 | |
| *** vnogin has joined #openstack-ansible | 20:10 | |
| *** hybridpollo has quit IRC | 20:12 | |
| *** hamza21 has quit IRC | 20:12 | |
| *** hw_wutianwei has quit IRC | 20:14 | |
| *** gkadam has quit IRC | 20:17 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Use Ansible 2.4 https://review.openstack.org/522778 | 20:20 |
| *** marst has joined #openstack-ansible | 20:25 | |
| *** savvas_ has joined #openstack-ansible | 20:27 | |
| *** savvas has quit IRC | 20:30 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible master: [TEST] Update Ansible to 2.4.3.0-0.1.beta1 https://review.openstack.org/501814 | 20:37 |
| *** poopcat1 is now known as poopcat | 20:45 | |
| *** chyka has joined #openstack-ansible | 21:01 | |
| *** chyka has quit IRC | 21:05 | |
| *** liangqi has quit IRC | 21:08 | |
| *** cshen_ has quit IRC | 21:10 | |
| *** indistylo has quit IRC | 21:11 | |
| *** cshen_ has joined #openstack-ansible | 21:11 | |
| *** askb has joined #openstack-ansible | 21:18 | |
| *** smatzek has quit IRC | 21:23 | |
| *** woodard_ has joined #openstack-ansible | 21:25 | |
| *** woodard has quit IRC | 21:27 | |
| *** chyka has joined #openstack-ansible | 21:29 | |
| *** woodard_ has quit IRC | 21:29 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Use correct name for networking-odl git group https://review.openstack.org/521260 | 21:31 |
| *** threestrands has joined #openstack-ansible | 21:41 | |
| *** threestrands has quit IRC | 21:41 | |
| *** threestrands has joined #openstack-ansible | 21:41 | |
| *** Pingue_ has joined #openstack-ansible | 21:51 | |
| *** shan5464_ has joined #openstack-ansible | 21:55 | |
| *** homerp_ has joined #openstack-ansible | 21:55 | |
| *** lhinds|out has joined #openstack-ansible | 21:55 | |
| *** EmilienM has quit IRC | 21:56 | |
| *** galstrom_zzz has quit IRC | 21:56 | |
| *** lkoranda has quit IRC | 21:56 | |
| *** brad[] has quit IRC | 21:56 | |
| *** lhinds has quit IRC | 21:56 | |
| *** captaindave has quit IRC | 21:56 | |
| *** Pingue has quit IRC | 21:56 | |
| *** shananigans has quit IRC | 21:56 | |
| *** homerp has quit IRC | 21:56 | |
| *** Bico_Fino has quit IRC | 21:56 | |
| *** hybridpollo has joined #openstack-ansible | 21:57 | |
| *** galstrom_zzz has joined #openstack-ansible | 21:57 | |
| *** Bico_Fino has joined #openstack-ansible | 21:57 | |
| *** EmilienM has joined #openstack-ansible | 21:57 | |
| *** lkoranda has joined #openstack-ansible | 21:57 | |
| *** tlbr has quit IRC | 21:57 | |
| *** captaindave has joined #openstack-ansible | 21:59 | |
| *** PramodJ has joined #openstack-ansible | 22:00 | |
| *** gouthamr has joined #openstack-ansible | 22:00 | |
| *** Pramod has joined #openstack-ansible | 22:00 | |
| *** PramodJ has quit IRC | 22:01 | |
| *** Pramod has quit IRC | 22:01 | |
| *** pramodrj07 has quit IRC | 22:01 | |
| *** MasterOfBugs has quit IRC | 22:01 | |
| *** cshen_ has quit IRC | 22:02 | |
| *** tlbr has joined #openstack-ansible | 22:02 | |
| *** brad[] has joined #openstack-ansible | 22:08 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible stable/pike: Allow disabling container restarts https://review.openstack.org/527256 | 22:33 |
| *** dave-mccowan has quit IRC | 22:34 | |
| *** gouthamr has quit IRC | 22:40 | |
| *** rodolof has quit IRC | 22:53 | |
| *** SimAloo has quit IRC | 22:57 | |
| *** woodard has joined #openstack-ansible | 23:01 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Converge neutron agents onto Baremetal https://review.openstack.org/454450 | 23:02 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add nspawn container driver https://review.openstack.org/477017 | 23:02 |
| cloudnull | mhayden: ^ I had to rebase for a merge conflict, I also addressed the nits | 23:02 |
| *** threestrands_ has joined #openstack-ansible | 23:04 | |
| *** threestrands has quit IRC | 23:06 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_keystone stable/ocata: Add security headers to web accessable services. https://review.openstack.org/513974 | 23:10 |
| *** jappleii__ has joined #openstack-ansible | 23:20 | |
| *** jappleii__ has quit IRC | 23:21 | |
| *** jappleii__ has joined #openstack-ansible | 23:22 | |
| *** threestrands_ has quit IRC | 23:22 | |
| *** phalmos_ has quit IRC | 23:23 | |
| *** marst has quit IRC | 23:24 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_glance master: Update glance NFS for systemd https://review.openstack.org/526930 | 23:26 |
| openstackgerrit | Merged openstack/openstack-ansible-os_keystone master: Updated from OpenStack Ansible Tests https://review.openstack.org/525815 | 23:26 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add nspawn container driver https://review.openstack.org/477017 | 23:39 |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron stable/ocata: Ensure LBaaSv2 deployment is tested https://review.openstack.org/525656 | 23:50 |
| *** xdfil has quit IRC | 23:55 | |
| *** fdegir has quit IRC | 23:55 | |
| *** xdfil has joined #openstack-ansible | 23:55 | |
| *** johnsom has quit IRC | 23:56 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_barbican master: Updated from OpenStack Ansible Tests https://review.openstack.org/525806 | 23:56 |
| *** xar- has quit IRC | 23:57 | |
| *** portdirect has quit IRC | 23:57 | |
| *** MeltedLux has quit IRC | 23:57 | |
| *** kysse has quit IRC | 23:57 | |
| *** Adri2000 has quit IRC | 23:57 | |
| *** toan has quit IRC | 23:57 | |
| *** johnsom has joined #openstack-ansible | 23:58 | |
| *** kysse has joined #openstack-ansible | 23:59 | |
| *** Adri2000 has joined #openstack-ansible | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
