| *** gyee has quit IRC | 00:02 | |
| *** markvoelker has quit IRC | 00:05 | |
| *** tosky has quit IRC | 00:05 | |
| *** hamzy has joined #openstack-ansible | 00:09 | |
| *** cmart has quit IRC | 00:55 | |
| *** cjloader has quit IRC | 00:55 | |
| *** markvoelker has joined #openstack-ansible | 01:05 | |
| *** spsurya has joined #openstack-ansible | 01:14 | |
| *** markvoelker has quit IRC | 01:40 | |
| *** klamath has joined #openstack-ansible | 01:52 | |
| *** armaan has joined #openstack-ansible | 02:13 | |
| *** armaan has quit IRC | 02:14 | |
| idlemind | ... k updated the cell's mysql paths as suggested in os-nova irc ... now i get http://paste.openstack.org/show/720333/ still .. i can ping the host in question from hc1 so i truly don't understand the could not resolve hostname and i can't clearly see why or where a compute node (hc1, for me) is trying to ssh to the conductor node | 02:22 |
|---|---|---|
| idlemind | seems my conductor containers are not getting the ssh authorized keys file setup in /var/lib/nova/.ssh | 02:27 |
| *** markvoelker has joined #openstack-ansible | 02:37 | |
| *** esberglu has quit IRC | 02:47 | |
| *** cjloader has joined #openstack-ansible | 03:04 | |
| *** markvoelker has quit IRC | 03:10 | |
| *** dave-mccowan has quit IRC | 03:14 | |
| *** cjloader has quit IRC | 03:42 | |
| *** cjloader has joined #openstack-ansible | 03:42 | |
| *** nicolasbock has quit IRC | 03:44 | |
| *** cjloader has quit IRC | 03:47 | |
| *** klamath has quit IRC | 03:50 | |
| *** mwarad has joined #openstack-ansible | 03:57 | |
| *** dxiri has quit IRC | 03:58 | |
| *** dxiri_ has joined #openstack-ansible | 04:01 | |
| *** poopcat has quit IRC | 04:04 | |
| *** markvoelker has joined #openstack-ansible | 04:08 | |
| *** radeks has joined #openstack-ansible | 04:38 | |
| *** mma has quit IRC | 04:38 | |
| *** markvoelker has quit IRC | 04:41 | |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia stable/queens: Adds certificate generation https://review.openstack.org/565845 | 04:45 |
| *** cjloader has joined #openstack-ansible | 05:01 | |
| *** cjloader has quit IRC | 05:06 | |
| *** ianw has quit IRC | 05:06 | |
| *** berendt has quit IRC | 05:15 | |
| *** mubix has quit IRC | 05:15 | |
| *** bauruine has quit IRC | 05:15 | |
| *** sdake has quit IRC | 05:15 | |
| *** Chealion has quit IRC | 05:15 | |
| *** hogepodge has quit IRC | 05:15 | |
| *** charz has quit IRC | 05:15 | |
| *** udesale has joined #openstack-ansible | 05:15 | |
| *** sdake has joined #openstack-ansible | 05:15 | |
| *** sdake has quit IRC | 05:15 | |
| *** sdake has joined #openstack-ansible | 05:15 | |
| *** charz has joined #openstack-ansible | 05:15 | |
| *** mubix has joined #openstack-ansible | 05:15 | |
| *** Chealion has joined #openstack-ansible | 05:16 | |
| *** hogepodge has joined #openstack-ansible | 05:16 | |
| *** berendt_ has joined #openstack-ansible | 05:16 | |
| *** bauruine has joined #openstack-ansible | 05:21 | |
| *** mma has joined #openstack-ansible | 05:30 | |
| *** mma has quit IRC | 05:35 | |
| *** markvoelker has joined #openstack-ansible | 05:38 | |
| *** Taseer has quit IRC | 05:41 | |
| *** threestrands has quit IRC | 05:47 | |
| *** klamath has joined #openstack-ansible | 05:51 | |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia stable/queens: Adds certificate generation https://review.openstack.org/565845 | 05:52 |
| *** ianw has joined #openstack-ansible | 05:55 | |
| openstackgerrit | Olivier Bourdon proposed openstack/openstack-ansible-lxc_hosts stable/ocata: Fix proper versions pinning for test https://review.openstack.org/566216 | 05:55 |
| *** klamath has quit IRC | 05:56 | |
| *** cjloader has joined #openstack-ansible | 06:01 | |
| *** cjloader has quit IRC | 06:06 | |
| *** markvoelker has quit IRC | 06:11 | |
| *** evin has joined #openstack-ansible | 06:12 | |
| nsingh | hi evrardjp goodmorning.. :) | 06:22 |
| *** klamath has joined #openstack-ansible | 06:29 | |
| *** DanyC has quit IRC | 06:31 | |
| *** ppetit has joined #openstack-ansible | 06:37 | |
| *** dmsimard has quit IRC | 06:39 | |
| *** dmsimard has joined #openstack-ansible | 06:39 | |
| *** yolanda_ has joined #openstack-ansible | 06:42 | |
| *** pcaruana has joined #openstack-ansible | 06:43 | |
| *** yolanda has quit IRC | 06:45 | |
| *** cjloader has joined #openstack-ansible | 06:59 | |
| *** radeks has quit IRC | 07:00 | |
| *** mma has joined #openstack-ansible | 07:00 | |
| *** cjloader has quit IRC | 07:04 | |
| evrardjp | nsingh: good mornin | 07:04 |
| evrardjp | let's do this | 07:04 |
| evrardjp | what's your status for now? | 07:05 |
| nsingh | i installed masakarimonitor and pacemaker services. | 07:07 |
| nsingh | evrardjp: I was asking about this file http://paste.openstack.org/show/720274/ | 07:08 |
| evrardjp | and everything works fine on your machine? | 07:08 |
| evrardjp | in my case I will push infra for the base jobs let me do that now | 07:08 |
| *** markvoelker has joined #openstack-ansible | 07:08 | |
| evrardjp | use a ini file for now | 07:08 |
| nsingh | i can see the servces running on my machine not tested functionally yet | 07:08 |
| evrardjp | that will be easier | 07:08 |
| evrardjp | for now you can have an ini file into /etc/openstack_deploy/inventory.ini | 07:09 |
| evrardjp | ok | 07:09 |
| nsingh | what about the services on compute nodes. how to configure this part. | 07:10 |
| *** epalper has joined #openstack-ansible | 07:13 | |
| evrardjp | not sure what you meant there | 07:14 |
| *** cjloader has joined #openstack-ansible | 07:16 | |
| *** mbuil has joined #openstack-ansible | 07:18 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: [DNM] Test integrated build https://review.openstack.org/565025 | 07:18 |
| nsingh | we discussed about the masakari monitor services will be installed on compute hosts only right. My question is how can we setup files under /opt/openstack-ansible/inventory/env.d/ so that the services will install on compute hosts. How to setup containers there. | 07:19 |
| *** cjloader has quit IRC | 07:21 | |
| evrardjp | nsingh: if you are configuring a static inventory.ini you shouldn't worry | 07:21 |
| evrardjp | about conf.d/env.d | 07:22 |
| evrardjp | but I think you're mistaken again | 07:22 |
| evrardjp | I know you want to work on the integration of masakari | 07:22 |
| evrardjp | but the first step, as we discussed already, is making sure the role test is fine | 07:22 |
| evrardjp | the role test is far simpler | 07:22 |
| evrardjp | did you read the documentation I sent you? | 07:23 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_glance master: Add support for using distribution packages for OpenStack services https://review.openstack.org/566093 | 07:26 |
| nsingh | yes. i gone through that. | 07:29 |
| evrardjp | ok let me introduce the base jobs, so that we can start patching the role | 07:30 |
| *** klamath has quit IRC | 07:30 | |
| evrardjp | nsingh: so you mean running, on a new clean machine, ./run_tests.sh functional deploys fine? | 07:31 |
| nsingh | evrardjp: i didn't run ./run_tests.sh on a clean machine. | 07:38 |
| nsingh | i will do that. | 07:38 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: [Docs] Clarify the testing page https://review.openstack.org/566227 | 07:38 |
| evrardjp | it's on the page I gave you :) | 07:38 |
| nsingh | yes. | 07:38 |
| evrardjp | run a clean machine, git clone openstack-ansible-os_masakari, and then run all the tests there | 07:38 |
| evrardjp | linting, functional. | 07:39 |
| evrardjp | very separate things | 07:39 |
| evrardjp | because that's what will be tested by gates | 07:39 |
| evrardjp | we basically have two levels: One to make sure the role itself works, and one to integrate it with the rest in a flexible way | 07:39 |
| nsingh | ok. | 07:40 |
| *** gkadam has joined #openstack-ansible | 07:41 | |
| *** markvoelker has quit IRC | 07:42 | |
| nsingh | evardjp: sorry but where to clone openstack-ansible-os_masakari. | 07:43 |
| nsingh | is it means that i need to add masakari role under /et/ansible/roles/ and run the test | 07:45 |
| *** kukacz has quit IRC | 07:46 | |
| *** tosky has joined #openstack-ansible | 07:48 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible stable/pike: Bump lxc_host role https://review.openstack.org/566229 | 07:49 |
| evrardjp | nsingh: if you have a really new machine you don't have all of that :) | 07:50 |
| evrardjp | so you just git clone it | 07:50 |
| evrardjp | anywhere | 07:50 |
| evrardjp | I do it in /opt/ or in my /home/user/ folder | 07:50 |
| evrardjp | then run the bash script | 07:50 |
| evrardjp | simple as that | 07:50 |
| *** gkadam has quit IRC | 07:52 | |
| evrardjp | andymccr: hey, do you remember when we discussed with ttx about delaying all the milestones of two weeks because we are cycle-trailing? | 07:52 |
| *** gkadam has joined #openstack-ansible | 07:52 | |
| evrardjp | I think it was done about during Pike time | 07:52 |
| *** cjloader has joined #openstack-ansible | 07:52 | |
| *** kukacz_ has joined #openstack-ansible | 07:55 | |
| *** cjloader has quit IRC | 07:57 | |
| *** robert___ has joined #openstack-ansible | 08:00 | |
| *** DanyC has joined #openstack-ansible | 08:00 | |
| *** kukacz_ has quit IRC | 08:01 | |
| *** kukacz_ has joined #openstack-ansible | 08:01 | |
| *** DanyC has quit IRC | 08:04 | |
| *** kukacz_ has quit IRC | 08:07 | |
| *** kukacz_ has joined #openstack-ansible | 08:11 | |
| *** kukacz_ has quit IRC | 08:14 | |
| *** kukacz_ has joined #openstack-ansible | 08:14 | |
| *** vnogin has joined #openstack-ansible | 08:17 | |
| *** DanyC has joined #openstack-ansible | 08:17 | |
| *** lirt has joined #openstack-ansible | 08:22 | |
| lirt | Hello guys, can somebody give a little time for this issue https://bugs.launchpad.net/openstack-ansible/+bug/1752540? I try to find out what is problem with that, but unfortunately I am not able to troubleshoot it. | 08:24 |
| openstack | Launchpad bug 1752540 in openstack-ansible "os_nova cell_v2 discover failure" [Undecided,In progress] - Assigned to git-harry (git-harry) | 08:24 |
| *** robert___ has quit IRC | 08:25 | |
| *** vnogin has quit IRC | 08:30 | |
| *** markvoelker has joined #openstack-ansible | 08:39 | |
| openstackgerrit | fupingxie proposed openstack/openstack-ansible-os_congress stable/queens: Do not log passwords https://review.openstack.org/566240 | 08:42 |
| *** gkadam has quit IRC | 08:42 | |
| Tahvok | Hey guys! | 08:42 |
| Tahvok | I've just found a regression with my latest patch that was merged | 08:43 |
| Tahvok | https://review.openstack.org/#/c/566119/ | 08:43 |
| Tahvok | First, with this patch, the version should always be checked, even when rabbitmq_upgrade is set to True | 08:44 |
| Tahvok | Secondly, even if this is fixed, rabbitmq will still fail, as for some reason, the same stop procedure is running in the playbook itself as well (as I've just found out): | 08:44 |
| *** lbragstad has quit IRC | 08:45 | |
| Tahvok | https://github.com/openstack/openstack-ansible/blob/master/playbooks/rabbitmq-install.yml#L39 | 08:45 |
| *** lbragstad has joined #openstack-ansible | 08:45 | |
| Tahvok | I could create a fast review to remove this stop tasks, the question is, why is it there? | 08:46 |
| openstackgerrit | Albert Mikaelyan proposed openstack/openstack-ansible-rabbitmq_server master: Always check for rabbitmq version https://review.openstack.org/566243 | 08:54 |
| hwoarang | hello. friday again so the usual weekly ping for the distro packages blueprint reviews https://review.openstack.org/#/q/status:open+branch:master+topic:bp/openstack-distribution-packages | 08:55 |
| hwoarang | we now have glance into the mix | 08:55 |
| evrardjp | wow that's getting traction | 08:58 |
| evrardjp | I was busy reviewing some of those earlier today | 08:58 |
| evrardjp | :) | 08:58 |
| evrardjp | thanks for the work hwoarang | 08:58 |
| openstackgerrit | Albert Mikaelyan proposed openstack/openstack-ansible-rabbitmq_server master: Always check for rabbitmq version https://review.openstack.org/566243 | 08:58 |
| hwoarang | no problem i want to get at least one of the roles in so we have a point of reference | 08:59 |
| hwoarang | for the implementation | 08:59 |
| * hwoarang is moving to cinder | 08:59 | |
| evrardjp | hwoarang: do you mind editing this: https://review.openstack.org/#/c/562607/6 ? | 09:03 |
| evrardjp | or I can do it | 09:03 |
| hwoarang | i dont understand all the comments to be honest | 09:08 |
| hwoarang | so if you want to edit it please go ahead :) | 09:10 |
| *** markvoelker has quit IRC | 09:12 | |
| evrardjp | I think it helps reading the commit message. I was confused at first. :) | 09:16 |
| evrardjp | ok cool | 09:16 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_keystone master: Do not pip install keystoneclient on the host https://review.openstack.org/562607 | 09:21 |
| evrardjp | hwoarang: reapplied votes ^ if you don't like the commit message, please -1. | 09:21 |
| evrardjp | -w* | 09:21 |
| hwoarang | yeah should be ok | 09:23 |
| hwoarang | i also think that virtualenv can be removed too | 09:23 |
| hwoarang | because it's being installed by pip_install | 09:24 |
| hwoarang | but lets do that later on | 09:24 |
| evrardjp | yeah that's a fair point | 09:24 |
| evrardjp | or remove it from the pip_install :p | 09:24 |
| hwoarang | i deliberately put it there because it's a common dep on all roles | 09:24 |
| evrardjp | pip_install will be close to an empty shell | 09:24 |
| evrardjp | it's fine for now | 09:25 |
| evrardjp | it's clear | 09:25 |
| hwoarang | yep | 09:25 |
| evrardjp | we can slowly phase that out | 09:25 |
| evrardjp | I have on my todolist to use SC to make sure everyone is using a recent enough version of pip/setuptools/virtualenv/wheel from packages | 09:25 |
| evrardjp | so we'd never need anything pip installed on the host. | 09:26 |
| * hwoarang gives evrardjp a belgian chocolate to move it higher up in the todo list | 09:26 | |
| evrardjp | inventory still has precedence | 09:26 |
| evrardjp | but I can move that above 18.04 | 09:26 |
| evrardjp | above "use tempest venv for resource creation" | 09:27 |
| evrardjp | :) | 09:27 |
| evrardjp | hwoarang: if you could vote on https://review.openstack.org/#/c/566229/ | 09:27 |
| evrardjp | and https://review.openstack.org/#/c/565029/ | 09:27 |
| evrardjp | that would make the gates less red. | 09:28 |
| *** udesale_ has joined #openstack-ansible | 09:29 | |
| evrardjp | hwoarang: https://review.openstack.org/#/c/557896/ is also something we could bring in for more reliability. As you can see it works by extending the delays, but some delays are sometimes not good enough. | 09:31 |
| *** udesale__ has joined #openstack-ansible | 09:31 | |
| hwoarang | yep on it | 09:31 |
| evrardjp | We still have a race condition to fix that causes cinder-backup to consume 100% cpu. That would be SUPER NICE if someone could work on it, right d34dh0r53 ? :p | 09:31 |
| evrardjp | hwoarang: thanks man | 09:31 |
| *** udesale has quit IRC | 09:32 | |
| *** udesale_ has quit IRC | 09:34 | |
| *** blinkiz has quit IRC | 09:35 | |
| evrardjp | who has seen arrested development here? | 09:36 |
| Tahvok | Is it possible to revert a change? | 09:36 |
| evrardjp | yes | 09:36 |
| evrardjp | Tahvok: click on revert :p | 09:36 |
| evrardjp | but what's the issue? | 09:36 |
| hwoarang | ah these clicky clicky git UI things... | 09:37 |
| evrardjp | git revert works too :p | 09:37 |
| hwoarang | git revert <sha> && git review | 09:37 |
| evrardjp | better: | 09:37 |
| evrardjp | git checkout -b revert_topic; git revert <sha> ; git review -f | 09:37 |
| evrardjp | works on fish! | 09:37 |
| evrardjp | if you want same "successs" behavior, add "and" or replace ; with && when using bash! | 09:38 |
| evrardjp | fish is great! and there are plenty of fish in the C | 09:39 |
| evrardjp | (OMG I am seriously lacking some sleep) | 09:39 |
| hwoarang | lol | 09:39 |
| hwoarang | evrardjp: how can i convince you to get the tempest https://review.openstack.org/#/c/562606/ thing in? it's a dependency for all the distro work | 09:41 |
| Tahvok | evrardjp: do you see my messages an hour ago? | 09:43 |
| Tahvok | I can paste them here again.. | 09:44 |
| evrardjp | Tahvok: I will read them, no need to repaste. I might have missed them, yes :p | 09:44 |
| evrardjp | hwoarang: you just convinced me, if it's required and I have no alternative, there is no way to block this. | 09:45 |
| evrardjp | it's not bad, but it's not good either :p I have to tackle this when I can | 09:45 |
| *** foutatoro has joined #openstack-ansible | 09:46 | |
| evrardjp | but lower in my todo list, we just discussed this :) | 09:47 |
| mma | Hi, just noticed that keystone_ssl_protocol defaults to "{{ ssl_protocol | default('ALL -SSLv2 -SSLv3') }}". If I check nginx docs at "http://nginx.org/en/docs/http/ngx_http_ssl_module.html#ssl_protocols" there doesn't seem to be any support for "ALL", also the SSLvX shouldn't include a leading '-' character. When I tried to use 'keystone_ssl: True' this resulted in an nginx error 'nginx: [warn] invalid value "ALL"'. Easy enough | 09:48 |
| mma | for me to just override the default keystone_ssl_protocol so that is fine. Just curious if this is something that should be changed? | 09:48 |
| mma | Of if it is just me doing something wrong :) | 09:48 |
| evrardjp | Tahvok: that's very old. | 09:49 |
| evrardjp | The playbook I mean | 09:49 |
| evrardjp | for me I am fine with a later patch to remove remnants | 09:50 |
| Tahvok | I just wanted to make sure before submitting a review... | 09:50 |
| evrardjp | go ahead. If you think it's worth it, let's do it | 09:50 |
| Tahvok | It's a dependency actually. If the patch of the role merges before the playbook, the role will fail when running with rabbitmq_upgrade | 09:51 |
| foutatoro | hi all, I get this error while running the setup-hosts but the /etc/resolv.conf already exists in the host | 09:51 |
| foutatoro | fatal: [infra]: FAILED! => {"changed": false, "cmd": ["chroot", "/var/lib/lxc/LXC_NAME/rootfs", "/usr/local/bin/cache-prep-commands.sh"], "delta": "0:00:00.142970", "end": "2018-05-04 11:25:55.207465", "failed": true, "rc": 1, "start": "2018-05-04 11:25:55.064495", "stderr": "+ mkdir -p /etc/ansible/facts.d/\n+ '[' -a /etc/resolv.conf ']'\n+ echo 'nameserver 10.0.3.1'\n/usr/local/bin/cache-prep-commands.sh: line 8: /et | 09:51 |
| evrardjp | mma: you shouldn't use SSLv2 SSLv3 | 09:51 |
| foutatoro | what could be the cause of the error ? | 09:51 |
| evrardjp | mma: you can override keystone_ssl_protocol with what you want that's nginx valid | 09:52 |
| evrardjp | I don't know nginx, so you have to check their docs | 09:52 |
| evrardjp | foutatoro: sadly you're not alone. | 09:52 |
| evrardjp | the lxc images have changed. | 09:52 |
| evrardjp | so we are adapting there. | 09:52 |
| evrardjp | we have patches pending | 09:53 |
| evrardjp | foutatoro: sorry for the inconvenience | 09:53 |
| evrardjp | see https://bugs.launchpad.net/openstack-ansible/+bug/1768592 | 09:53 |
| openstack | Launchpad bug 1768592 in openstack-ansible "/usr/local/bin/cache-prep-commands.sh failing to write to /etc/resolv.conf when resolvconf used on Xenial" [Undecided,New] | 09:53 |
| *** klamath has joined #openstack-ansible | 09:54 | |
| foutatoro | evrardjp: thanks for this information | 09:54 |
| foutatoro | Is this only for ocata release ? | 09:54 |
| hwoarang | foutatoro: all releases are affected | 09:55 |
| evrardjp | we are using the same images in all our branches. | 09:56 |
| evrardjp | so we've changed master, queens is bumping shas at the moment, same for Pike. Ocata didn't get the first patch to lxc_hosts yet, so we'd need a bump of the role afterwards. | 09:56 |
| *** klamath has quit IRC | 09:59 | |
| odyssey4me | well, that's nice - the ocata one is failing due to some openssl issue... I think we fixed that in later branches | 10:01 |
| ArchiFleKs | Hi, I have some issues on master, apparently repo build is trying to acess a local repository (/var/www/repo) to locally clone ceilometer repo, but this repo is not normally cloned (skipped by ansibe when cloning the upstream repo) because I do not use ceilometer, but requierement still appear in os-release folder | 10:02 |
| odyssey4me | perhaps set the job to -nv for now, find the other patch to backport? | 10:02 |
| odyssey4me | ArchiFleKs argh, that might be one of the repositories pulling in bad requirements | 10:02 |
| odyssey4me | we've frozen some specifically for that reason | 10:02 |
| *** hamza21 has joined #openstack-ansible | 10:03 | |
| openstackgerrit | Albert Mikaelyan proposed openstack/openstack-ansible master: Rabbitmq during upgrade is stop already in role https://review.openstack.org/566269 | 10:03 |
| odyssey4me | for example: https://github.com/openstack/openstack-ansible/commit/d2e1065e5b7277cbae895a359bc4085462c5122c#diff-470e013a5a7dbed983e7e29788d3fd81 | 10:03 |
| ArchiFleKs | odyssey4me: thanks, i'll have a look, do you have a bug or an idea or where that might be happening ? You mean another project specifying ceilometer as a requierement ? | 10:03 |
| *** armaan has joined #openstack-ansible | 10:04 | |
| odyssey4me | ArchiFleKs see the note in the patch, it points to the problem | 10:04 |
| odyssey4me | and yes, the upstream projects have done some rather silly requirements specs | 10:04 |
| openstackgerrit | Albert Mikaelyan proposed openstack/openstack-ansible-rabbitmq_server master: Always check for rabbitmq version https://review.openstack.org/566243 | 10:06 |
| openstackgerrit | Albert Mikaelyan proposed openstack/openstack-ansible master: Rabbitmq during upgrade is stopped already in role https://review.openstack.org/566269 | 10:08 |
| evrardjp | odyssey4me: sorry, I didn't get the chance to fix that too... | 10:08 |
| evrardjp | adding this one that disappeared from todolist | 10:08 |
| *** markvoelker has joined #openstack-ansible | 10:08 | |
| Tahvok | Please review this 2: https://review.openstack.org/#/c/566269/ and https://review.openstack.org/#/c/566243/ | 10:09 |
| Tahvok | This is important as it fixes regression mentioned by Mohammed here: https://review.openstack.org/#/c/566119/ | 10:10 |
| odyssey4me | evrardjp if you can explain the problem and the suggested solution to ArchiFleKs then perhaps he can do the implementation in the projects :) | 10:10 |
| evrardjp | tha's fair | 10:11 |
| evrardjp | I don't have much to explain, it's case by case and explained in the commits | 10:12 |
| ArchiFleKs | odyssey4me I think it is the same issue you described | 10:12 |
| ArchiFleKs | source branch updater must have updated odl | 10:12 |
| *** Taseer has joined #openstack-ansible | 10:12 | |
| odyssey4me | ArchiFleKs yes, it will do that | 10:12 |
| ArchiFleKs | I'll check and revert the SHA1 to the one you pinned | 10:12 |
| ArchiFleKs | should be good | 10:12 |
| evrardjp | ArchiFleKs: check bgpvpn too | 10:12 |
| odyssey4me | the issue is that they've added requirements which are direct git SHA's for things that are actually test requirements, not real requirements | 10:13 |
| odyssey4me | in some cases they might be real requirements, but they should be optional | 10:13 |
| *** udesale_ has joined #openstack-ansible | 10:13 | |
| evrardjp | whether its tests or real requirements (for example bgpvpn), they should be surgical | 10:13 |
| evrardjp | odyssey4me: yes exactly. Faster to type it! | 10:13 |
| evrardjp | odyssey4me: :) | 10:13 |
| odyssey4me | so ideally we need to work with the upstream projects to change them to be specified as extras, so the requirement can be something like 'foo[bar]' if you want support for bar included | 10:14 |
| *** udesale__ has quit IRC | 10:16 | |
| olivierb- | morning everyone | 10:17 |
| ArchiFleKs | I see, thanks for the input evrardjp odyssey4me | 10:17 |
| olivierb- | evrardjp odyssey4me , concerning the issue with ocata backport https://review.openstack.org/#/c/566054/ I am wondering if there are some missing backports to ocata branch in openstack-ansible-pip_install ? | 10:18 |
| olivierb- | just a guess as I am not at all familiar with all these | 10:18 |
| odyssey4me | as far as I remember, centos on ocata is broken anyway - trying to unwind it to make it all work right is not going to be easy given all the hcanges that were made in pike to get it right | 10:20 |
| odyssey4me | in ocata we still used yum, in pike we switched to dnf | 10:21 |
| olivierb- | ok I see, no pb | 10:22 |
| odyssey4me | hmm, wait a minute | 10:22 |
| odyssey4me | https://bugs.launchpad.net/openstack-ansible/+bug/1708493 | 10:22 |
| openstack | Launchpad bug 1708493 in openstack-ansible "Bindep files expect python2-pyOpenSSL to be present" [High,Fix released] - Assigned to Major Hayden (rackerhacker) | 10:22 |
| odyssey4me | yep, that fix went into pike but not ocata | 10:23 |
| odyssey4me | lemme do a port quick | 10:23 |
| olivierb- | sure thanks | 10:23 |
| odyssey4me | hmm, it appears the problem package doesn't get used by ocata through bindep | 10:24 |
| odyssey4me | so that's not exactly it | 10:24 |
| *** radeks has joined #openstack-ansible | 10:25 | |
| *** ethfci has joined #openstack-ansible | 10:27 | |
| ethfci | Hello, I have the following problem since today: | 10:28 |
| ethfci | TASK [lxc_hosts : Ensure that the LXC cache has been prepared] ***************** | 10:28 |
| ethfci | Friday 04 May 2018 09:59:39 +0000 (0:00:00.351) 0:09:38.572 ************ | 10:28 |
| ethfci | fatal: [infra1]: FAILED! => {"ansible_job_id": "430350509737.8837", "attempts": 1, "changed": true, "cmd": "chroot /var/lib/machines/ubuntu-xenial-amd64 /usr/local/bin/cache-prep-commands.sh > /var/log/lxc-cache-prep-commands.log 2>&1", "delta": "0:00:00.028542", "end": "2018-05-04 09:59:37.935611", "failed": true, "finished": 1, "msg": "non-zero return code", "rc": 1, "start": "2018-05-04 09:59:37.907069", "stderr | 10:28 |
| ethfci | ": "", "stderr_lines": [], "stdout": "", "stdout_lines": []} | 10:28 |
| ethfci | fatal: [log1]: FAILED! => {"ansible_job_id": "369551530620.8473", "attempts": 1, "changed": true, "cmd": "chroot /var/lib/machines/ubuntu-xenial-amd64 /usr/local/bin/cache-prep-commands.sh > /var/log/lxc-cache-prep-commands.log 2>&1", "delta": "0:00:00.022634", "end": "2018-05-04 09:59:38.412639", "failed": true, "finished": 1, "msg": "non-zero return code", "rc": 1, "start": "2018-05-04 09:59:38.390005", "stderr": | 10:28 |
| ethfci | "", "stderr_lines": [], "stdout": "", "stdout_lines": []} | 10:28 |
| ethfci | on stable/queens | 10:28 |
| odyssey4me | olivierb- it looks like https://review.openstack.org/502489 is the thing which changed things up significantly to resolve this issue | 10:29 |
| openstackgerrit | Merged openstack/openstack-ansible-os_masakari master: Updated from OpenStack Ansible Tests https://review.openstack.org/566007 | 10:30 |
| *** vnogin has joined #openstack-ansible | 10:30 | |
| olivierb- | Indeed, seems like it, does this mean that it should be backported to ocata branch ? | 10:30 |
| *** radeks has quit IRC | 10:32 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-pip_install stable/ocata: Don't install any packages via pip https://review.openstack.org/566278 | 10:34 |
| *** vnogin has quit IRC | 10:35 | |
| odyssey4me | olivierb- it's a fairly significant change to backport - I hope there's no other fall out... but it's either this or removing centos support | 10:37 |
| odyssey4me | evrardjp I think https://review.openstack.org/566278 might be needed to get the lxc_hosts change for ocata passing | 10:38 |
| evrardjp | if you remove centos support, please update the meta | 10:38 |
| evrardjp | ethfci: yes we are on it | 10:38 |
| evrardjp | a new patch should land very quick | 10:38 |
| odyssey4me | removing centos support would require removal across all roles, and I don't have time to execute that right now | 10:39 |
| evrardjp | maybe just mark it nv if this one hurts more than fixes? | 10:39 |
| evrardjp | let's see how this bp part goes | 10:40 |
| odyssey4me | well, if lxc_hosts won't even build for centos - then everything is broken :p | 10:40 |
| evrardjp | fair fair :) | 10:40 |
| odyssey4me | by the way - perhaps it'd be good to add a reno about this issue, and share the workaround which can be implemented for earlier releases? | 10:40 |
| evrardjp | well today I saw orange in gates everywhere | 10:40 |
| odyssey4me | I can put something together now | 10:40 |
| evrardjp | not really encouraging isn't it? :p | 10:41 |
| evrardjp | odyssey4me: oh I didn't think of that | 10:41 |
| evrardjp | adding resolvers manually or something? | 10:41 |
| odyssey4me | lemme put one together for each branch | 10:41 |
| odyssey4me | I think you can just add this to user_variables.yml: lxc_cache_prep_pre_commands: rm -f /etc/resolv.conf | 10:42 |
| evrardjp | better you than me at writing english, indeed. | 10:42 |
| odyssey4me | given that's what we do now with https://review.openstack.org/#/c/566046/1/vars/redhat-7.yml | 10:42 |
| evrardjp | isn't the fault due to no resolv.conf being there? | 10:42 |
| evrardjp | oh | 10:42 |
| *** markvoelker has quit IRC | 10:42 | |
| odyssey4me | once the role SHA's merge, are you going to rush a release? | 10:43 |
| ethfci | resolv.conf is auto put by libc!s resolver | 10:43 |
| evrardjp | today I can't odyssey4me . But yes I'd like to do so. | 10:43 |
| evrardjp | I will probably only get the time to work on this on Sunday evening, for personal reasons. | 10:43 |
| evrardjp | but that's still acceptable. | 10:44 |
| evrardjp | if a reno is published | 10:44 |
| odyssey4me | ethfci if you add this to your user_variables.yml file, then retry your build, it should work: lxc_cache_prep_pre_commands: "rm /etc/resolv.conf || true" | 10:44 |
| odyssey4me | yeah, that's fine - the release won't action until next week anyway | 10:44 |
| evrardjp | yeah | 10:44 |
| ethfci | thanks guys so much | 10:44 |
| evrardjp | but we need many of those in for sunday, which I can't drive | 10:45 |
| odyssey4me | ethfci could you try that please to confirm for me while I prepare a release note to advise | 10:45 |
| evrardjp | I think queens is on its way | 10:45 |
| evrardjp | Pike too | 10:45 |
| ethfci | oke give me 2 hours | 10:45 |
| evrardjp | but O and N aren't | 10:45 |
| evrardjp | not that N matters, as N should be EOL. | 10:45 |
| odyssey4me | I can do a reno for N, but that's all - all roles are EOL | 10:46 |
| evrardjp | Maybe I should re-ping tonyb :) | 10:46 |
| odyssey4me | I'll work on prepping a patch to switch master to using diskimage-builder on Monday. This kind of stuff is getting old. | 10:48 |
| odyssey4me | It's time we took control. | 10:48 |
| evrardjp | that's a big step. | 10:48 |
| evrardjp | I understand, but I also see the risk of maintaining this. | 10:49 |
| evrardjp | I know it's not recommended to use infra images, but it's SO tempting. | 10:50 |
| evrardjp | anyway, I am fine with that. | 10:51 |
| evrardjp | python build can be delayed a little more | 10:51 |
| odyssey4me | well, I'm not sure exactly what you mean by 'use infra images' - if you mean we add the code to produce out images, then use that to publish our own images... then perhaps we can do that to help ci builds go faster... but there are no existing images we can use in openstack-infra at this point | 10:52 |
| evrardjp | odyssey4me: slightly related question: do you think that the process to build images with dib could be extended for operators to manage their clouds' glance images? | 10:54 |
| odyssey4me | yes | 10:54 |
| *** epalper has quit IRC | 10:54 | |
| evrardjp | in other words: We have a process, and we output rootfs for lxc/nspawn + something for glance | 10:55 |
| evrardjp | because if images can be built in a CI way, it's definitely positive for cloud operators | 10:55 |
| odyssey4me | I won't target that to start with, but it can be extended to do that. | 10:55 |
| evrardjp | because that's a good operational story. | 10:56 |
| *** spsurya has quit IRC | 10:56 | |
| evrardjp | "we have a process to ensure you have rotating, up to date, glance images" | 10:56 |
| odyssey4me | yeah, perhaps I'll add a user story to the TODO to make use of diskimage-builder for image building and publishing into glance. | 10:56 |
| evrardjp | I think it can be even further than a user story | 10:56 |
| evrardjp | be part of the operations guide. | 10:57 |
| evrardjp | but that's maybe later, as first it could be a user story | 10:57 |
| evrardjp | but that's very long term | 10:57 |
| ethfci | odyssey4me: your idea's checking on its way | 10:58 |
| evrardjp | ethfci: thanks. | 10:58 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/queens: Add release note for resolvconf cache prep failure https://review.openstack.org/566287 | 10:59 |
| evrardjp | ethfci: please review the release note odyssey4me just published here ^ | 10:59 |
| odyssey4me | thank ethfci - if you can confirm that it makes things work then I'll port that back to each stable branch | 10:59 |
| ethfci | odyssey4me i will vote on your review okay? | 11:06 |
| odyssey4me | ethfci yes, perfect thank you | 11:09 |
| *** cjloader has joined #openstack-ansible | 11:12 | |
| *** dave-mccowan has joined #openstack-ansible | 11:14 | |
| evrardjp | cores, could you vote on https://review.openstack.org/#/c/566278/1 please? | 11:14 |
| odyssey4me | do we have a https://review.openstack.org/566229 for queens? | 11:15 |
| odyssey4me | I see that https://review.openstack.org/565933 has merged, so unless I'm missing something the integrated sha bump is missing | 11:15 |
| *** cjloader has quit IRC | 11:17 | |
| evrardjp | odyssey4me: yes, the sha bump of next queens didn't merge yet | 11:18 |
| evrardjp | odyssey4me: https://review.openstack.org/#/c/565029/ | 11:18 |
| evrardjp | it's checking right now | 11:18 |
| odyssey4me | evrardjp ah I see - and the patch for lxc_hosts is in that a-r-r update? | 11:18 |
| evrardjp | yes | 11:19 |
| openstackgerrit | Andy McCrae proposed openstack/ansible-config_template master: Allow top section in ini to be specified https://review.openstack.org/566292 | 11:19 |
| evrardjp | free! | 11:19 |
| odyssey4me | ok, it might have been nice to mark the related bug in the commit msg - but meh | 11:19 |
| evrardjp | yeah I thought of that ... then... well it's already in :p | 11:21 |
| *** cjloader has joined #openstack-ansible | 11:21 | |
| *** udesale__ has joined #openstack-ansible | 11:22 | |
| *** udesale__ has quit IRC | 11:22 | |
| *** udesale_ has quit IRC | 11:24 | |
| *** cjloader has quit IRC | 11:26 | |
| *** pcaruana has quit IRC | 11:27 | |
| *** jwitko has quit IRC | 11:36 | |
| *** markvoelker has joined #openstack-ansible | 11:39 | |
| *** pcaruana has joined #openstack-ansible | 11:39 | |
| *** nicolasbock has joined #openstack-ansible | 11:42 | |
| *** foutatoro has quit IRC | 11:53 | |
| *** vnogin has joined #openstack-ansible | 11:54 | |
| *** klamath has joined #openstack-ansible | 11:55 | |
| *** vnogin has quit IRC | 11:56 | |
| *** pcaruana has quit IRC | 11:59 | |
| *** klamath has quit IRC | 12:00 | |
| *** gkadam has joined #openstack-ansible | 12:09 | |
| *** pcaruana has joined #openstack-ansible | 12:12 | |
| odyssey4me | ethfci has your test passed the lxc_hosts execution stage yet? | 12:12 |
| *** markvoelker has quit IRC | 12:12 | |
| odyssey4me | lol, I see you just reviewed - excellent, thanks! | 12:12 |
| ethfci | odyssey4me yes, voted on your review :D | 12:13 |
| *** cjloader has joined #openstack-ansible | 12:13 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Add release note for resolvconf cache prep failure https://review.openstack.org/566298 | 12:14 |
| mgariepy | morning everyone | 12:15 |
| odyssey4me | o/ mgariepy how're things? | 12:17 |
| openstackgerrit | Olivier Bourdon proposed openstack/openstack-ansible master: Add option to change fs type on bootstrap device https://review.openstack.org/566299 | 12:17 |
| *** cjloader has quit IRC | 12:17 | |
| openstackgerrit | Olivier Bourdon proposed openstack/openstack-ansible master: Factorize code for better readbility https://review.openstack.org/566300 | 12:18 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/ocata: Add release note for resolvconf cache prep failure https://review.openstack.org/566301 | 12:19 |
| openstackgerrit | Olivier Bourdon proposed openstack/openstack-ansible master: Add more meaningful/user readable failure https://review.openstack.org/566302 | 12:19 |
| mnaser | morning folks | 12:20 |
| *** ethfci has quit IRC | 12:21 | |
| odyssey4me | o/ mnaser | 12:22 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/newton: Add release note for resolvconf cache prep failure https://review.openstack.org/566304 | 12:23 |
| *** gkadam has quit IRC | 12:24 | |
| *** markvoelker has joined #openstack-ansible | 12:24 | |
| odyssey4me | cores, if we could get https://review.openstack.org/#/q/I7549e6710727d71f05dbc4082f543536280e9d85 reviewed asap so that it's published soon, that'd be swell - this issue is hitting a lot of people, so the sooner we provide the workaround, the better | 12:24 |
| *** electrofelix has quit IRC | 12:26 | |
| nsingh | i ran ./run_tests for masakari. but it is failing for "Could not find a version that satisfies the requirement pbr===4.0.2" and i also check for other roles like nova, cinder. these are also failing for the same. | 12:27 |
| openstackgerrit | Merged openstack/openstack-ansible-pip_install stable/ocata: Don't install any packages via pip https://review.openstack.org/566278 | 12:27 |
| openstackgerrit | Olivier Bourdon proposed openstack/openstack-ansible master: Add more meaningful/user readable failure https://review.openstack.org/566302 | 12:28 |
| mgariepy | odyssey4me, reviewed. lgtm | 12:30 |
| odyssey4me | thx mgariepy | 12:30 |
| mgariepy | thing are not too bad, i have enough project to keep me busy ... haha | 12:31 |
| odyssey4me | nsingh hmm, that's odd - can you paste a complete output to where it fails? | 12:31 |
| mgariepy | how about you ? | 12:31 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova master: Define lxd.pool in nova.conf based on lxd_storage_pool https://review.openstack.org/565891 | 12:31 |
| odyssey4me | yeah, my day job's keeping me busy - building out nodepool infrastructure so that we use it for all our jenkins testing | 12:32 |
| mnaser | odyssey4me: lgtm | 12:34 |
| mnaser | odyssey4me: sorry, had to abandon/restore to get it out of ci to verify something first - https://review.openstack.org/#/c/566304/ | 12:37 |
| mnaser | the change https://review.openstack.org/#/q/I3769053956b958a0ac7cab499ea90af2a40de850 never went to stable/newton | 12:37 |
| mnaser | do i understand that the workaround doesnt dpened on that change? | 12:37 |
| *** cmart has joined #openstack-ansible | 12:38 | |
| *** armaan has quit IRC | 12:38 | |
| odyssey4me | mnaser yep, that's why the workaround in the release note says to apply to all newton releases | 12:38 |
| mnaser | odyssey4me: ok, thats what i thought | 12:38 |
| odyssey4me | there's no dependency | 12:38 |
| *** armaan has joined #openstack-ansible | 12:38 | |
| odyssey4me | thee newton roles are EOL, so the only option for a deployer is to fork and apply the patch to the role - or apply the workaround | 12:38 |
| mnaser | oh i see | 12:40 |
| odyssey4me | the newton integrated repo should also be eol by now, but I guess the release team is busy - so I pushed the additional reno as it may be handy for someone to reference | 12:41 |
| mnaser | maybe if evrardjp is around he can give a poke at #openstack-release | 12:42 |
| odyssey4me | yeah, it's on his TODO list and he pokes tonyb from time to time | 12:43 |
| evrardjp[m] | ok | 12:43 |
| *** hamza21 has quit IRC | 12:44 | |
| nsingh | odyssey4me: http://paste.openstack.org/show/720363/ | 12:44 |
| odyssey4me | nsingh ah, I see the issue - your host is configured to use the repo server for some reason... role tests run against pypi directly, and won't work in an integrated build setupp | 12:45 |
| odyssey4me | the repo server doesn't have that package | 12:46 |
| *** geb has quit IRC | 12:47 | |
| nsingh | odyssey4m: ooh so how can i fix this. | 12:48 |
| odyssey4me | nsingh remove /root/.pip/pip.conf - or use a clean host | 12:50 |
| *** geb has joined #openstack-ansible | 12:51 | |
| *** hamza21 has joined #openstack-ansible | 12:53 | |
| *** udesale has joined #openstack-ansible | 12:54 | |
| nsingh | odyessey4me: it works :) | 12:57 |
| odyssey4me | hwoarang evrardjp[m] we discovered the other day that pip_install is never run in the role tests unless the role test specifically does it, right? would there be any issues running the role in https://github.com/openstack/openstack-ansible-tests/blob/master/test-setup-host.yml so that it's almost universally executed? | 12:57 |
| hwoarang | odyssey4me: i reverted the patch and now pip_install is executed on openstack_hosts as before | 12:58 |
| odyssey4me | hwoarang huh? | 12:58 |
| hwoarang | odyssey4me: https://review.openstack.org/#/c/565873/ | 12:59 |
| odyssey4me | oh, we do it for centos - but we use packages for opensuse/ubuntu, right? | 12:59 |
| hwoarang | is that what you mean? | 13:00 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts stable/ocata: Fix lxc cache prep resolvers https://review.openstack.org/566054 | 13:00 |
| odyssey4me | hwoarang oh, so now we're installing the same packages in openstack_hosts and in pip_install | 13:01 |
| odyssey4me | oh, looks like we aren't | 13:01 |
| hwoarang | they should be similar | 13:02 |
| odyssey4me | we just skip the get-pip stuff in pip_install now because openstack_hosts has already put the packages there | 13:02 |
| hwoarang | get-pip only happens on centos | 13:02 |
| hwoarang | because it comes with such an old pip | 13:02 |
| hwoarang | ubuntu also needs setuptools from pip because i said yesterday the setuptools from ubuntu splits pkg-resources to a separate package, so using 'virtualenv' from distro with our index-url pointing to the repo-build or openstack-ci mirror does not have pkg-resources there and pip fails | 13:04 |
| odyssey4me | ah, so that was the issue - awesome, thanks for figuring that out! | 13:04 |
| hwoarang | it's not easy to explain all i can do is to point you at this http://logs.openstack.org/06/562606/12/check/openstack-ansible-functional-ubuntu-xenial/de2aedd/job-output.txt.gz#_2018-05-02_11_04_29_831359 | 13:05 |
| hwoarang | you see a 404 http error there basically saying that pkg-resources is missing from our index | 13:05 |
| odyssey4me | hwoarang I guess that opensuse doesn't auto-start nginx if I install the package... much like centos - only ubuntu does that, right? | 13:05 |
| hwoarang | yep | 13:05 |
| odyssey4me | yeah, I saw that issue some time ago, but didn't get time to dig into it - I'm grateful that you have managed to! | 13:06 |
| odyssey4me | ok, so https://review.openstack.org/565873 is moving along now and should merge in a bit | 13:06 |
| odyssey4me | with that I can do my revisions to https://review.openstack.org/564778 with the assumption that openstack_hosts is taking care of pip/virtualenv correctly for each distro | 13:07 |
| hwoarang | right | 13:07 |
| evrardjp | that would be the plan | 13:08 |
| evrardjp | openstack_hosts is doing what's right all accross the board | 13:08 |
| odyssey4me | evrardjp https://review.openstack.org/566054 is now merged, so we can do an integrated build patch to bring that and https://review.openstack.org/566278 in which seems necessary | 13:11 |
| evrardjp | I have to go for today, but I can help with the reviews on the way | 13:11 |
| evrardjp | cool it passed | 13:11 |
| odyssey4me | ok, lemme push a patch up now then | 13:11 |
| evrardjp | thanks | 13:12 |
| evrardjp | if you ping me I will review in the train. | 13:12 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/ocata: Update lxc_host/pip_install role SHA's https://review.openstack.org/566309 | 13:15 |
| odyssey4me | evrardjp ^ done | 13:15 |
| odyssey4me | cores - please track and babysit https://review.openstack.org/#/q/I8523c34af069d097c9ffb46cbc9a16f87cc84fb1 and https://review.openstack.org/565029 which need to merge asap to get a release out early next week which solves the resolvconf related cache prep failure for new builds | 13:16 |
| *** vnogin has joined #openstack-ansible | 13:17 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/pike: Bump lxc_host role https://review.openstack.org/566229 | 13:17 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: Add release note for resolvconf cache prep failure https://review.openstack.org/566287 | 13:17 |
| openstackgerrit | Merged openstack/openstack-ansible stable/pike: Add release note for resolvconf cache prep failure https://review.openstack.org/566298 | 13:17 |
| openstackgerrit | Merged openstack/openstack-ansible stable/ocata: Add release note for resolvconf cache prep failure https://review.openstack.org/566301 | 13:17 |
| openstackgerrit | Merged openstack/openstack-ansible stable/newton: Add release note for resolvconf cache prep failure https://review.openstack.org/566304 | 13:17 |
| odyssey4me | heh, ok one done :) | 13:17 |
| hwoarang | evrardjp: btw re https://review.openstack.org/#/c/562606/ the experimental job fails on repo-build prob due to an unrelated issue http://logs.openstack.org/06/562606/16/experimental/openstack-ansible-integrated-deploy-aio/4840096/logs/openstack/aio1_repo_container-1fd5c0fd/repo/wheel_build.log | 13:22 |
| *** klamath has joined #openstack-ansible | 13:22 | |
| *** dxiri_ has quit IRC | 13:23 | |
| odyssey4me | hwoarang argh, that looks painful | 13:25 |
| hwoarang | yeah can't decipher it yet | 13:26 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_cinder master: defaults: Allow uwsgi to autoload required plugins. https://review.openstack.org/566314 | 13:26 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_cinder master: Add support for using distribution packages for OpenStack services https://review.openstack.org/566315 | 13:26 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Revise the role to properly cater to use-cases https://review.openstack.org/564778 | 13:27 |
| *** esberglu has joined #openstack-ansible | 13:28 | |
| openstackgerrit | Merged openstack/openstack-ansible-openstack_hosts master: task: main: Run the pip_install role on all distros https://review.openstack.org/565873 | 13:31 |
| evrardjp[m] | hwoarang: do you have the requirements file for that build ? | 13:37 |
| d34dh0r53 | anyone seeing repo-build only putting wheels in one container? third mnaio I've build and /var/www/repo/pools is empty on infra1 and infra2 repo containers. (stable/queens) | 13:39 |
| d34dh0r53 | *built | 13:39 |
| *** esberglu has quit IRC | 13:39 | |
| *** esberglu has joined #openstack-ansible | 13:39 | |
| d34dh0r53 | evrardjp: I may be able to look at that today but my bucket is rather full right now, and now I can't even get a mnaio to build properly :/ | 13:41 |
| *** vnogin has quit IRC | 13:43 | |
| evrardjp[m] | check the rsyncd | 13:44 |
| *** mma has quit IRC | 13:50 | |
| *** mma has joined #openstack-ansible | 13:51 | |
| *** armaan has quit IRC | 13:52 | |
| odyssey4me | d34dh0r53 nope, lsync is supposed to push the content to the other repo containers | 13:54 |
| odyssey4me | lsync executes rsync | 13:54 |
| *** cjloader has joined #openstack-ansible | 13:55 | |
| d34dh0r53 | k, rsync is running but nothing in the logs, looking at lsync now | 13:55 |
| odyssey4me | it's been a long time since any of that changed - so maybe lsync, rsync or something in openssh changed :/ | 13:55 |
| odyssey4me | it's plausible that lsync changed in some way that breaks our lua script | 13:56 |
| hwoarang | evrardjp: prob not. we do not seem to collect such info on the logs | 13:56 |
| d34dh0r53 | odyssey4me: so this https://github.com/openstack/openstack-ansible-repo_build/blob/stable/queens/tasks/repo_package_sync.yml is outdated? | 13:57 |
| odyssey4me | d34dh0r53 That's for pulling multi-arch changes from secondary repo hosts back to the master for redistribution from there - so that is unlikely to be executing unless you have a multi-arch/multi-distro environment. | 13:58 |
| d34dh0r53 | ahh, ok | 13:58 |
| odyssey4me | this is the lua script: https://github.com/openstack/openstack-ansible-repo_server/blob/master/templates/lsyncd.lua.j2 | 13:59 |
| d34dh0r53 | odyssey4me: is there an easy way for me to manually test the lua script? | 13:59 |
| *** cjloader has quit IRC | 13:59 | |
| odyssey4me | these are the tasks setting up the sync management: https://github.com/openstack/openstack-ansible-repo_server/blob/master/tasks/repo_sync_manager.yml | 13:59 |
| *** mma has quit IRC | 13:59 | |
| odyssey4me | d34dh0r53 change something in one of the folders under management, I guess? | 13:59 |
| d34dh0r53 | I can't force a sync? | 14:00 |
| odyssey4me | That stuff hasn't changed all that much since it was initially built in kilo - I think we did some small changes in liberty to take the secondary repo servers offline wwhile syncing | 14:00 |
| odyssey4me | d34dh0r53 no idea | 14:00 |
| d34dh0r53 | yeah, it's really weird, this was working fine 2 days ago | 14:01 |
| odyssey4me | I've never executed a multi-node build, so I've never touched any of that. | 14:01 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Revise the role to properly cater to use-cases https://review.openstack.org/564778 | 14:02 |
| evrardjp[m] | nope we do not | 14:08 |
| *** armaan has joined #openstack-ansible | 14:08 | |
| evrardjp[m] | it was if you had a host handy with those. | 14:08 |
| evrardjp[m] | systemctl restart lsyncd.service ? | 14:09 |
| odyssey4me | I don't think that'll do anything. IIRC lsync keeps a cache of checksums, and reacts to file changes. restarting the service won't likely do a thing. | 14:10 |
| d34dh0r53 | yeah, it didn't seem to do much | 14:10 |
| *** evin has quit IRC | 14:12 | |
| *** cjloader has joined #openstack-ansible | 14:13 | |
| *** kstev has joined #openstack-ansible | 14:14 | |
| odyssey4me | d34dh0r53 I guess you have an opportunity now to become the expert on lsyncd. ;) | 14:15 |
| d34dh0r53 | lol, yay! | 14:15 |
| *** cjloader has quit IRC | 14:17 | |
| *** hamza21 has quit IRC | 14:19 | |
| *** pylot_ has quit IRC | 14:22 | |
| *** vnogin has joined #openstack-ansible | 14:22 | |
| *** pylot has joined #openstack-ansible | 14:24 | |
| idlemind | so to get my cloud back-online (stable/pike) i had to manually copy the $NOVA_HOME/.ssh/authorized_users from compute nodes to the conductor containers ... the "Perform cell_v2 discovery task" would then pass successfully on the compute nodes ... however the os-nova.yml play would remove the authorized_keys file from the conductor containers so now subsequent runs fail ... is this a known bug w/pike-osa? | 14:29 |
| idlemind | re-word 1: should the "perform cell_2 discovery" task run on compute nodes and result in ssh'ing to a (or possibly all) conductor containers? | 14:30 |
| idlemind | re-word 2: if 1 is true, shouldn't the nova-key-distribution mechanism add / maintain the "authorized_keys" file on the conductor containers then? | 14:31 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Revise the role to properly cater to use-cases https://review.openstack.org/564778 | 14:31 |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia stable/queens: Adds certificate generation https://review.openstack.org/565845 | 14:32 |
| odyssey4me | idlemind this patch fixed an issue recently where the dns name was used in appropriately: https://github.com/openstack/openstack-ansible-os_nova/commit/aa5446fd26b542937b7a4ed64188281c3321fe2c | 14:33 |
| odyssey4me | that was introduced by https://github.com/openstack/openstack-ansible-os_nova/commit/c581c2af7b2bfdf60fb4467d7d5b17bed32fc85f | 14:33 |
| idlemind | odyssey4me ahh ok so that could explain why some runs i was seeing 1 conductor container and others it was a different one | 14:34 |
| idlemind | looks like that was pulled in on apr 12 to stable/pike | 14:35 |
| idlemind | so i should be able to grab that | 14:35 |
| odyssey4me | the keys are dropped in this: https://github.com/openstack/openstack-ansible-os_nova/blob/stable/pike/tasks/nova_compute_key_distribute.yml | 14:35 |
| odyssey4me | which uses https://github.com/openstack/openstack-ansible-os_nova/blob/stable/pike/templates/nova-key-insert.sh.j2 - that relies on the fact cache being up to date for *all* nova hosts, otherwise things will go missing | 14:36 |
| idlemind | so could you verify for me if the conductor containers are supposed to get the authorized keys? it seems like they'd be needed at least on conductor[0] to ensure that the ssh into it works for cell_v2 discovery | 14:36 |
| idlemind | oh so it's possibly something is up w/my fact caches? should that be something i can force an update on? | 14:37 |
| odyssey4me | I don't know whether conductors need the keys - I think only the computes do, because the keys are used for live migrations IIRC... | 14:37 |
| odyssey4me | the delegation should be handled by ansible, and be using ansible's keys - not nova's keys | 14:37 |
| idlemind | odyssey4me ya that's what i had thought but w/o the authorized_keys file on the conductors when "perform cell_v2 discovery" runs on a compute node it errors out saying it can't ssh into the conductor | 14:37 |
| odyssey4me | I suspect that's because of the bug that was fixed. | 14:39 |
| idlemind | so i should grab at least those commits mentioned if not just bring my stable/pike up to date and re-run | 14:39 |
| *** cjloader has joined #openstack-ansible | 14:40 | |
| idlemind | cool cool it'd be nice to get back on solid footing lol fought w/that little guy for a few days | 14:40 |
| odyssey4me | idlemind probably better to bring pike up to day as a whole | 14:41 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_glance master: Add support for using distribution packages for OpenStack services https://review.openstack.org/566093 | 14:41 |
| odyssey4me | *date | 14:41 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_glance master: Add support for using distribution packages for OpenStack services https://review.openstack.org/566093 | 14:42 |
| idlemind | now to get cisco's fmc / ftd boxes to boot on openstack ... their quick-start has instructions but they seem to to be failing atm | 14:42 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_cinder master: Add support for using distribution packages for OpenStack services https://review.openstack.org/566315 | 14:43 |
| openstackgerrit | Olivier Bourdon proposed openstack/openstack-ansible master: Add more meaningful/user readable failure https://review.openstack.org/566302 | 14:43 |
| *** armaan has quit IRC | 14:44 | |
| *** armaan has joined #openstack-ansible | 14:45 | |
| *** spine55 has quit IRC | 14:47 | |
| *** vnogin has quit IRC | 14:49 | |
| *** weezS has joined #openstack-ansible | 14:52 | |
| *** spine55 has joined #openstack-ansible | 14:54 | |
| *** mma has joined #openstack-ansible | 14:55 | |
| *** kstev1 has joined #openstack-ansible | 14:56 | |
| *** kstev has quit IRC | 14:57 | |
| *** mma has quit IRC | 15:00 | |
| *** armaan has quit IRC | 15:02 | |
| *** armaan has joined #openstack-ansible | 15:02 | |
| odyssey4me | hwoarang on opensuse, what's the default documentroot for a fresh nginx install | 15:07 |
| odyssey4me | mnaser any idea what it is for centos? | 15:08 |
| hwoarang | lets see | 15:08 |
| hwoarang | odyssey4me: /srv/www/htdocs/ | 15:09 |
| odyssey4me | ah, awesome, thanks | 15:09 |
| hwoarang | i assume that documentroot is 'root /srv/www/htdocs/;' | 15:09 |
| odyssey4me | centos appears to use /usr/share/nginx/html/ | 15:10 |
| odyssey4me | I guess it's designed to have you override that to something else | 15:10 |
| odyssey4me | ugh, trying to do something simple and it ends up getting messy | 15:13 |
| weezS | its life | 15:13 |
| *** pcaruana has quit IRC | 15:13 | |
| * mnaser looks | 15:15 | |
| mnaser | odyssey4me: oh you already have your answer | 15:15 |
| mnaser | sorry i was writing up something and didnt get back quickly | 15:15 |
| *** dxiri has joined #openstack-ansible | 15:18 | |
| *** cjloader has quit IRC | 15:19 | |
| odyssey4me | hwoarang what's the ansible_distribution value for opensuse? | 15:19 |
| *** cjloader has joined #openstack-ansible | 15:20 | |
| *** dave-mccowan has quit IRC | 15:20 | |
| odyssey4me | apologies, I'm not able to build opensuse instances in the cloud I use for testing :/ | 15:21 |
| odyssey4me | actually, never mind - I can just use else - no need for a conditional | 15:21 |
| hwoarang | no problem | 15:21 |
| hwoarang | ti's actually 'opeSUSE Leap' | 15:22 |
| hwoarang | but these changes between suse versions (or ansible versions) so i normally use ansible_pkg_mgr to capture all suse releases ;p | 15:22 |
| d34dh0r53 | evrardjp: when are you next bumping SHAs for stable/queens? | 15:26 |
| *** nicolasbock has quit IRC | 15:26 | |
| *** dave-mccowan has joined #openstack-ansible | 15:27 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Revise the role to properly cater to use-cases https://review.openstack.org/564778 | 15:28 |
| odyssey4me | d34dh0r53 there's already a patch in to do so - it failed gate earlier and is in recheck | 15:29 |
| odyssey4me | https://review.openstack.org/565029 | 15:30 |
| d34dh0r53 | tys | 15:30 |
| openstackgerrit | Merged openstack/openstack-ansible-openstack_hosts master: Add IPv6 kernel module https://review.openstack.org/565940 | 15:31 |
| *** cjloader has quit IRC | 15:34 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_cinder master: Add support for using distribution packages for OpenStack services https://review.openstack.org/566315 | 15:38 |
| *** nicolasbock has joined #openstack-ansible | 15:40 | |
| odyssey4me | oh bother, there's no systemd-resolved in the base image... I think those lxC_hosts changes to fix the issue might be bad patches | 15:48 |
| odyssey4me | does anyone know where I can find dates for package releases for ubuntu? | 15:48 |
| *** armaan has quit IRC | 15:49 | |
| *** armaan has joined #openstack-ansible | 15:49 | |
| jrosser | odyssey4me: you can do that on packages.ubuntu.com | 15:49 |
| jrosser | https://packages.ubuntu.com/artful/apt <- for example there is a link to the changelog on there | 15:50 |
| odyssey4me | hmm, am I missing something here? https://packages.ubuntu.com/xenial-updates/bash | 15:50 |
| odyssey4me | ah, changelog - I see it | 15:50 |
| *** kstev has joined #openstack-ansible | 15:56 | |
| *** kstev1 has quit IRC | 15:59 | |
| openstackgerrit | Andy McCrae proposed openstack/ansible-config_template master: Allow default section in ini to be specified https://review.openstack.org/566292 | 16:00 |
| *** gyee has joined #openstack-ansible | 16:00 | |
| odyssey4me | hmm, so no new bash then | 16:04 |
| odyssey4me | odd | 16:04 |
| odyssey4me | the workaround does seem to work for me, as does the patch | 16:04 |
| *** hamza21 has joined #openstack-ansible | 16:08 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_cinder master: Add support for using distribution packages for OpenStack services https://review.openstack.org/566315 | 16:08 |
| *** EmilienM is now known as EvilienM | 16:09 | |
| *** armaan has quit IRC | 16:11 | |
| *** armaan has joined #openstack-ansible | 16:11 | |
| *** lirt has quit IRC | 16:12 | |
| *** kyblik has joined #openstack-ansible | 16:12 | |
| kyblik | hi. any idea how to configure OSA to use totally different IP to connect via ssh to host than defined in https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/openstack_user_config.yml.test.example ? | 16:14 |
| *** cjloader has joined #openstack-ansible | 16:20 | |
| *** cjloader has quit IRC | 16:20 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Add a playbook to set local facts on newton redeploy https://review.openstack.org/566360 | 16:20 |
| *** cjloader has joined #openstack-ansible | 16:21 | |
| odyssey4me | looking for another review of https://review.openstack.org/565999 from the cores please | 16:23 |
| *** radeks has joined #openstack-ansible | 16:29 | |
| cloudnull | d34dh0r53: any anyone else, mind having a look at https://review.openstack.org/#/c/565896/ | 16:30 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Provide an example for switch port configurations https://review.openstack.org/566120 | 16:31 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Convert rsyslog to an include_task with group_vars https://review.openstack.org/556002 | 16:38 |
| openstackgerrit | Merged openstack/openstack-ansible stable/ocata: Update lxc_host/pip_install role SHA's https://review.openstack.org/566309 | 16:38 |
| *** radeks has quit IRC | 16:43 | |
| *** armaan has quit IRC | 16:45 | |
| odyssey4me | cloudnull want to tweak that little thing I commented on - otherwise your next patch will be another mixup :p | 16:45 |
| *** udesale has quit IRC | 16:50 | |
| odyssey4me | meh, on its way | 16:50 |
| *** jra has joined #openstack-ansible | 16:51 | |
| *** cmart has quit IRC | 16:55 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Correct cluster properties https://review.openstack.org/565896 | 16:55 |
| *** armaan has joined #openstack-ansible | 16:55 | |
| *** kstev1 has joined #openstack-ansible | 16:56 | |
| *** kstev has quit IRC | 16:57 | |
| jra | I'm trying to have all syslog output from all OSA-managed containers and container hosts sent to my upstream syslog aggregator. The containers can be configured to send some logs to a user-defined location, but only local7.* are included; and the container hosts themselves don't seem to be configurable this way at all. How should I go about this? | 16:59 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Revise the role to properly cater to use-cases https://review.openstack.org/564778 | 17:04 |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia stable/queens: Adds certificate generation https://review.openstack.org/565845 | 17:04 |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia stable/queens: Adds certificate generation https://review.openstack.org/565845 | 17:05 |
| *** cmart has joined #openstack-ansible | 17:10 | |
| jra | I could submit changes for review that would either remove or make configurable the syslog facility filter, and extend the same remote logging pattern to the container hosts as well; would those changes be consistent with the goals of the project? | 17:12 |
| odyssey4me | jra there is already the capability to configure rsyslog to go to a sinle rsyslog host... or to configure it to go wherever you want | 17:13 |
| jra | Yes, but as above: only for the containers, and only for local7.* | 17:14 |
| odyssey4me | if you have a log host in the inventory, it'll get done automagically for you - but if it's a host managed by something else, then just add something like this: https://github.com/openstack/openstack-ansible-rsyslog_client/blob/master/defaults/main.yml#L66-L83 | 17:14 |
| odyssey4me | jra are you working with master, or a stable branch? | 17:15 |
| jra | Stable, 16.0.10 | 17:15 |
| jra | But this behaviour appears in master | 17:16 |
| jra | https://github.com/openstack/openstack-ansible-rsyslog_client/blob/master/templates/remote-logging.j2#L34 | 17:16 |
| odyssey4me | hmm, I don't totally understand how it's all done - it's been a long time since I worked with it - but IIRC there are services set to log locally, then everything is piped through one facility to the remote host | 17:17 |
| *** esberglu has quit IRC | 17:17 | |
| jra | That's true for openstack services, that's why local7 is sufficient for them | 17:17 |
| jra | but that doesn't capture other system logs | 17:17 |
| odyssey4me | hmm, ok - well I do expect that anyone using this will want more funnelled through too - so yeah, that would fit the project I guess | 17:18 |
| jra | Okay, I'll file a bug and start working on a changeset for this | 17:18 |
| odyssey4me | in master, though the containers now journal straight to the host - then the hosts push to the remote journal server | 17:19 |
| jra | as in systemd journal? | 17:19 |
| odyssey4me | yup | 17:19 |
| jra | where's the bit that pushes the journal data to the remote server? | 17:19 |
| odyssey4me | looking for it now | 17:19 |
| jra | that might be entirely sufficient | 17:19 |
| odyssey4me | far more efficient that rsyslog that way | 17:20 |
| odyssey4me | there's also a bunch of work in the ops repo to implement the elastic stack for log shipping and analysis | 17:20 |
| *** AR2018 has joined #openstack-ansible | 17:20 | |
| jra | I have the benefit of a fairly mature log management environment in my organization, so would like to take advantage of it, but it's nice to know that there's some focus being put on log management in the OS community | 17:21 |
| odyssey4me | jra https://review.openstack.org/555982 | 17:21 |
| jra | That looks super promising! | 17:22 |
| odyssey4me | there's another patch in flight to switch rsyslog off https://review.openstack.org/556002 | 17:22 |
| jra | So will a change that's in master now make it into a Queens release? | 17:23 |
| odyssey4me | that relies on some work done in the lxc hosts and containers roles, and those changes were significant and scattered... so no, that is not going to be backported | 17:23 |
| odyssey4me | it's not worth destablizing | 17:24 |
| jra | Seems reasonable | 17:24 |
| jra | Those changes tell me that my proposed changes wouldn't fit with the project's direction; no point updating rsyslog's behaviour if it's going away anyway | 17:24 |
| *** kyblik has quit IRC | 17:24 | |
| jra | I'll create a local overlay to handle this in our environment | 17:24 |
| odyssey4me | you could, of course, fork and give it a go yourself | 17:24 |
| jra | We're pretty focused on trying to get a new cloud deployment out the door, so I don't think I can put that time in right now | 17:25 |
| odyssey4me | ja, understood | 17:25 |
| jra | Thanks for your help and advice on this | 17:25 |
| cloudnull | odyssey4me: sorry was not paying attention | 17:26 |
| cloudnull | which on e? | 17:26 |
| jra | I'm going to be attending the OSA onboarding at Summit in Vancouver this month, hopefully I can learn a bit about where the project is going and see where I can be helpful in offering changes | 17:26 |
| *** kstev has joined #openstack-ansible | 17:26 | |
| odyssey4me | sure thing, no problem at all - you're welcome to pop patches up if it'll help reduce your maintenance going forward - given opensuse doesn't have the remote journaling functionality, it still needs to use rsyslog, so it'll still be handy | 17:26 |
| *** tosky has quit IRC | 17:27 | |
| cloudnull | jra: awesome ! i'll look forward to seeing you there :) | 17:27 |
| odyssey4me | more help is always good, especially from operators :) | 17:27 |
| *** kstev1 has quit IRC | 17:27 | |
| cloudnull | ++ | 17:27 |
| odyssey4me | anyway, I gotta run - have a great day! | 17:28 |
| cloudnull | odyssey4me: what should I change ? | 17:28 |
| odyssey4me | cloudnull it merged long ago :p | 17:29 |
| cloudnull | oh. | 17:29 |
| cloudnull | ok. never mind :) | 17:29 |
| * cloudnull winning | 17:29 | |
| cloudnull | in that case, have a wonderful day :) | 17:29 |
| *** tosky has joined #openstack-ansible | 17:30 | |
| cloudnull | jra: I've done a mess of work with rsyslog and the journald things, if you have any questions don't hesitate to reach out | 17:32 |
| openstackgerrit | Merged openstack/openstack-ansible-os_tempest stable/queens: Install os-testr https://review.openstack.org/565999 | 17:36 |
| jra | @cloudnull: thanks! | 17:40 |
| *** ppetit has quit IRC | 17:41 | |
| *** evin has joined #openstack-ansible | 17:52 | |
| *** pcaruana has joined #openstack-ansible | 17:54 | |
| *** kstev1 has joined #openstack-ansible | 17:57 | |
| *** kstev has quit IRC | 17:59 | |
| *** jra has quit IRC | 18:07 | |
| *** DanyC has quit IRC | 18:13 | |
| *** DanyC has joined #openstack-ansible | 18:13 | |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia stable/queens: [WIP] Adds certificate generation https://review.openstack.org/565845 | 18:16 |
| *** DanyC has quit IRC | 18:18 | |
| *** klamath has quit IRC | 18:49 | |
| *** klamath has joined #openstack-ansible | 18:50 | |
| *** cmart has quit IRC | 18:52 | |
| Tahvok | I would like to have your opinion on this: https://review.openstack.org/#/c/566269/2 | 19:06 |
| *** pcaruana has quit IRC | 19:10 | |
| dmsimard | cloudnull, odyssey4me: it was one of you two that told me about unreliability or performance things in 'ara-report', this should now be fixed -- let me know if you notice anything odd | 19:14 |
| odyssey4me | dmsimard yeah, it was doing 500's earlier - thanks for solving that! I was a little lost without ARA. :p | 19:22 |
| odyssey4me | trawling console logs is so old school | 19:22 |
| *** radeks has joined #openstack-ansible | 19:23 | |
| *** cmart has joined #openstack-ansible | 19:26 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-rabbitmq_server master: Convert self-signed SSL to use the openssl modules https://review.openstack.org/566391 | 19:26 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-rabbitmq_server master: Convert self-signed SSL to use the openssl modules https://review.openstack.org/566391 | 19:31 |
| *** radeks has quit IRC | 19:33 | |
| *** radeks has joined #openstack-ansible | 19:33 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: Update all SHAs for 17.0.4 https://review.openstack.org/565029 | 19:34 |
| *** DanyC has joined #openstack-ansible | 19:38 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-haproxy_server master: Convert self-signed SSL to use the openssl modules https://review.openstack.org/566394 | 19:39 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-haproxy_server master: Convert self-signed SSL to use the openssl modules https://review.openstack.org/566394 | 19:42 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-rabbitmq_server master: Convert self-signed SSL to use the openssl modules https://review.openstack.org/566391 | 19:43 |
| rschulman | I'm trying to create my own AIO (rather than using the built in one) so I know how it works and I can expand it later. Running setup-hosts is complaining, however, that infra1 and compute1 have the same IP addresses. | 19:50 |
| cloudnull | rschulman: in the openstack_user_config file the host names need to be the same if the ip is the same | 19:50 |
| rschulman | ahh, ok | 19:51 |
| rschulman | so I need to put everything on infra1 for now. | 19:51 |
| cloudnull | now these are not the physical host name, its the ansible inventory host name | 19:51 |
| cloudnull | yes. | 19:51 |
| rschulman | Got it, thank you. | 19:52 |
| cloudnull | anytime | 19:53 |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia stable/queens: Adds certificate generation https://review.openstack.org/565845 | 20:03 |
| *** bgmccollum_ is now known as bgmccollum | 20:08 | |
| rschulman | Running `openstack-ansible setup-hosts.yml` the script fails on "Ensure LXC cache has been prepared". The log files indicate that cache-prep-commands.sh failed because `/etc/resolv.conf` doesn't exist. | 20:14 |
| rschulman | I'm not seeing a similar bug in Google. Anyone have thoughts on what's going wrong there? | 20:14 |
| shananigans | rschulman: https://bugs.launchpad.net/openstack-ansible/+bug/1768592 | 20:20 |
| openstack | Launchpad bug 1768592 in openstack-ansible "/usr/local/bin/cache-prep-commands.sh failing to write to /etc/resolv.conf when resolvconf used on Xenial" [Critical,Fix committed] | 20:20 |
| shananigans | I just tried the workaround and it didn't work. It got past the cache-prep-commands.sh part, but failed later on the repo server as /etc/resolv.conf didn't exist at all. | 20:21 |
| *** armaan has quit IRC | 20:23 | |
| *** armaan has joined #openstack-ansible | 20:23 | |
| rschulman | shananigans: Ahh, ok, thanks. I'll track that bug. | 20:23 |
| odyssey4me | shananigans hmm, that's troubling | 20:28 |
| odyssey4me | I suspect that we should actually not remove the file - I noticed that the file check didn't work, so perhaps that file check should rather be fixed. | 20:29 |
| odyssey4me | shananigans were you doing that on newton, or later? | 20:30 |
| shananigans | newton | 20:30 |
| shananigans | was failing on pike, so went back to a known working(or so i thought) This was before I saw the bug report. | 20:30 |
| odyssey4me | ok, it's late for me right now so if you're able to do some tests and put info in the bug or submit a changed approach review that'd be awesome | 20:31 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-rabbitmq_server master: Convert self-signed SSL to use the openssl modules https://review.openstack.org/566391 | 20:31 |
| odyssey4me | in the image that comes down, it has resolvconf installed - so there's a symlink on /etc/resolv.conf which points at /run/... | 20:32 |
| odyssey4me | in the chroot, obviously, /run doesn't exist - so the file check here doesn't work: https://github.com/openstack/openstack-ansible-lxc_hosts/blob/newton-eol/vars/ubuntu-16.04.yml#L59-L61 | 20:32 |
| odyssey4me | however, using 'if -h' works.. and should allow the symlink to be moved aside, and moved back later if that file test is also fixed | 20:33 |
| odyssey4me | so, perhaps the right workaround is this | 20:34 |
| shananigans | If I get to the root of it and its not something stupid I'm doing, I'll update with the details. | 20:34 |
| odyssey4me | check for the file in the pre commands using the right check and move it to some other name, then revert the name in the post commands | 20:35 |
| shananigans | This round I'm adding the link back in with the post to see if it helps. | 20:36 |
| shananigans | just waiting on it to spin up. | 20:36 |
| odyssey4me | hmm, that said the changes that went in does place a resolv.conf there: https://github.com/openstack/openstack-ansible-lxc_hosts/commit/e77c3a923bc8aae4a264917f592b58b5d1c79aed | 20:36 |
| odyssey4me | oh of course - the workaround doesn't have that part | 20:36 |
| odyssey4me | perhaps the workaround should include using this in the post commands: echo "nameserver {{ lxc_net_address }}" > /etc/resolv.conf | 20:37 |
| odyssey4me | Thanks for digging into it. I trust you'll find a solution and am keen to see what it is. | 20:38 |
| *** radeks has quit IRC | 20:45 | |
| *** mbuil has quit IRC | 21:11 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-rabbitmq_server master: Convert self-signed SSL to use the openssl modules https://review.openstack.org/566391 | 21:15 |
| *** hamza21 has quit IRC | 21:18 | |
| *** dave-mccowan has quit IRC | 21:19 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-galera_server master: Convert self-signed SSL to use the openssl modules https://review.openstack.org/566408 | 21:36 |
| *** kstev1 has quit IRC | 21:47 | |
| *** DanyC has quit IRC | 21:49 | |
| *** brad[] has quit IRC | 21:50 | |
| *** hamzy_ has joined #openstack-ansible | 21:51 | |
| *** hamzy has quit IRC | 21:54 | |
| *** AR2018 has quit IRC | 21:55 | |
| *** kstev has joined #openstack-ansible | 22:02 | |
| *** kstev1 has joined #openstack-ansible | 22:04 | |
| *** kstev has quit IRC | 22:06 | |
| *** weezS has quit IRC | 22:06 | |
| *** AR2018 has joined #openstack-ansible | 22:16 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-rabbitmq_server master: Convert self-signed SSL to use the openssl modules https://review.openstack.org/566391 | 22:16 |
| *** evin has quit IRC | 22:17 | |
| *** AR2018 has quit IRC | 22:19 | |
| *** AR2018 has joined #openstack-ansible | 22:24 | |
| *** markvoelker has quit IRC | 22:29 | |
| *** AR2018 has quit IRC | 22:29 | |
| *** AR2018 has joined #openstack-ansible | 22:30 | |
| *** brad[] has joined #openstack-ansible | 22:33 | |
| *** AR2018 has quit IRC | 22:34 | |
| *** kstev1 has quit IRC | 22:53 | |
| *** cmart has quit IRC | 23:32 | |
| *** cjloader has quit IRC | 23:35 | |
| *** dxiri has quit IRC | 23:54 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!