| *** odyssey4me has quit IRC | 00:11 | |
| *** odyssey4me has joined #openstack-ansible | 00:11 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Converge distro interfaces to systemd-networkd https://review.openstack.org/552721 | 00:12 |
|---|---|---|
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add playbook to ship journals from hosts https://review.openstack.org/555982 | 00:12 |
| *** cjloader has joined #openstack-ansible | 00:17 | |
| *** hamzy has quit IRC | 00:19 | |
| *** hamzy has joined #openstack-ansible | 00:20 | |
| *** cjloader has quit IRC | 00:22 | |
| *** chyka has joined #openstack-ansible | 00:25 | |
| *** chyka has quit IRC | 00:30 | |
| *** throwsb1 has quit IRC | 00:44 | |
| *** fatdragon has quit IRC | 01:08 | |
| *** Guest44281 has quit IRC | 01:10 | |
| *** chhavi__ has joined #openstack-ansible | 01:11 | |
| *** chhavi__ has quit IRC | 01:16 | |
| *** Guest44281 has joined #openstack-ansible | 01:16 | |
| *** Guest44281 has quit IRC | 01:27 | |
| *** tinwood has quit IRC | 02:10 | |
| *** tinwood has joined #openstack-ansible | 02:11 | |
| *** cjloader has joined #openstack-ansible | 02:14 | |
| *** cjloader has quit IRC | 02:19 | |
| *** vnogin has joined #openstack-ansible | 02:24 | |
| *** vnogin has quit IRC | 02:29 | |
| *** chyka has joined #openstack-ansible | 02:33 | |
| *** chyka has quit IRC | 02:38 | |
| *** flemingo_ has quit IRC | 03:07 | |
| *** lbragstad has joined #openstack-ansible | 03:11 | |
| *** Guest44281 has joined #openstack-ansible | 03:29 | |
| *** armaan has quit IRC | 03:45 | |
| *** armaan has joined #openstack-ansible | 03:46 | |
| *** chhavi__ has joined #openstack-ansible | 03:54 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Add playbook to ship journals from hosts https://review.openstack.org/555982 | 04:04 |
| *** poopcat has quit IRC | 04:06 | |
| *** Guest44281 has quit IRC | 04:10 | |
| *** Guest44281 has joined #openstack-ansible | 04:11 | |
| *** udesale has joined #openstack-ansible | 04:12 | |
| *** udesale has quit IRC | 04:13 | |
| *** udesale has joined #openstack-ansible | 04:13 | |
| *** lbragstad has quit IRC | 04:18 | |
| *** indistylo has joined #openstack-ansible | 04:36 | |
| *** armaan has quit IRC | 04:43 | |
| *** armaan has joined #openstack-ansible | 04:52 | |
| *** cjloader has joined #openstack-ansible | 05:14 | |
| *** cjloader has quit IRC | 05:18 | |
| *** sar has quit IRC | 05:24 | |
| *** aruns has joined #openstack-ansible | 05:42 | |
| *** indistylo has quit IRC | 05:42 | |
| *** gkadam_ has joined #openstack-ansible | 05:48 | |
| *** DanyC has joined #openstack-ansible | 06:19 | |
| *** vnogin has joined #openstack-ansible | 06:26 | |
| *** sar has joined #openstack-ansible | 06:28 | |
| *** vnogin has quit IRC | 06:31 | |
| *** gkadam_ has quit IRC | 06:42 | |
| *** gkadam has joined #openstack-ansible | 06:42 | |
| jrosser | cloudnull: this also looks unused http://git.openstack.org/cgit/openstack/openstack-ansible/tree/etc/openstack_deploy/openstack_user_config.yml.aio#n28 | 06:43 |
| *** markvoelker has quit IRC | 06:43 | |
| *** markvoelker has joined #openstack-ansible | 06:44 | |
| *** DanyC has quit IRC | 06:45 | |
| *** mbuil has joined #openstack-ansible | 06:46 | |
| *** pcaruana has joined #openstack-ansible | 06:47 | |
| admin0 | morning | 06:48 |
| *** fghaas has joined #openstack-ansible | 06:48 | |
| *** markvoelker has quit IRC | 06:48 | |
| evrardjp | good morning everyone | 06:51 |
| evrardjp | andymccr: are you there? | 06:52 |
| evrardjp | andymccr: I'm trying to streamline our group usage in order to do inventory cleanups -- I seen a lot of groups, I'd like to discuss what we can remove | 06:52 |
| evrardjp | odyssey4me: cloudnull logan- jmccrory your opinion on the groups would be welcomed too ofc :) | 06:53 |
| evrardjp | andymccr: I guess you add an opinion when you started with the tests repo | 06:53 |
| *** DanyC has joined #openstack-ansible | 06:54 | |
| *** DanyC has quit IRC | 06:57 | |
| *** DanyC has joined #openstack-ansible | 06:58 | |
| openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_neutron master: DNM: Testing ODL gates https://review.openstack.org/559065 | 06:58 |
| *** arbrandes1 has joined #openstack-ansible | 07:01 | |
| *** DanyC has quit IRC | 07:02 | |
| *** arbrandes has quit IRC | 07:03 | |
| *** ivve has joined #openstack-ansible | 07:05 | |
| *** DanyC has joined #openstack-ansible | 07:05 | |
| *** aruns has quit IRC | 07:07 | |
| evrardjp | mbuil: I wrote a message for you yesterday, please check irc logs on eavesdrop.openstack.org | 07:09 |
| mbuil | evrardjp: ok! thanks for the heads up :) | 07:10 |
| mbuil | good morning by the way | 07:10 |
| *** ivve has quit IRC | 07:10 | |
| *** haad1 has joined #openstack-ansible | 07:12 | |
| *** DanyC has quit IRC | 07:15 | |
| *** epalper has joined #openstack-ansible | 07:26 | |
| *** fghaas has quit IRC | 07:27 | |
| *** armaan has quit IRC | 07:33 | |
| *** armaan has joined #openstack-ansible | 07:33 | |
| *** threestrands has quit IRC | 07:36 | |
| *** fghaas has joined #openstack-ansible | 07:43 | |
| andymccr | evrardjp: yo | 07:46 |
| evrardjp | heyya | 07:46 |
| evrardjp | andymccr: please let me give you a link | 07:47 |
| evrardjp | http://paste.openstack.org/show/718567/ | 07:47 |
| evrardjp | andymccr: could you explain what these swift groups are for? | 07:48 |
| evrardjp | the swift_all gathers acc/obj/cont I guess | 07:48 |
| evrardjp | and proxy | 07:48 |
| evrardjp | but what is remote_all ? | 07:48 |
| andymccr | remote is for multi_region swift | 07:49 |
| *** electrofelix has joined #openstack-ansible | 07:49 | |
| andymccr | its basically all hosts that arent in your immediate swift region but you still need to care about their rings, and sync to them etc. | 07:49 |
| evrardjp | ok | 07:49 |
| evrardjp | how do you wire them generally? It's some hosts you add to your inventory manually? | 07:50 |
| evrardjp | (wondering now HOW it is used) | 07:51 |
| *** DanyC has joined #openstack-ansible | 07:51 | |
| andymccr | you would specify swift-remote_hosts in user_config which would be the hosts in the other inventory, proxy & storage hosts, but just their accesible ip | 07:52 |
| andymccr | the role then uses those hosts only when syncing the ring and SSH keys for the swift user | 07:52 |
| jrosser | evrardjp: provider_networks arrggghh https://etherpad.openstack.org/p/osa-provider-networks | 07:53 |
| andymccr | but wont adjust the config files or change the devices in the ring for those hosts, or anything else really, aside from just sync the necessary files | 07:53 |
| *** fghaas has quit IRC | 07:55 | |
| evrardjp | prometheanfire: I misunderstood you. I thought you had openstack-ansible jobs in the cross jobs. But you don't, you simply have timeouts. I'd say don't update your job timeouts generally, but maybe think about tweaking the nova job (likely one of the slowest ones) on ovh. | 07:57 |
| evrardjp | andymccr: understood. Thanks! | 07:57 |
| evrardjp | that group can probably be dynamically created | 07:57 |
| evrardjp | if we have a user variable with hosts | 07:58 |
| evrardjp | you don't need the full power of the inventory there | 07:58 |
| evrardjp | it was I guess just a convenience group | 07:58 |
| evrardjp | oh you mean openstack_user_config | 07:58 |
| evrardjp | gotcha | 07:58 |
| evrardjp | jrosser: opening the link rn :) | 07:59 |
| evrardjp | I am answering inline jrosser | 08:04 |
| *** rmart04 has joined #openstack-ansible | 08:04 | |
| *** haad1 has quit IRC | 08:07 | |
| *** armaan has quit IRC | 08:08 | |
| *** armaan has joined #openstack-ansible | 08:08 | |
| *** fghaas has joined #openstack-ansible | 08:11 | |
| *** shardy has joined #openstack-ansible | 08:16 | |
| *** eumel8 has joined #openstack-ansible | 08:17 | |
| *** haad1 has joined #openstack-ansible | 08:19 | |
| eumel8 | o/ | 08:21 |
| openstackgerrit | Frank Kloeker proposed openstack/openstack-ansible-os_horizon master: Rename pull_catalog args from version to branch https://review.openstack.org/559290 | 08:21 |
| openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_neutron master: DNM: Testing ODL gates https://review.openstack.org/559065 | 08:21 |
| eumel8 | odyssey4me: ^^ | 08:21 |
| evrardjp | eumel8: will there be other pull_catalog implementations? Else we can hardcode horizon I think | 08:27 |
| evrardjp | it's a horizon-manage call anyway | 08:27 |
| evrardjp | except if in your test you have horizon-test or something like that | 08:28 |
| evrardjp | :) | 08:28 |
| evrardjp | not sure | 08:28 |
| eumel8 | evrardjp: I'm also not sure :) But from my understanding we must also pull translation strings from other projects like heat-dashboard to put this in there. | 08:29 |
| evrardjp | oh | 08:29 |
| evrardjp | that's harder | 08:30 |
| eumel8 | the pull of other projects works with option -p but I'm not sure what are the next steps. Maybe I have to grib some details from DevStack | 08:35 |
| evrardjp | eumel8: I think for all the projects that are enabled for deployment under horizon: (we have vars for that, like horizon_enable_heat_ui | 08:36 |
| evrardjp | ) | 08:36 |
| evrardjp | we should be using it | 08:36 |
| evrardjp | so adding a task with condition horizon_enable_heat_ui | bool | 08:36 |
| evrardjp | and do the right thing for zanata + heat | 08:37 |
| evrardjp | but I am not sure how it works, | 08:37 |
| evrardjp | can we run the command https://github.com/openstack/openstack-ansible-os_horizon/blob/master/tasks/horizon_translations_update.yml#L17 multiple times? | 08:37 |
| evrardjp | should we? | 08:37 |
| evrardjp | else we can replace this task with a loop, passing the right things | 08:38 |
| eumel8 | you mean with_items? could be | 08:38 |
| evrardjp | yeah | 08:38 |
| evrardjp | let me draft you something | 08:38 |
| evrardjp | do you mind if I edit your patch? | 08:38 |
| evrardjp | Maybe you can revert if you don't like it? | 08:39 |
| eumel8 | sure, go ahead | 08:39 |
| evrardjp | thanks | 08:39 |
| evrardjp | b means branch right? | 08:43 |
| *** markvoelker has joined #openstack-ansible | 08:44 | |
| eumel8 | yes | 08:45 |
| eumel8 | in Zanata its "version", but somebody used this arg in other place, so now there is a conflict ;) | 08:45 |
| evrardjp | is there a chance the branch would be overriden per component? | 08:46 |
| eumel8 | mhm? which component? | 08:47 |
| *** hamza21 has joined #openstack-ansible | 08:49 | |
| *** armaan has quit IRC | 08:49 | |
| *** armaan has joined #openstack-ansible | 08:50 | |
| *** vnogin has joined #openstack-ansible | 08:51 | |
| evrardjp | eumel8: I mean is there for example a case where heat-dashboard would have to be fetched from master, where horizon would have to fetched from a stable branch for example | 08:51 |
| evrardjp | or the other way around | 08:51 |
| *** vnogin has quit IRC | 08:52 | |
| *** gkadam has quit IRC | 08:52 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_horizon master: Rename pull_catalog args from version to branch https://review.openstack.org/559290 | 08:53 |
| evrardjp | eumel8: ^ want me to update with all the dashboards? | 08:53 |
| *** gkadam has joined #openstack-ansible | 08:53 | |
| eumel8 | depends on the installed version. We implemented this for stable/queens to pull in a OSA master installation. It's possible you have another dashboard plugin in another version. | 08:54 |
| eumel8 | master should be default all the time | 08:54 |
| hwoarang | good morning | 08:54 |
| evrardjp | good mornign hwoarang | 08:54 |
| evrardjp | eumel8: ok | 08:54 |
| evrardjp | will update the patch then | 08:55 |
| eumel8 | ok | 08:55 |
| evrardjp | eumel8: do you agree that the version is probably linked to the git installed version of the dashboard? | 08:57 |
| evrardjp | (sorry I am defacing your patch :/) | 08:58 |
| eumel8 | evardjp: I think this is a must. Other constellation will not work out | 08:58 |
| evrardjp | so let me get this straight -- you have queens, but you are pulling OSA master horizon and master dashboards | 08:59 |
| evrardjp | for example master heat-ui | 09:00 |
| evrardjp | so you want to backport this change to queens | 09:01 |
| eumel8 | we had master OSA and were required to pull stable/queens translation files. This is the situation in release week | 09:01 |
| eumel8 | when I realized OSA stable/queens comes very later after OpenStack release | 09:02 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_horizon master: Rename pull_catalog args from version to branch https://review.openstack.org/559290 | 09:05 |
| evrardjp | eumel8: we have stable/queens for a while now -- we were there directly when it was official, as a cycle trailing project... Not sure what you mean there :) | 09:06 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/queens: Work around nested virt issues on OVH test nodes https://review.openstack.org/559295 | 09:06 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/ocata: Work around nested virt issues on OVH test nodes https://review.openstack.org/559296 | 09:07 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_horizon master: Rename pull_catalog args from version to branch https://review.openstack.org/559290 | 09:07 |
| evrardjp | dammit gerrit | 09:07 |
| evrardjp | eumel8: sorry for the pebkac there | 09:08 |
| evrardjp | will issue next one real quick | 09:08 |
| ArchiFleKs | Hi, I have an issue on pike with nova playbook, I have one host failing (everytime so not just a temporarly failure) because of a name resolution of a container (http://paste.openstack.org/show/718581/) but everything is OK on /etc/hosts and I can ping the container from the compute node, any idea where to look ? | 09:09 |
| eumel8 | evrardjp: I think that's not a real problem, because codebase is already the same beside backports. Only the translation strings are in the other branch on Zanata. | 09:09 |
| ArchiFleKs | the task is " Perform a cell_v2 discover" | 09:09 |
| *** mgrewy has joined #openstack-ansible | 09:09 | |
| mgrewy | service containers not able to reach other infra host (& containers on it) over br-mgmt network. both infra host can reach each other on br-mgmt network...any pointer here | 09:11 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_horizon master: Rename pull_catalog args from version to branch https://review.openstack.org/559290 | 09:11 |
| odyssey4me | ArchiFleKs we just resolved that with the latest release - make sure you have https://github.com/openstack/openstack-ansible/commit/77a219af292070b94b14b9ce0134c8157fa8b1c7 in your checkout and ensure you bootstrap-ansible after you update the checkout | 09:11 |
| evrardjp | odyssey4me: oh that was solved? | 09:12 |
| odyssey4me | evrardjp yup | 09:12 |
| evrardjp | good | 09:12 |
| evrardjp | oh I expected the other nova issue with cells. | 09:12 |
| evrardjp | Sad. | 09:12 |
| odyssey4me | evrardjp with https://review.openstack.org/#/q/I4f7328e1aff79eb94f278999b97fbad8f44c6a3e | 09:13 |
| odyssey4me | what other issue? | 09:13 |
| evrardjp | the upstream nova change | 09:14 |
| evrardjp | on master | 09:14 |
| evrardjp | I didn't see this was a "pike" issue | 09:14 |
| evrardjp | :) | 09:14 |
| evrardjp | ok let's fix gates today | 09:14 |
| evrardjp | my march experiment is over after all | 09:15 |
| *** markvoelker has quit IRC | 09:18 | |
| odyssey4me | evrardjp we resolved that in master already, but didn't touch the stable branches | 09:19 |
| evrardjp | ok | 09:19 |
| evrardjp | hwoarang: thanks for the recheck on the voting jobs. I haven't got the chance to read your spec yesterday after all. Will do today. | 09:25 |
| evrardjp | odyssey4me: gotcha | 09:25 |
| hwoarang | evrardjp: yeah no worries i am working on a prototype for the keystone role | 09:25 |
| hwoarang | how messed up things are going to be | 09:25 |
| odyssey4me | bbiab, afk for a while | 09:26 |
| evrardjp | to talk about the issues pending, I think we have a few ones that need attention. | 09:27 |
| evrardjp | jrosser: has expressed the problem that the documentation is definitely not in-line with what's happening since we moved the neutron agents to baremetal. | 09:27 |
| evrardjp | a few examples are there: https://etherpad.openstack.org/p/osa-provider-networks | 09:28 |
| evrardjp | I think we should fix this before including new features. | 09:28 |
| evrardjp | On top of that, our upgrade jobs are now massively failing all accross the board. | 09:28 |
| evrardjp | We should stabilize things. Help is definitely wanted there. | 09:29 |
| ArchiFleKs | odyssey4me: thanks, I'm two commit after that because I needed that: https://github.com/openstack/openstack-ansible-os_nova/commit/5d66142e461fec400262509f7480fb3274019cf0 | 09:31 |
| ArchiFleKs | so I should have the cell commit, but it only happened for one compute node, not the other, but because it is the first one in the compute node list | 09:32 |
| *** sep has quit IRC | 09:33 | |
| *** sar has quit IRC | 09:34 | |
| hwoarang | evrardjp: are there bugs about these pending issues? | 09:34 |
| odyssey4me | ArchiFleKs it should only happen on one compute node - perhaps you haven't pulled in the role updates? | 09:35 |
| evrardjp | hwoarang: that's what I will do today -- probe what's going on, try to fix it quickly if I can, else file a bug, and hope someone else would fix it. | 09:37 |
| evrardjp | like everybody does. | 09:37 |
| evrardjp | :p | 09:37 |
| *** sep has joined #openstack-ansible | 09:38 | |
| *** jillr has quit IRC | 09:40 | |
| *** jillr has joined #openstack-ansible | 09:42 | |
| *** armaan has quit IRC | 09:42 | |
| hwoarang | ;p | 09:42 |
| *** armaan has joined #openstack-ansible | 09:43 | |
| odyssey4me | evrardjp FYI I'm working on P->Q upgrades, specifically on the cleanup stuff | 09:43 |
| odyssey4me | I've also noticed that cinder-volume/cinder-backup goes to 100% CPU again, so I think we might have to ditch the reload and do restart... I'll look into that too. | 09:44 |
| ArchiFleKs | odyssey4me: yes i happened on only one | 09:45 |
| ArchiFleKs | I just tried running the nova playbook on a specific compute node, and I get the same error, but yes if ran it on all my host, I get the error on only one compute node | 09:46 |
| ArchiFleKs | the task that failed has been changed by : https://github.com/openstack/openstack-ansible-os_nova/commit/c581c2af7b2bfdf60fb4467d7d5b17bed32fc85f | 09:47 |
| ArchiFleKs | it is the "Perform a cell_v2 discover" | 09:48 |
| evrardjp | odyssey4me: thanks for the work. once again. | 09:48 |
| evrardjp | you get the MVP | 09:48 |
| *** savvas has joined #openstack-ansible | 09:50 | |
| *** savvas has quit IRC | 09:55 | |
| *** esberglu has joined #openstack-ansible | 09:59 | |
| *** esberglu has quit IRC | 10:04 | |
| *** ibmko_ has joined #openstack-ansible | 10:12 | |
| *** savvas has joined #openstack-ansible | 10:12 | |
| *** savvas has quit IRC | 10:13 | |
| *** savvas has joined #openstack-ansible | 10:13 | |
| *** savvas has quit IRC | 10:14 | |
| *** savvas has joined #openstack-ansible | 10:14 | |
| *** markvoelker has joined #openstack-ansible | 10:15 | |
| *** sep has quit IRC | 10:15 | |
| *** savvas_ has joined #openstack-ansible | 10:15 | |
| *** savvas__ has joined #openstack-ansible | 10:17 | |
| *** savvas has quit IRC | 10:19 | |
| *** savvas_ has quit IRC | 10:20 | |
| *** armaan has quit IRC | 10:23 | |
| *** armaan has joined #openstack-ansible | 10:24 | |
| *** sar has joined #openstack-ansible | 10:24 | |
| openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_neutron master: DNM: Testing ODL gates https://review.openstack.org/559065 | 10:25 |
| *** mgrewy has quit IRC | 10:26 | |
| *** armaan has quit IRC | 10:28 | |
| *** armaan has joined #openstack-ansible | 10:28 | |
| hwoarang | mbuil: for this^ | 10:38 |
| hwoarang | you can edit tests/test.yml and add just two tasks | 10:38 |
| hwoarang | one for the apt_repository and one to cat source.list | 10:38 |
| hwoarang | so you dont have to wait a long time for the results | 10:38 |
| hwoarang | i mean, just have 2 tasks in tests/test.yml. one to add the repo, and one to see what ansible actually added there | 10:39 |
| mbuil | hwoarang: oh, that is very helpful! I'll try that :) | 10:43 |
| *** sep has joined #openstack-ansible | 10:44 | |
| *** gkadam has quit IRC | 10:47 | |
| evrardjp | yeah and make it fail, so you have a fast answer. | 10:49 |
| evrardjp | and remove the other builds of the project | 10:49 |
| *** markvoelker has quit IRC | 10:49 | |
| evrardjp | boom | 10:49 |
| odyssey4me | ArchiFleKs check whether your os_nova role has https://github.com/openstack/openstack-ansible-os_nova/commit/6321cf90304023c20dc6abb87e2fa20107b7051d included | 10:54 |
| odyssey4me | cores, reviews appreciated for https://review.openstack.org/#/q/I54681092adf9c0647bf65583176ddc21c99494f3 which will help improve gate success | 10:59 |
| odyssey4me | also, please review https://review.openstack.org/#/q/I6d7a9af234b7a8a6e83f9a730ffe1c4fd698786a to improve the ARA reporting | 11:00 |
| evrardjp | odyssey4me: do you have any knowledge of angularJS by any chance? | 11:19 |
| evrardjp | or anyone here | 11:20 |
| evrardjp | do we know who is heading to Vancouver? jmccrory logan- hwoarang ? | 11:25 |
| hwoarang | im not | 11:25 |
| evrardjp | ok | 11:25 |
| evrardjp | I hope cloudnull would be able to go | 11:26 |
| *** markvoelker has joined #openstack-ansible | 11:46 | |
| *** ibmko_ has quit IRC | 11:51 | |
| *** armaan has quit IRC | 11:54 | |
| *** armaan has joined #openstack-ansible | 12:05 | |
| *** udesale has quit IRC | 12:05 | |
| mhayden | y'all gotta get mattt to vancouver, evrardjp ;) | 12:17 |
| evrardjp | yes I tried! | 12:19 |
| evrardjp | apparently it was denied | 12:19 |
| *** markvoelker has quit IRC | 12:19 | |
| *** markvoelker has joined #openstack-ansible | 12:24 | |
| *** sep has quit IRC | 12:33 | |
| *** armaan has quit IRC | 12:39 | |
| *** armaan has joined #openstack-ansible | 12:39 | |
| *** hamza21 has quit IRC | 12:42 | |
| *** armaan_ has joined #openstack-ansible | 12:42 | |
| *** armaan_ has quit IRC | 12:44 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Added monitoring tag to monitoring user creation task. https://review.openstack.org/559012 | 12:45 |
| openstackgerrit | Merged openstack/openstack-ansible stable/ocata: Work around nested virt issues on OVH test nodes https://review.openstack.org/559296 | 12:45 |
| *** armaan_ has joined #openstack-ansible | 12:45 | |
| *** armaan has quit IRC | 12:45 | |
| *** sep has joined #openstack-ansible | 12:46 | |
| ArchiFleKs | odyssey4me: yes I have it | 12:47 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible stable/queens: Added monitoring tag to monitoring user creation task. https://review.openstack.org/559320 | 12:47 |
| *** odyssey4me has quit IRC | 12:47 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible stable/pike: Added monitoring tag to monitoring user creation task. https://review.openstack.org/559321 | 12:47 |
| *** odyssey4me has joined #openstack-ansible | 12:47 | |
| *** hamza21 has joined #openstack-ansible | 12:52 | |
| *** sawblade6 has joined #openstack-ansible | 13:04 | |
| evrardjp | prometheanfire: are you there? | 13:04 |
| evrardjp | ping me when you can it's about refstack -- wondering the purpose of the two same images. | 13:04 |
| odyssey4me | refstack, as far as I know, just needs two images | 13:05 |
| odyssey4me | not two of the same | 13:05 |
| evrardjp | yeah so | 13:05 |
| odyssey4me | it's just easier to use two of the same with a small image like cirros | 13:05 |
| evrardjp | here we do two images upload | 13:05 |
| evrardjp | so that they have different ids | 13:06 |
| evrardjp | but tempest would be just fine with the _alt pointing to the same id | 13:06 |
| evrardjp | it would skip those runs | 13:06 |
| evrardjp | so I wonder what should we do | 13:06 |
| odyssey4me | yes, that's how it was before prometheanfire's patch | 13:06 |
| evrardjp | I think we should keep the _alt | 13:06 |
| odyssey4me | but apparently that's not enough for refstack | 13:06 |
| evrardjp | ok | 13:06 |
| evrardjp | so we need two ids | 13:06 |
| odyssey4me | yes | 13:06 |
| evrardjp | that's silly | 13:06 |
| evrardjp | but ok | 13:07 |
| odyssey4me | we could just provide a list, but have the list contain the same image details twice? | 13:07 |
| odyssey4me | (by default) | 13:07 |
| evrardjp | that's what we do | 13:07 |
| evrardjp | ok let me continue my patch | 13:07 |
| evrardjp | you'll understand | 13:07 |
| odyssey4me | what we're doing now is assuming two images, which we shouldn't - and I see that jmccrory is working on a patch there | 13:07 |
| evrardjp | oh | 13:07 |
| *** fdsa has quit IRC | 13:07 | |
| evrardjp | maybe I am working on the same thing | 13:07 |
| evrardjp | I should have known | 13:08 |
| odyssey4me | there should be a list provided to the role with no assumptions about the size of the list | 13:08 |
| evrardjp | nah | 13:08 |
| evrardjp | he takes the last one | 13:08 |
| evrardjp | well | 13:08 |
| boxrick | Howdy guys | 13:08 |
| evrardjp | in fact it's not that bad | 13:08 |
| evrardjp | I had something like | 13:08 |
| odyssey4me | I haven't looked at the patch, just saw it flying by. ;) | 13:08 |
| *** fdsa has joined #openstack-ansible | 13:09 | |
| evrardjp | and tempest_images | length > 1 | 13:09 |
| boxrick | Version 10 pip beta seems to get installed and break things. | 13:09 |
| evrardjp | boxrick: yes, take the last version of OSA and you should have the fix included | 13:09 |
| odyssey4me | boxrick yep, look out for the pbr and pip constraints patches - they're in some releases, but pike I think it still outstanding, but you can use the head of the branch | 13:10 |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible-os_nova master: (WIP) Update to use oslo.messaging services for RPC and Notify https://review.openstack.org/558620 | 13:11 |
| *** gokhan__ has quit IRC | 13:11 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_tempest master: Allow for single tempest image upload https://review.openstack.org/559146 | 13:15 |
| evrardjp | ansmith: woot :) | 13:15 |
| ansmith | evrardjp: still working a few things out | 13:16 |
| *** lvdombrkr has joined #openstack-ansible | 13:16 | |
| evrardjp | ansmith: That's generally implied when WIP :p | 13:17 |
| ansmith | evrardjp: sure does | 13:17 |
| evrardjp | :) | 13:17 |
| ansmith | evrardjp: actually a quick question, I noticed that vhosts names are pre-pended with '/' | 13:17 |
| ansmith | so in url it ends up '//' | 13:18 |
| odyssey4me | that's likely just an oversight | 13:18 |
| ansmith | ok, I may remove the '/' from the vhost names as I work through the services | 13:19 |
| evrardjp | odyssey4me: omg changing that will be a nightmare | 13:19 |
| ansmith | or we can leave it as is | 13:19 |
| evrardjp | wouldn't that be a problem during upgrades? | 13:19 |
| odyssey4me | I think the assumption is that // resolves to / in most URL refs? | 13:19 |
| odyssey4me | if that's not the case, then indeed that'd be a pain in a major upgrade | 13:19 |
| evrardjp | I thought that meant the vhost was nmed differently "/something" instead of "something" | 13:20 |
| ansmith | in the url it should resolve to / but it does change the vhost name added to rabbit | 13:20 |
| *** armaan_ has quit IRC | 13:20 | |
| evrardjp | yeah so maybe worth leaving as is for now, filing a bug for those who want to tackle it :D | 13:20 |
| ansmith | ack, will leave as s | 13:20 |
| ansmith | is | 13:20 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_nova master: Update paste, policy and rootwrap configurations 2018-03-31 https://review.openstack.org/558078 | 13:21 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Update all SHAs for master https://review.openstack.org/558082 | 13:22 |
| *** Guest44281 has quit IRC | 13:22 | |
| *** vakuznet has joined #openstack-ansible | 13:23 | |
| *** jwitko has joined #openstack-ansible | 13:25 | |
| *** jwitko has quit IRC | 13:25 | |
| *** jwitko has joined #openstack-ansible | 13:26 | |
| *** lbragstad has joined #openstack-ansible | 13:28 | |
| mbuil | hwoarang: do you understand why the task - name: debugging apt is not executed in the controllers? http://logs.openstack.org/65/559065/7/check/openstack-ansible-opendaylight-ubuntu-xenial-nv/6e0388a/job-output.txt.gz | 13:29 |
| *** sar has quit IRC | 13:30 | |
| evrardjp | odyssey4me: so no rc this time? | 13:30 |
| *** gkadam has joined #openstack-ansible | 13:30 | |
| evrardjp | I mean we do rcs but we are not in sync with the code | 13:30 |
| odyssey4me | evrardjp Can you give me some context to your statement please? | 13:31 |
| *** throwsb1 has joined #openstack-ansible | 13:31 | |
| evrardjp | no | 13:32 |
| evrardjp | :p | 13:32 |
| evrardjp | https://review.openstack.org/#/c/559031 | 13:32 |
| evrardjp | I find myself funny sumetimez | 13:32 |
| *** esberglu has joined #openstack-ansible | 13:32 | |
| odyssey4me | evrardjp we've never done an rc openstack_release version | 13:32 |
| odyssey4me | well, not to my knowledge | 13:33 |
| odyssey4me | that would add more patches to merge unnecessarily | 13:33 |
| odyssey4me | I'd like to actually do away with openstack_release entirely - or at least have it automatically derived from the git tag or something | 13:34 |
| evrardjp | with setuptools? | 13:35 |
| evrardjp | we can do that | 13:35 |
| odyssey4me | anything other than a static value IMO | 13:35 |
| *** lbragstad has quit IRC | 13:35 | |
| evrardjp | setuptools can give you next version | 13:35 |
| evrardjp | I can do a patch for that | 13:35 |
| evrardjp | it would be a mess for master | 13:35 |
| evrardjp | but the rest is fine | 13:35 |
| odyssey4me | or pbr or whatever | 13:36 |
| evrardjp | that's what pbr is using behind the scenes | 13:36 |
| evrardjp | iir | 13:36 |
| evrardjp | iirc* | 13:36 |
| evrardjp | let me still update that page | 13:36 |
| hwoarang | mbuil: which line in particular? | 13:37 |
| hwoarang | mbuil: please give me the url for the specific timestamp that fails | 13:38 |
| evrardjp | odyssey4me: I am thinking about listing the repos like you suggested. | 13:38 |
| evrardjp | It's not really smart. | 13:38 |
| evrardjp | because it would drift | 13:38 |
| hwoarang | mbuil: i see debuggin apt was executed everywhere | 13:38 |
| evrardjp | should we put a link to governance instead? | 13:38 |
| hwoarang | mbuil: status is 'changed' | 13:38 |
| odyssey4me | evrardjp clearly I'm missing something | 13:39 |
| mbuil | hwoarang: line 72 here ==> https://review.openstack.org/#/c/559065/7/tasks/dependent_neutron_roles.yml | 13:39 |
| hwoarang | mbuil: it appears it was executed everywhere | 13:39 |
| mbuil | or line 69 | 13:39 |
| *** lbragstad has joined #openstack-ansible | 13:39 | |
| odyssey4me | evrardjp I wasn't suggesting you list all the repositories, although linking to the 'independent release' repositories in governance may be a way to do that | 13:39 |
| mbuil | hwoarang: it executed on localhost, agents1 and agents 2 or? | 13:40 |
| hwoarang | mbuil: http://logs.openstack.org/65/559065/7/check/openstack-ansible-opendaylight-ubuntu-xenial-nv/6e0388a/job-output.txt.gz#_2018-04-06_11_00_16_256842 | 13:40 |
| evrardjp | odyssey4me: I think I misread your comment. You didn't mean to add the list of repos to manually create branch | 13:40 |
| odyssey4me | I was mentioning that there are tasks *for the repositories we branch* | 13:40 |
| evrardjp | ok | 13:40 |
| mbuil | hwoarang: agh, thanks, it was somewhere else | 13:41 |
| mbuil | grrrr deb http://ppa.launchpad.net/odl-team/nitrogen/ubuntu sid main. All repos have ubuntu xenial except this one which for whatever reason is converted into ubuntu sid | 13:41 |
| hwoarang | mbuil: fwiw if you want to 'workaround' the issue you might want to use this kind of notation | 13:42 |
| hwoarang | https://github.com/openstack/openstack-ansible-openstack_hosts/blob/master/defaults/main.yml#L139 | 13:42 |
| hwoarang | to add ubuntu repositories and drop the weird ppa: syntax | 13:42 |
| hwoarang | but it would be nice to find out why ppa: broke of course | 13:43 |
| mbuil | hwoarang: could I use that workaround with apt_repository module as we use it? https://github.com/opendaylight/integration-packaging-ansible-opendaylight/blob/master/tasks/add_odl_deb_repo.yml | 13:43 |
| hwoarang | mbuil: no but you can do it here https://github.com/opendaylight/integration-packaging-ansible-opendaylight/blob/master/vars/main.yml#L78 | 13:44 |
| hwoarang | instead of "ppa:odl-team/oxygen" just use 'deb http://ppa.launchpad.net/odl-team/oxygen/ubuntu/ {{ ansible_lsb.codename }}' | 13:45 |
| hwoarang | or something like that... | 13:45 |
| mbuil | hwoarang: ok thanks. How can I fetch the VM which is being used in openstack infra? The XCI VM gets deb http://ppa.launchpad.net/odl-team/oxygen/ubuntu xenial main | 13:46 |
| hwoarang | mbuil: the images are not public as far as i know | 13:46 |
| hwoarang | so you have to build one yourself | 13:46 |
| hwoarang | or ask openstack-infra to reserve a node for you to ssh and look at it | 13:47 |
| mbuil | hwoarang: ok :) let's see how friendly the infra guys are | 13:47 |
| hwoarang | they are usually pretty friendly on that front | 13:47 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Add periodic work for release preparations https://review.openstack.org/559031 | 13:49 |
| *** hwoarang has quit IRC | 13:51 | |
| *** ell0log has left #openstack-ansible | 13:52 | |
| *** hwoarang has joined #openstack-ansible | 13:55 | |
| evrardjp | mbuil: hwoarang you can probably override that deb_repo_url when calling the role. | 14:05 |
| hwoarang | yeah i was proposing the change to the role as a permanent 'workaround' | 14:06 |
| mbuil | evrardjp: yes, we can :). I am just really keen on finding the cause of the problem. When I get desperate, I'll add hwoarang's workaround | 14:06 |
| mbuil | it is sooo weird. Ubuntu should pick up the correct Ubuntu version but chooses another on instead. And only happens in the infra VMs. I should play lottery this weekend | 14:07 |
| evrardjp | I meant in os_neutron | 14:07 |
| evrardjp | :) | 14:07 |
| evrardjp | https://github.com/openstack/openstack-ansible-os_neutron/blob/5c0f020bc6d80a3dd10be3724b80b76b29ca7a0d/tasks/dependent_neutron_roles.yml#L42 | 14:08 |
| evrardjp | but yeah, not needed if you find out the root cause | 14:08 |
| *** adriano has quit IRC | 14:09 | |
| *** adriano__ has joined #openstack-ansible | 14:10 | |
| *** ell0log has joined #openstack-ansible | 14:11 | |
| *** DanyC has quit IRC | 14:12 | |
| *** hachi has joined #openstack-ansible | 14:12 | |
| *** DanyC has joined #openstack-ansible | 14:12 | |
| hwoarang | mbuil: another solution would be to create a simple reproducer (like i said before by modifying tests/test.yml with 2 tasks) so infra can easily look at it themselves | 14:14 |
| hwoarang | this will also rule out any OSA specific brokeness | 14:14 |
| hwoarang | since we haven't touched much up to that point | 14:14 |
| *** hamza21 has quit IRC | 14:15 | |
| *** DanyC has quit IRC | 14:17 | |
| cloudnull | mornings | 14:18 |
| cloudnull | jrosser: it looks like you're right, "tunnel_bridge" is only ever referenced in our config examples. | 14:20 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Add periodic work for release preparations https://review.openstack.org/559031 | 14:20 |
| evrardjp | hwoarang: odyssey4me cloudnull if you could vote on https://review.openstack.org/#/c/559146/2 that would be great | 14:21 |
| evrardjp | two of you should do the trick :p | 14:21 |
| evrardjp | sorry for the overkill ping :p | 14:21 |
| *** lvdombrkr has quit IRC | 14:22 | |
| evrardjp | jmccrory: I've listed all the groups, and that's freaking scary. | 14:23 |
| odyssey4me | evrardjp done | 14:23 |
| evrardjp | thanks | 14:23 |
| *** haad1 has quit IRC | 14:23 | |
| evrardjp | odyssey4me: not the same patch I suggested to vote :) | 14:23 |
| evrardjp | but thanks! | 14:24 |
| odyssey4me | oh, that tempest/glance patch looks like we're just digging a deeper hole | 14:25 |
| evrardjp | mnaser: if you got two minutes, I'd be super happy if we could discuss the telemetry groups used for deploy | 14:25 |
| evrardjp | odyssey4me: yes. | 14:25 |
| evrardjp | but that should be fine | 14:25 |
| *** armaan has joined #openstack-ansible | 14:26 | |
| *** cjloader has joined #openstack-ansible | 14:26 | |
| evrardjp | if 0 image and ask for glance, fail. If 1 image, use that one for the upload, and refer to its id in the two. If using refstack, and have 2 images or more, use the first and last one. | 14:26 |
| *** cjloader has quit IRC | 14:26 | |
| *** cjloader has joined #openstack-ansible | 14:27 | |
| jrosser | cloudnull: ah ok, i had a really good dig and couldnt find anything | 14:27 |
| adcpdk | Morning everyone! | 14:28 |
| adcpdk | Is there are any solution of adding ctrl-alt-delete button to the Spice console? | 14:28 |
| adcpdk | if this is what needs to be done? https://bugs.launchpad.net/openstack-ansible/icehouse/+bug/1423669 | 14:28 |
| adcpdk | I am not sure where should I edit that CSS and html files | 14:28 |
| openstack | Launchpad bug 1423669 in spice-html5 (Ubuntu) "CTRL+ALT+DEL Button missing" [Undecided,New] | 14:28 |
| jrosser | cloudnull: btw i braindumped about provider_networks into here https://etherpad.openstack.org/p/osa-provider-networks | 14:28 |
| odyssey4me | evrardjp yeah, better than it was | 14:29 |
| evrardjp | adcpdk: is it fixed in spice? | 14:29 |
| evrardjp | we are just consuming it | 14:30 |
| *** gokhan has joined #openstack-ansible | 14:30 | |
| odyssey4me | adcpdk switch to using novnc, or use a spice fork | 14:30 |
| *** armaan has quit IRC | 14:30 | |
| odyssey4me | unless spice has implemented something themselves - not sure | 14:30 |
| adcpdk | Hi evrardjp__ I don't see crtl-alt-delete button | 14:30 |
| *** hamzy has quit IRC | 14:30 | |
| *** udesale has joined #openstack-ansible | 14:31 | |
| adcpdk | I though maybe to use novnc instead, but not sure what playbooks to run and what changes to make | 14:31 |
| evrardjp | it should be documented in the nova role I guess | 14:33 |
| evrardjp | nova role documentation | 14:33 |
| evrardjp | https://docs.openstack.org/openstack-ansible-os_nova/latest/app-powervm.html#enabling-vnc-console | 14:33 |
| evrardjp | it's not for powervm but you get the gist :) | 14:34 |
| adcpdk | right, thanks | 14:34 |
| adcpdk | what playbooks should I run afterwards? | 14:34 |
| odyssey4me | if you have existing instances, though, then they don't just get novnc | 14:34 |
| odyssey4me | you'll have to do some database wrangling, then hard reboot/migrate them | 14:35 |
| odyssey4me | or delete/rebuild | 14:35 |
| evrardjp | adcpdk: start clean :p | 14:35 |
| evrardjp | to avoid what odyssey4me said | 14:36 |
| *** hamzy has joined #openstack-ansible | 14:36 | |
| adcpdk | thanks | 14:36 |
| adcpdk | by starting clean you mean re-run all three playbooks? | 14:36 |
| odyssey4me | well, I'd love to see a patch for us to set a primary, but support all console type - this will be best for multi arch environments | 14:36 |
| adcpdk | hosts, infrustructure, openstack? | 14:36 |
| evrardjp | odyssey4me: is that even possible? | 14:36 |
| odyssey4me | evrardjp yes, of course | 14:36 |
| evrardjp | how would nova know which one to use? | 14:37 |
| evrardjp | how would you point to the proper one in libvirt? | 14:37 |
| odyssey4me | evrardjp it would require the use of host aggregates though, as a hypervisor, I think, can only do one | 14:37 |
| evrardjp | That's interesting | 14:37 |
| evrardjp | yeah | 14:37 |
| evrardjp | that's complex :p | 14:37 |
| odyssey4me | it is possible though, but our console container doesn't allow running multiple | 14:37 |
| evrardjp | spice is almost not maintained | 14:37 |
| evrardjp | I think people should stick to novnc | 14:37 |
| *** udesale has quit IRC | 14:38 | |
| cloudnull | jrosser: yea looking back that key was used in the liberty-ish timeframe, just was never cleaned up from config . | 14:38 |
| odyssey4me | so I'd prefer to see something like the glance storage back-ends... this is the default, and these are also supported... then the combined list is installed on the console container, and a deployer can then do host aggregates, etc to support serial for the s390, vnc for windows and spice for whatever spice is good for | 14:38 |
| *** kstev has joined #openstack-ansible | 14:38 | |
| cloudnull | good find that | 14:38 |
| odyssey4me | well, perhaps we should switch the default to novnvc ;) | 14:38 |
| evrardjp | odyssey4me: OH | 14:39 |
| evrardjp | interesting | 14:39 |
| evrardjp | the github mirror is not in sync with their upstream | 14:39 |
| evrardjp | https://cgit.freedesktop.org/spice/spice-html5/ | 14:39 |
| evrardjp | odyssey4me: it's still lagging behind novnc indeed | 14:40 |
| evrardjp | we should probably switch back | 14:40 |
| evrardjp | worth asking the other ppl | 14:40 |
| *** haad1 has joined #openstack-ansible | 14:40 | |
| evrardjp | let me add that to our meetings page | 14:40 |
| odyssey4me | well, if we can cater to multiple consoles in the console container, then switching the default will be easy | 14:40 |
| odyssey4me | if we persist with only catering for a single console type, then we'd have to declare big warnings that it's changing | 14:41 |
| evrardjp | I don't know -- I'd rather abstain talking about it. | 14:42 |
| adcpdk | so currently there is no way of just changing the value in nova yml file and re-run that playbook? | 14:43 |
| evrardjp | adcpdk: that's not how you should do configuration management :p | 14:44 |
| evrardjp | configuraton data should go into your user_* variables, or anywhere to override things | 14:44 |
| evrardjp | then only run the playbooks | 14:45 |
| evrardjp | manually editing a configuration file seems bad to me | 14:45 |
| evrardjp | but up to you :) | 14:45 |
| evrardjp | d34dh0r53: are you there? | 14:49 |
| *** armaan has joined #openstack-ansible | 14:49 | |
| cloudnull | jrosser: nice write up | 14:50 |
| d34dh0r53 | evrardjp: yep | 14:51 |
| *** gkadam_ has joined #openstack-ansible | 14:51 | |
| *** epalper has quit IRC | 14:52 | |
| evrardjp | odyssey4me: cloudnull do you remember when we deployed ceilometer long ago, we had mongo deployed right? | 14:52 |
| evrardjp | Did we have clustering ? | 14:52 |
| odyssey4me | no | 14:52 |
| cloudnull | nope | 14:52 |
| odyssey4me | I mean, yes I remember - and no we didn't cluster. | 14:52 |
| mnaser | evrardjp: i have 2 minutes to make you super happy to discuss telemetry groups :) | 14:52 |
| evrardjp | great mnaser | 14:53 |
| odyssey4me | We said 'you build your cluster, then you point config at it'. | 14:53 |
| *** gkadam has quit IRC | 14:53 | |
| cloudnull | ^ | 14:53 |
| mnaser | odyssey4me: because we're openstack, we don't do web scale | 14:53 |
| mnaser | :( | 14:53 |
| odyssey4me | in our gates we simply installed mongodb on the host and did nothing much else | 14:53 |
| *** sar has joined #openstack-ansible | 14:53 | |
| odyssey4me | mnaser well, mongodb === /dev/null ') | 14:53 |
| cloudnull | magodb is where its at, https://github.com/dcramer/mangodb | 14:54 |
| cloudnull | all the data retention of mongodb , but HYPER Web Scale | 14:54 |
| odyssey4me | go serverless: https://devnull-as-a-service.com/features/ | 14:54 |
| cloudnull | hahahahaha! | 14:55 |
| prometheanfire | evrardjp: odyssey4me more refstack? | 14:55 |
| cloudnull | https://devnull-as-a-service.com/pricing/ - made me LOL | 14:55 |
| odyssey4me | you can even pay $5k per year for your own appliance :) https://devnull-as-a-service.com/code/ | 14:55 |
| odyssey4me | it has an api of course | 14:55 |
| evrardjp | prometheanfire: don't worry about it | 14:56 |
| evrardjp | check the other ping I sent you :p | 14:56 |
| cloudnull | odyssey4me: the gift to the opensource community is priceless. | 14:56 |
| cloudnull | `curl https://devnull-as-a-service.com/one-less-to-go.sh | bash` ftw! | 14:57 |
| evrardjp | cloudnull: hahah | 14:57 |
| prometheanfire | evrardjp: tweaking how? | 14:57 |
| evrardjp | "Dedicated Trashbin" | 14:57 |
| evrardjp | HAHAH one less to go! | 14:58 |
| evrardjp | prometheanfire: don't know? what do you mean? | 14:58 |
| Bico_Fino | Morning! | 14:59 |
| prometheanfire | evrardjp: you are the one that suggested tweaking ovh nova jobs | 15:02 |
| prometheanfire | :P | 15:02 |
| evrardjp | prometheanfire: I didn't know what you were talking about | 15:03 |
| evrardjp | yes | 15:03 |
| evrardjp | so | 15:03 |
| evrardjp | if your cross repo tests are running virt things, like nova maybe | 15:04 |
| evrardjp | you might want to check if you're running on ovh | 15:04 |
| evrardjp | but because you're only triggering nova gates, I am sure it's probably to check with them | 15:04 |
| evrardjp | that's a weird sentence | 15:04 |
| evrardjp | I meant | 15:05 |
| evrardjp | for ovh nodes, run qemu without kvm. | 15:05 |
| evrardjp | if you don't have the power to change that, check the jobs you trigger are doing it | 15:05 |
| evrardjp | if you don't think it matters, and you still hit timeouts, check at the reason :P | 15:05 |
| evrardjp | we didn't have to extend our timeouts -- so you probably shouldn't need it | 15:06 |
| prometheanfire | evrardjp: ya, got it :P | 15:08 |
| devx | \query cloudnull | 15:09 |
| cloudnull | o/ | 15:09 |
| devx | lol my irc skills are off this morning | 15:09 |
| *** gkadam_ has quit IRC | 15:10 | |
| *** gokhan has quit IRC | 15:10 | |
| *** rmart04 has quit IRC | 15:10 | |
| *** armaan has quit IRC | 15:11 | |
| *** armaan has joined #openstack-ansible | 15:11 | |
| *** gokhan has joined #openstack-ansible | 15:11 | |
| *** tomtomtom has joined #openstack-ansible | 15:15 | |
| *** hachi has quit IRC | 15:17 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova master: Re-order database migrations https://review.openstack.org/559354 | 15:17 |
| tomtomtom | Hello, currently running into this issue: http://paste.openstack.org/show/718607/ anyone got any ideas what may be the issue. I've run the chroot command by hand and found that repo.keys is not copied to the base machine image in the chroot. Does that make any sense? What am I missing here? | 15:17 |
| tomtomtom | I'm running a ocata->pike upgrade btw | 15:17 |
| *** armaan has quit IRC | 15:18 | |
| mnaser | so has anyone experimented with moving more and more things to is_metal ? | 15:19 |
| mnaser | for example, unbound and memcached seems like two really good targets | 15:19 |
| odyssey4me | evrardjp we helped find a nova bug, hence https://review.openstack.org/559354 (see the related bug) | 15:19 |
| *** yolanda_ is now known as yolanda | 15:20 | |
| mnaser | oooo | 15:21 |
| mnaser | evrardjp: thanks for https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/env.d/aio_metal.yml.example | 15:21 |
| mnaser | aha | 15:21 |
| mnaser | i'm going to try adopting that | 15:21 |
| adcpdk | Ok, instead of re-running all the playbooks just to add the ctrl-alt-delete functionality I chose to edit the HTML file and got the desired result: | 15:21 |
| adcpdk | https://bugs.launchpad.net/openstack-ansible/icehouse/+bug/1423669 | 15:21 |
| openstack | Launchpad bug 1423669 in spice-html5 (Ubuntu) "CTRL+ALT+DEL Button missing" [Undecided,New] | 15:21 |
| adcpdk | root@node01:/var/lib/lxc/infra1_nova_console_container-e5931327/rootfs/usr/share/spice-html5# sudo vim spice_auto.html | 15:21 |
| evrardjp | mnaser: make it simpler | 15:22 |
| evrardjp | that's just to show :) | 15:22 |
| evrardjp | you can use yaml anchors | 15:22 |
| evrardjp | odyssey4me: will check at this after the meeting | 15:22 |
| mnaser | i need to get around learning to use yaml anchors | 15:22 |
| *** sawblade6 has quit IRC | 15:23 | |
| *** sawblade6 has joined #openstack-ansible | 15:23 | |
| adcpdk | that spice_auto.html changes should be added to the main repository for sure | 15:24 |
| mnaser | evrardjp: broke baremetal on my first try (unbound) aha | 15:26 |
| mnaser | fatal: [controller-1-dev]: FAILED! => {"msg": "'log_dirs' is undefined"} | 15:26 |
| * mnaser is checking | 15:26 | |
| evrardjp | oh great interesting | 15:28 |
| evrardjp | that's why I wanted to have it voting | 15:28 |
| mnaser | i dont think unbound deploys in gate too | 15:28 |
| mnaser | im checking | 15:28 |
| adcpdk | evrardjp are you talking about the spice_auto.html changes? | 15:29 |
| mnaser | https://github.com/openstack/openstack-ansible/blob/master/playbooks/os-heat-install.yml#L45-L49 | 15:29 |
| evrardjp | adcpdk: ? | 15:30 |
| mnaser | so it looks like other projects have added the log_dirs variable | 15:30 |
| mnaser | and i guess unbound hasnt | 15:30 |
| evrardjp | mnaser: OMG these are my nightmares | 15:30 |
| tomtomtom | This is also consistently failing without any reason: http://paste.openstack.org/show/718609/ Anyone got any ideas? | 15:30 |
| mnaser | :D | 15:30 |
| evrardjp | ppl just don't stop changing them and breaking gates | 15:30 |
| mnaser | i will go over a patch which cleans up as much as i can | 15:30 |
| evrardjp | for me these whole folder linking is nonsense | 15:30 |
| mnaser | i can probably hack around it with group_vars until these changes merge | 15:31 |
| evrardjp | just use the damn /var/log folder | 15:31 |
| evrardjp | tomtomtom: there is probably a reason, and you should check the build logs | 15:31 |
| adcpdk | nevermind, I though you said "oh great interesting" is referred to the message I send about the spice and that I were able to add the ctrl-alt-delete button | 15:31 |
| odyssey4me | evrardjp we used to, then someone complained that container logs were taking up too much space | 15:32 |
| evrardjp | odyssey4me: I see. | 15:32 |
| odyssey4me | then someone complained that logs were in different places for different things | 15:32 |
| evrardjp | I see too. | 15:32 |
| odyssey4me | then we bind-mounted log folders from containers onto the host | 15:32 |
| odyssey4me | and all hell broke loose | 15:32 |
| evrardjp | then complained it's not central logged? | 15:32 |
| odyssey4me | so, why don't we just make everything journal instead | 15:33 |
| evrardjp | we should just get rid of id | 15:33 |
| evrardjp | it* | 15:33 |
| evrardjp | yeah | 15:33 |
| *** armaan has joined #openstack-ansible | 15:33 | |
| evrardjp | just log to syslog | 15:33 |
| mnaser | 34 playbooks with -install in name, only 22 have 'log_dirs' in them | 15:33 |
| odyssey4me | journal everything, then kill the bind mounts, the synlinks and everything | 15:33 |
| evrardjp | mnaser: \o/ | 15:33 |
| evrardjp | odyssey4me: yes | 15:33 |
| *** openstackgerrit has quit IRC | 15:33 | |
| evrardjp | ship to host if you want (nspawn) | 15:33 |
| evrardjp | for lxc deal with it. | 15:33 |
| evrardjp | ship to graylog :p | 15:34 |
| mnaser | https://bugs.launchpad.net/openstack-ansible/+bug/1761781 going to work on something now | 15:34 |
| openstack | Launchpad bug 1761781 in openstack-ansible "log_dirs variable missing from roles" [Undecided,New] | 15:34 |
| odyssey4me | given queens is now doing the journal forwarding, we've got all the pieces in place - we just need the openstack services to log to the journal now | 15:34 |
| odyssey4me | sorry, not queens - rocky | 15:34 |
| tomtomtom | evrardjp, yes build log only shows this error: http://paste.openstack.org/show/718610/ then continues on, but setup-hosts.yml fails on the error | 15:35 |
| mnaser | ironic doesnt even have os-log-dir-setup.yml | 15:36 |
| mnaser | woo lol | 15:36 |
| evrardjp | odyssey4me: a few of them already do | 15:36 |
| evrardjp | apache doesn't | 15:36 |
| evrardjp | not sure about nginx | 15:36 |
| *** Leo_m has joined #openstack-ansible | 15:36 | |
| evrardjp | mnaser: it's not because they have it that it fails | 15:37 |
| evrardjp | we've done some hpfbm to make sure it works in the role without failures | 15:37 |
| evrardjp | odyssey4me: that's litterally the best acronym I will remember from you. | 15:37 |
| evrardjp | :D | 15:38 |
| mnaser | wait so | 15:39 |
| mnaser | if i understand correctly | 15:39 |
| mnaser | include: common-tasks/os-log-dir-setup.yml should not be ran at all for unbound if it has no log_dirs because we're literally not "centralizing" anything, right? | 15:39 |
| mnaser | so the bug is that it's there, rather than the fact it doesnt have log_dirs? | 15:39 |
| evrardjp | wait. it depends on what you're running in containers or not | 15:40 |
| odyssey4me | um, yes I think | 15:40 |
| mnaser | https://github.com/openstack/openstack-ansible/blob/stable/queens/playbooks/unbound-install.yml#L22 | 15:40 |
| mnaser | https://github.com/openstack/openstack-ansible/blob/stable/queens/playbooks/common-tasks/os-log-dir-setup.yml | 15:40 |
| mnaser | which really means its doing nothing but create those folders | 15:40 |
| evrardjp | https://github.com/openstack/openstack-ansible/blob/stable/queens/playbooks/common-tasks/os-log-dir-setup.yml#L31 | 15:40 |
| mnaser | and if we're not giving it any, it skips on containers (hence not failing) | 15:41 |
| mnaser | how do we feel about with_items: "{{ log_dirs }}" => with_items: "{{ log_dirs|default([]) }}" | 15:41 |
| mnaser | until that whole centralized log stuff fades away i guess | 15:41 |
| evrardjp | oh yes | 15:41 |
| *** openstackgerrit has joined #openstack-ansible | 15:41 | |
| openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_neutron master: DNM: Testing ODL gates https://review.openstack.org/559065 | 15:41 |
| evrardjp | I didn't see that was the issue in your play | 15:41 |
| odyssey4me | that common task set creates a log folder in /openstack/logs if the service is on metal, otherwise it'd be in a container and doesn't need the symlink | 15:41 |
| odyssey4me | however, that creates a problem in roles when it's trying to create the log dir - because ansible wants the state to be 'directory', but the state is 'link' | 15:42 |
| evrardjp | yes that's the main cause of issues | 15:42 |
| evrardjp | but here mnaser has found a bug I didn't fix on my on metal work | 15:42 |
| odyssey4me | so each role needs to be adjusted to use the realpath filter, like so: https://github.com/openstack/openstack-ansible-os_tempest/blob/b4140c9eedaedc86ca2675695ff241da7644cc81/tasks/tempest_post_install.yml#L42 | 15:42 |
| evrardjp | because I never used unbound | 15:43 |
| mnaser | likely because unbound probably isn't being deployed | 15:43 |
| odyssey4me | well, of course - unbound is not gated | 15:43 |
| mnaser | unless you add unbound_hosts: | 15:43 |
| mnaser | ^ | 15:43 |
| evrardjp | yeah | 15:43 |
| mnaser | so you don't have to feel bad :) | 15:43 |
| evrardjp | I know | 15:43 |
| evrardjp | but still | 15:43 |
| evrardjp | :p | 15:43 |
| evrardjp | you patch it real quick? or should I do it? | 15:43 |
| odyssey4me | we cannot hope to gate absolutely every combination - some things will always have to be second tier | 15:43 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: Add default value for log_dirs in os-log-dir-setup.yml https://review.openstack.org/559361 | 15:44 |
| mnaser | evrardjp, odyssey4me ^ | 15:44 |
| odyssey4me | note that some roles will have a bash script, or a stat dance... but they all could be simplified to use the realpath filter like so: https://github.com/openstack/openstack-ansible-os_tempest/commit/b4140c9eedaedc86ca2675695ff241da7644cc81 | 15:44 |
| evrardjp | thnaks | 15:44 |
| mnaser | and ill likely backport to queens after | 15:44 |
| mnaser | until then ill just run the unbound playbook with -elog_dirs=[] or something | 15:45 |
| odyssey4me | of course I'd prefer us to stop using weird folders like /openstack and do normal things like use /opt instead | 15:45 |
| *** haad1 has quit IRC | 15:45 | |
| mnaser | or maybe leave logs in their place and let me people fix their partition schemes or whatever :) | 15:46 |
| openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_neutron master: DNM: Testing ODL gates https://review.openstack.org/559065 | 15:46 |
| mnaser | found anoooooother bug woo | 15:47 |
| mnaser | we're cleaning up unbound today | 15:47 |
| mnaser | listens by default to 0.0.0.0, lxc-dnsmasq listens on 10.0.3.1 => error: bind: address already in use | 15:47 |
| mnaser | thankfully logan- has already setup unbound_listen_interface so i think all we need to do is make it listen on the ip of the host/container rather than an open 0.0.0.0 | 15:48 |
| mbuil | hwoarang: this is what you meant, right? https://review.openstack.org/#/c/559065/9/tests/test.yml | 15:49 |
| *** pcaruana has quit IRC | 15:49 | |
| evrardjp | agreed with mnaser on the fix your partitions | 15:50 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova master: Re-order database migrations https://review.openstack.org/559354 | 15:50 |
| evrardjp | clean you hosts | 15:50 |
| *** chyka has joined #openstack-ansible | 15:50 | |
| mnaser | quickly before i start digging, is there a go-to ansible variable for getting the ip of the container/host/etc? | 15:51 |
| evrardjp | mnaser: the fun begins | 15:51 |
| mnaser | do we rely on ansible_host or? | 15:51 |
| evrardjp | mnaser: why do you run lxc though? | 15:51 |
| evrardjp | If you're running things on metal? | 15:51 |
| odyssey4me | and yes, I meant that we should use normal paths for things like /var/cache or /var/log | 15:51 |
| odyssey4me | right now we do kinda like to litter the hosts | 15:51 |
| evrardjp | odyssey4me: It's a party over there! | 15:51 |
| evrardjp | the more folder we touch, the more fun we have ! | 15:51 |
| evrardjp | wooot | 15:51 |
| mnaser | evrardjp: transitionary right now, ideally i'd like to have all metal | 15:51 |
| mnaser | buuut im sure there will be roles that might be breaking | 15:52 |
| mnaser | so i'd like to have the option of both open | 15:52 |
| evrardjp | the on metal job skips many things on an AIO | 15:52 |
| evrardjp | for example it doesn't run haproxy | 15:52 |
| evrardjp | because the services run already on the host | 15:52 |
| evrardjp | so if you have multiple on metal nodes, you should probably pay attention to where you place your haproxy | 15:53 |
| evrardjp | (either in a container, or outside) | 15:53 |
| *** armaan has quit IRC | 15:53 | |
| mnaser | yeah but haproxy is already on metal i thinkkkk | 15:53 |
| evrardjp | mnaser: mmmm | 15:54 |
| evrardjp | so you'll have issues | 15:54 |
| mnaser | guess ill have to fix them | 15:54 |
| mnaser | :) | 15:54 |
| evrardjp | you can't bind on 0.0.0.0:port1 multiple times | 15:54 |
| mnaser | well the same exact issue as unbound here | 15:54 |
| mnaser | the way we would work with that was that haproxy was bound to the vip only | 15:54 |
| mnaser | and services were bound to <hostip> | 15:55 |
| mnaser | and that way, containers or not, you never have a clash | 15:55 |
| evrardjp | that is fair, but what is hostip? | 15:55 |
| evrardjp | you can't assume the interface name | 15:55 |
| evrardjp | you can't assume the bridge name | 15:55 |
| mnaser | so ctrl1 is 1.1.1.1, ctrl2 is 2.2.2.2, ctrl3 is 3.3.3.3 and vip is 5.5.5.5 | 15:55 |
| mnaser | keystone listens on 1.1.1.1:5000 | 15:56 |
| evrardjp | I know how to do it for a specific deploy, don't get me wrong | 15:56 |
| evrardjp | :p | 15:56 |
| mnaser | oh yes but i dont care about interface or bridge name | 15:56 |
| mnaser | im just relying on the interface/ip | 15:56 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova master: Re-order database migrations https://review.openstack.org/559354 | 15:56 |
| evrardjp | what I mean is that it depends on ansible_host for example , but for some, they won't be using that interface for the trafic of API/data/whatever | 15:56 |
| mnaser | evrardjp: okay, you're right, but for that case, we're talking mostly api<->haproxy interaction and that usually goes through the | 15:57 |
| mnaser | whatever runs br-mgmt | 15:57 |
| evrardjp | well that's not what I have seen in some deploys | 15:57 |
| evrardjp | so I didn't want to break things | 15:57 |
| mnaser | so they have br-api or something i guess | 15:57 |
| evrardjp | well I think if you change that, make sure it's something that's overridable | 15:58 |
| mnaser | absolutely | 15:58 |
| mnaser | there's other haproxy stuff i wanted to touch (i.e. we don't run anything on ports, we use virtualhosts in haproxy) | 15:58 |
| evrardjp | if possible defaulting to current behavior, or with a reno if behavior changes | 15:58 |
| mnaser | not gonna make the world suffer dealing with dns :D | 15:58 |
| mnaser | yeah, i don't want to change the default behavior honestly | 15:59 |
| evrardjp | mnaser: dns is fine for haproxy | 15:59 |
| evrardjp | virtualhosts? | 15:59 |
| evrardjp | What do you mean? | 15:59 |
| evrardjp | you use different uris for different services? | 15:59 |
| mnaser | so like, auth.domain.com rather than domain.com:5000 | 15:59 |
| mnaser | yeah | 15:59 |
| evrardjp | instead of ports? | 15:59 |
| evrardjp | well that's fine | 15:59 |
| evrardjp | I strongly encourage you to run your own haproxy role if ours is not good enough | 16:00 |
| mnaser | if anything though, for the interface changes, maybe make it when: is_metal | 16:00 |
| evrardjp | maybe worth asking to logan | 16:00 |
| evrardjp | mnaser: I think it doesn't matter if you're on metal or not | 16:00 |
| evrardjp | if you want more "security" you shouldn't listen everywhere anyway | 16:00 |
| mnaser | tru.dat | 16:01 |
| evrardjp | it makes sense to be more surgical | 16:01 |
| evrardjp | but taking assumptions is hard | 16:01 |
| evrardjp | in our project at least | 16:01 |
| evrardjp | because we've been very open to anything | 16:01 |
| mnaser | i see, so a change making unbound listen on ansible_host rather than 0.0.0.0 wouldn't be ideal | 16:01 |
| evrardjp | that's changing the default | 16:02 |
| evrardjp | :p | 16:02 |
| evrardjp | For master it's fine | 16:02 |
| evrardjp | for backport it won't be as fine | 16:02 |
| mnaser | i guess that can just be a local override then in our stuff not to break $world | 16:02 |
| mnaser | i think ideally i'll override things and see how it all works together, once its all working good, then start bringing stuff up | 16:03 |
| mnaser | rather than like bringing up a bunch of patches and then getting them merged and realizing "oops, this doesn't work" | 16:03 |
| odyssey4me | :) that'd be nice | 16:03 |
| mnaser | cause i'm sure we'll run into wild things down the line of baremetal | 16:03 |
| xgerman_ | evrardjp: https://review.openstack.org/#/c/553630/ - wonder what you think of this version | 16:03 |
| evrardjp | mnaser: yes, I think so | 16:04 |
| odyssey4me | fork the repositories - implement fixes you like, then when you have something that's a working set - propose upstream in the hopes that you don't have to carry that fork once you upgrade | 16:04 |
| *** chyka has quit IRC | 16:04 | |
| mnaser | odyssey4me: well the nice thing is we're not forking as much as just having a bigger extra vars file :D | 16:04 |
| mnaser | aha | 16:04 |
| odyssey4me | not using containers is nice in some ways, but I find myself wondering what the driver is | 16:04 |
| *** chyka has joined #openstack-ansible | 16:05 | |
| odyssey4me | right now you can blow away a container and replace it wholesale, whic for something like rabbitmq is damned useful | 16:05 |
| mnaser | odyssey4me: thats why i want to balance things out, like unbound in a container is overkill imho | 16:05 |
| mnaser | but i'd like to keep something like the repo container still its own thing cause i (feel) its pretty intrusive and does a lot of things in the system | 16:06 |
| odyssey4me | oh? I'm curious why you think that is | 16:06 |
| evrardjp | xgerman_: I've marked it to my to review list | 16:06 |
| evrardjp | thanks! | 16:06 |
| mnaser | i mean carrying sshd, a bunch of consoles, dhcp client, rsyslog, crond, etc | 16:06 |
| mnaser | all just for a single service that can run alone | 16:06 |
| evrardjp | odyssey4me: I heard there exist a technology call virtualization that works quite right. | 16:07 |
| evrardjp | :p | 16:07 |
| odyssey4me | ah, well - we'll sort that all out soon enough - nspawn paves the way for us to do app containers if we want to | 16:07 |
| evrardjp | also lxd works | 16:08 |
| evrardjp | :) | 16:08 |
| mnaser | yeah app containers would be much better | 16:08 |
| odyssey4me | only on ubuntu, but yeah it's pretty nice | 16:08 |
| hwoarang | mbuil: yes | 16:08 |
| *** savvas__ has quit IRC | 16:08 | |
| *** savvas has joined #openstack-ansible | 16:09 | |
| mnaser | also i'm not a fan of running a bridge and all those containers and ips that come with them, management becomes much harder tbh | 16:09 |
| odyssey4me | yeah, that's a thing | 16:09 |
| odyssey4me | I would like to see us treat an all metal and containerised deployment equally, personally. | 16:10 |
| odyssey4me | Mainly to cater for the people who need an easier entrypoint. | 16:10 |
| openstackgerrit | Amy Marrich (spotz) proposed openstack/openstack-ansible master: [DOC] Update for Newton EOL https://review.openstack.org/551316 | 16:11 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova master: Re-order database migrations https://review.openstack.org/559354 | 16:11 |
| *** savvas has quit IRC | 16:13 | |
| *** hachi has joined #openstack-ansible | 16:13 | |
| *** savvas has joined #openstack-ansible | 16:16 | |
| *** savvas has quit IRC | 16:20 | |
| *** savvas has joined #openstack-ansible | 16:26 | |
| *** savvas has quit IRC | 16:30 | |
| *** savvas has joined #openstack-ansible | 16:35 | |
| mbuil | hwoarang: I don't think your idea worked :( https://review.openstack.org/#/c/559065/9 | 16:39 |
| *** savvas has quit IRC | 16:39 | |
| hwoarang | mbuil: test.yml is supposed to be a playbook | 16:39 |
| hwoarang | you have no hosts there, no tasks | 16:39 |
| hwoarang | what do you expect ansible to do | 16:40 |
| mbuil | hwoarang: right, sorry. I meant it because of the 47m 21s it took | 16:40 |
| mbuil | the upgrade must be defined somewhere else | 16:40 |
| hwoarang | why does the upgrade matter? | 16:41 |
| hwoarang | you only care about the openstack-ansible-functional-ubuntu-xenial results | 16:41 |
| hwoarang | no? | 16:41 |
| mbuil | yes, but I had to wait 47 minutes | 16:41 |
| mbuil | I thought I could just wait 5 minutes and then get the results | 16:42 |
| hwoarang | ok then patch zuul.d/jobs to remove all the jobs except the one you need :) | 16:42 |
| vakuznet | is Marc Gariepy here? | 16:43 |
| odyssey4me | mgariepy appears to have disappeared off the map, unfortunately vakuznet | 16:44 |
| *** armaan has joined #openstack-ansible | 16:44 | |
| openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_neutron master: DNM: Testing ODL gates https://review.openstack.org/559065 | 16:46 |
| tomtomtom | So I set the delay to be much longer, and now the lxc_cache_create says: invoke-rc.d: Could not determine runlevel. So the cache never gets created and the setup-hosts fails and does not continue | 16:46 |
| tomtomtom | It seems that the chroot'ed machines that are responsible for this process die whenever they have an issue.... I do not see them completing | 16:48 |
| vakuznet | https://github.com/openstack/openstack-ansible-os_nova/blob/master/tasks/nova_selinux.yml#L32 - I think "not in" should be "in", if you want to delete what is in. | 16:50 |
| *** electrofelix has quit IRC | 16:50 | |
| *** kstev1 has joined #openstack-ansible | 16:50 | |
| vakuznet | mhayden ^ | 16:51 |
| mhayden | vakuznet: i believe you are correct | 16:52 |
| mhayden | vakuznet: would you like to submit the patch? | 16:52 |
| *** kstev has quit IRC | 16:52 | |
| vakuznet | sure. (just need to finish upgrade) | 16:53 |
| *** savvas has joined #openstack-ansible | 16:53 | |
| *** mbuil has quit IRC | 16:53 | |
| mhayden | thanks for finding that ;) | 16:55 |
| *** savvas has quit IRC | 16:58 | |
| *** savvas has joined #openstack-ansible | 17:01 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: Do not generate the html ARA report any more https://review.openstack.org/559078 | 17:09 |
| *** savvas has quit IRC | 17:10 | |
| *** savvas has joined #openstack-ansible | 17:15 | |
| *** tomtomtom has quit IRC | 17:18 | |
| *** savvas has quit IRC | 17:20 | |
| *** armaan has quit IRC | 17:20 | |
| cloudnull | tomtomtom, best thing to do in this case is to blow away the image cache and allow the playbooks to start fresh | 17:24 |
| openstackgerrit | Merged openstack/openstack-ansible master: Reinstate SUSE testing in periodics https://review.openstack.org/559041 | 17:24 |
| *** savvas has joined #openstack-ansible | 17:25 | |
| *** savvas has quit IRC | 17:30 | |
| openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_neutron master: DNM: Testing ODL gates https://review.openstack.org/559065 | 17:31 |
| *** savvas has joined #openstack-ansible | 17:35 | |
| *** flemingo has joined #openstack-ansible | 17:37 | |
| *** flemingo_ has joined #openstack-ansible | 17:38 | |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia master: Downloads amphora images from artifact storage https://review.openstack.org/555084 | 17:39 |
| *** savvas has quit IRC | 17:40 | |
| *** flemingo has quit IRC | 17:42 | |
| *** fleming__ has joined #openstack-ansible | 17:43 | |
| *** savvas has joined #openstack-ansible | 17:43 | |
| *** flemingo_ has quit IRC | 17:44 | |
| xgerman_ | cores wonder if I can get some eyes on https://review.openstack.org/#/c/553866/ | 17:46 |
| *** fleming__ has quit IRC | 17:46 | |
| *** flemingo_ has joined #openstack-ansible | 17:47 | |
| *** DanyC has joined #openstack-ansible | 17:47 | |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible stable/queens: Fixes typo to enable log rotate for Octavia https://review.openstack.org/559383 | 17:47 |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible stable/pike: Fixes typo to enable log rotate for Octavia https://review.openstack.org/559384 | 17:48 |
| *** savvas has quit IRC | 17:48 | |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia stable/queens: Fixes mgmt net and flavor fact setting https://review.openstack.org/559385 | 17:48 |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia stable/pike: Fixes mgmt net and flavor fact setting https://review.openstack.org/559386 | 17:48 |
| *** savvas has joined #openstack-ansible | 17:53 | |
| *** savvas has quit IRC | 17:59 | |
| *** Leo_m has quit IRC | 17:59 | |
| *** Leo_m has joined #openstack-ansible | 18:00 | |
| *** armaan has joined #openstack-ansible | 18:02 | |
| *** shardy has quit IRC | 18:03 | |
| *** brad[] has quit IRC | 18:03 | |
| *** brad[] has joined #openstack-ansible | 18:04 | |
| *** nicolasbock has joined #openstack-ansible | 18:04 | |
| *** savvas has joined #openstack-ansible | 18:04 | |
| *** hw_wutianwei_ has quit IRC | 18:07 | |
| *** savvas has quit IRC | 18:09 | |
| *** armaan has quit IRC | 18:11 | |
| *** adriano__ has quit IRC | 18:12 | |
| openstackgerrit | Amy Marrich (spotz) proposed openstack/openstack-ansible master: [DOC] Update for Newton EOL https://review.openstack.org/551316 | 18:13 |
| *** tomtom001 has joined #openstack-ansible | 18:13 | |
| *** savvas has joined #openstack-ansible | 18:14 | |
| tomtom001 | hello | 18:15 |
| tomtom001 | +Zi | 18:18 |
| *** savvas has quit IRC | 18:18 | |
| *** shardy has joined #openstack-ansible | 18:21 | |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible-os_nova master: (WIP) Update to use oslo.messaging services for RPC and Notify https://review.openstack.org/558620 | 18:21 |
| *** adcpdk has quit IRC | 18:22 | |
| *** tomtom001 has quit IRC | 18:23 | |
| *** savvas has joined #openstack-ansible | 18:24 | |
| *** tomtom001 has joined #openstack-ansible | 18:24 | |
| *** savvas has quit IRC | 18:30 | |
| *** savvas has joined #openstack-ansible | 18:32 | |
| *** savvas has quit IRC | 18:37 | |
| *** savvas has joined #openstack-ansible | 18:41 | |
| *** savvas has quit IRC | 18:46 | |
| mnaser | guys | 18:47 |
| mnaser | big oopsie | 18:47 |
| mnaser | https://github.com/openstack/openstack-ansible/blob/17.0.1/inventory/group_vars/all/all.yml#L17 | 18:47 |
| mnaser | notice the tag. notice the openstack_release | 18:47 |
| mnaser | cc evrardjp ^ | 18:47 |
| goldenfri | yea I just ran into that | 18:47 |
| mnaser | goldenfri: have a bug filed? | 18:49 |
| goldenfri | no, like literally just now | 18:49 |
| mnaser | aha | 18:49 |
| mnaser | i was really confused, i was pretty sure i checked out 17.0.1 | 18:49 |
| goldenfri | me too | 18:49 |
| goldenfri | haha | 18:49 |
| mnaser | okay so i guess this means 17.0.2 should be released to fix this | 18:50 |
| mnaser | and somehow someway 17.0.1 should be marked as bad or something | 18:50 |
| *** jona4302 has joined #openstack-ansible | 18:54 | |
| *** savvas has joined #openstack-ansible | 18:54 | |
| *** jsisson has joined #openstack-ansible | 18:55 | |
| *** jsisson has quit IRC | 18:55 | |
| *** logan- has quit IRC | 18:56 | |
| *** logan- has joined #openstack-ansible | 18:56 | |
| *** savvas has quit IRC | 18:59 | |
| *** Leo_m_ has joined #openstack-ansible | 18:59 | |
| *** Leo_m_ has quit IRC | 19:00 | |
| *** Leo_m_ has joined #openstack-ansible | 19:00 | |
| *** Leo_m has quit IRC | 19:00 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Run the on metal job https://review.openstack.org/556521 | 19:02 |
| *** chhavi__ has quit IRC | 19:02 | |
| *** Leo_m_ has quit IRC | 19:02 | |
| cloudnull | mnaser: opps, thats a bummer. | 19:03 |
| mnaser | happens! | 19:04 |
| cloudnull | it does | 19:04 |
| mnaser | https://bugs.launchpad.net/openstack-ansible/+bug/1761843 | 19:05 |
| openstack | Launchpad bug 1761843 in openstack-ansible "17.0.1 release has openstack_release pointing to 17.0.0" [Undecided,New] | 19:05 |
| cloudnull | evrardjp: ^ | 19:05 |
| mnaser | i dont trust myself with the release process but i assume releasing 17.0.2 with things fixed would be nice | 19:05 |
| mnaser | (and i'd say this is fairly critical, the earlier this is done, the less likely people will have 17.0.1 on their hands) | 19:05 |
| mnaser | evrardjp: also, when you're back, i'd like to make the metal jobs voting and i will personally take responsibility in making sure to fix them if something breaks (outside a change obviously) | 19:10 |
| mnaser | and they can go back to non voting if i'm slowing things down but i dont think it'll break that often | 19:10 |
| evrardjp | mnaser: that's too late we can't change a release | 19:11 |
| mnaser | evrardjp: yeah i know, hence releasing 17.0.2 asap :) | 19:12 |
| evrardjp | I forgot to bump the code inside on first release | 19:12 |
| evrardjp | mnaser: the next code has 17.0.2 included as openstack_release | 19:12 |
| mnaser | usually people will go to checking out the latest release so if we make the period of time shorter that 17.0.1 is latest, less people might have it broken | 19:12 |
| evrardjp | and will be marked as 17.0.2 | 19:12 |
| evrardjp | sorry for that | 19:12 |
| *** savvas has joined #openstack-ansible | 19:12 | |
| evrardjp | yes I can issue another one | 19:13 |
| mnaser | yep, that was the suggestion :) | 19:13 |
| evrardjp | I planned to have that done next week friday | 19:13 |
| evrardjp | but I can make it more urgently if you prefer | 19:13 |
| evrardjp | the tag is out so I will mark this as won't fix | 19:13 |
| evrardjp | I am sorry for the confusion | 19:13 |
| mnaser | i mean i think operationally it'll be easier. people won't be checking out a broken release | 19:13 |
| evrardjp | the release isn't broken | 19:13 |
| mnaser | hey it's fine, the release process for something like this is super complex, easy to miss things | 19:13 |
| evrardjp | there is just a mismatch between 17.0.1 tag and the version shown | 19:14 |
| mnaser | well if you're using 17.0.0 and try using 17.0.1 things like os-releases will all be located in 17.0.0 ? | 19:14 |
| mnaser | i dont know the specifics of what it could/would affect | 19:14 |
| evrardjp | yes you're right | 19:14 |
| *** eumel8 has quit IRC | 19:14 | |
| mnaser | that's how i noticed it in the first place | 19:14 |
| evrardjp | the problem was just the first bump of the stable branch | 19:14 |
| evrardjp | for that to never happen again (funny story, it happened last cycle too) | 19:15 |
| mnaser | ahaha | 19:15 |
| evrardjp | I finally wrote it in the docs. | 19:15 |
| mnaser | third time is the charm | 19:15 |
| mnaser | next release will be good | 19:15 |
| evrardjp | well next release is the standard release process | 19:15 |
| evrardjp | so it should work :p | 19:15 |
| evrardjp | for the onmetal story | 19:15 |
| evrardjp | yes I have a patch up for that | 19:16 |
| evrardjp | well it just merged | 19:16 |
| evrardjp | oh it's still nv | 19:16 |
| evrardjp | let me fix that quick | 19:16 |
| *** savvas has quit IRC | 19:17 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Make the on metal job voting https://review.openstack.org/559404 | 19:18 |
| evrardjp | mnaser: ^ | 19:18 |
| evrardjp | sorry for the issue with the tag number mnaser | 19:18 |
| evrardjp | is that urgent for you to have that mismatch? | 19:18 |
| evrardjp | or can this wait next friday? | 19:19 |
| mnaser | evrardjp: not that urgent | 19:19 |
| evrardjp | ok | 19:19 |
| mnaser | https://www.irccloud.com/pastebin/y2qvd7v5/ | 19:19 |
| mnaser | :) | 19:19 |
| evrardjp | haha | 19:19 |
| evrardjp | ok | 19:19 |
| mnaser | now i'll complain "WHY IS 17.0.2 TAGGED AS 17.0.1???" :P | 19:20 |
| evrardjp | I return to my other activities. | 19:20 |
| *** shardy has quit IRC | 19:20 | |
| mnaser | so you can yell PEBKAC in advance | 19:20 |
| mnaser | take care | 19:20 |
| evrardjp | haha if that happen, we can laugh about it together | 19:20 |
| evrardjp | shared pebkacs | 19:20 |
| evrardjp | :) | 19:20 |
| evrardjp | take care, don't hesitate to ping me if you have another issue -- I don't promise anything, but I am always willing to help! | 19:21 |
| mnaser | sounds good :) | 19:21 |
| *** savvas has joined #openstack-ansible | 19:22 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-galera_client master: Use UNIX socket instead of TCP socket https://review.openstack.org/559408 | 19:28 |
| *** savvas has quit IRC | 19:29 | |
| *** DanyC_ has joined #openstack-ansible | 19:33 | |
| *** DanyC has quit IRC | 19:36 | |
| *** fghaas has quit IRC | 19:44 | |
| *** savvas has joined #openstack-ansible | 19:46 | |
| *** savvas has quit IRC | 19:52 | |
| *** DanyC has joined #openstack-ansible | 19:54 | |
| *** jbadiapa_ has joined #openstack-ansible | 19:55 | |
| *** DanyC__ has joined #openstack-ansible | 19:56 | |
| *** DanyC_ has quit IRC | 19:56 | |
| *** jbadiapa has quit IRC | 19:57 | |
| *** DanyC has quit IRC | 19:59 | |
| *** fghaas has joined #openstack-ansible | 20:00 | |
| *** hachi has quit IRC | 20:08 | |
| *** esberglu has quit IRC | 20:11 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia master: Fixes mgmt net and flavor fact setting https://review.openstack.org/553866 | 20:16 |
| openstackgerrit | Merged openstack/openstack-ansible stable/pike: Update all SHAs for 16.0.11 https://review.openstack.org/559039 | 20:21 |
| *** fghaas has quit IRC | 20:27 | |
| *** esberglu has joined #openstack-ansible | 20:28 | |
| *** esberglu has quit IRC | 20:34 | |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia stable/queens: Fixes mgmt net and flavor fact setting https://review.openstack.org/559385 | 20:38 |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia stable/pike: Fixes mgmt net and flavor fact setting https://review.openstack.org/559386 | 20:39 |
| *** kstev1 has quit IRC | 20:40 | |
| vakuznet | https://github.com/openstack/openstack-ansible-galera_server/blob/stable/queens/tasks/galera_post_install.yml#L39 - I can't find "Restart all mysql" handler. | 20:54 |
| vakuznet | There is "Restart mysql (All)" handler - https://github.com/openstack/openstack-ansible-galera_server/blob/stable/queens/handlers/main.yml#L71 | 20:54 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_octavia stable/queens: Fixes mgmt net and flavor fact setting https://review.openstack.org/559385 | 20:55 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_octavia stable/pike: Fixes mgmt net and flavor fact setting https://review.openstack.org/559386 | 20:56 |
| vakuznet | cluster is dying because of that. | 20:56 |
| evrardjp | odyssey4me: gift for you: http://zuul.openstack.org/builds.html now has branches ;) | 20:58 |
| evrardjp | Well it was more a gift done to myself, but still :p | 20:58 |
| *** throwsb1 has quit IRC | 20:59 | |
| vakuznet | cloudnull ^ | 21:07 |
| *** haad1 has joined #openstack-ansible | 21:18 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/pike: Do not generate the html ARA report any more https://review.openstack.org/559079 | 21:24 |
| *** jwitko_ has joined #openstack-ansible | 21:33 | |
| *** jwitko has quit IRC | 21:36 | |
| *** fghaas has joined #openstack-ansible | 21:39 | |
| *** fghaas has quit IRC | 21:39 | |
| *** fghaas has joined #openstack-ansible | 21:39 | |
| *** fghaas has quit IRC | 21:42 | |
| *** haad1 has quit IRC | 21:43 | |
| *** haad1 has joined #openstack-ansible | 21:51 | |
| *** KeithMnemonic has quit IRC | 21:51 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Fixes typo to enable log rotate for Octavia https://review.openstack.org/555631 | 22:05 |
| *** vadim__ has joined #openstack-ansible | 22:19 | |
| *** sar has quit IRC | 22:20 | |
| *** vakuznet has quit IRC | 22:23 | |
| *** vadim__ has quit IRC | 22:27 | |
| *** vakuznet has joined #openstack-ansible | 22:28 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/queens: Fixes typo to enable log rotate for Octavia https://review.openstack.org/559383 | 22:50 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Fixes typo to enable log rotate for Octavia https://review.openstack.org/559384 | 22:50 |
| *** tomtom001 has quit IRC | 22:57 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-tests master: Resolve common roles required for test https://review.openstack.org/558227 | 23:04 |
| *** lbragstad has quit IRC | 23:34 | |
| *** chyka has quit IRC | 23:39 | |
| *** haad1 has quit IRC | 23:43 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!