| *** DanyC has quit IRC | 00:01 | |
| *** DanyC_ has quit IRC | 00:27 | |
| *** nurdie has joined #openstack-ansible | 00:27 | |
| *** nurdie has quit IRC | 00:31 | |
| *** irclogbot_0 has quit IRC | 00:38 | |
| *** nurdie has joined #openstack-ansible | 00:48 | |
| *** cshen has joined #openstack-ansible | 00:50 | |
| *** nurdie has quit IRC | 00:52 | |
| *** cshen has quit IRC | 00:55 | |
| *** jonher has quit IRC | 01:37 | |
| *** jonher has joined #openstack-ansible | 01:40 | |
| *** jonher has quit IRC | 01:46 | |
| *** nurdie has joined #openstack-ansible | 01:49 | |
| *** jonher has joined #openstack-ansible | 01:51 | |
| *** nurdie has quit IRC | 01:54 | |
| *** markvoelker has quit IRC | 02:00 | |
| *** markvoelker has joined #openstack-ansible | 02:00 | |
| *** markvoelker has quit IRC | 02:02 | |
| *** cmart has quit IRC | 02:06 | |
| *** dave-mccowan has joined #openstack-ansible | 02:16 | |
| *** DanyC has joined #openstack-ansible | 02:27 | |
| *** dave-mccowan has quit IRC | 02:27 | |
| *** nurdie has joined #openstack-ansible | 02:30 | |
| *** DanyC has quit IRC | 02:32 | |
| *** nurdie has quit IRC | 02:34 | |
| *** mmercer has quit IRC | 02:44 | |
| ravirjn | jrosser: I have deployed cinder successfully. it was haproxy. | 02:50 |
|---|---|---|
| *** nurdie has joined #openstack-ansible | 02:51 | |
| *** nurdie has quit IRC | 02:55 | |
| *** cmart has joined #openstack-ansible | 03:06 | |
| *** dave-mccowan has joined #openstack-ansible | 03:24 | |
| *** cmart has quit IRC | 03:24 | |
| *** dave-mccowan has quit IRC | 03:36 | |
| *** radeks has joined #openstack-ansible | 03:47 | |
| *** cmart has joined #openstack-ansible | 03:50 | |
| *** nurdie has joined #openstack-ansible | 03:52 | |
| *** nurdie has quit IRC | 03:57 | |
| *** nurdie has joined #openstack-ansible | 04:13 | |
| *** nurdie has quit IRC | 04:17 | |
| *** cmart has quit IRC | 04:23 | |
| *** ThiagoCMC has quit IRC | 04:33 | |
| *** cmart has joined #openstack-ansible | 04:43 | |
| *** cshen has joined #openstack-ansible | 04:45 | |
| *** cshen has quit IRC | 04:50 | |
| *** cshen has joined #openstack-ansible | 04:56 | |
| *** faizy98 has joined #openstack-ansible | 04:56 | |
| *** cshen has quit IRC | 05:00 | |
| *** cmart has quit IRC | 05:02 | |
| *** nurdie has joined #openstack-ansible | 05:14 | |
| *** hamzy has joined #openstack-ansible | 05:18 | |
| *** med_ has quit IRC | 05:29 | |
| openstackgerrit | Corey Wright proposed openstack/openstack-ansible master: Fix blazar-nova Git field names https://review.openstack.org/616073 | 05:47 |
| *** vnogin has joined #openstack-ansible | 05:52 | |
| *** vnogin has quit IRC | 05:57 | |
| *** cshen has joined #openstack-ansible | 06:04 | |
| *** cshen has quit IRC | 06:09 | |
| *** rgogunskiy has joined #openstack-ansible | 06:39 | |
| *** radeks has quit IRC | 06:49 | |
| *** nurdie has quit IRC | 06:55 | |
| *** nurdie has joined #openstack-ansible | 07:15 | |
| *** nurdie has quit IRC | 07:20 | |
| *** cshen has joined #openstack-ansible | 07:28 | |
| *** nurdie has joined #openstack-ansible | 07:36 | |
| *** pcaruana has joined #openstack-ansible | 07:36 | |
| *** hamzaachi has joined #openstack-ansible | 07:40 | |
| *** nurdie has quit IRC | 07:40 | |
| *** fghaas has joined #openstack-ansible | 07:48 | |
| *** nurdie has joined #openstack-ansible | 07:57 | |
| *** nurdie has quit IRC | 08:01 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible master: Imported Translations from Zanata https://review.openstack.org/615770 | 08:08 |
| *** blinkiz has joined #openstack-ansible | 08:17 | |
| blinkiz | Hello. Software requirements for targets hosts include a working Ubuntu 18.04 installation. I have a bunch of Supermicro X11 computers that will be the target hosts. Am looking for recommendation about automating the deployment of these target hosts into a running system that OpenStack-Ansible likes. Something like Canonical Metal as a Service (MAAS) for Ubuntu. | 08:21 |
| *** rgogunskiy has quit IRC | 08:25 | |
| jrosser | blinkiz: there will be as many different answers to that as people you ask..... :) | 08:32 |
| blinkiz | jrosser, it sure is. | 08:35 |
| jrosser | I spent time with cobbler and other “black box” provisioning tools, but eventually ditched anything like that in favour of just running plain vanilla dhcpd and tftpd. | 08:35 |
| blinkiz | Apparently what am looking for is called Linux Server Provisioning Software. With that keyword I have now found a couple of alternatives | 08:35 |
| jrosser | If you are doing anything with reasonable scale or wanting repeatability imho you should first consider what the “source of truth” is going to be for the data which describes everything you provision | 08:37 |
| *** nurdie has joined #openstack-ansible | 08:37 | |
| blinkiz | After I have searched a bit it seems that Foreman and Ubuntu MAAS is best if I want Ubuntu 18.04 as operating system. | 08:39 |
| jrosser | You are probably going to need some subsequent configuration after pxebootling, like an additional set of ansible that you must run | 08:40 |
| jrosser | In order to meet the prerequisites of OSA | 08:40 |
| blinkiz | jrosser, yeah. you are right. Actually currently looking into that. How Foreman/Ubuntu MAAS can launch Ansible playbooks within AWX. I hope one of these two products will be a clear winner so I do not need to install and test them both ;) | 08:42 |
| *** nurdie has quit IRC | 08:42 | |
| jrosser | You need to pick whatever suits your requirements, the products you mention will have a defined workflow and set of behaviour that may/may not be acceptable. Also consider what would happen if your mass/foreman server blew up, can you trivially reprovision it and put the data back? | 08:46 |
| *** radeks has joined #openstack-ansible | 08:52 | |
| openstackgerrit | KimMinsik proposed openstack/openstack-ansible-os_cinder master: Add resource_filter field in cinder.conf to retrive volume & snapshot information with filter for for non-admin user https://review.openstack.org/616113 | 08:58 |
| *** nurdie has joined #openstack-ansible | 08:58 | |
| *** nurdie has quit IRC | 09:03 | |
| *** kukacz has quit IRC | 09:04 | |
| *** kukacz has joined #openstack-ansible | 09:05 | |
| *** vnogin has joined #openstack-ansible | 09:15 | |
| openstackgerrit | wangqi proposed openstack/openstack-ansible-plugins master: 'os' imported but unused https://review.openstack.org/616119 | 09:18 |
| *** shardy has joined #openstack-ansible | 09:19 | |
| *** vnogin has quit IRC | 09:20 | |
| *** DanyC has joined #openstack-ansible | 09:20 | |
| *** django has joined #openstack-ansible | 09:26 | |
| *** electrofelix has joined #openstack-ansible | 09:29 | |
| odyssey4me | morning folks | 09:30 |
| fghaas | blinkiz: you could stand up an AIO OSA on your laptop, use that to deploy Ubuntu to your proper cloud hardware with ironic, then use OSA again to deploy OpenStack on that... | 09:33 |
| odyssey4me | jrosser hwoarang could you give https://review.openstack.org/615258 a review please - it'll unblock stable/rocky's integrated build | 09:33 |
| blinkiz | fghaas, aa, cool solution | 09:34 |
| odyssey4me | blinkiz you may want to take a look at the very simple dhcpd/tftpboot implementation here: https://github.com/openstack/openstack-ansible-ops/tree/master/pxelinux-provisioning | 09:34 |
| odyssey4me | smoke87 I think the vpnaas dashboard is implemented - which series & release are you using? | 09:36 |
| noonedeadpunk | morning everyone | 09:36 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: Adds option to enable ARA plugin install https://review.openstack.org/616004 | 09:38 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: Restart cinder-scheduler instead of reloading https://review.openstack.org/615645 | 09:38 |
| hwoarang | odyssey4me: checking | 09:39 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: zuul: Make openSUSE distro jobs voting https://review.openstack.org/614463 | 09:39 |
| hwoarang | ah one of these again | 09:39 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: Ensure ceph server apt pinning is well defined https://review.openstack.org/614777 | 09:39 |
| odyssey4me | hwoarang yep, this is the final one that pulls it all together again | 09:40 |
| arxcruz | odyssey4me: https://review.openstack.org/#/c/612968/ pass, can you check so i can set the tempestconf value to false? | 09:44 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_neutron master: Refactor pre-install checks into neutron_check.yml https://review.openstack.org/615262 | 09:58 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-openstack_hosts master: refactor: remove unused proxy settings file https://review.openstack.org/615337 | 09:59 |
| *** blinkiz has quit IRC | 09:59 | |
| *** nurdie has joined #openstack-ansible | 10:00 | |
| odyssey4me | arxcruz so tempestconf puts its result into the workspace/etc/tempest.conf ? | 10:03 |
| *** nurdie has quit IRC | 10:04 | |
| openstackgerrit | wangqiang-bj proposed openstack/openstack-ansible master: fix typo https://review.openstack.org/616137 | 10:07 |
| openstackgerrit | wangqi proposed openstack/openstack-ansible-os_sahara master: Trivial: Fix the pep8 warning https://review.openstack.org/589057 | 10:08 |
| arxcruz | odyssey4me: yes, i tried to put directly on /etc/tempest but didn’t work, i don’t know the reason | 10:08 |
| odyssey4me | arxcruz alright - to add tests with that enabled would be a few lines in tox.ini and zuul.d/ - but I guess you're wanting to work that out next, right? | 10:09 |
| openstackgerrit | Arx Cruz proposed openstack/openstack-ansible-os_tempest master: Enable python-tempestconf support https://review.openstack.org/612968 | 10:10 |
| arxcruz | odyssey4me: yup | 10:10 |
| arxcruz | odyssey4me: done, patch with the value set to false | 10:10 |
| noonedeadpunk | folks, am I right, that keystone needs this https://github.com/openstack/openstack-ansible-os_keystone/blob/master/templates/keystone.conf.j2#L25-L27 only when keystone_ceilometer_enabled: true? | 10:11 |
| noonedeadpunk | As I'm going to patch this | 10:11 |
| noonedeadpunk | oh, it's just not backported to rocky... | 10:12 |
| odyssey4me | noonedeadpunk port it back then :) | 10:12 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-lxc_hosts master: Drop COPR priority settings https://review.openstack.org/615349 | 10:16 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-lxc_hosts master: lxc: stop mirroring COPR locally on CentOS 7 https://review.openstack.org/615351 | 10:16 |
| openstackgerrit | wangqi proposed openstack/openstack-ansible-os_sahara master: Trivial: Fix the pep8 warning https://review.openstack.org/589057 | 10:17 |
| openstackgerrit | wangqi proposed openstack/openstack-ansible-os_sahara master: Trivial: Fix the pep8 warning https://review.openstack.org/589057 | 10:19 |
| *** nurdie has joined #openstack-ansible | 10:21 | |
| mnaser | does anyone know off the top of their minds why the endpoint would report as missing here? https://review.openstack.org/#/c/614556/ | 10:21 |
| mnaser | trying to give dcdamien a hand with his patch | 10:21 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_keystone stable/rocky: Update messaging configuration https://review.openstack.org/616142 | 10:22 |
| mnaser | i cant imagine why it wouldnt be there.. | 10:23 |
| mnaser | oh wait | 10:23 |
| mnaser | mistral trove murano and designate tests running there | 10:23 |
| mnaser | did we break something with os_tempest maybe? | 10:24 |
| mnaser | http://logs.openstack.org/56/614556/5/check/openstack-ansible-functional-distro_install-centos-7/2b12218/logs/openstack/keystone/stestr_results.html arxcruz any ideas? | 10:24 |
| jrosser | mnaser: the barbican role tests look like they have blown up co-incident with the tempest changes too | 10:24 |
| *** nurdie has quit IRC | 10:25 | |
| arxcruz | mnaser: well, my patch isn’t merged yet, the endpoint not found error is related to keystone, perhaps the server didn’t started properly ? | 10:28 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_hosts stable/rocky: Combined backport from master to ensure nspawn functionality in rocky https://review.openstack.org/611905 | 10:38 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_hosts stable/rocky: Combined backport from master to ensure nspawn functionality in rocky https://review.openstack.org/611905 | 10:38 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_hosts stable/rocky: Combined backport from master to ensure nspawn functionality in rocky https://review.openstack.org/611905 | 10:39 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_hosts stable/rocky: Install the timezone package by default https://review.openstack.org/616150 | 10:40 |
| noonedeadpunk | Does anybody knows, why do we need it here? https://github.com/openstack/openstack-ansible/blob/master/inventory/group_vars/nova_all.yml#L65-L68 Just curiosity | 10:40 |
| *** nurdie has joined #openstack-ansible | 10:41 | |
| odyssey4me | noonedeadpunk that was a way of trying to optimise the deployment - we could move that to using a sensible default in the role | 10:41 |
| odyssey4me | same with these: https://github.com/openstack/openstack-ansible/blob/master/inventory/group_vars/nova_all.yml#L44-L49 | 10:42 |
| odyssey4me | in fact, many of the group vars can be moved into role defaults instead - and it'd reduce memory consumption and speed deploys up a bit | 10:42 |
| noonedeadpunk | ok, thanks for explanation) | 10:43 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_hosts stable/rocky: Ensure libpython is included in base container builds https://review.openstack.org/616151 | 10:44 |
| noonedeadpunk | I've just found, that every group has "<service>_ceilometer_enabled", except ceilometer_all, so decided to dig a bit into it | 10:44 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_hosts stable/rocky: Ensure development libs are installed in containers https://review.openstack.org/616152 | 10:45 |
| DoberMann | instead of splitting load hi, i was wondering if there is a reason to use haproxy in active backup mode for the galera cluster by default ? ends up with all connections active on the same galera node | 10:45 |
| noonedeadpunk | So we've got situation, that keystone thinks, that ceilometer is enabled, but ceilometer itself doesn't think so and don't read messages from rabbitmq queue | 10:45 |
| *** nurdie has quit IRC | 10:45 | |
| DoberMann | hi, i was wondering if there is a reason to use haproxy in active backup mode for the galera cluster by default ? ends up with all connections active on the same galera node instead of splitting load | 10:45 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_hosts stable/rocky: Document adding "extra" networks https://review.openstack.org/616154 | 10:46 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_hosts stable/rocky: Add documentation for routed interfaces https://review.openstack.org/616155 | 10:47 |
| *** hamzaachi has quit IRC | 10:48 | |
| noonedeadpunk | odyssey4me: so should I stick to the rules and patch group_vars, or place condition in ceilometer role itself in main.yml? | 10:49 |
| odyssey4me | noonedeadpunk we've been trying to move things out of group vars and into role defaults where possible, so I'd suggest doing that with the appropriate precautions | 10:52 |
| odyssey4me | assuming it's sensible to do that | 10:52 |
| DoberMann | answering to myself : it seems related to https://openstack.nimeyo.com/35943/openstack-operators-openstack-and-mysql-galera-with-haproxy | 10:52 |
| *** priteau has joined #openstack-ansible | 10:53 | |
| odyssey4me | DoberMann we have found that there's little benefit from a performance standpoint, but more risk of transactions collisions - most OSA consumers are happy with it as-is and have found no real incentive to change it | 10:54 |
| *** coolj has quit IRC | 10:57 | |
| *** nurdie has joined #openstack-ansible | 11:02 | |
| *** gokhan has joined #openstack-ansible | 11:02 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_hosts stable/rocky: Add documentation for routed interfaces https://review.openstack.org/616155 | 11:03 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_hosts stable/rocky: Add missing variable for nspawn networking https://review.openstack.org/616159 | 11:03 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_hosts stable/rocky: Add host only network option https://review.openstack.org/616160 | 11:03 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_hosts stable/rocky: Modify the network to ensure their more restrictive https://review.openstack.org/616161 | 11:03 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_hosts stable/rocky: Correct host only routes https://review.openstack.org/616162 | 11:03 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova master: Clean up a-r-r role for upgrade test removal https://review.openstack.org/613330 | 11:04 |
| *** nurdie has quit IRC | 11:06 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_container_create stable/rocky: Add test for host_only networks and ensure network ordering https://review.openstack.org/616163 | 11:08 |
| jrosser | odyssey4me: thanks for the tip yesterday about how to run the lxc_hosts role test locally | 11:09 |
| jrosser | it did of course pass, somewhat in contrast to my across-the-board failure here https://review.openstack.org/#/c/613078/ | 11:10 |
| jrosser | something seems to clobber the container startup there which isnt obvious, have you seen anything like that? | 11:11 |
| DoberMann | odyssey4me: thx | 11:11 |
| DoberMann | we're hitting this issue because in iddle (dur mainly to monitoring) we have about 1450 connections and the limit is (by default) 1600 the too many connections is not that far ;) | 11:12 |
| DoberMann | *due | 11:13 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_container_create stable/rocky: Add missing user defined interface support https://review.openstack.org/616164 | 11:13 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_container_create stable/rocky: Add missing user defined interface support https://review.openstack.org/616164 | 11:14 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_container_create stable/rocky: Add a guard so we don't allow for duplicate config https://review.openstack.org/616165 | 11:14 |
| odyssey4me | jrosser hwoarang these are all the backports needed to get nspawn up to speed for rocky: https://review.openstack.org/#/q/topic:nspawn-rocky-fixes+(status:open+OR+status:merged) | 11:16 |
| odyssey4me | If you could verify that they're matching backports I'd appreciate it. | 11:17 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_ceilometer master: Added designate and trove services https://review.openstack.org/616167 | 11:19 |
| mnaser | arxcruz: those services aren’t even deployed though. It’s trying to test for things like murano | 11:21 |
| arxcruz | mnaser: so keystone isn’t deployed? Then, that’s the problem huh? | 11:22 |
| mnaser | No, that link was testing murano too | 11:23 |
| arxcruz | tempest.lib.exceptions.EndpointNotFound: Endpoint not found | 11:23 |
| arxcruz | Something went wrong with whatever designate_tempest_plugin endpoint uses | 11:30 |
| arxcruz | And it’s only happening in this job right ? | 11:33 |
| arxcruz | Other jobs i see it’s passing | 11:33 |
| arxcruz | odyssey4me: mnaser https://review.openstack.org/612968 ready for review/merge :) | 11:34 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_ceilometer master: Added designate and trove services https://review.openstack.org/616167 | 11:40 |
| odyssey4me | DoberMann it's probably best to up the number of connections then? | 11:41 |
| *** nurdie has joined #openstack-ansible | 11:42 | |
| DoberMann | odyssey4me: yes will do | 11:42 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_ceilometer master: Added designate and trove services https://review.openstack.org/616167 | 11:44 |
| *** nurdie has quit IRC | 11:47 | |
| noonedeadpunk | jrosser: I've moved "ssl=0" out of this patch. So should I split another (all) parts of it into separate patches? | 11:48 |
| jrosser | noonedeadpunk: in the spirit of one patch solving one just problem it would be better to split them. Particularly having the fix for something broken in it's own patch means it can easily be backported. | 11:53 |
| odyssey4me | folks, I need another vote to remove this tech debt: https://review.openstack.org/#/q/topic:remove-dragonflow+status:open | 11:53 |
| * jrosser looks | 11:53 | |
| jrosser | odyssey4me: done the one i can...... can i swap for some advice on lxc_hosts? | 11:55 |
| odyssey4me | jrosser of course | 11:56 |
| jrosser | i was digging into this https://review.openstack.org/#/c/613078/ full house of test failures there | 11:56 |
| jrosser | and i ran it locally via the lxc_container_create role test, which of course has passed | 11:57 |
| jrosser | is there anything obvious there i'm missing? | 11:57 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_ceilometer master: Added designate and trove services https://review.openstack.org/616167 | 11:58 |
| odyssey4me | jrosser looking | 12:00 |
| *** nurdie has joined #openstack-ansible | 12:03 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible stable/rocky: [DNM] Testing openSUSE jobs https://review.openstack.org/601518 | 12:04 |
| *** nurdie has quit IRC | 12:07 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_ceilometer master: <service>_ceilometer_enabled defaults based on services existence https://review.openstack.org/616176 | 12:09 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_neutron master: Update legacy neutron L3HA tool support https://review.openstack.org/616179 | 12:17 |
| mnaser | fire drill | 12:18 |
| mnaser | gah | 12:18 |
| mnaser | centos 7 rabbitmq gpg keys are borked? | 12:18 |
| mnaser | https://review.openstack.org/#/c/615258/ at least there it's complaining | 12:19 |
| mnaser | Public key for erlang-20.3.8.10-1.el7.centos.x86_64.rpm is not installed | 12:19 |
| mnaser | https://www.rabbitmq.com/rabbitmq-release-signing-key.asc is down | 12:19 |
| mnaser | oh | 12:20 |
| mnaser | looks like they changed their site | 12:20 |
| mnaser | https://github.com/rabbitmq/signing-keys/releases/download/2.0/rabbitmq-release-signing-key.asc is the new path | 12:20 |
| mnaser | boo | 12:20 |
| * mnaser will hardcode key into role to avoid further breakages | 12:21 | |
| odyssey4me | jrosser the discussion in https://review.openstack.org/615338 - I'm of the opinion that disabling caching like that when using a proxy is too brutal too, and think perhaps what's done there should be opt-in, or should be left to the deployer to implement on their own | 12:21 |
| noonedeadpunk | cores, your reviews on https://review.openstack.org/#/c/584629/ will be appreciated) | 12:21 |
| jrosser | odyssey4me: it feels like fixing it in the wrong place, a properly done proxy should not break upstream repos | 12:23 |
| odyssey4me | agreed, although I kinda see how it might have got there - because not all proxies are configured well | 12:23 |
| jrosser | if a deployer finds themselves behind a broken proxy then they can add an apt config fragment as part of host prep | 12:24 |
| odyssey4me | mnaser yes, please do | 12:25 |
| odyssey4me | we should actually do that for all the gpg keys | 12:25 |
| odyssey4me | jrosser agreed, and those would get copied into the container in the greenfield deploy | 12:25 |
| jrosser | odyssey4me: here is an example of doing it properly with squid http://www.rushiagr.com/blog/2015/06/05/cache-apt-packages-with-squid-proxy/ | 12:25 |
| odyssey4me | jrosser perhaps push up a patch to remove it, but add a release note informing people of the removal and link to where to do it right | 12:26 |
| odyssey4me | maybe also add that to the limited connectivity docs? | 12:26 |
| jrosser | sounds like a good plan | 12:28 |
| mnaser | jrosser: feel free to restore my patch | 12:28 |
| mnaser | we could leave this in there with opt in | 12:28 |
| mnaser | but i feel like just by default ripping it out might be meh | 12:28 |
| mnaser | at this rate we might as well bundle the damn debs | 12:30 |
| mnaser | :P | 12:30 |
| odyssey4me | mnaser I'd prefer to remove it - but we need to notify people about it and how to work with proxies the right way. | 12:31 |
| jrosser | mnaser: it sounds like we are all agreed that it should be removed | 12:31 |
| odyssey4me | We actually have a bunch of stuff to disable cert checks for a similar reason and could do much the same if we're able to articulate how to make things work the right way. | 12:32 |
| mnaser | yeah | 12:32 |
| jrosser | tbh i never understood that | 12:33 |
| jrosser | unless there are a bunch of wierd MITM proxies that people have | 12:34 |
| odyssey4me | I think it's transparent proxies that cause trouble. | 12:34 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_hosts master: Minimise distro packages installed into the lxc image https://review.openstack.org/613078 | 12:36 |
| odyssey4me | jrosser I don't see why that's failing - trying a rebase to see whether it makes a difference. | 12:36 |
| jrosser | ah good plan - my local test would have been very much with latest version of everything | 12:37 |
| odyssey4me | jrosser I wonder if the lxc-container-create role does some thing which rely on some of those packages | 12:38 |
| odyssey4me | in which case, perhaps it should install them | 12:38 |
| jrosser | it's odd, becasue the lxc-hosts role test which invokes the lxc_container_create tests fails there | 12:38 |
| odyssey4me | (the packages it requires) | 12:38 |
| jrosser | but locally the lxc_container_create one on it's own is ok, but like you say a rebase is a good idea | 12:38 |
| jrosser | it's certainly needing ping as the lxc_container_create role test uses that | 12:39 |
| jrosser | but that shouldnt stop the containers starting up | 12:39 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts master: Drop COPR priority settings https://review.openstack.org/615349 | 12:39 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts master: lxc: stop mirroring COPR locally on CentOS 7 https://review.openstack.org/615351 | 12:39 |
| fghaas | jrosser: do you happen to know why we're doing this? https://github.com/openstack/openstack-ansible/blob/master/inventory/group_vars/ceph-rgw.yml#L3 in other words: why doesn't this simply use "client.rgw.{{ ansible_hostname }}"? | 12:42 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-rabbitmq_server master: Use vendored RPM GPG keys for RabbitMQ https://review.openstack.org/616186 | 12:46 |
| mnaser | odyssey4me: ^ | 12:46 |
| mnaser | may be of interest to hwoarang too as it touches a bit of the suse stuff | 12:47 |
| noonedeadpunk | jrosser: splitted into 2 commits https://review.openstack.org/#/q/topic:fix/missing_services+(status:open+OR+status:merged) | 12:49 |
| mnaser | ok did we dig into tempest failures on centos? | 12:49 |
| mnaser | sorry i had to step out | 12:49 |
| hwoarang | mnaser: for suse so far we are using the distro packages from the cloud repos. maybe centos can do the same from rdo ? | 12:53 |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron master: Remove dragonflow ML2 driver https://review.openstack.org/615576 | 12:53 |
| jrosser | fghaas: i'm really not sure - i was just fiddling with some ansible ad-hoc commands to see what that all comes out as | 12:53 |
| fghaas | jrosser: All right I'm officially stumped because I fail to grok how user_variables and inventory and whatnot works together in this instance. If I override ceph_conf_overrides_rgw in user_variables.yml, run the ceph-rgw-install.yml playbook, then remove the override again from user_variables.yml, and rerun the playbook, I end up with this mess and my rgw's are borked (because they no longer run on the correct ports): http://p | 12:57 |
| fghaas | aste.openstack.org/show/734349/ ([global] section omitted for brevity) | 12:57 |
| fghaas | Don't even know where to start looking here. :( | 12:57 |
| odyssey4me | mnaser mind if I tweak to respect the existing data structure which allows people to override if they wish - or are you intentionally removing the ability to override? | 12:58 |
| odyssey4me | overriding is useful for offline builds, or using your own packages | 12:58 |
| mnaser | odyssey4me: all yours | 12:58 |
| odyssey4me | ^ with rgards to https://review.openstack.org/#/c/616186 | 12:59 |
| mnaser | Yeah it was just easier to put through and I guess I didn’t think of the use casss | 12:59 |
| jrosser | fghaas: unfortunatley my ceph is not deployed with OSA, but i did try this test http://paste.openstack.org/show/734351/ | 13:02 |
| fghaas | jrosser: right, so those two out to be equivalent then — or did I miss something? | 13:03 |
| jrosser | fghaas: ansible_hostname may not be available if facts have not been gathered properly, that could be a problem | 13:03 |
| jrosser | for me it looks OK, but perhaps try that in your environment, changing the hosts: to your rgqw | 13:03 |
| jrosser | *rgw | 13:03 |
| fghaas | well I've been using just ansible_hostname in my overrides and that worked just fine, hence why I asked | 13:04 |
| jrosser | this may be one for an ansible jedi....... odyssey4me ? | 13:04 |
| fghaas | but why I am getting this [client.rgw.] section for things I *didn't* override I have no idea... | 13:05 |
| *** nurdie has joined #openstack-ansible | 13:05 | |
| odyssey4me | mnaser bother - it looks like it didn't work in the first place anyway, ok will do a review as is | 13:06 |
| jrosser | fghaas: oh hmm - this could be related to default sections with config_template | 13:09 |
| *** nurdie has quit IRC | 13:09 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server master: Use vendored RPM GPG keys for RabbitMQ https://review.openstack.org/616186 | 13:10 |
| odyssey4me | mnaser so here's the kicker - neither redhat nor suse use the upstream packages any more for stein, they use the 'distro' packages (ie OBS (suse) and RDO (redhat)) | 13:11 |
| odyssey4me | the issue for stein is that the repo has that URL for gpg checks | 13:11 |
| odyssey4me | but yeah, if we can get that merged pronto I'll backport a proper change for the stable branches | 13:12 |
| mnaser | odyssey4me: oh thats silly, thats what i thought | 13:14 |
| mnaser | odyssey4me: we desperately need to figure out the upgrade path for that though | 13:14 |
| odyssey4me | mnaser well, it'll just take a test to verify what happens | 13:15 |
| jrosser | fghaas: so it's going wrong somewhere here https://github.com/ceph/ceph-ansible/blob/stable-3.2/roles/ceph-config/templates/ceph.conf.j2#L172-L174 | 13:19 |
| jrosser | which probably means some debug around this point would be useful https://github.com/ceph/ceph-ansible/blob/stable-3.2/roles/ceph-defaults/tasks/facts.yml#L228-L241 | 13:21 |
| fghaas | jrosser: I have absolutely no idea. | 13:23 |
| fghaas | I also don't know how I got to this point. | 13:23 |
| fghaas | I'll probably nuke my rgw containers and try deploying them anew | 13:23 |
| *** nurdie has joined #openstack-ansible | 13:26 | |
| vollman | odyssey4me When you have time can you give this another look? https://review.openstack.org/#/c/614210/ | 13:27 |
| *** nurdie has quit IRC | 13:30 | |
| arxcruz | mnaser: hey, when you have time, can you review https://review.openstack.org/#/c/612968/ ? | 13:30 |
| *** ansmith has joined #openstack-ansible | 13:30 | |
| odyssey4me | jrosser fghaas that is a bit odd - why reach into the cached facts when you can just use the facts at hand... perhaps logan- can shed light, but it seems to me better to use {{ ansible_hostname }} as you suggest | 13:31 |
| odyssey4me | vollman LGTM - need a second review on https://review.openstack.org/614210 from cloudnull or d34dh0r53 | 13:34 |
| fghaas | odyssey4me: right. I've given up on this one and I'm spinning a whole new reference environment. Might be able to get back to it tomorrow. | 13:34 |
| vollman | odyssey4me tyvm | 13:34 |
| *** fghaas has quit IRC | 13:34 | |
| guilhermesp | thanks hwoarang https://review.openstack.org/#/c/615608/ :) | 13:38 |
| smoke87 | odyssey4me: thanks! openstack-ansible queens 17.1.2 | 13:42 |
| mnaser | arxcruz: checking | 13:43 |
| *** nurdie has joined #openstack-ansible | 13:43 | |
| mnaser | arxcruz, odyssey4me: i don't want to be a pain and we've done a lot of reiterating on the tempestconf support but is there any reason we're not prefixing variables with the role name | 13:44 |
| mnaser | aka no tempest_ ? | 13:44 |
| *** pcaruana has quit IRC | 13:45 | |
| mnaser | https://review.openstack.org/#/c/616186/ could get some love | 13:48 |
| *** nurdie has quit IRC | 13:49 | |
| mnaser | hwoarang: is opensuse 15 the release that comes after 42 or? | 13:50 |
| mnaser | aka once we add opensuse 15 support can we drop the jobs for 42? | 13:51 |
| odyssey4me | smoke87 it was remove, because the dashboard as it was then was deprecated: https://github.com/openstack/openstack-ansible-os_horizon/commit/6d4412889123516d4cc2ec9f55a961659257eb13#diff-7eeda618087b49ae876084ab6c73fdbb | 13:51 |
| odyssey4me | smoke87 if there's a new dashboard, then it'll need to be implemented in the horizon role again | 13:52 |
| *** strattao has joined #openstack-ansible | 13:52 | |
| smoke87 | odyssey4me: ah ok.. I tried to install the neutron-vpnaas-dashboard but I messed up the horizon container .. . | 13:54 |
| odyssey4me | smoke87 just destroy and recreate the container | 13:55 |
| smoke87 | Yes.. I did | 13:55 |
| smoke87 | anyway.. thanks for your help :) | 13:55 |
| odyssey4me | hwoarang are you happy with https://review.openstack.org/616186 - it doesn't really affect the default path for suse, but it does touch the code | 13:56 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-openstack_hosts master: Do not disable caching for apt when http proxies are configured https://review.openstack.org/615338 | 13:57 |
| jrosser | odyssey4me: ^ theres some different words in the commit message and a release note | 13:57 |
| odyssey4me | mnaser is the lerlang vendored key missing in here on purpose? https://review.openstack.org/#/c/616186/2/vars/redhat.yml | 13:58 |
| *** pcaruana has joined #openstack-ansible | 14:00 | |
| arxcruz | mnaser: well, i wasn’t aware of that rule, if that’s a thing, I’m glad to change the patch | 14:05 |
| arxcruz | In fact, i prefer | 14:06 |
| *** radeks_ has joined #openstack-ansible | 14:06 | |
| *** nurdie has joined #openstack-ansible | 14:06 | |
| *** radeks has quit IRC | 14:07 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ops master: MNAIO: Add ability to deploy Ceph https://review.openstack.org/607338 | 14:08 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ops master: MNAIO: Add ability to deploy Ceph https://review.openstack.org/607338 | 14:09 |
| arxcruz | mnaser: hmmm checking now the name of the variable is tempest_use_tempestconf ... | 14:09 |
| *** nurdie has quit IRC | 14:10 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ops master: MNAIO: Add ability to deploy Ceph https://review.openstack.org/607338 | 14:12 |
| hwoarang | mnaser: yes 15 is after 42... | 14:16 |
| hwoarang | odyssey4me: https://review.openstack.org/616186 don't know this code path is never actually executed. it's just a placeholder | 14:17 |
| odyssey4me | hwoarang yep, I know | 14:17 |
| odyssey4me | hwoarang the same is true for redhat on stein - I wonder whether the code path should just be removed rather | 14:19 |
| hwoarang | i am not familiar with this role to judge whether it makes sense to use the upstream packages rather than the distro cloud ones :/ i would say that using the distro ones is probably safer | 14:20 |
| *** nurdie has joined #openstack-ansible | 14:26 | |
| *** pcaruana has quit IRC | 14:33 | |
| *** pcaruana has joined #openstack-ansible | 14:34 | |
| *** nurdie has quit IRC | 14:35 | |
| mnaser | arxcruz: as much as it sounds a bit silly | 14:35 |
| mnaser | i'd prefer tempest_use_tempestconf | 14:35 |
| arxcruz | mnaser: but it is tempest_use_tempestconf :) | 14:36 |
| *** strattao has quit IRC | 14:36 | |
| *** DanyC has quit IRC | 14:36 | |
| arxcruz | mnaser: oh, got it, the other variables, sorry, my bad, fixing | 14:36 |
| *** DanyC has joined #openstack-ansible | 14:37 | |
| openstackgerrit | Arx Cruz proposed openstack/openstack-ansible-os_tempest master: Enable python-tempestconf support https://review.openstack.org/612968 | 14:40 |
| openstackgerrit | Merged openstack/openstack-ansible-openstack_hosts master: refactor: remove unused proxy settings file https://review.openstack.org/615337 | 14:41 |
| *** DanyC has quit IRC | 14:42 | |
| openstackgerrit | Arx Cruz proposed openstack/openstack-ansible-os_tempest master: Enable python-tempestconf support https://review.openstack.org/612968 | 14:42 |
| mnaser | ok | 14:43 |
| mnaser | i found out why it's broken | 14:43 |
| mnaser | arxcruz: the distro install support in centos7 installs all of the tempest plugins | 14:44 |
| mnaser | arxcruz: and we run smoke tests here -- https://github.com/openstack/openstack-ansible-os_cinder/blob/master/tests/os_cinder-overrides.yml | 14:44 |
| mnaser | so thats why that role is broken | 14:44 |
| odyssey4me | mnaser but tempest is supposed to be smart and skip tests for services which aren't installed | 14:45 |
| *** strattao has joined #openstack-ansible | 14:45 | |
| odyssey4me | I guess some plugins might be dumb. | 14:45 |
| arxcruz | no, not plugin | 14:45 |
| mnaser | http://logs.openstack.org/56/614556/5/check/openstack-ansible-functional-distro_install-centos-7/d11af04/logs/ara-report/result/94c415ef-cf80-4998-93e6-3db4b56bb8e6/ | 14:45 |
| mnaser | yeah | 14:45 |
| arxcruz | problem is | 14:45 |
| arxcruz | plugins have their own config.py | 14:45 |
| odyssey4me | In which case we're going to have to re-do how the plugins are installed for pretty much all install methods. | 14:45 |
| arxcruz | and the value to run particular test/service is set to true by default | 14:45 |
| mnaser | http://logs.openstack.org/56/614556/5/check/openstack-ansible-functional-distro_install-centos-7/d11af04/logs/openstack/keystone/stestr_results.html | 14:46 |
| mnaser | ah i see | 14:46 |
| mnaser | so maybe we need some toggle on/off? | 14:46 |
| arxcruz | tempestconf is able to check for the service and set it to false or true btw | 14:46 |
| arxcruz | ;) | 14:46 |
| mnaser | i knew it | 14:46 |
| mnaser | arxcruz: purposely broke our gate to make us use tempestconf | 14:46 |
| arxcruz | IMHO, we should not install all plugins | 14:46 |
| arxcruz | LOL, that patch wasn’t mine :P | 14:46 |
| odyssey4me | sure, but then we need to do a bunch of conditional logic and maintain the enable/disable bits - which isn't fun | 14:47 |
| arxcruz | yes, i agree | 14:47 |
| arxcruz | we should not install all plugins, we had similar problems on tripleo | 14:47 |
| odyssey4me | FML | 14:47 |
| openstackgerrit | Panagiotis Karalis proposed openstack/openstack-ansible-os_glance master: Turn the filesystem directory to be configurable https://review.openstack.org/613924 | 14:47 |
| arxcruz | and we decided to install the plugin only if it’s required | 14:47 |
| arxcruz | and later, using tempestconf we were able to fix these services set to true/false | 14:48 |
| mnaser | http://logs.openstack.org/56/614556/5/check/openstack-ansible-functional-distro_install-centos-7/d11af04/logs/ara-report/result/f2397549-e670-4acb-a689-e963e7d0fea0/ | 14:50 |
| mnaser | this is installing the world | 14:50 |
| mnaser | looks like its trying to install openstack-tempest-all | 14:51 |
| dcdamien | hm, I know this review ^^ | 14:51 |
| mnaser | dcdamien: hehe yeah trying to see how your stuff is broken | 14:51 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/rocky: Use vendored RPM GPG keys for RabbitMQ https://review.openstack.org/616218 | 14:51 |
| mnaser | oh thats weird | 14:52 |
| mnaser | https://github.com/openstack/openstack-ansible-os_tempest/commit/7e6e614ff035c428fa34384169da7dfc00ccb103 | 14:52 |
| mnaser | nothing here | 14:52 |
| dcdamien | mnaser: cool, it looked for me like a unrelated issue | 14:52 |
| mnaser | oh | 14:53 |
| odyssey4me | mnaser https://github.com/openstack/openstack-ansible-os_tempest/commit/2dbd375b44dbc2ae8788482500785be0e1579550 | 14:53 |
| mnaser | there we go | 14:53 |
| mnaser | https://github.com/openstack/openstack-ansible-os_tempest/commit/2dbd375b44dbc2ae8788482500785be0e1579550 | 14:53 |
| mnaser | arxcruz: what do you think we should do about this? | 14:53 |
| odyssey4me | mnaser arxcruz we have https://github.com/openstack/openstack-ansible-os_tempest/blob/master/vars/main.yml#L23 which does all the source plugins, we could extend that to also include the plugin package names for each distro package, then use that var to extract the enabled plugin list | 14:55 |
| odyssey4me | I was thinking about doing that yesterday, then got bogged down in other work | 14:55 |
| odyssey4me | another option is to just have things like "{{ (<service_name>_enabled | bool) | ternary('<service_plugin_package_name', '') }}" in the vars list for each distro, assuming that it allows empty list entries | 14:56 |
| logan- | odyssey4me fghaas: yeah, not sure. those 2 statements should generally be interchangeable, i can't think of any reason to use one vs the other there | 14:56 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova master: Add support to kvm nested virt https://review.openstack.org/615608 | 14:57 |
| *** vnogin has joined #openstack-ansible | 14:57 | |
| *** priteau has quit IRC | 15:01 | |
| *** vnogin has quit IRC | 15:02 | |
| *** priteau has joined #openstack-ansible | 15:02 | |
| arxcruz | mnaser: odyssey4me to unblock the gate, we should remove the tempest-all package, and then, install the plugins per service with an override in that variable | 15:03 |
| odyssey4me | arxcruz adding that unblocked a different set of gates, so we're going to have to do this all at once | 15:04 |
| arxcruz | sorry, didn’t get, add what? | 15:05 |
| mnaser | arxcruz: i think odyssey4me meant that if we revert that, we break the gate in another way | 15:06 |
| * arxcruz doesn’t speak English very well, sometimes it’s hard to follow | 15:06 | |
| mnaser | so that patch itself was a gate unbreaker | 15:06 |
| odyssey4me | arxcruz mnaser apologies - let me be clearer | 15:06 |
| mnaser | so if we revert it, we take it from broken to broken :D | 15:06 |
| mnaser | i think | 15:06 |
| odyssey4me | before that patch, the gates in neutron were broken because they use the neutron tempest plugins for testing | 15:07 |
| odyssey4me | so if we revert that patch, we break the os_neutron role tests | 15:07 |
| arxcruz | lol | 15:07 |
| odyssey4me | so - in order to get things working right, we should implement a fix which selectively installs the packages as you suggested | 15:07 |
| arxcruz | okay, I understand now | 15:07 |
| odyssey4me | it seems that for source builds we do https://github.com/openstack/openstack-ansible-os_tempest/blob/master/defaults/main.yml#L84 | 15:07 |
| mnaser | odyssey4me: im thinking something similar to https://github.com/openstack/openstack-ansible-os_tempest/commit/bf077bc1936432d3677cc67eb28f7b268505dd79 .. where we have two vars files | 15:08 |
| odyssey4me | well, I guess I've been meaning to figure this all out for some time - so maybe I should just get it done now :p | 15:08 |
| *** cmart has joined #openstack-ansible | 15:08 | |
| mnaser | or i was thinking having var/source.yml and vars/distro.yml | 15:09 |
| odyssey4me | mnaser I started working out how to make that do pip installs from pypi yesterday, so that we don't always have to install from git... so yeah, I'm thinking much the same thing | 15:09 |
| odyssey4me | lemme see if I can work out a way to stop the bleeding - that'll buy some time to figure out a decent holistic solution | 15:09 |
| mnaser | odyssey4me: thanks and sorry for the extra work you just got signed up for | 15:10 |
| mnaser | ;< | 15:10 |
| odyssey4me | ok, I have an idea which I think will work | 15:12 |
| arxcruz | are the rpm package equals to the name in vars/main.yml ? | 15:14 |
| *** gkadam has joined #openstack-ansible | 15:20 | |
| arxcruz | we can parse zuul job name to figure out which package to install | 15:23 |
| arxcruz | for example, for os_keystone, we can parse this keystone, and get the package required from the vars/main.yml | 15:23 |
| noonedeadpunk | Have a bit strange question. It seems, that currently service and rpc user password are updated every time when playbok is launched. So why don't we use "update_password: always" for mysql_user module, so that changed password inside user_secrets.yml lead to password rotation? | 15:25 |
| noonedeadpunk | oh, it's default. sorry | 15:26 |
| noonedeadpunk | So than it's rather safe now to change password inside user_secrets.yml, isn't it? | 15:27 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest master: Selectively install tempest plugin distro packages https://review.openstack.org/615869 | 15:40 |
| odyssey4me | mnaser arxcruz ^ | 15:40 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-openstack_hosts master: Do not disable caching for apt when http proxies are configured https://review.openstack.org/615338 | 15:40 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-openstack_hosts master: Do not disable caching for apt when http proxies are configured https://review.openstack.org/615338 | 15:41 |
| *** nurdie has joined #openstack-ansible | 15:41 | |
| odyssey4me | jrosser I just removed some trailing whitespace | 15:41 |
| jrosser | doh! thanks | 15:42 |
| mnaser | I wonder why it didn’t get caught in ci | 15:42 |
| mnaser | Oh we don’t run smoke tests in cI | 15:42 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest master: Selectively install tempest plugin distro packages https://review.openstack.org/615869 | 15:43 |
| *** DanyC has joined #openstack-ansible | 15:43 | |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible master: Add qdrouterd role for oslo.messaging backend deployment https://review.openstack.org/570261 | 15:43 |
| *** ivve has joined #openstack-ansible | 15:44 | |
| *** DanyC has quit IRC | 15:47 | |
| arxcruz | odyssey4me: do you mind if later I change it a little bit? but just for the future :) | 15:48 |
| odyssey4me | arxcruz could you suggest a change in the review as a comment? | 15:49 |
| arxcruz | odyssey4me: nah, right now we need to stop the bleeding, i want to have like install from source, from pip, and from rpm splited on their own files/tasks | 15:50 |
| arxcruz | but it’s something for the future | 15:50 |
| *** cmart has quit IRC | 15:50 | |
| odyssey4me | noonedeadpunk - maybe, I think you'll find that changing the password will break any related services until the playbook for the service is run to update the template and restart the service | 15:51 |
| odyssey4me | arxcruz yep, I'll be doing a follow up to do something similar for source builds - the current mechanism is somewhat inadequate | 15:51 |
| arxcruz | it’s pretty complicated imho | 15:52 |
| arxcruz | but hey, let me work on that, i need to help more on this :) | 15:52 |
| * arxcruz wants to be core on os_tempest in the future :P | 15:52 | |
| odyssey4me | arxcruz the current source build thing - it is complicated, yep, but it will need to cater to git and pypi-based installs... but as part of the python build simplification I've been wanting to simplify it anyway | 15:53 |
| odyssey4me | but if you'd like to have a go at it - please do :) | 15:53 |
| arxcruz | sure | 15:54 |
| odyssey4me | arxcruz there are a few things I'm thinking would be good to do - one is to move these defaults into the role, and remove them from the integrated build's group_vars: https://github.com/openstack/openstack-ansible/blob/master/inventory/group_vars/utility_all.yml#L57-L66 | 15:55 |
| odyssey4me | that way we don't have to maintain or set them in two places | 15:55 |
| odyssey4me | then have this dict have another attribute - perhaps 'enabled', which enables/disables the presence of the plugin based on that same key var from defaults: https://github.com/openstack/openstack-ansible-os_tempest/blob/master/vars/main.yml#L23-L91 | 15:57 |
| *** DanyC has joined #openstack-ansible | 15:57 | |
| odyssey4me | then also have the source install determine whether it should install from git or pypi, and adapt accordingly | 15:57 |
| odyssey4me | does all of that make much sense? | 15:57 |
| odyssey4me | popping into a meeting now - be back in a bit | 15:58 |
| *** DanyC_ has joined #openstack-ansible | 15:58 | |
| *** gyee has joined #openstack-ansible | 15:58 | |
| *** DanyC has quit IRC | 16:01 | |
| openstackgerrit | Panagiotis Karalis proposed openstack/openstack-ansible-os_glance master: Turn the filesystem directory to be configurable https://review.openstack.org/613924 | 16:11 |
| openstackgerrit | wangqi proposed openstack/openstack-ansible-os_sahara master: fix Unexpected indentation https://review.openstack.org/616232 | 16:17 |
| openstackgerrit | wangqi proposed openstack/openstack-ansible-os_sahara master: fix Unexpected indentation https://review.openstack.org/616235 | 16:20 |
| *** cshen has quit IRC | 16:27 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Fix blazar-nova Git field names https://review.openstack.org/616073 | 16:32 |
| arxcruz | odyssey4me: yeah, and keep the role self contained | 16:34 |
| odyssey4me | arxcruz I'm happy to work on that last bit (install from git/pypi) if you'd like to tackle the first two bits? | 16:35 |
| arxcruz | sure | 16:35 |
| arxcruz | odyssey4me: can we talk tomorrow? pretty late here now :) | 16:36 |
| odyssey4me | arxcruz of course :) I'm winding down for the day too, and tomorrow morning's better for new information :) | 16:36 |
| arxcruz | yup, after a big glass of coffee | 16:36 |
| odyssey4me | ++ | 16:37 |
| *** pcaruana has quit IRC | 16:38 | |
| *** shardy has quit IRC | 16:46 | |
| openstackgerrit | Guilherme Steinmuller Pimentel proposed openstack/openstack-ansible-os_nova stable/rocky: Add support to kvm nested virt https://review.openstack.org/616246 | 16:52 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server master: Use vendored RPM GPG keys for RabbitMQ https://review.openstack.org/616186 | 16:57 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/rocky: Use vendored RPM GPG keys for RabbitMQ https://review.openstack.org/616218 | 16:58 |
| *** irclogbot_0 has joined #openstack-ansible | 17:04 | |
| *** gkadam has quit IRC | 17:04 | |
| jrosser | odyssey4me: what do you reckon? missing python lib? http://paste.openstack.org/show/734361/ | 17:16 |
| odyssey4me | jrosser yep, that seems likely | 17:18 |
| jrosser | that's why bionic/nova-lxd is broken | 17:18 |
| odyssey4me | perhaps a class path has changed or something? | 17:18 |
| jrosser | would i expect to see that in the nova requirements? | 17:19 |
| jrosser | i.e which lib it should need | 17:19 |
| odyssey4me | https://github.com/openstack/openstack-ansible-os_nova/blob/master/defaults/main.yml#L173-L177 | 17:19 |
| jrosser | https://github.com/openstack/nova-lxd/blob/master/requirements.txt#L13 | 17:19 |
| odyssey4me | https://github.com/openstack/openstack-ansible-os_nova/blob/master/defaults/main.yml#L594-L597 | 17:20 |
| jrosser | wierd, can't find much that looks like nova-lxd | 17:23 |
| dcdamien | mnaser, odyssey4me https://review.openstack.org/#/c/561266/ - this is not passing on ceph because there is no additional volume types defined and association fails. I'm for skipping qos association when expected volume type doesn't exist. What do you think about? | 17:24 |
| odyssey4me | dcdamien that seems sensible | 17:25 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest master: Selectively install tempest plugin distro packages https://review.openstack.org/615869 | 17:26 |
| odyssey4me | jrosser what do you mean, nova-lxd is at https://git.openstack.org/openstack/nova-lxd | 17:30 |
| odyssey4me | that's not browsable, but this is: https://github.com/openstack/nova-lxd | 17:30 |
| odyssey4me | jrosser you can also reach out to tinwood or jamespage - although I think they're uk and may not be around right now, the #openstack-charms folks may also be able to help | 17:32 |
| odyssey4me | jrosser although I wonder if we broke it somehow given that module's not there, but should be | 17:33 |
| *** cmart has joined #openstack-ansible | 17:34 | |
| jrosser | odyssey4me: sorry was trying to find the files - which are there in the venv | 17:39 |
| jrosser | so they are installed | 17:39 |
| odyssey4me | I know some of the projects now require a full classpath when doing the import - maybe nova has changed too | 17:40 |
| jrosser | i'm also wondering if that is the nova service started before the proper config is present, and just log noise | 17:40 |
| odyssey4me | fair point | 17:41 |
| tinwood | odyssey4me, jrosser is that a py2 py3 issue? | 17:41 |
| odyssey4me | well, the venv was prepped with py2, but that is perhaps an issue | 17:41 |
| jrosser | just a mo..... | 17:42 |
| tinwood | i.e. are you loading from ubuntu packages? | 17:42 |
| tinwood | Or from source (pip) | 17:42 |
| odyssey4me | tinwood nope, this is pip installed from source | 17:42 |
| odyssey4me | from git, specifically | 17:42 |
| jrosser | 2018-11-07 17:29:05.542 1461 DEBUG oslo_service.service [req-dc5d1e5c-e38e-4a1e-89ce-17fa6b4da737 - - - - -] lxd.pool = None log_opt_values /openstack/venvs/nova-testing/lib/python2.7/site-packages/oslo_config/cfg.py:2575 | 17:42 |
| jrosser | ^ is that a problem? | 17:43 |
| tinwood | jrosser, is that on bionic? | 17:43 |
| odyssey4me | jrosser look in the venv itself, is there a python3 lib folder? | 17:43 |
| odyssey4me | tinwood yes | 17:43 |
| dcdamien | odyssey4me: yesterday you've written that horizon's firewall and vpn options was removed because they're deprecated. As I see there are separate projects for them. Is there any reason to not integrate them in OSA? | 17:44 |
| jrosser | odyssey4me: this is all i have root@ubuntu-bionic:/openstack/venvs/nova-testing/lib/python2.7 | 17:44 |
| odyssey4me | dcdamien no, assuming they're up to date and useful - they're just waiting for someone to add a patch! | 17:44 |
| tinwood | jrosser, odyssey4me, so lxd 3 is the default on bionic and it *has* to have a storage pool set up (even if it's just a dir). -- however, that might not be the issue here | 17:45 |
| dcdamien | odyssey4me: okay, thanks | 17:45 |
| odyssey4me | tinwood good heads up - I dunno if that's done in our case | 17:45 |
| dcdamien | smoke87: stay tuned :D | 17:46 |
| tinwood | odyssey4me, jrosser the only time I've seen an error similar to the one you posted in the paste is when the nova-lxd module isn't getting installed - in my case it was a mismatch between a python3 nova and a python2 nova-lxd ... but your doing pip, so that seems unlikely??? | 17:47 |
| jrosser | tinwood: the lxd storage pool is created, i can see it with lxc storage list | 17:49 |
| tinwood | jrosser, okay, that's good. | 17:50 |
| jrosser | but this looks bad from the nova log "lxd.pool = None" | 17:50 |
| jrosser | i expect thats a config option which our default test doesnt account for | 17:50 |
| tinwood | jrosser, ah, that's a config option that needs to be set in the nova config. | 17:50 |
| * tinwood goes looking | 17:50 | |
| tinwood | jrosser, so in /etc/nova/nova.conf, in the [lxd] section, you need a pool = <name-of-pool> | 17:52 |
| jrosser | i'll try - /me failing with google to find the config reference | 17:53 |
| tinwood | jrosser, Interestingly, I raised this bug: https://bugs.launchpad.net/openstack-ansible/+bug/1800837 | 17:54 |
| openstack | Launchpad bug 1800837 in openstack-ansible "openstack-ansible-nova-lxd test fails due to lxd storage pool not set up on bionic host" [Low,Confirmed] | 17:54 |
| jrosser | ah well after that got discussed yesterday i fired off a speculative local build to see if it was anything obvious | 17:55 |
| tinwood | Sadly, I don't know enough ansible to fix it, although I'm reading up on it! | 17:55 |
| jrosser | and just getting round to poking at that | 17:55 |
| *** vnogin has joined #openstack-ansible | 18:00 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server master: Ensure that RedHat/SUSE only use a distro install https://review.openstack.org/616186 | 18:02 |
| odyssey4me | hwoarang mnaser rather than constantly muck about - perhaps https://review.openstack.org/616186 is the better way to handle it for stein.... there can be only one | 18:03 |
| odyssey4me | for the rocky implementation, the current issue is http://logs.openstack.org/18/616218/2/check/openstack-ansible-functional-centos-7/d25eb22/logs/ara-report/result/770acdc1-87aa-412a-ae6c-8d46cf54e6d3/ | 18:04 |
| *** vnogin has quit IRC | 18:04 | |
| jrosser | tinwood: i have to run to get my train now, but this is as far as i've got, i see the lxd section in the nova.conf template here https://github.com/openstack/openstack-ansible-os_nova/blob/master/templates/nova.conf.j2#L283-L286 | 18:05 |
| jrosser | but when i ran the test locally the [lxd] section is missing from nova.conf, so that needs figuring out why | 18:05 |
| tinwood | jrosser, kk, we can catch up tomorrow? Just ping me. | 18:06 |
| jrosser | sure, i'd like to get this sorted out | 18:07 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/rocky: Use vendored RPM GPG keys for RabbitMQ https://review.openstack.org/616218 | 18:18 |
| *** radeks_ has quit IRC | 18:29 | |
| *** mmercer has joined #openstack-ansible | 18:36 | |
| *** cshen has joined #openstack-ansible | 18:40 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server master: Ensure that RedHat/SUSE only use a distro install https://review.openstack.org/616186 | 18:58 |
| odyssey4me | cores - https://review.openstack.org/616218 should unblock rocky in terms of the rabbitmq installation failures for centos | 18:59 |
| odyssey4me | mnaser ^ | 18:59 |
| *** nurdie has quit IRC | 19:02 | |
| *** priteau has quit IRC | 19:10 | |
| guilhermesp | mnaser: same rabbit gpg key error in https://review.openstack.org/#/c/616246/ | 19:11 |
| guilhermesp | http://logs.openstack.org/46/616246/1/check/openstack-ansible-functional-centos-7/da2e599/job-output.txt.gz#_2018-11-07_17_14_02_852530 | 19:11 |
| *** priteau has joined #openstack-ansible | 19:12 | |
| guilhermesp | I saw that you hardcoded the role earlier to avoid further breakage | 19:12 |
| guilhermesp | odyssey4me mnaser it depends on this? https://review.openstack.org/#/c/616186/ | 19:19 |
| *** radeks_ has joined #openstack-ansible | 19:20 | |
| guilhermesp | ah ok just saw above https://review.openstack.org/#/c/616218/ :) | 19:20 |
| *** faizy_ has joined #openstack-ansible | 19:30 | |
| *** faizy98 has quit IRC | 19:33 | |
| *** nurdie has joined #openstack-ansible | 19:33 | |
| *** electrofelix has quit IRC | 19:47 | |
| *** DanyC_ has quit IRC | 19:47 | |
| *** radeks_ has quit IRC | 19:51 | |
| *** nurdie_ has joined #openstack-ansible | 19:52 | |
| *** radeks_ has joined #openstack-ansible | 19:54 | |
| *** nurdie has quit IRC | 19:54 | |
| *** DanyC has joined #openstack-ansible | 20:07 | |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible master: Add qdrouterd role for oslo.messaging backend deployment https://review.openstack.org/570261 | 20:07 |
| odyssey4me | guilhermesp for master, https://review.openstack.org/616186 - for rocky, https://review.openstack.org/616218 | 20:15 |
| odyssey4me | guilhermesp given it's a blocking issue - fast approving those reviews, once they merge, just recheck your patch(es) | 20:19 |
| guilhermesp | thanks odyssey4me ! The only thing I can do now is +1 :P | 20:19 |
| guilhermesp | let's wait then | 20:20 |
| *** radeks_ has quit IRC | 20:20 | |
| odyssey4me | guilhermesp you can +1 if you think it's a good patch, it also helps me know that I'm not doing work without someone else seeing it ;) | 20:20 |
| guilhermesp | odyssey4me: ++ | 20:21 |
| odyssey4me | cloudnull if you're around, https://review.openstack.org/614210 needs a second review with someone more in touch with that tooling | 20:34 |
| *** guilhermesp has quit IRC | 20:47 | |
| *** jungleboyj has quit IRC | 20:47 | |
| *** hogepodge has quit IRC | 20:47 | |
| *** fdegir has quit IRC | 20:47 | |
| *** neith has quit IRC | 20:47 | |
| *** mnaser has quit IRC | 20:48 | |
| *** mwhahaha has quit IRC | 20:48 | |
| *** guilhermesp has joined #openstack-ansible | 20:48 | |
| *** hogepodge has joined #openstack-ansible | 20:48 | |
| *** fdegir has joined #openstack-ansible | 20:48 | |
| *** mwhahaha has joined #openstack-ansible | 20:49 | |
| *** mnaser has joined #openstack-ansible | 20:49 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_nova master: Switch LXD job from xenial to bionic https://review.openstack.org/616301 | 20:50 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_nova master: Ensure LXD pool name is always specified in nova config https://review.openstack.org/616302 | 20:50 |
| openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server master: Ensure that RedHat/SUSE only use a distro install https://review.openstack.org/616186 | 20:51 |
| odyssey4me | guilhermesp there we go - master should be unblocked | 20:53 |
| *** jungleboyj has joined #openstack-ansible | 21:35 | |
| *** ansmith has quit IRC | 21:37 | |
| *** strattao has quit IRC | 21:37 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_nova master: Ensure LXD pool name is always specified in nova config https://review.openstack.org/616302 | 21:45 |
| *** ivve has quit IRC | 21:46 | |
| openstackgerrit | Merged openstack/openstack-ansible master: SUSE: Add support for openSUSE Leap 15 https://review.openstack.org/604366 | 21:47 |
| openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server stable/rocky: Use vendored RPM GPG keys for RabbitMQ https://review.openstack.org/616218 | 22:03 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: Ensure pip/setuptools/wheel are up to date in the utility venv https://review.openstack.org/615258 | 22:05 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: zuul: Make openSUSE distro jobs voting https://review.openstack.org/614463 | 22:06 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: Ensure ceph server apt pinning is well defined https://review.openstack.org/614777 | 22:06 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: Adds option to enable ARA plugin install https://review.openstack.org/616004 | 22:06 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: Restart cinder-scheduler instead of reloading https://review.openstack.org/615645 | 22:06 |
| odyssey4me | jrosser mnaser unfortunately had to update https://review.openstack.org/615258 again due to another circumstance that changed, but that should now unblock stable/rocky's integrated build again cc guilhermesp | 22:07 |
| *** vnogin has joined #openstack-ansible | 22:11 | |
| *** vnogin has quit IRC | 22:15 | |
| guilhermesp | thanks odyssey4me ! Rechecked | 22:16 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest master: Selectively install tempest plugin distro packages https://review.openstack.org/615869 | 22:24 |
| *** priteau has quit IRC | 22:47 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_nova master: Ensure LXD pool name is always specified in nova config https://review.openstack.org/616302 | 22:49 |
| *** DanyC_ has joined #openstack-ansible | 22:59 | |
| *** dcdamien has quit IRC | 23:00 | |
| *** dcdamien has joined #openstack-ansible | 23:00 | |
| *** DanyC has quit IRC | 23:02 | |
| *** dcdamien has quit IRC | 23:02 | |
| *** dcdamien has joined #openstack-ansible | 23:03 | |
| *** dcdamien has joined #openstack-ansible | 23:03 | |
| *** DanyC_ has quit IRC | 23:04 | |
| *** faizy_ has quit IRC | 23:13 | |
| *** faizy_ has joined #openstack-ansible | 23:13 | |
| *** cshen has quit IRC | 23:15 | |
| *** nurdie_ has quit IRC | 23:18 | |
| *** nurdie has joined #openstack-ansible | 23:19 | |
| *** nurdie has quit IRC | 23:23 | |
| openstackgerrit | Damian Cikowski (dcdamien) proposed openstack/openstack-ansible-os_cinder master: tasks: cinder_qos: Ommit nonexistent volume types https://review.openstack.org/616376 | 23:42 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-plugins stable/rocky: Modify provider_network plugin to compare group_binds to group_names https://review.openstack.org/616377 | 23:43 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible stable/rocky: Provides group_names for use with provider_network plugin https://review.openstack.org/616378 | 23:43 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible stable/rocky: Provides group_names for use with provider_network plugin https://review.openstack.org/616378 | 23:44 |
| *** dhellmann has quit IRC | 23:45 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible stable/queens: Provides group_names for use with provider_network plugin https://review.openstack.org/616379 | 23:45 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-plugins stable/queens: Modify provider_network plugin to compare group_binds to group_names https://review.openstack.org/616380 | 23:45 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible stable/queens: Provides group_names for use with provider_network plugin https://review.openstack.org/616379 | 23:46 |
| *** nurdie has joined #openstack-ansible | 23:56 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!