| *** ansmith has joined #openstack-ansible | 00:00 | |
| *** gyee has quit IRC | 00:25 | |
| *** cshen has joined #openstack-ansible | 00:38 | |
| *** cshen has quit IRC | 00:43 | |
| *** ram5391 has quit IRC | 00:44 | |
| *** mino_ has joined #openstack-ansible | 01:36 | |
| *** mmercer has quit IRC | 02:00 | |
| *** ianychoi has quit IRC | 02:10 | |
| *** ianychoi has joined #openstack-ansible | 02:10 | |
| *** nicolasbock has quit IRC | 02:17 | |
| *** lbragstad has quit IRC | 03:16 | |
| *** harry_huang has joined #openstack-ansible | 03:18 | |
| *** faizy_ has joined #openstack-ansible | 04:13 | |
| *** faizy98 has quit IRC | 04:16 | |
| *** dave-mccowan has joined #openstack-ansible | 04:28 | |
| ravirjn | ravirjn | 04:32 |
|---|---|---|
| *** nurdie has joined #openstack-ansible | 04:37 | |
| openstackgerrit | Kaio Kassiano Moura Oliveira proposed openstack/openstack-ansible master: [Docs] Update role maturity matrix https://review.openstack.org/615722 | 04:46 |
| *** dave-mccowan has quit IRC | 04:53 | |
| *** ivve has joined #openstack-ansible | 04:55 | |
| *** udesale has joined #openstack-ansible | 04:55 | |
| *** udesale has quit IRC | 04:58 | |
| *** udesale has joined #openstack-ansible | 04:58 | |
| *** udesale has quit IRC | 04:58 | |
| *** udesale has joined #openstack-ansible | 05:05 | |
| openstackgerrit | 98k proposed openstack/ansible-hardening master: fix tox python3 overrides https://review.openstack.org/608571 | 05:22 |
| *** udesale has quit IRC | 05:40 | |
| *** udesale has joined #openstack-ansible | 05:53 | |
| *** faizy98 has joined #openstack-ansible | 06:32 | |
| *** faizy_ has quit IRC | 06:34 | |
| *** udesale has quit IRC | 06:54 | |
| ravirjn | Thanks odyssey4me! now I am able to create cinder-api container. when I am running os-cinder-install.yml , getting again no hosts matched under Install cinder services. And, at the end of the playbook getting some checksum error. http://paste.openstack.org/show/734178/ | 07:19 |
| *** cshen has joined #openstack-ansible | 07:26 | |
| *** udesale has joined #openstack-ansible | 07:35 | |
| *** pcaruana has joined #openstack-ansible | 07:36 | |
| *** gkadam_ has joined #openstack-ansible | 07:45 | |
| *** nurdie has quit IRC | 07:50 | |
| *** rgogunskiy has joined #openstack-ansible | 07:59 | |
| jrosser | ravirjn: if you’ve only just introduced the cinder services it could be that the repo server has not built the venvs required for cinder. You could try running the repo-build playbook to ensure everything you need is present on the repo server. | 08:06 |
| *** fghaas has joined #openstack-ansible | 08:06 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible master: Imported Translations from Zanata https://review.openstack.org/615770 | 08:18 |
| *** priteau has joined #openstack-ansible | 08:39 | |
| *** nurdie has joined #openstack-ansible | 08:51 | |
| *** nurdie has quit IRC | 08:55 | |
| evrardjp | o/ | 08:57 |
| jrosser | good morning evrardjp | 08:57 |
| evrardjp | good morning jrosser | 08:58 |
| evrardjp | how are things? | 08:58 |
| jrosser | good :) | 08:58 |
| jrosser | we should fix some more things today | 08:59 |
| openstackgerrit | Florian Haas proposed openstack/openstack-ansible-os_nova master: Expose the Libvirt cpu_model and cpu_model_extra_flags options https://review.openstack.org/615788 | 09:00 |
| jrosser | i think some of the recent changes to tempest have broken a few of our role tests | 09:00 |
| openstackgerrit | Florian Haas proposed openstack/openstack-ansible-os_nova master: Expose the Libvirt cpu_model and cpu_model_extra_flags options https://review.openstack.org/615788 | 09:07 |
| fghaas | ^^ this one is related to https://review.openstack.org/#/c/615608/, if anyone is inclined to take a look. If it lands, it'll make guilhermesp's patch much easier. | 09:10 |
| *** DanyC has joined #openstack-ansible | 09:13 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder stable/rocky: tasks: cinder_qos: Use the new OpenStack CLI client https://review.openstack.org/614496 | 09:16 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder stable/queens: tasks: cinder_qos: Use the new OpenStack CLI client https://review.openstack.org/614499 | 09:17 |
| openstackgerrit | Florian Haas proposed openstack/openstack-ansible-os_nova master: Expose the Libvirt cpu_model and cpu_model_extra_flags options https://review.openstack.org/615788 | 09:17 |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder master: RHEL: Move zlib-devel to cinder_devel_distro_packages https://review.openstack.org/615526 | 09:18 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: Docs: Fix VLAN bridge interface in example setup https://review.openstack.org/615795 | 09:20 |
| *** priteau has quit IRC | 09:20 | |
| *** Emine has joined #openstack-ansible | 09:20 | |
| *** priteau has joined #openstack-ansible | 09:23 | |
| *** DanyC_ has joined #openstack-ansible | 09:28 | |
| *** DanyC has quit IRC | 09:28 | |
| hwoarang | good morning | 09:31 |
| arxcruz | Morning | 09:31 |
| *** aedc has joined #openstack-ansible | 09:33 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_masakari master: Add missing docs/releasenotes tests https://review.openstack.org/615622 | 09:49 |
| ravirjn | jrosser: Now I am getting another error. Also, I observed cinder-scheduler container is not running. I am trying to make controller node as storage node as well. http://paste.openstack.org/show/734187/ | 09:55 |
| jrosser | ravirjn: you need to trace that api failure back through the loadbalancer, to the cinder containers and then to the services themselves | 09:58 |
| jrosser | you can use curl to try the endpoint on the LB, and directly to the backend service, that'll tell you if it is a loadbalancer problem or the backend | 09:59 |
| *** lbragstad has joined #openstack-ansible | 10:00 | |
| *** arxcruz has quit IRC | 10:03 | |
| *** aedc has quit IRC | 10:05 | |
| *** aedc has joined #openstack-ansible | 10:05 | |
| *** electrofelix has joined #openstack-ansible | 10:10 | |
| *** mino_ has quit IRC | 10:11 | |
| *** asettle is now known as Guest64088 | 10:22 | |
| *** asettle has joined #openstack-ansible | 10:23 | |
| *** arxcruz has joined #openstack-ansible | 10:36 | |
| *** DanyC_ has quit IRC | 10:36 | |
| noonedeadpunk | morning everyone | 10:38 |
| noonedeadpunk | folks, is there a way to get suse 15 image for openstack "without registration and sms"?:) | 10:38 |
| noonedeadpunk | I just need to test ansible roles on it | 10:39 |
| odyssey4me | noonedeadpunk I just use vagrant for my testing - hwoarang usually ensures the vagrantfile includes the right things. | 10:40 |
| odyssey4me | noonedeadpunk yep, there is the 'opensuse150' machine in the role test vagrantfile | 10:41 |
| noonedeadpunk | odyssey4me: it seems, that I need to read smth about vagrant, as I never used it before... | 10:42 |
| *** hamzaachi has joined #openstack-ansible | 10:42 | |
| hwoarang | who called | 10:43 |
| hwoarang | noonedeadpunk: vagrant up ubuntu1604 will run all the tests on ubuntu1604. similar for opensuse150, opensuse423 and centos7 | 10:44 |
| hwoarang | if you only want to bring up a machine just use up --no-provision | 10:44 |
| hwoarang | that's it | 10:44 |
| noonedeadpunk | hwoarang: oh, great, will try it now) | 10:44 |
| hwoarang | and ubuntu1804 of course | 10:45 |
| noonedeadpunk | hwoarang: I've catched the idea, thanks) | 10:46 |
| *** faizy98 has quit IRC | 10:47 | |
| hwoarang | noonedeadpunk: re suse 15, you can use opensuse leap 15 if it's only for ansible testing. the core codebase is the same | 10:47 |
| arxcruz | odyssey4me: hey, any idea why this http://logs.openstack.org/68/612968/15/check/openstack-ansible-functional-distro_install-opensuse-150/8ed78c5/job-output.txt.gz#_2018-11-05_15_29_30_786606 ? | 10:51 |
| *** nurdie has joined #openstack-ansible | 10:51 | |
| arxcruz | I mean, i know the error, but how can i fix it :) | 10:51 |
| odyssey4me | arxcruz bother, I need to change things up to be rid of that thing, well, for now perhaps do the same as is done for stackviz - ie use https://github.com/openstack/openstack-ansible-os_tempest/blob/master/tasks/stackviz.yml#L21 | 10:54 |
| odyssey4me | apologies for the run around | 10:55 |
| noonedeadpunk | hwoarang: ok, thanks. I'm trying to understand, why suse 15 fails on command "rpm -ql libvirt-python | egrep '^.*python2.*/(site|dist)-packages/.*'", as the same is used in nova role. | 10:55 |
| arxcruz | odyssey4me: no problem, lessons learned, i need to get more into openstack-ansible internals | 10:55 |
| *** faizy98 has joined #openstack-ansible | 10:56 | |
| evrardjp | is there anyone interested by giving opinions on the project update slides for Berlin? | 10:56 |
| evrardjp | I am sure I have missed things | 10:56 |
| evrardjp | https://docs.google.com/presentation/d/1b1sWeYWLFjMbg9b4hWS7ly4iZH6zwwQcLczWf4f1pUo/edit?usp=sharing | 10:56 |
| hwoarang | noonedeadpunk: prob because leap 15 comes with python3 by default? | 10:57 |
| odyssey4me | arxcruz that's part of an old mechanism for builds which I'm trying to get rid of to simplify everything - unfortunately that's slow moving - every time I change something, something else breaks | 10:57 |
| hwoarang | noonedeadpunk: i am guessing that libvirt-python is only python3. you can try rpm -ql libvirt-python to find out what files it contains | 10:57 |
| *** nurdie has quit IRC | 10:57 | |
| noonedeadpunk | hwoarang: thanks for the tip. But it shouldn't finish with return code 1 then... | 10:58 |
| hwoarang | why not? grep returns 1 if nothing matches | 10:59 |
| noonedeadpunk | oh, really. yep, than you're right | 11:00 |
| noonedeadpunk | never checked this out to be honest) | 11:00 |
| *** lbragstad has quit IRC | 11:02 | |
| odyssey4me | arxcruz chandankumar the os_neutron centos distro tests are failing because there are no tests to run, which is due to there being no plugin package installed... are there plugin packages available yet to add? | 11:04 |
| evrardjp | noonedeadpunk: that's what I like the most in grep! | 11:04 |
| *** lbragstad has joined #openstack-ansible | 11:04 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder stable/rocky: tasks: cinder_qos: Use the new OpenStack CLI client https://review.openstack.org/614496 | 11:05 |
| arxcruz | odyssey4me: chandankumar is the one handling tempest plugins | 11:07 |
| chandankumar | odyssey4me: logs? | 11:08 |
| openstackgerrit | Arx Cruz proposed openstack/openstack-ansible-os_tempest master: Enable python-tempestconf support https://review.openstack.org/612968 | 11:09 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder stable/queens: tasks: cinder_qos: Use the new OpenStack CLI client https://review.openstack.org/614499 | 11:10 |
| odyssey4me | chandankumar http://logs.openstack.org/58/612958/1/check/openstack-ansible-functional-distro_install-centos-7/04e3cee/job-output.txt.gz#_2018-11-05_18_11_22_328598 | 11:10 |
| odyssey4me | that's because https://github.com/openstack/openstack-ansible-os_neutron/blob/master/tests/os_neutron-overrides.yml#L27 | 11:10 |
| odyssey4me | so either we need to shift to using another whitelist (we could just use the standard whitelist from the role), or we need to ensure we have plugins installed | 11:11 |
| *** hamzaachi has quit IRC | 11:11 | |
| *** hamzaachi has joined #openstack-ansible | 11:12 | |
| *** aedc has quit IRC | 11:13 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder stable/pike: tasks: cinder_qos: Use the new OpenStack CLI client https://review.openstack.org/614502 | 11:13 |
| odyssey4me | dcdamien in the interests of getting that fix in asap, I've edited the commit msg accordingly | 11:14 |
| *** faizy98 has quit IRC | 11:15 | |
| dcdamien | odyssey4me: thanks a lot, I've no time yet | 11:15 |
| *** hamzaachi has quit IRC | 11:20 | |
| noonedeadpunk | hwoarang: it seems, that nova uses the same approach here https://github.com/openstack/openstack-ansible-os_nova/blob/master/tasks/drivers/kvm/nova_compute_kvm_install_source.yml#L37-L40 but it doesn't cause problem (or this path is not tested currently) | 11:23 |
| noonedeadpunk | hwoarang: no, the problem here, is that package can be installed by name "libvirt-python", but it's actual name is python2-libvirt-python, which is not shown by rpm -ql | 11:30 |
| chandankumar | odyssey4me: looking | 11:30 |
| *** DanyC has joined #openstack-ansible | 11:30 | |
| chandankumar | odyssey4me: let me push two patches | 11:30 |
| odyssey4me | thanks chandankumar | 11:31 |
| hwoarang | noonedeadpunk: yeah makes sense good catch | 11:34 |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: Added task to list tempest plugins https://review.openstack.org/615837 | 11:37 |
| chandankumar | odyssey4me: arxcruz ^^ | 11:38 |
| *** kaiokmo has joined #openstack-ansible | 11:39 | |
| kaiokmo | evrardjp: around? | 11:40 |
| odyssey4me | chandankumar posed a question in review | 11:40 |
| evrardjp | kaiokmo: yes I am | 11:41 |
| evrardjp | quite a busy day, but don't hesitate :) | 11:41 |
| odyssey4me | chandankumar FYI, it seems that the list here should be extended? https://github.com/openstack/openstack-ansible-os_tempest/blob/master/vars/redhat-7.yml#L24-L25 | 11:42 |
| kaiokmo | evrardjp: last time we did chat, we agreed to retire the os_monasca-ui role, right? | 11:43 |
| noonedeadpunk | So it seems, that we need to separate vars for suse 42 and 15 again? | 11:44 |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_neutron master: Install tempest from distro on CentOS Jobs https://review.openstack.org/615838 | 11:45 |
| chandankumar | odyssey4me: ^^ looking I am not sure about this one | 11:45 |
| evrardjp | kaiokmo: agreed indeed | 11:45 |
| evrardjp | ui can be dealt with in horizon probably | 11:46 |
| kaiokmo | yup | 11:46 |
| evrardjp | or else where | 11:46 |
| *** gkadam_ has quit IRC | 11:46 | |
| evrardjp | is there a problem with that? | 11:46 |
| kaiokmo | not at all | 11:46 |
| evrardjp | :) | 11:46 |
| *** gkadam has joined #openstack-ansible | 11:46 | |
| kaiokmo | do you mind if I open a patch retiring the role? | 11:46 |
| evrardjp | not at all | 11:46 |
| kaiokmo | or it's needed to be done by an osa-core? | 11:47 |
| evrardjp | do you know the procedure? | 11:47 |
| kaiokmo | nah :) | 11:47 |
| evrardjp | retiring a repo is to be validated by the ptl | 11:47 |
| evrardjp | it has multiple steps. You can help on a few :) | 11:47 |
| odyssey4me | chandankumar I'd rather see the os_tempest role extended to install all the relevant plugin packages, instead of overriding defaults all over the place. | 11:48 |
| *** noonedeadpunk has quit IRC | 11:48 | |
| kaiokmo | thanks. I have some things to do. maybe we can chat after the meeting? | 11:48 |
| chandankumar | odyssey4me: yes make sense | 11:48 |
| *** noonedeadpunk has joined #openstack-ansible | 11:48 | |
| evrardjp | kaiokmo: https://docs.openstack.org/infra/manual/drivers.html#retiring-a-project | 11:49 |
| kaiokmo | evrardjp: thank you. also, about role maturity, I opened https://review.openstack.org/#/c/615722/ | 11:49 |
| kaiokmo | I don't know if this is an meeting topic, so I appreciate if you can take a look at it later | 11:50 |
| evrardjp | I will review this | 11:50 |
| evrardjp | I think I want to change the maturity generator | 11:50 |
| evrardjp | we have adapted our code and practices, we can now have more automated generation | 11:51 |
| kaiokmo | yeah. atm, I thought that the status needed to be updated, so this is what the patch concerns about | 11:53 |
| hwoarang | noonedeadpunk: you can add a conditional like this https://review.openstack.org/#/c/615535/4/vars/suse.yml | 11:54 |
| chandankumar | odyssey4me: http://paste.openstack.org/show/734208/ you mean this? | 11:58 |
| chandankumar | because we need something to check where it is os_neutron is used or enabled? | 11:58 |
| odyssey4me | chandankumar do we really need to? if you're adding the plugins, they don't get used unless the whitelist specifies it, surely/ | 11:58 |
| odyssey4me | ? | 11:59 |
| chandankumar | odyssey4me: yes | 11:59 |
| odyssey4me | chandankumar I think for simplicity's sake we can forget about going lean and simply just install the plugins for everything, then in the whitelist enable/disable based on the services around (if necessary) | 12:00 |
| odyssey4me | as far as I know, tempest won't run the test if the service is not found anyway | 12:00 |
| noonedeadpunk | hwoarang: actually I was thinking about the same, but had some doubts if this approach is used somewhere else due to it's non-trivial readability | 12:01 |
| chandankumar | odyssey4me: in RDO, we maintain a package openstack-tempest-all which installs all the tempest plugin | 12:01 |
| chandankumar | odyssey4me: yes through whitelist we can control the tests what we want to run | 12:01 |
| odyssey4me | chandankumar that sounds like a great way to keep things simpler to maintain | 12:01 |
| hwoarang | noonedeadpunk: it is not soooo unreadable ;p | 12:01 |
| noonedeadpunk | agree) | 12:02 |
| chandankumar | odyssey4me: let me add the tempest-all package there | 12:02 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_masakari master: Basic implementation of masakari-monitors https://review.openstack.org/584629 | 12:02 |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: Added task to list tempest plugins https://review.openstack.org/615837 | 12:04 |
| chandankumar | odyssey4me: I am still confuse, on os_neutron side how tempest_install_method will get overridden ? does distro_vars will take care of that? | 12:06 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_neutron master: zuul: Switch to distro job template https://review.openstack.org/612958 | 12:06 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_neutron master: Remove dragonflow ML2 driver https://review.openstack.org/615576 | 12:06 |
| odyssey4me | chandankumar ok, let's see if those neutron repo tests pass now | 12:06 |
| odyssey4me | hwoarang is there a similar suse package for all tempest distro packages, including plugins? | 12:06 |
| hwoarang | evrardjp: ^ maybe? | 12:07 |
| odyssey4me | if not maybe we could get one made up so that we make the management of all the plugins simpler | 12:07 |
| odyssey4me | I dunno if the spec for it is in the rpm project chandankumar ? | 12:08 |
| hwoarang | odyssey4me: which one is on centos? i lost track of the discussion | 12:08 |
| odyssey4me | hwoarang evrardjp see https://review.openstack.org/615837 | 12:08 |
| hwoarang | ok let me ask some people | 12:09 |
| chandankumar | odyssey4me: https://github.com/rdo-packages/tempest-distgit/blob/rpm-master/openstack-tempest.spec#L191 | 12:09 |
| chandankumar | odyssey4me: it is not on rpm-packaging project side as most of the temepst plugins are still not packaged in rpm-packaging project | 12:09 |
| *** fghaas has quit IRC | 12:09 | |
| hwoarang | hmm i do not see such a package. i can see individual tempest plugin packages | 12:10 |
| chandankumar | hwoarang: feel free to create one :-) | 12:10 |
| odyssey4me | hwoarang we could add them all into the os_tempest list, although that'd be somewhat tedious to maintain | 12:11 |
| hwoarang | what's the alternative? pulling all the separate plugin packages? | 12:11 |
| odyssey4me | hwoarang yep, adding them all into https://github.com/openstack/openstack-ansible-os_tempest/blob/master/vars/suse.yml#L20 | 12:11 |
| odyssey4me | tinwood does ubuntu have any packages for the tempest plugins yet, and is there some sort of metapackage to include them all, similar to RDO's https://github.com/rdo-packages/tempest-distgit/blob/rpm-master/openstack-tempest.spec#L191 | 12:14 |
| chandankumar | odyssey4me: ubuntu temepst packages are not maintained | 12:15 |
| chandankumar | odyssey4me: https://review.openstack.org/#/c/615514/ here I am struggling to make it working | 12:16 |
| openstackgerrit | Merged openstack/openstack-ansible-os_ceilometer master: Usage of systemd services is mapped to the host group https://review.openstack.org/613977 | 12:16 |
| hwoarang | odyssey4me: that's the list http://paste.opensuse.org/61430527 but i doubt a 'meta' package can be created soon... especially for Rocky | 12:17 |
| odyssey4me | chandankumar if the tempest version is a bit old, it won't accept a blacklist with a whitelist - we saw that for the pike release IIRC | 12:17 |
| chandankumar | odyssey4me: yes | 12:17 |
| hwoarang | you need to convince evrardjp or nicolasbock | 12:17 |
| tinwood | odyssey4me, no, I don't think we use packages for tempest; we use the git repos directly in our tempest charm as well. | 12:20 |
| dcdamien | hwoarang: Can you please review? https://review.openstack.org/#/c/561266/ | 12:20 |
| odyssey4me | tinwood ok, thanks - good to know :) | 12:20 |
| *** markvoelker has joined #openstack-ansible | 12:27 | |
| *** dave-mccowan has joined #openstack-ansible | 12:29 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_masakari master: Basic implementation of masakari-monitors https://review.openstack.org/584629 | 12:36 |
| *** ansmith has quit IRC | 12:39 | |
| *** hamzaachi has joined #openstack-ansible | 12:40 | |
| *** aedc has joined #openstack-ansible | 12:53 | |
| *** nurdie has joined #openstack-ansible | 12:53 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova master: Expose the Libvirt cpu_model and cpu_model_extra_flags options https://review.openstack.org/615788 | 12:54 |
| *** fghaas has joined #openstack-ansible | 12:58 | |
| *** nurdie has quit IRC | 12:58 | |
| fghaas | whoa I think that was the fastest patch I've ever landed anywhere in OpenStack land. Thanks odyssey4me and jrosser for the lightning quick reviews! | 12:59 |
| *** Emine has quit IRC | 12:59 | |
| odyssey4me | fghaas we used to have those settings, and they caused some trouble - but I liked the way you only enabled them if a custom cpu mode was set | 12:59 |
| odyssey4me | it makes more sense to do what we're doing now - let the nova defaults take effect unless you set your own things | 13:00 |
| fghaas | well I had a discussion with the Nova folks about that just last week. cpu_model only makes sense if you set cpu_mode=custom. cpu_model_extra_flags, however, is also legit with cpu_mode=host-model or cpu_mode=none, because host-model may not pass through some flags you might still want. | 13:01 |
| odyssey4me | great, thanks for doing that :) | 13:02 |
| fghaas | Such as vmx, which guilhermesp's patch is about. Though I think most people will really care most about pcid, with Spectre and Meltdown mitigation patches and whatnot. | 13:02 |
| fghaas | odyssey4me: any objections to stable/rocky and stable/queens backports for that? | 13:07 |
| odyssey4me | fghaas should be fine as far as I can see | 13:08 |
| fghaas | they're both clean cherry-picks. guilhermesp will be able to enable vmx in stable/rocky only, as vmx is not an allowed cpu_model_extra_flags option there. | 13:08 |
| *** tlunkw has joined #openstack-ansible | 13:08 | |
| fghaas | gah, that was confusing. Let me say that again: | 13:08 |
| fghaas | they're both clean cherry-picks. guilhermesp will be able to enable vmx in stable/rocky only, as vmx is not an allowed cpu_model_extra_flags option in Queens. | 13:08 |
| openstackgerrit | Florian Haas proposed openstack/openstack-ansible-os_nova stable/rocky: Expose the Libvirt cpu_model and cpu_model_extra_flags options https://review.openstack.org/615863 | 13:12 |
| openstackgerrit | Florian Haas proposed openstack/openstack-ansible-os_nova stable/queens: Expose the Libvirt cpu_model and cpu_model_extra_flags options https://review.openstack.org/615864 | 13:12 |
| *** Emine has joined #openstack-ansible | 13:15 | |
| *** emine__ has joined #openstack-ansible | 13:19 | |
| *** Emine has quit IRC | 13:20 | |
| *** markvoelker has quit IRC | 13:21 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_masakari master: Basic implementation of masakari-monitors https://review.openstack.org/584629 | 13:28 |
| guilhermesp | morning fghaas ! I didn't have a chance to see your comments in the patch. Lemme see now | 13:29 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest master: Add SUSE tempest plugin packages https://review.openstack.org/615869 | 13:30 |
| odyssey4me | hwoarang ^ thanks for the list :) | 13:30 |
| hwoarang | let see | 13:34 |
| hwoarang | dcdamien: LGTM thanks | 13:34 |
| *** mwe1 has quit IRC | 13:36 | |
| *** mwe1 has joined #openstack-ansible | 13:38 | |
| *** radeks has joined #openstack-ansible | 13:41 | |
| openstackgerrit | Arx Cruz proposed openstack/openstack-ansible-os_tempest master: Enable python-tempestconf support https://review.openstack.org/612968 | 13:42 |
| evrardjp | chandankumar: that package doesn't seem to come from rpm-packaging, it's an rdo only thing? | 13:44 |
| evrardjp | shouldn't that be upstreamed? | 13:44 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build stable/rocky: Ensure the venv lib64 symlink is relative https://review.openstack.org/615876 | 13:46 |
| odyssey4me | hwoarang ^ that's the fixer for the culprit blocking the rocky integrated build | 13:48 |
| odyssey4me | or at least another one | 13:48 |
| fghaas | guilhermesp: I think what you now want to do in your patch is to simply inject "vmx" into the nova_cpu_model_extra_flags list. That'll do the right thing for cpu_mode=custom and cpu_mode=host-model, and be a no-op for cpu_mode=host-passthrough. | 13:49 |
| guilhermesp | fghaas you PR https://review.openstack.org/#/c/615788/ really sounds reasonable. Thanks | 13:50 |
| hwoarang | odyssey4me: interesting bug | 13:50 |
| *** nurdie has joined #openstack-ansible | 13:51 | |
| guilhermesp | just something: shouldn't we define nova_cpu_mode and cpu_model_extra_flags with defaults? does it make sense? | 13:51 |
| odyssey4me | hwoarang yeah, in master we've removed the venv packaging - largely due to these sorts of things | 13:51 |
| odyssey4me | hwoarang as we move towards py3, packaging a venv is just becoming more and more cumbersome | 13:52 |
| fghaas | guilhermesp: no. nova_cpu_mode already has a defined behavior for when it's empty or "none", which is to use the default. And cpu_model_extra_flags should have no default. | 13:52 |
| fghaas | sorry, meant to say "which is to use the *libvirt* default". | 13:52 |
| guilhermesp | k fghaas so it makes sense your comment in my PR. I'm going to submit the change | 13:53 |
| odyssey4me | hwoarang I'm busy testing whether that patch does the trick using an infra node | 13:53 |
| *** strattao has joined #openstack-ansible | 13:55 | |
| guilhermesp | fghaas: this shouldn't be 'custom' and 'host-model'? https://github.com/openstack/openstack-ansible-os_nova/commit/58bad8e6b11931ee09fdc0bec07e426d911dac07#diff-ca98b38be47a1d270f7d2d87697fac8fR261 | 14:04 |
| odyssey4me | hwoarang ok, https://review.openstack.org/615876 is confirmed to work for opensuse at least - if we can get that merged, I can update the other patch to include the SHA bump for it | 14:04 |
| guilhermesp | as both set nova_cpu_model_extra_flags | 14:05 |
| odyssey4me | cores - https://review.openstack.org/615876 will be part of unblocking stable/rocky's integrated build on non-container hosts - please review asap | 14:05 |
| fghaas | guilhermesp: you only get to set cpu_model when cpu_mode == "custom". But you do get to set cpu_model_extra_flags if cpu_mode is "custom" or "host-model". (And if you set it in cpu_mode=host-passthrough, it's a no-op) | 14:06 |
| *** ansmith has joined #openstack-ansible | 14:07 | |
| mnaser | fghaas: i think it's best that we leave these decisions up to the deployer imho | 14:08 |
| mnaser | as i commented and added in my review, centos actually doesnt need extra flags | 14:08 |
| mnaser | so we might be overwriting things for users when they dont want them overwritten | 14:08 |
| mnaser | we don't know what the actual cpu model defined in libvirt looks like across operating systems so it's best to leave that out of it | 14:09 |
| mnaser | imho | 14:09 |
| fghaas | mnaser: and as I just commented in my review, I wholeheartedly disagree with you on this point :) | 14:11 |
| odyssey4me | fghaas mnaser there is an added complexity that centos/rhel uses the latest libvirt every time, while suse/ubuntu use the libvirt which is set out in openstack's upper constraints | 14:12 |
| odyssey4me | that's why we have to do funky venv symlinks to the host | 14:13 |
| mnaser | fghaas: ok, well, your change breaks users having custom user flags | 14:14 |
| mnaser | based on your way of doing that | 14:14 |
| fghaas | huh? how so? | 14:14 |
| mnaser | if i have custom cpu flags configured such as pcid and enable nested virt | 14:14 |
| mnaser | i am no longer able to use my custom cpu flags | 14:14 |
| mnaser | if i dont want to add extra configs because my distribution actually supports it natively, now i have them there | 14:15 |
| *** emine__ has quit IRC | 14:15 | |
| mnaser | ubuntu wants to be special and use a custom cpu_map.xml that doesnt have vmx natively | 14:15 |
| fghaas | wait what, prior to *my* change the only way you could set it in nova_nova_conf_overrides, and you can still totally do that, so I'm not sure how my change breaks anything here. | 14:16 |
| odyssey4me | yep, conf overrides trump *everything* in the template | 14:16 |
| mnaser | yes, your change is ok, but your comments regarding enforcing vmx breaks things | 14:16 |
| mnaser | and also if i have custom cpu flag, and enabled nested virt, and the vmx flag disappears, and now i'm in a sad spot where i dont know why things arent working | 14:17 |
| mnaser | so it comes out to the same thing | 14:17 |
| mnaser | except one allows you to avoid editing configs and sticking more to defaults, and the other overrrides things possibly breaking you down the line for $reasons | 14:17 |
| mnaser | i am all for documenting it though, that's why i added in the release note a comment to point towards the document you added | 14:17 |
| fghaas | mnaser: hold on there. I suggested to *inject* vmx *into* cpu_model_extra_flags. Meaning add it to the list. Not override it as in https://review.openstack.org/#/c/615608/2/templates/nova.conf.j2. | 14:18 |
| mnaser | why would we do that on systems that don't need it | 14:19 |
| odyssey4me | mnaser is the contention here the reno, not necessarily the patch's added options? | 14:19 |
| fghaas | well hold on a bit. guilhermesp, can you please rebase your patch on current master? I think that's part of the confusion here. | 14:20 |
| mnaser | odyssey4me: sorry, i am ok with fghaas patch adding that option -- i'm talking about https://review.openstack.org/#/c/615608/2 and adding "vmx" there automatically if someone sets nova_nested_virt_enabled: true | 14:20 |
| odyssey4me | yeah, I'd be concerned with mucking with the extra options too - if the distro doesn't do the right things, then I'd rather we added a note in docs - like a known issue or something | 14:21 |
| mnaser | yeah. for example in centos7 + rocky in sjc1 we didn't have to do any of that | 14:21 |
| mnaser | it was just enabled kvm_intel nested options and off you go | 14:22 |
| odyssey4me | on ubuntu kvm_nested is enabled by default, so that should carry through too? | 14:22 |
| mnaser | odyssey4me: except libvirt cpu_map is funky in ubuntu | 14:22 |
| fghaas | odyssey4me: it's the question of whether the vmx flag gets passed through to the guest by default | 14:22 |
| mnaser | and doesnt include vmx by default | 14:23 |
| mnaser | vs centos/rhel which does include it if it detects it | 14:23 |
| odyssey4me | interesting - we're using it for MNAIO tests, and we do nothing special there - and it's definitely doing nested virt by default | 14:23 |
| odyssey4me | that said, we're not using nova to manage the VM's - just libvirt | 14:23 |
| mnaser | so really: ubuntu always enables nested virt, but only pushes it when you add as extra flag -- centos never enables nested virt by default, but only pushes the flag when it detects nested virt is available | 14:24 |
| fghaas | right. My point is this: *if* your distro already enables vmx, then adding the vmx flag if the user has configured OSA to enable nested virt is a no-op and does no harm. However if it does not, then *not* adding vmx means you don't get nested virt despite setting a toggle that ostensibly enables it. And _that_ I'd find confusing. | 14:24 |
| *** aedc has quit IRC | 14:25 | |
| * mnaser doesn't like adding unnecessary config values and feels our config files are already way too bloated | 14:25 | |
| *** aedc has joined #openstack-ansible | 14:25 | |
| mnaser | in an ideal world: our configs are minimums with things just being overridden using overrides | 14:25 |
| odyssey4me | yep, except where we decide that something is a sensible default - and we can't get that sensible default applied upstream for whatever reason | 14:26 |
| fghaas | also if we introduce a toggle that says "do everything so that I can run nested virt" then it had better do that. | 14:26 |
| mnaser | it's not "do everything" .. it's "do something that isn't done by the os" .. otherwise downstream users will start to have to do things manually and we obviously want to avoid that | 14:27 |
| mnaser | and again, reno will add a note to the docs you added so its just to do one of the steps to reduce out of band configs for OSA deploys | 14:27 |
| mnaser | odyssey4me: ++ | 14:28 |
| mnaser | oh boy | 14:28 |
| mnaser | i think i found a big oops | 14:28 |
| odyssey4me | fyi this is largely the script which learned from what is done in the ubuntu package: https://github.com/openstack/openstack-ansible-ops/blob/master/multi-node-aio/playbooks/setup-host.yml#L112-L146 | 14:29 |
| odyssey4me | I found that sometimes nested virt isn't enabled, so I opted to make sure. | 14:30 |
| odyssey4me | That doesn't take care of AMD, but that'd be a simple extension. | 14:30 |
| mnaser | echo "Intel CPU not found." | 14:30 |
| mnaser | shoulda added another clause for no cpu found | 14:30 |
| mnaser | :D | 14:30 |
| fghaas | mnaser: what's the oops? | 14:30 |
| mnaser | i think we're publishing non-accessible urls in some keystone configs | 14:31 |
| * mnaser is digging in | 14:31 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_masakari master: Basic implementation of masakari-monitors https://review.openstack.org/584629 | 14:33 |
| mnaser | yeah magnum is returning the internal auth uri for users which is not accessible | 14:34 |
| mnaser | sad face | 14:34 |
| odyssey4me | mnaser I thought there was already a patch up for that | 14:35 |
| *** aedc has quit IRC | 14:35 | |
| *** mwe1 has quit IRC | 14:36 | |
| mnaser | odyssey4me: seems like the same with neutron too, we're pointing www_authenticate_uri to the internaluri | 14:36 |
| mnaser | try curl -vvv https://some.osa.cloud:9696/v2.0 | 14:36 |
| mnaser | header points to internal url | 14:36 |
| mnaser | so this is gonna be something we're gonna have to go over a whole bunch of roles | 14:37 |
| odyssey4me | mnaser some things need to talk internally, and some need to be given to public clients as a reference... I don't know which are the right things | 14:37 |
| odyssey4me | we want the internal chatter to use the internal endpoint, but the public stuff to be given the public endpoint | 14:38 |
| mnaser | odyssey4me: i believe auth_uri which became www_authenticate_uri is -- by docs -- Complete "public" Identity API endpoint. | 14:38 |
| mnaser | https://docs.openstack.org/keystonemiddleware/latest/middlewarearchitecture.html | 14:38 |
| mnaser | auth_url is the one *actually* used | 14:39 |
| mnaser | "Reason: The auth_uri option is deprecated in favor of www_authenticate_uri and will be removed in the S release." | 14:39 |
| mnaser | we should probably catch up on that too | 14:39 |
| guilhermesp | odyssey4me: so you comment implicitly suggests enabling the modules live then? https://review.openstack.org/#/c/615608/2/tasks/drivers/kvm/nova_compute_kvm.yml | 14:41 |
| odyssey4me | guilhermesp I'm not suggesting anything - just providing some background and other similar work | 14:41 |
| odyssey4me | mnaser jrosser could you take a peek at https://review.openstack.org/615876 to help unblock stable/rocky? | 14:42 |
| jrosser | odyssey4me: ok, will look | 14:44 |
| odyssey4me | jrosser tyvm | 14:44 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_magnum master: Drop identity_uri usage in keystone_authtoken https://review.openstack.org/615895 | 14:46 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_magnum master: Switch to www_authenticate_uri instead of auth_uri https://review.openstack.org/615896 | 14:46 |
| mnaser | odyssey4me: any reason this is straight to stable/rocky? | 14:46 |
| odyssey4me | mnaser yes, because we don't build venvs in the repo on master | 14:46 |
| odyssey4me | that entire body of code no longer exists in stein | 14:47 |
| mnaser | oh ok this affected stuff build in the repo container | 14:47 |
| odyssey4me | yep | 14:47 |
| mnaser | +w | 14:47 |
| jrosser | hah beat me to it | 14:48 |
| odyssey4me | moar workflow! | 14:48 |
| jrosser | :) | 14:49 |
| guilhermesp | so odyssey4me I think would be reasonable then to add the check for both intel and amd, does it sounds good? | 14:49 |
| mnaser | guilhermesp: you can just create the file as is, without worrying about intel and amd | 14:49 |
| mnaser | odyssey4me: was just providing context to similar work but it's a bit dangerous to do the same in osa because it might do baaad things to vms | 14:50 |
| guilhermesp | I see. Also, I think that there is something still missing in my PR, like load the module, they are placed but not loaded | 14:50 |
| odyssey4me | guilhermesp as far as I remember, if you try and loa the module while vm's are running - bad things happen... you'd need to confirm | 14:50 |
| odyssey4me | in that case, just placing the file and waiting for a host reboot would be better | 14:51 |
| guilhermesp | yes indeed odyssey4me | 14:51 |
| mnaser | guilhermesp: yes, we actually are kinda missing a step.. but it involves a host reboot and it would be a bit out of the scope | 14:51 |
| mnaser | the idea is at least to have that file put in place and then the operator (which we assume is an advanced user) should know where to go from there | 14:51 |
| odyssey4me | whatever the case, I would hope that this is disabled by default so that it's opt-in... quite honestly, this is the sort of thing that belongs in an upstream package | 14:52 |
| mnaser | odyssey4me: yes, it's going to be 100% disabled by default | 14:52 |
| odyssey4me | but then every distro has its own philosophy, which is why we have so much stuff inOSA to make things simpler and more consistent | 14:52 |
| mnaser | odyssey4me: another argument for "the openstack distro" | 14:53 |
| mnaser | :P | 14:53 |
| guilhermesp | thanks mnaser and odyssey4me also, regarding mnaser 's comment in the releasenote file..would we just link the docs https://docs.openstack.org/nova/latest/admin/configuration/hypervisor-kvm.html#nested-guest-support or do kinda of a summary of what the user should do | 14:53 |
| *** nurdie has quit IRC | 14:58 | |
| chandankumar | evrardjp: yes it is created in RDO | 14:58 |
| mnaser | i think a link is easier | 14:59 |
| guilhermesp | k mnaser addressing all comments | 14:59 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_magnum master: Switch to www_authenticate_uri instead of auth_uri https://review.openstack.org/615896 | 15:04 |
| openstackgerrit | Guilherme Steinmuller Pimentel proposed openstack/openstack-ansible-os_nova master: Add support to kvm nested virt https://review.openstack.org/615608 | 15:05 |
| *** strattao has quit IRC | 15:05 | |
| openstackgerrit | Guilherme Steinmuller Pimentel proposed openstack/openstack-ansible-os_nova master: Add support to kvm nested virt https://review.openstack.org/615608 | 15:05 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_magnum master: Switch to www_authenticate_uri instead of auth_uri https://review.openstack.org/615896 | 15:05 |
| arxcruz | odyssey4me: http://logs.openstack.org/68/612968/17/check/openstack-ansible-functional-distro_install-centos-7/7706a45/job-output.txt.gz#_2018-11-06_14_34_39_669748 | 15:07 |
| arxcruz | chandankumar: ^ | 15:07 |
| arxcruz | did you guys change something, it’s complaining there’s no .stestr file | 15:07 |
| chandankumar | arxcruz: checking | 15:08 |
| chandankumar | arxcruz: nope | 15:08 |
| chandankumar | arxcruz: https://review.openstack.org/615837 is about to land | 15:09 |
| odyssey4me | arxcruz that's a new one to me | 15:09 |
| *** aedc has joined #openstack-ansible | 15:11 | |
| arxcruz | :( | 15:11 |
| *** strattao has joined #openstack-ansible | 15:11 | |
| arxcruz | honestly, i don’t know what’s happening, tempest init is passing | 15:12 |
| chandankumar | arxcruz: http://logs.openstack.org/68/612968/17/check/openstack-ansible-functional-centos-7/d037a46/job-output.txt.gz#_2018-11-06_14_46_24_534363 | 15:12 |
| chandankumar | happing in source job also | 15:12 |
| chandankumar | let me take a look at patch | 15:13 |
| openstackgerrit | Arx Cruz proposed openstack/openstack-ansible-os_tempest master: Enable python-tempestconf support https://review.openstack.org/612968 | 15:16 |
| arxcruz | It must be a problem in the source | 15:17 |
| openstackgerrit | Merged openstack/openstack-ansible-os_tempest master: Added task to list tempest plugins https://review.openstack.org/615837 | 15:18 |
| chandankumar | arxcruz: ^^ in above patch it was working fine, if it will still trouble I am going to add more info stuff | 15:19 |
| chandankumar | * debug info | 15:19 |
| openstackgerrit | Arx Cruz proposed openstack/openstack-ansible-os_tempest master: Enable python-tempestconf support https://review.openstack.org/612968 | 15:23 |
| arxcruz | chandankumar: is tempest list-plugins work outside tempest directory ? | 15:23 |
| *** aedc has quit IRC | 15:26 | |
| chandankumar | arxcruz: yes | 15:27 |
| chandankumar | arxcruz: it just cares about executables | 15:27 |
| arxcruz | Ok | 15:28 |
| openstackgerrit | Benoît Knecht proposed openstack/openstack-ansible master: haproxy: allow status 200 on ceph-rgw backend https://review.openstack.org/615912 | 15:30 |
| *** gkadam has quit IRC | 15:30 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build stable/rocky: Ensure the venv lib64 symlink is relative https://review.openstack.org/615876 | 15:31 |
| odyssey4me | cores - sad to see this happen, but dragonflow is dead, so there's no need for us to carry the code: https://review.openstack.org/#/q/project:%22%255Eopenstack/openstack-ansible.*%22+topic:remove-dragonflow+(status:open+OR+status:merged) | 15:32 |
| *** nurdie has joined #openstack-ansible | 15:34 | |
| 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 | 15:37 |
| jrosser | odyssey4me: do we need a release note for removing dragonflow - i see one for it's introduction? | 15:40 |
| odyssey4me | jrosser https://review.openstack.org/#/c/615576/3/releasenotes/notes/dragonflow-removed-6285225b5525cd50.yaml | 15:40 |
| odyssey4me | it will be pulled in at the next milestone | 15:40 |
| jrosser | ooooh right i see it :) | 15:41 |
| openstackgerrit | Guilherme Steinmuller Pimentel proposed openstack/openstack-ansible-os_nova master: Add support to kvm nested virt https://review.openstack.org/615608 | 15:44 |
| *** mwe1 has joined #openstack-ansible | 15:44 | |
| openstackgerrit | Guilherme Steinmuller Pimentel proposed openstack/openstack-ansible-os_nova master: Add support to kvm nested virt https://review.openstack.org/615608 | 15:45 |
| *** priteau has quit IRC | 15:49 | |
| evrardjp | cloudnull: it would be nice if you could show up during this session at Berlin summit, to give your opinion with nspawn: https://etherpad.openstack.org/p/BER-container-security | 15:51 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_heat master: Switch to using public endpoint for URIs https://review.openstack.org/615916 | 15:54 |
| *** cshen has quit IRC | 15:55 | |
| mnaser | cloudnull, DimGR, d34dh0r53, hughsaunders, b3rnard0, palendae, odyssey4me, serverascode, rromans, erikmwilson, mancdaz, _shaps_, BjoernT, claco, echiu, dstanek, jwagner, ayoung, prometheanfire, evrardjp, arbrandes, scarlisle, luckyinva, ntt, javeriak, spotz, vdo, jmccrory, alextricity25, jasondotstar, admin0, michaelgugino, ametts, bgmccollum, darrenc, JRobinson__, colinmcnamara, thorst, adreznec, eil397, | 15:59 |
| mnaser | qwang,nishpatwa_, cathrichardson, drifterza, hwoarang, cshen, ullbeking, mnaser, nicolasbock, jrosser, cjloader, antonym | 15:59 |
| mnaser | friendly heads up for meeting :-) | 15:59 |
| mnaser | #startmeeting openstack_ansible_meeting | 16:00 |
| prometheanfire | sup | 16:00 |
| openstack | Meeting started Tue Nov 6 16:00:55 2018 UTC and is due to finish in 60 minutes. The chair is mnaser. Information about MeetBot at http://wiki.debian.org/MeetBot. | 16:00 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 16:00 |
| *** openstack changes topic to " (Meeting topic: openstack_ansible_meeting)" | 16:00 | |
| openstack | The meeting name has been set to 'openstack_ansible_meeting' | 16:00 |
| mnaser | #topic roll call | 16:01 |
| *** openstack changes topic to "roll call (Meeting topic: openstack_ansible_meeting)" | 16:01 | |
| mnaser | o/ | 16:01 |
| arxcruz | o/ | 16:01 |
| jrosser | o/ | 16:01 |
| evrardjp | o/ | 16:01 |
| prometheanfire | o/ | 16:01 |
| chandankumar | \o/ | 16:01 |
| mnaser | yay | 16:01 |
| hwoarang | o/ | 16:01 |
| odyssey4me | o/ but in another meeting too | 16:02 |
| evrardjp | we definitely have the quorum here. | 16:02 |
| mnaser | #topic Last week highlights | 16:02 |
| *** openstack changes topic to "Last week highlights (Meeting topic: openstack_ansible_meeting)" | 16:02 | |
| evrardjp | no highlights? Nothing big to share? :( | 16:02 |
| *** cloudnull has quit IRC | 16:02 | |
| *** d34dh0r53 has quit IRC | 16:02 | |
| *** eglute has quit IRC | 16:02 | |
| mnaser | everyone will be in berlinnnnn | 16:02 |
| evrardjp | I was off last week,so i don't know what happened last week | 16:03 |
| mnaser | yeah it was just a get boring stuff done type of week | 16:03 |
| evrardjp | ok | 16:03 |
| jrosser | i poked a bit more at mitogen and got our first AIO to build sucessfully, that was quite cool | 16:03 |
| evrardjp | then next? :D | 16:03 |
| evrardjp | jrosser: nice! | 16:03 |
| mnaser | yes, i think we need to hack around more with that | 16:03 |
| mnaser | the speed ups were great | 16:04 |
| jrosser | indeed - anyone interested can take a look at my patch and play https://review.openstack.org/#/c/591236/ | 16:04 |
| mnaser | lets work with dw_ to make it happen, i'd be excited to see the speeds up both in gates and overall deployments | 16:04 |
| *** rgogunskiy has quit IRC | 16:05 | |
| mnaser | #topic Bug triage | 16:05 |
| *** openstack changes topic to "Bug triage (Meeting topic: openstack_ansible_meeting)" | 16:05 | |
| mnaser | #link https://bugs.launchpad.net/openstack-ansible/+bug/1801657 | 16:05 |
| openstack | Launchpad bug 1801657 in openstack-ansible "fix tox python3 overrides" [Undecided,New] | 16:05 |
| mnaser | does this affect us (i feel like this is a global "add a bug in every project" type of thing | 16:06 |
| evrardjp | yup I have the impression | 16:06 |
| evrardjp | I think we changed tox.ini for the docs/reno, right? | 16:06 |
| mnaser | duplicates too | 16:06 |
| mnaser | we did that | 16:06 |
| mnaser | and also for linters i think? i mean the thing that mordred once mentioned was | 16:07 |
| evrardjp | yeah | 16:07 |
| mnaser | if our project focuses on deploying somewhere where the norm is... python2 | 16:07 |
| evrardjp | it seems testenv still has 2.7 | 16:07 |
| mnaser | then we should adapt to that, because we're actually tied to the base OS | 16:07 |
| evrardjp | https://github.com/openstack/openstack-ansible/blob/master/tox.ini but I think it's fine until we decide a whole move towards python3 | 16:07 |
| mnaser | and i kinda agree.. until we decide to do python3 everywhere (which might be in a while with rhel in there :() | 16:07 |
| *** cloudnull has joined #openstack-ansible | 16:08 | |
| evrardjp | mnaser: remember I had pushed for python3 earlier | 16:08 |
| dw_ | biggest mitogen blocker for OSA is all the delegate_to bugs, i think | 16:08 |
| evrardjp | and at the end, I dropped the ball. So many little things | 16:08 |
| *** nicolasbock has joined #openstack-ansible | 16:08 | |
| dw_ | i'm trying to fit a rewrite of delegate_to into 0.2.4 to fix it once and for all | 16:08 |
| evrardjp | so we moved to use basic python2 instead | 16:08 |
| mnaser | nice dw_ ! | 16:08 |
| evrardjp | dw_: nice! | 16:08 |
| mnaser | evrardjp: it feels like we should stick to python2 maybe once all projects are python3 first of all | 16:08 |
| evrardjp | thanks for joining our meeting :D | 16:08 |
| evrardjp | mnaser: that was what we decided | 16:08 |
| mnaser | and once all supported operating systems are python3 we can look into it | 16:09 |
| mnaser | so for now i can say | 16:09 |
| evrardjp | we don't own that much code that's python2 only, and it can be moved to python3 quite easily. The rest of the stack where we deploy, that's where the pain is. | 16:09 |
| evrardjp | mnaser: agreed there | 16:09 |
| mnaser | thats confirmed/medium? | 16:09 |
| mnaser | ugh im getting a really big blob of red text | 16:09 |
| mnaser | when i try to set it to confirmed | 16:09 |
| mnaser | launchpad PLZ | 16:10 |
| evrardjp | well I guess it depends... tox.ini just has testenv on python2 the rest is python3 already | 16:10 |
| evrardjp | (linters/doc/reno) | 16:10 |
| mnaser | <title>Error: Launchpad system error</title> | 16:10 |
| evrardjp | mnaser: refresh? | 16:10 |
| evrardjp | oh yeah the 5:10 thing. | 16:10 |
| evrardjp | refresh in a minute or so | 16:10 |
| evrardjp | :D | 16:10 |
| mnaser | oh this is common at this time? | 16:10 |
| mnaser | lol | 16:10 |
| mnaser | ok lets get start triaging next bug | 16:11 |
| mnaser | and we can get back when launchpad wakes up | 16:11 |
| mnaser | #link https://bugs.launchpad.net/openstack-ansible/+bug/1800837 | 16:11 |
| openstack | Launchpad bug 1800837 in openstack-ansible "openstack-ansible-nova-lxd test fails due to lxd storage pool not set up on bionic host" [Undecided,New] | 16:11 |
| *** strattao has quit IRC | 16:12 | |
| mnaser | this is a bit above my knowledge level | 16:12 |
| evrardjp | likewise | 16:13 |
| mnaser | evrardjp: who contributed/maintained this initially? | 16:13 |
| evrardjp | maybe odyssey4me can refresh my memory there. | 16:13 |
| evrardjp | :D | 16:14 |
| evrardjp | IIRC jrosser played with it at some point | 16:14 |
| evrardjp | but no idea who is using it in production. cloudnull had it in a lab at some point too | 16:14 |
| jrosser | not exactly - but I do a ton of LXD outside of OSA | 16:14 |
| evrardjp | that's the cause of my confusion :) | 16:15 |
| evrardjp | IIRC what cloudnull said in the past, he had a multi-type environment with kvm/LXD/ironic | 16:15 |
| jrosser | this is all really to do with the release of lxd 3 as the bug says | 16:15 |
| *** strattao has joined #openstack-ansible | 16:15 | |
| evrardjp | jrosser: the point is to find who will be owning said code | 16:15 |
| evrardjp | if nobody cares when it breaks, why should we? | 16:16 |
| evrardjp | let me check in code real quick | 16:17 |
| mnaser | thanks evrardjp | 16:17 |
| nicolasbock | Does lxd bring advantages over lxc? | 16:18 |
| jrosser | i dont have any nova-lxd, but have planned to. just not enough time | 16:18 |
| evrardjp | jmccrory: are you there? | 16:18 |
| mnaser | could this be a labour of love thing if we have to do a very small change? | 16:18 |
| evrardjp | nicolasbock: lxd was only used for nova backend | 16:19 |
| *** hamzy_ has quit IRC | 16:19 | |
| evrardjp | mnaser: it's probably nothing | 16:19 |
| nicolasbock | Ah, sorry, I should have realized from the name `nova-lxd` ;) | 16:19 |
| evrardjp | mnaser: we can mark it as confirmed and medium : ) | 16:19 |
| evrardjp | or low | 16:19 |
| evrardjp | and wait for someone to do this labour | 16:19 |
| evrardjp | (while we are busy with so many other things) | 16:20 |
| *** gyee has joined #openstack-ansible | 16:21 | |
| evrardjp | should we move to next bug? | 16:22 |
| mnaser | Yep. I will mark it in a second | 16:22 |
| evrardjp | (you can probably skip the bug talking about the example, as there is no update, except maybe to encourage again someone to contribute?) | 16:23 |
| evrardjp | mnaser: still issues with LP? | 16:23 |
| mnaser | Sorry, just needed to move actually | 16:23 |
| mnaser | apologies for that | 16:25 |
| mnaser | just needed to relocate | 16:26 |
| mnaser | ok so | 16:26 |
| mnaser | confirmed low | 16:26 |
| cloudnull | o/ all | 16:26 |
| evrardjp | lgtm | 16:26 |
| * cloudnull super late | 16:26 | |
| mnaser | hey cloudnull | 16:26 |
| evrardjp | cloudnull: just in time for talking about nova-lxd | 16:26 |
| mnaser | i still cant change state of https://bugs.launchpad.net/openstack-ansible/+bug/1801656 for some reason | 16:26 |
| openstack | Launchpad bug 1801656 in openstack-ansible "fix tox python3 overrides" [Undecided,New] | 16:26 |
| * mnaser shrugs | 16:26 | |
| mnaser | #link https://bugs.launchpad.net/openstack-ansible/+bug/1799514 | 16:27 |
| openstack | Launchpad bug 1799514 in openstack-ansible "Test environment example in openstack-ansible" [Undecided,New] | 16:27 |
| evrardjp | cloudnull: were you the one to contribute to it? Or was it jmccrory ? | 16:27 |
| evrardjp | I can't seem to remember | 16:27 |
| mnaser | ah damn | 16:27 |
| mnaser | i think someone fixed this | 16:27 |
| cloudnull | I worked on it some | 16:27 |
| *** tlunkw has quit IRC | 16:27 | |
| cloudnull | however I think it was jmccrory and crew that did the original work | 16:27 |
| evrardjp | ok | 16:27 |
| mnaser | https://review.openstack.org/#/c/614379/ fixed it! | 16:28 |
| evrardjp | then the bug we discussed just before this one was about it, if you or jmccrory could have a look that would be great. It's a new LXD issue that needs probably an extra task or so | 16:28 |
| evrardjp | mnaser: great you found it :) | 16:28 |
| evrardjp | Mark it as fix released then? | 16:28 |
| evrardjp | or fix committed | 16:28 |
| evrardjp | depending on the situation :) | 16:28 |
| mnaser | released cause pushed to master i guess | 16:29 |
| mnaser | no closes-bug so it wont ever make it down thechain | 16:29 |
| *** cshen has joined #openstack-ansible | 16:29 | |
| mnaser | #link https://bugs.launchpad.net/openstack-ansible/+bug/1704161 | 16:29 |
| openstack | Launchpad bug 1704161 in openstack-ansible "open-iscsi post-installation script fails in cinder-api-container" [High,New] - Assigned to Damian Cikowski (dcdamien) | 16:29 |
| *** hamzaachi has quit IRC | 16:29 | |
| mnaser | marked as confirmed | 16:30 |
| evrardjp | dcdamien: has worked on this, right? | 16:30 |
| dcdamien | yep | 16:30 |
| evrardjp | it's "related bug" in the commits, so not sure what's missing | 16:30 |
| evrardjp | do you think we can close it dcdamien ? | 16:30 |
| mnaser | https://review.openstack.org/#/c/614556/ probably needs a rebase right | 16:30 |
| mnaser | its not resolved/merged yet | 16:30 |
| evrardjp | or there are extra patches that need to happen? | 16:30 |
| mnaser | ^ that one probably needs to land | 16:30 |
| evrardjp | mnaser: but would it be a "closes-bug" ? | 16:31 |
| evrardjp | I haven't followed sorry | 16:31 |
| mnaser | i think it would close that bug | 16:31 |
| mnaser | barring some wild stuff | 16:31 |
| dcdamien | we can close when mentioned change merge | 16:31 |
| evrardjp | ok so we probably need to rebase and update the commit message | 16:31 |
| dcdamien | Okay, I'll do it | 16:32 |
| mnaser | dcdamien: if you cahnge Related-Bug to Closes-Bug it will close bug automatically when done :) | 16:32 |
| evrardjp | dcdamien: thanks! | 16:32 |
| mnaser | dcdamien: i have some reviews, ill push them shortly for that | 16:32 |
| evrardjp | dcdamien: you can ping us for review when you're ready :) | 16:32 |
| dcdamien | kk | 16:32 |
| evrardjp | thanks dcdamien ! | 16:32 |
| *** udesale has quit IRC | 16:32 | |
| mnaser | dcdamien: left some comments :D | 16:33 |
| mnaser | and thank you so muuch for your contribs :D | 16:33 |
| mnaser | they haven't gone unnoticed | 16:33 |
| mnaser | that's it for bug triage | 16:33 |
| mnaser | changed the last one to confirmed | 16:33 |
| mnaser | #topic open discussion | 16:33 |
| *** openstack changes topic to "open discussion (Meeting topic: openstack_ansible_meeting)" | 16:33 | |
| mnaser | berlin stuff: https://etherpad.openstack.org/p/OSA-berlin-planning as usual | 16:34 |
| mnaser | we have a google hangout, if you're not there, ping any of us and we'll add ya to it :) everyone is welcome | 16:34 |
| *** cshen has quit IRC | 16:34 | |
| prometheanfire | mnaser: spotz wanted to talk about dinner plans, current proposal is thursday night | 16:34 |
| prometheanfire | ~7:30 PM | 16:34 |
| evrardjp | mnaser: the project update and project onboarding are not listed there? | 16:34 |
| noonedeadpunk | mnaser: I'd probably would like to take part) | 16:34 |
| mnaser | evrardjp: oh they're not?! kmaybe there too obvious to us not to have there | 16:35 |
| evrardjp | yeah the - "space" thing | 16:35 |
| mnaser | pm'd you the link noonedeadpunk | 16:35 |
| evrardjp | added a link for the second search :p | 16:35 |
| mnaser | ah sillyness | 16:36 |
| *** andymccr has quit IRC | 16:36 | |
| mnaser | ok make sure y'all go to day 2 keynotes! | 16:36 |
| mnaser | something cool happening using an OSA cloud, JUST SAYING | 16:36 |
| mnaser | ok, and dinner plans | 16:36 |
| arxcruz | mnaser: cool :) | 16:37 |
| arxcruz | mnaser: at the end i would like to talk about os_tempest :) | 16:37 |
| cloudnull | ^ ++ | 16:37 |
| mnaser | unfortunately i have to be in rehearsals almost every night and catching up with family who lives in berlin on thursday before leaving friday, so i dont think i'll be able to make much plans :( | 16:37 |
| evrardjp | arxcruz: go ahead :0 | 16:38 |
| mnaser | but yeah, amy wanted opinions on what prometheanfire suggested | 16:38 |
| mnaser | but we can hash that out in a hangout | 16:38 |
| arxcruz | So... | 16:38 |
| prometheanfire | not been to berlin, so no clue what's good, I could ask a friend if you want | 16:38 |
| evrardjp | mnaser: agreed, let's put that into hangout :) | 16:39 |
| arxcruz | We are about to add python-tempesconf initial patch | 16:39 |
| prometheanfire | about how many do we expect? | 16:39 |
| evrardjp | arxcruz: woot! | 16:39 |
| arxcruz | I would like to propose a job to test it | 16:39 |
| arxcruz | Even though it’s non-vote | 16:39 |
| mnaser | i support that | 16:39 |
| evrardjp | what's the issue ? you can if you want :) | 16:39 |
| arxcruz | Just saying :) | 16:39 |
| evrardjp | haha great :) | 16:39 |
| arxcruz | Also, i would like in the summit we talk about the collaboration between tripleo and os_tempest role | 16:40 |
| evrardjp | I support that too, so, here you are with your votes :D | 16:40 |
| arxcruz | My manager wants more visibility from other developers | 16:40 |
| evrardjp | sounds good -- do you have something in mind? | 16:40 |
| arxcruz | Not yet, but we can discuss later | 16:40 |
| arxcruz | We just realize that we talked about it in ptg | 16:40 |
| arxcruz | But nobody else knows what’s going on | 16:41 |
| arxcruz | We plan to have a report weekly with the progress | 16:41 |
| arxcruz | Just to let people aware | 16:41 |
| evrardjp | thanks for the info | 16:41 |
| arxcruz | Since this is the beginning to collaboration between tripleo and osa team | 16:41 |
| evrardjp | I think mnaser said it already in the past, we welcome said collaboration :) | 16:42 |
| arxcruz | I don’t know who will be presenting osa in the summit, but mention the collaboration would be nice :) | 16:42 |
| arxcruz | Yes :) but for my boss, need to be brooding spread | 16:42 |
| arxcruz | Hehehehe | 16:42 |
| mnaser | yes | 16:42 |
| mnaser | ill add that to our project update and project onboarding | 16:42 |
| * mnaser takes note | 16:42 | |
| evrardjp | arxcruz: RedHat contributions slowly grow across cycles. I can highlight it in the project update slides | 16:43 |
| openstackgerrit | Damian Cikowski (dcdamien) proposed openstack/openstack-ansible-os_cinder master: Make open-iscsi as requirement just for volume and backup services https://review.openstack.org/614556 | 16:43 |
| evrardjp | mnaser: good :) | 16:43 |
| evrardjp | mnaser: you should write access on said slides btw :p | 16:44 |
| *** hamzy_ has joined #openstack-ansible | 16:44 | |
| mnaser | yes, i just added it now so i dont forget | 16:44 |
| evrardjp | arxcruz: redhat is 4th in terms of commits! | 16:44 |
| mnaser | and ill emphasize on it later | 16:44 |
| evrardjp | yeah I saw the slide change, great! | 16:45 |
| arxcruz | evrardjp: cool, but i was looking more for os_tempest, that is our first collaboration on tripleo | 16:45 |
| arxcruz | But we can talk about it outside this meeting, i don’t want to go on on this in this meeting | 16:45 |
| evrardjp | for other people interested about the slides we are talking about, it's here: https://docs.google.com/presentation/d/1b1sWeYWLFjMbg9b4hWS7ly4iZH6zwwQcLczWf4f1pUo/edit?usp=sharing | 16:45 |
| mnaser | yeah i added that as a note as well | 16:45 |
| evrardjp | (WIP) | 16:45 |
| evrardjp | I have nothing else to add for open discussion. | 16:46 |
| arxcruz | Added on my bookmark to see later, it’s almost 6 here and toddler is already asking for attention :P | 16:46 |
| * prometheanfire copies them down as final slides | 16:46 | |
| mnaser | hehe | 16:46 |
| mnaser | anyone has any other subjects? | 16:47 |
| mnaser | oh also | 16:47 |
| mnaser | we'll cancel next weeks meeting? | 16:47 |
| evrardjp | yup | 16:47 |
| prometheanfire | I haven't even started on my project update, but it's going to just be 5 min for me :| | 16:47 |
| evrardjp | next week* | 16:47 |
| openstackgerrit | Damian Cikowski (dcdamien) proposed openstack/openstack-ansible-os_cinder master: Make open-iscsi as requirement just for volume and backup services https://review.openstack.org/614556 | 16:48 |
| jrosser | i put an agenda item on the wiki about coordinaation services | 16:48 |
| jrosser | specifically we have deployment tooling for two things which might want one | 16:49 |
| jrosser | designate, seems ok without but would benefit, gnocchi seems all a bit broken without | 16:49 |
| jrosser | wondered what opinions were on that | 16:50 |
| evrardjp | could you clarify a little more? | 16:50 |
| mnaser | i think its pretty much something like etcd | 16:51 |
| evrardjp | what do you mean by "coordination service" ? | 16:51 |
| mnaser | or redis | 16:51 |
| mnaser | things to be consumed by tooz | 16:51 |
| evrardjp | oh god if we can avoid etcd I'd do it | 16:51 |
| jrosser | some openstack things are semi-expecting to hook into this https://docs.openstack.org/tooz/latest/user/index.html | 16:51 |
| evrardjp | but I guess etcd is in base services | 16:51 |
| evrardjp | I see | 16:51 |
| jrosser | and there are many back ends for that, with a whole pile of caveats/warnings that come with those various choices | 16:52 |
| evrardjp | can we use memcached? | 16:52 |
| mnaser | "The memcached driver is a basic implementation and provides little resiliency, though it’s much simpler to setup. A lot of the features provided in tooz are based on timeout (heartbeats, locks, etc) so are less resilient than other backends." | 16:52 |
| mnaser | "Lacks certain primitives (compare and delete) so certain functionality is fragile and/or broken due to this." | 16:52 |
| evrardjp | OMG mysql is even scarier: https://docs.openstack.org/tooz/latest/user/drivers.html | 16:52 |
| odyssey4me | I think a few folks have already built roles - I can't recall who exactly. Maybe noonedeadpunk ? | 16:52 |
| jrosser | yes, some stuff went into the documentation for gnocchi/redis/zookeeper | 16:53 |
| noonedeadpunk | I tried to use zookeeper as a coordination service | 16:53 |
| odyssey4me | memcached is not about resiliency - it's used as a cache only... there are other options like kafka which ansmith has proposed for messaging that requires more resilience... not sure if kafka fits the bill for this requirement | 16:54 |
| openstackgerrit | Damian Cikowski (dcdamien) proposed openstack/openstack-ansible-os_cinder master: Make open-iscsi as requirement just for volume and backup services https://review.openstack.org/614556 | 16:54 |
| evrardjp | odyssey4me: doesn't seem to be listed in tooz drivers | 16:54 |
| odyssey4me | we don't really have to own the coordination service as a project - but we should recommend it and have a user story for it | 16:54 |
| mnaser | well the thing is | 16:55 |
| mnaser | we kinda already have etcd there.. | 16:55 |
| noonedeadpunk | and it works pretty ok. I may place some patches to https://github.com/openstack/ansible-role-zookeeper | 16:55 |
| mnaser | i mean for me the thing is we want to avoid making as much out of band things for the user | 16:55 |
| jrosser | is it OK that we deploy gnocchi but it pretty much doesnt work as we leave it? | 16:55 |
| jrosser | thats really where i started thinking about this | 16:55 |
| mnaser | i don't think that's ok. also, etcd happens to be an openstack base service | 16:55 |
| mnaser | (i'm not a fan) | 16:56 |
| mnaser | but hey | 16:56 |
| odyssey4me | for zookeeper, for nodepool, our team uses https://github.com/AnsibleShipyard/ansible-zookeeper which includes clustering functionality | 16:56 |
| odyssey4me | jrosser is it functional for a single node but broken for multi-node? | 16:56 |
| mnaser | odyssey4me: yes ^ | 16:56 |
| odyssey4me | yeah, thought as much | 16:57 |
| jrosser | i think that is true yes | 16:57 |
| mnaser | i can confirm it | 16:57 |
| odyssey4me | the trouble is that we barely have people maintaining those roles and understanding the services... now we want to stretch more? | 16:57 |
| mnaser | yeah i'm torn a bit, i' | 16:57 |
| mnaser | i'm hoping they're basic | 16:58 |
| odyssey4me | if we have folks like jrosser and mnaser owning the service integration and keeping it running, then we can take it on - but I don't have a cooking clue how any of that stuff works | 16:58 |
| noonedeadpunk | odyssey4me: it's functional for both, but it's not optiomal | 16:58 |
| ansmith | odyssey4me, kafka doesn't really appply | 16:58 |
| evrardjp | mnaser: I agree that it's a base service, and we should use it as appropriately, even if I am not a fan either | 16:58 |
| cloudnull | ^ evrardjp read my mind :) | 16:58 |
| odyssey4me | otherwise the simplest is to add a doc note that it's broken for multi-node and to add a user story for each different back-end and link to those - then deployers can decide for themselves | 16:59 |
| *** chandankumar is now known as chkumar|pto | 16:59 | |
| mnaser | i feel like we just need to shape up a bit of the etcd stuff | 16:59 |
| mnaser | the thing is the etcd role is actually owned by logan- on github | 16:59 |
| odyssey4me | given etcd is a base service - we could provide a betteries-included thing for it, and anyone not a fan can add alternative user stories | 16:59 |
| evrardjp | that sounds a good approach odyssey4me | 17:00 |
| odyssey4me | that etcd role is etcd2 IIRC, and this would need 3 IIRC | 17:00 |
| noonedeadpunk | folks, I really don't have any problems with gnocchi multi-node install without coordination service... Yeah, it has some drawbaks, as processing same metrics several times, but all-in-all statistic is correct | 17:00 |
| evrardjp | but I think at the end, those stories will not exist, because the general move is towards etcd | 17:00 |
| odyssey4me | but logan- does already need 3 for the new calico thing I think :) | 17:00 |
| logan- | o/ | 17:00 |
| noonedeadpunk | we've based our billing on gnocchi | 17:00 |
| mnaser | at scale it really hurts you though noonedeadpunk | 17:00 |
| evrardjp | well etcd as a base service was meant for etcd3 | 17:00 |
| logan- | i actually just refreshed the etcd role to do v3 yesterday | 17:00 |
| mnaser | oh what perfect timing | 17:01 |
| mnaser | :P | 17:01 |
| evrardjp | :) | 17:01 |
| evrardjp | Oh I have to run for another meeting | 17:01 |
| evrardjp | interesting topic though! | 17:01 |
| mnaser | i mean it's pretty straightforward, its just dropping in a file | 17:01 |
| *** asettle has quit IRC | 17:01 | |
| evrardjp | add etcd nodes in your inventory, run the role, point to it in gnocchi/whatever consumes it, pray. | 17:01 |
| evrardjp | haha | 17:02 |
| odyssey4me | evrardjp you forgot - hand over to someone else ;) | 17:02 |
| logan- | yea | 17:02 |
| evrardjp | odyssey4me: hahaha | 17:02 |
| mnaser | logan-: can we start by maybe adding the role to openstack gerrit if you're okay with that? | 17:02 |
| logan- | the base deployment is no TLS, the main improvement i would like to see is TLS by default in the clustered setup | 17:02 |
| logan- | something like we do with rabbitmq I guess | 17:02 |
| jrosser | so if noonedeadpunk says that multinode gnocchi works and mnaser and i are having trouble, there is something else that needs attention? | 17:03 |
| logan- | yeah I'm fine with importing it | 17:03 |
| odyssey4me | logan- should we perhaps import that role into the openstack namespace if this is going to get broader adoption? | 17:03 |
| logan- | fine with me | 17:03 |
| odyssey4me | I suspect it may be of interest to other projects too. | 17:03 |
| logan- | i just switched it to use molecule too btw | 17:03 |
| odyssey4me | Excellent. :) | 17:03 |
| * odyssey4me pens logan- down to rewrite role tests. | 17:03 | |
| logan- | :P | 17:03 |
| jrosser | are we happy that the etc backend for tooz is sufficient, with it only providing distributed locks? | 17:04 |
| mnaser | jrosser: i think that's all we need and its an openstack base service? | 17:04 |
| mnaser | i think it'll solve the gnocchi use case and others, doubt we need much more | 17:05 |
| odyssey4me | As far as I recall, tooz isn't the best library to use - native is better. tooz dumbs down a lot of functionality because it tries to have many back-ends. | 17:05 |
| odyssey4me | So if there's a native driver then use it. | 17:05 |
| noonedeadpunk | mnaser: I think it's not base service, as I don't heve it in installation | 17:05 |
| mnaser | noonedeadpunk: i meant in the openstack technical governance pov | 17:06 |
| noonedeadpunk | ah, ok. | 17:06 |
| mnaser | odyssey4me: tooz is a library that abstracts locks to different backends | 17:06 |
| mnaser | etcd being one of them | 17:06 |
| mnaser | i dunno if native lock exist in gnocchi? | 17:06 |
| odyssey4me | mnaser yep, I'm aware of that - but I also remember how everyone hated it and wanted etcd to be a base service so that we didn't have to bother with tooz | 17:07 |
| noonedeadpunk | not sure about this either | 17:07 |
| odyssey4me | maybe the world has changed | 17:07 |
| mnaser | it always is .. i think :P | 17:08 |
| mnaser | anyways, i think if we can get logan- to import to openstack namespace | 17:08 |
| mnaser | and then we can look at using tooz for a coordination driver | 17:08 |
| mnaser | using etcd | 17:08 |
| mnaser | sound bueno? | 17:08 |
| odyssey4me | fine for me | 17:09 |
| logan- | ++ | 17:10 |
| mnaser | sounds good to me | 17:11 |
| noonedeadpunk | just to get more acknowledged - what's wrong with zookeeper (except being written in Java and being part of Apache project)? It seems reliable enought and clusterization is also nice. Or maybe I'm just not common with etcd.... | 17:11 |
| mnaser | noonedeadpunk: one second | 17:12 |
| mnaser | ill do this outside meeting | 17:12 |
| mnaser | #endmeeting | 17:12 |
| *** openstack changes topic to "Launchpad: https://launchpad.net/openstack-ansible || Weekly Meetings: https://wiki.openstack.org/wiki/Meetings/openstack-ansible || Review Dashboard: http://bit.ly/2xA1eZC" | 17:12 | |
| openstack | Meeting ended Tue Nov 6 17:12:41 2018 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 17:12 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2018/openstack_ansible_meeting.2018-11-06-16.00.html | 17:12 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2018/openstack_ansible_meeting.2018-11-06-16.00.txt | 17:12 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2018/openstack_ansible_meeting.2018-11-06-16.00.log.html | 17:12 |
| mnaser | noonedeadpunk: https://governance.openstack.org/tc/reference/base-services.html | 17:12 |
| mnaser | if you feel like zookeeper is the better option, i would encourage pushing a tc change | 17:13 |
| odyssey4me | noonedeadpunk nothing wrong with it, etcd just happens to be passed in governance as a base service which means we need to prefer it for implementation | 17:13 |
| jrosser | it is odd (in those terms) that zookeeper is the reference implementation for tooz. that sounds somewhat contrary to the base services thing | 17:14 |
| noonedeadpunk | Ok, got it. I'm just not aware about some things. Just seems that zookeeper has more functionality with tooz, than etcd. | 17:15 |
| odyssey4me | jrosser tooz was written well before that time | 17:15 |
| odyssey4me | *that decision | 17:15 |
| jrosser | ah, ok | 17:15 |
| odyssey4me | noonedeadpunk yeah, while the initial implementation will use etcd - there's no reason why zookeeper could be an another option which you work on | 17:16 |
| odyssey4me | I suspect there will be deployers who prefer one or the other for whatever reason. | 17:16 |
| noonedeadpunk | sounds nice | 17:17 |
| *** cshen has joined #openstack-ansible | 17:24 | |
| nsmeds | if a PR has passed Zuul, am I allowed to bump here or simply be patient and someone will get to it at some point? | 17:25 |
| nsmeds | aka how annoying can I be ;) | 17:26 |
| odyssey4me | nsmeds there are often a lot of patches up for review, so asking one or twice a day is kinda expected if you need it reviewed | 17:27 |
| odyssey4me | but if there's already feedback on it, be sure to respond as quickly as you can to it in review | 17:27 |
| nsmeds | I can roll with that! https://review.openstack.org/#/c/613412/ <3 | 17:27 |
| nsmeds | there's feedback but its been addressed IMO | 17:28 |
| *** Leo_m has joined #openstack-ansible | 17:28 | |
| nsmeds | once I'm more familiar with OSA will try to contribute reviews/feedback as well, but still in the "what the hell is going on here" process | 17:28 |
| nsmeds | updated our openstack-ansible version forgetting the above PR wasn't merged and broke galera's SSL replication (just in development stages still, so not huge issue) | 17:30 |
| odyssey4me | nsmeds :) it happens - happy to see someone picking that up again... unfortunately the last guy who worked on that changed roles and hasn't continued contributing | 17:32 |
| odyssey4me | it's a little late in the day for me to work through that, hopefully jrosser will look again soon | 17:32 |
| nsmeds | no worries, no rush - I'll bump again in a few days if necessary | 17:32 |
| jrosser | i've starred it - obviously on my radar now | 17:33 |
| openstackgerrit | Arx Cruz proposed openstack/openstack-ansible-os_tempest master: Enable python-tempestconf support https://review.openstack.org/612968 | 17:35 |
| jrosser | dcdamien: does the ceph backend case need special treatment here? http://logs.openstack.org/66/561266/4/gate/openstack-ansible-deploy-ceph-ubuntu-bionic/34d0b7a/job-output.txt.gz#_2018-11-06_14_43_00_202404 | 17:37 |
| dcdamien | jrosser: I know that failed but not yet investigated. Will check. | 17:39 |
| jrosser | no worries | 17:41 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_masakari master: Basic implementation of masakari-monitors https://review.openstack.org/584629 | 17:47 |
| *** smoke87 has joined #openstack-ansible | 17:47 | |
| smoke87 | Hey.. someone is using vpnaas dashboard in horizon with openstack-ansible queens? I followed the instructions to enable but no new panel in horizon... | 17:48 |
| mnaser | nsmeds: please be as annoying as you want | 17:48 |
| mnaser | nsmeds: +2 from me and it looks like tests pass | 17:50 |
| *** vnogin has joined #openstack-ansible | 17:53 | |
| nsmeds | much appreciated | 17:55 |
| guilhermesp | if someone else could take a quick look at this https://review.openstack.org/#/c/615608/6 it would be really appreciated :) | 18:00 |
| *** fghaas has quit IRC | 18:02 | |
| openstackgerrit | Merged openstack/openstack-ansible-galera_server master: Fix Galera self-signed SSL functionality https://review.openstack.org/613412 | 18:28 |
| cloudnull | ^ very nice! | 18:28 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-galera_server stable/rocky: Fix Galera self-signed SSL functionality https://review.openstack.org/615971 | 18:29 |
| *** markvoelker has joined #openstack-ansible | 18:31 | |
| *** gyee has quit IRC | 18:34 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ops master: Add ability to deploy Ceph into a Multinode AIO https://review.openstack.org/607338 | 18:36 |
| *** hamzaachi has joined #openstack-ansible | 18:43 | |
| kaiokmo | mnaser: around? | 18:49 |
| *** hamzy_ is now known as hamzy | 18:53 | |
| odyssey4me | jrosser cloudnull are you using osprofiler? there's a chat in #openstack-keystone about potentially removing it due to bit rot and bad design | 18:56 |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible master: Add qdrouterd role for oslo.messaging backend deployment https://review.openstack.org/570261 | 18:56 |
| cloudnull | I am, however, I'm not married to it. | 18:57 |
| jrosser | The function it provides is very helpful | 18:57 |
| *** dw_ is now known as dw | 18:57 | |
| jrosser | But how it is done is less important, it’s the data that counts | 18:57 |
| odyssey4me | ok folks, I'm out for the night - my brain has decided it can't cope with data ingestion and processing no more... I think it's out of disk space | 18:59 |
| guilhermesp | heheh take care odyssey4me ! | 19:00 |
| *** shardy has quit IRC | 19:00 | |
| *** DanyC has quit IRC | 19:09 | |
| *** DanyC has joined #openstack-ansible | 19:10 | |
| *** rgogunskiy has joined #openstack-ansible | 19:11 | |
| *** electrofelix has quit IRC | 19:14 | |
| *** DanyC has quit IRC | 19:15 | |
| *** eglute has joined #openstack-ansible | 19:18 | |
| *** rgogunskiy has quit IRC | 19:25 | |
| *** rgogunskiy has joined #openstack-ansible | 19:26 | |
| *** rgogunskiy has quit IRC | 19:29 | |
| mnaser | kaiokmo: hiya | 19:30 |
| *** strattao has quit IRC | 19:31 | |
| *** strattao has joined #openstack-ansible | 19:32 | |
| kaiokmo | mnaser: so, I was talking to evrardjp earlier. a couple weeks ago he asked me if the os_monasca-ui really needed to be maintained | 19:32 |
| kaiokmo | we thought that would be better not to maintain the role, because what the role itself deploy it's only an section on horizon with "monitoring" label | 19:34 |
| kaiokmo | this could be included on the horizon role, and the os_monasca-ui role could be retired (it will be one less role to worry about, after all) | 19:37 |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible master: Add qdrouterd role for oslo.messaging backend deployment https://review.openstack.org/570261 | 19:42 |
| kaiokmo | we will continue good with monasca support as long os_monasca and os_monasca-agent roles are still maintained | 19:46 |
| kaiokmo | mnaser: sounds good? | 19:47 |
| mnaser | kaiokmo: i agree | 19:53 |
| mnaser | i tihnk that should live in horizon just like others | 19:54 |
| noonedeadpunk | Folks, check please this patch, once you have time for it https://review.openstack.org/#/c/584629/ | 19:55 |
| kaiokmo | mnaser: ok. so I will open a patch retiring the role. I think https://docs.openstack.org/infra/manual/drivers.html#retiring-a-project is the way to go, right? | 19:55 |
| noonedeadpunk | it's pretty big, but still... | 19:55 |
| *** gokhani has quit IRC | 19:57 | |
| mnaser | kaiokmo: yes sir | 20:02 |
| kaiokmo | mnaser: thanks :) | 20:03 |
| jrosser | theres a bunch more of these ready for more reviews https://review.openstack.org/#/q/topic:reduce-unnecessary-package-installs+(status:open) | 20:06 |
| jrosser | the barbican one fails on missing tempest tests, looks very much related to what happened in the neutron one | 20:07 |
| *** fghaas has joined #openstack-ansible | 20:12 | |
| *** fghaas has quit IRC | 20:23 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Adds option to enable ARA plugin install https://review.openstack.org/612775 | 20:23 |
| *** vnogin has quit IRC | 20:26 | |
| * dmsimard ara senses tingling | 20:26 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron master: zuul: Switch to distro job template https://review.openstack.org/612958 | 20:32 |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible master: Add qdrouterd role for oslo.messaging backend deployment https://review.openstack.org/570261 | 20:40 |
| arxcruz | odyssey4me: hey, if you’re still around... https://review.openstack.org/#/c/612968/ pass | 20:54 |
| arxcruz | odyssey4me: can I have your ack to set the use_tempestconf to false and +2+w ? | 20:54 |
| dcdamien | someone should rot in hell for this tempest output :D -> http://logs.openstack.org/56/614556/5/check/openstack-ansible-functional-distro_install-centos-7/2b12218/job-output.txt.gz#_2018-11-06_20_26_51_586594 | 20:54 |
| *** radeks has quit IRC | 20:55 | |
| dmsimard | dcdamien: http://logs.openstack.org/56/614556/5/check/openstack-ansible-functional-distro_install-centos-7/2b12218/logs/ara-report/result/da51ae9a-8d42-4474-a0a9-4ec78fc6493b/ is probably easier to read | 20:59 |
| dmsimard | (from http://logs.openstack.org/56/614556/5/check/openstack-ansible-functional-distro_install-centos-7/2b12218/logs/ara-report/ ) | 20:59 |
| dcdamien | dmsimard, you've saved my life, thank you | 21:01 |
| dmsimard | :) | 21:03 |
| openstackgerrit | Antony Messerli proposed openstack/openstack-ansible stable/rocky: Adds option to enable ARA plugin install https://review.openstack.org/616004 | 21:04 |
| dmsimard | dcdamien: for OSA there's two "layers" of "ara-report", one from the perspective of zuul and one from the perspective of OSA (under logs/) | 21:04 |
| openstackgerrit | Antony Messerli proposed openstack/openstack-ansible stable/queens: Adds option to enable ARA plugin install https://review.openstack.org/616005 | 21:05 |
| openstackgerrit | Antony Messerli proposed openstack/openstack-ansible stable/pike: Adds option to enable ARA plugin install https://review.openstack.org/616006 | 21:05 |
| dcdamien | dmsimardy: got it, thanks again | 21:06 |
| jrosser | dcdamien: for tempest results there is also this http://logs.openstack.org/56/614556/5/check/openstack-ansible-functional-distro_install-centos-7/2b12218/logs/openstack/keystone/stestr_results.html | 21:07 |
| dmsimard | oh that's actually much better :P | 21:07 |
| jrosser | it's super well hidden though :/ | 21:08 |
| *** hamzaachi has quit IRC | 21:08 | |
| dcdamien | this is incomparably easier to parse, thanks guys | 21:11 |
| dcdamien | see you tomorrow | 21:12 |
| cloudnull | take care dcdamien | 21:16 |
| smoke87 | Hey guys, vpnaas dashboard is working in openstack-ansible queens? I enabled it in neutron_plugins_base(vpnaas), but no horizon panel. | 21:20 |
| *** cshen has quit IRC | 21:22 | |
| mnaser | did you redeploy horizon smoke87 ? | 21:29 |
| *** pcaruana has quit IRC | 21:31 | |
| *** cshen has joined #openstack-ansible | 21:40 | |
| *** ansmith has quit IRC | 21:42 | |
| *** hamzy has quit IRC | 21:45 | |
| smoke87 | yes.. | 21:46 |
| smoke87 | I run the os-horizon-install.yml playbook | 21:46 |
| *** ArchiFleKs has joined #openstack-ansible | 21:49 | |
| *** strattao has quit IRC | 21:49 | |
| smoke87 | Searching the web I realized that neutron-vpnaas-dashboard is a separated project .. so I asked for you if it was implemented in openstack-ansible code | 21:50 |
| *** ThiagoCMC has joined #openstack-ansible | 21:52 | |
| *** noonedeadpunk[h] has joined #openstack-ansible | 21:53 | |
| noonedeadpunk[h] | cores, if someone have time for a review, you are welcome:) https://review.openstack.org/#/c/584629/ | 21:55 |
| *** devx has quit IRC | 21:55 | |
| *** geekinutah has quit IRC | 21:57 | |
| *** geekinutah has joined #openstack-ansible | 21:57 | |
| *** ivve has quit IRC | 22:07 | |
| *** DanyC has joined #openstack-ansible | 22:08 | |
| *** cmart has joined #openstack-ansible | 22:11 | |
| cmart | Hi OSA! I have a hopefully dumb question. How do we override config YAML files for multi-valued options? | 22:11 |
| cmart | For example, in the [pci] section of nova.conf, you can repeat the "alias" directive to specify aliases for multiple PCI devices | 22:12 |
| cmart | yet I cannot figure out syntax that will result in multiple values getting populated (on multiple lines) in nova.conf. | 22:13 |
| openstackgerrit | Merged openstack/openstack-ansible-os_horizon master: Remove unnecessary package install duplication https://review.openstack.org/614338 | 22:14 |
| openstackgerrit | Merged openstack/openstack-ansible-os_designate master: Remove unnecessary package install duplication https://review.openstack.org/614340 | 22:16 |
| cmart | ^ I think I figured it out :) Thank you odyssey4me for the nice release note https://github.com/openstack/openstack-ansible-plugins/blob/76bec48bc911e4b8d3e9f94b9129d4c71dc5cef7/releasenotes/notes/config_template-MultiStrOps-support-c28e33fd5044e14d.yaml | 22:18 |
| cmart | "If an override is used in an INI config file the set type is defined using the standard yaml construct of "?" as the item marker." (in other words use '?' instead of '-' to start each item in your list of directives). | 22:19 |
| openstackgerrit | Merged openstack/openstack-ansible-os_aodh master: Remove unnecessary package install duplication https://review.openstack.org/614335 | 22:20 |
| *** vnogin has joined #openstack-ansible | 22:27 | |
| *** vnogin has quit IRC | 22:31 | |
| *** mmercer has joined #openstack-ansible | 22:34 | |
| *** dave-mccowan has quit IRC | 22:35 | |
| *** nurdie has quit IRC | 22:38 | |
| *** cshen has quit IRC | 22:39 | |
| *** Leo_m has quit IRC | 23:00 | |
| *** nurdie has joined #openstack-ansible | 23:31 | |
| *** nurdie has quit IRC | 23:36 | |
| *** noonedeadpunk[h] has quit IRC | 23:55 | |
| *** DanyC_ has joined #openstack-ansible | 23:57 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!