| *** savvas has joined #openstack-ansible | 00:00 | |
| *** flemingo has joined #openstack-ansible | 00:01 | |
| *** sep_ has quit IRC | 00:04 | |
| *** savvas has quit IRC | 00:04 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Use production like interfaces in the gate https://review.openstack.org/553636 | 00:05 |
|---|---|---|
| *** sep has joined #openstack-ansible | 00:05 | |
| *** flemingo has quit IRC | 00:05 | |
| *** guhcampos has joined #openstack-ansible | 00:10 | |
| *** ibmko has quit IRC | 00:11 | |
| *** flemingo has joined #openstack-ansible | 00:22 | |
| *** flemingo has quit IRC | 00:27 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_tempest stable/newton: Add retries to apt cache update https://review.openstack.org/553357 | 00:30 |
| *** TxGirlGeek has quit IRC | 00:32 | |
| *** osnaya has joined #openstack-ansible | 00:34 | |
| *** odyssey4me has quit IRC | 00:34 | |
| *** odyssey4me has joined #openstack-ansible | 00:35 | |
| osnaya | @cloudnull @odyssey4me for osa deployment on host scenario (no containers)... what exact usage is for external_lb_vip_addr and internal_lb_vip_addr? | 00:38 |
| osnaya | @cloudnull @odyssey4me since I have only 1 network interface in the lab setup, I am providing same IP address for external and internal_lb_vip_addr. Do you foresee any issues? | 00:40 |
| *** savvas has joined #openstack-ansible | 00:47 | |
| *** savvas has quit IRC | 00:47 | |
| *** savvas has joined #openstack-ansible | 00:48 | |
| *** joseg has quit IRC | 00:54 | |
| *** hamzy__ has joined #openstack-ansible | 00:57 | |
| osnaya | @cloudnull @odyssey4me @evrardjp setup-infrastructure.yml error - Task Galera_server: Check that wsrep_ready is ready - failed (shows OFF).... any ideas/pointers on debug | 01:14 |
| *** poopcat has quit IRC | 01:14 | |
| *** poopcat has joined #openstack-ansible | 01:15 | |
| *** jra has quit IRC | 01:16 | |
| *** flemingo has joined #openstack-ansible | 01:27 | |
| *** flemingo has quit IRC | 01:31 | |
| *** osnaya has quit IRC | 01:33 | |
| *** chyka has joined #openstack-ansible | 01:34 | |
| *** savvas_ has joined #openstack-ansible | 01:37 | |
| *** chyka has quit IRC | 01:39 | |
| *** savvas has quit IRC | 01:39 | |
| *** flemingo has joined #openstack-ansible | 01:40 | |
| *** gokhan_ has quit IRC | 01:44 | |
| *** flemingo has quit IRC | 01:45 | |
| cloudnull | osnaya The public and internal vip will need to be that of your load balancer | 01:46 |
| *** zerick_ is now known as zerick | 01:46 | |
| cloudnull | if you're deploying on hosts it's assumed you have a separate load balacer host or device. if you put everything on a single machine you'll get into port conflicts | 01:46 |
| cloudnull | evrardjp: may have some thoughts on this | 01:47 |
| *** tasker has joined #openstack-ansible | 01:47 | |
| tasker | I have a homogenous cluster running ubuntu/trusty and I'm adding in some new compute hosts that are ubuntu/xenial. when I run openstack-ansible repo-build, it doesn't identify the need to build xenial packages. | 01:48 |
| tasker | the facts for the new compute hosts indicate that they are xenial. but why didn't the playbook notice the need? | 01:48 |
| tasker | skipping: [localhost] => (item=nova_compute) => {"changed": false, "item": "nova_compute", "skip_reason": "Conditional check failed", "skipped": true} | 01:51 |
| tasker | is there a way to instruct repo_build to make sure it builds packages for all my operating system versions? | 01:52 |
| *** flemingo has joined #openstack-ansible | 01:55 | |
| *** savvas_ has quit IRC | 01:56 | |
| *** savvas has joined #openstack-ansible | 01:57 | |
| *** guhcampos has quit IRC | 01:59 | |
| *** flemingo has quit IRC | 01:59 | |
| tasker | it looks like it doesn't. | 02:05 |
| tasker | going through the repo-build playbook, it only cares about what release / arch the "repo_server" themselves are made from. | 02:06 |
| tasker | it doesn't consider a mixed-architecture compute set-up. | 02:06 |
| *** flemingo has joined #openstack-ansible | 02:21 | |
| *** flemingo has quit IRC | 02:25 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Remove the "is_ssh_address" option from inventory https://review.openstack.org/548929 | 02:36 |
| *** flemingo has joined #openstack-ansible | 02:41 | |
| *** chyka has joined #openstack-ansible | 02:42 | |
| *** flemingo has quit IRC | 02:46 | |
| *** chyka has quit IRC | 02:46 | |
| *** savvas has quit IRC | 02:53 | |
| *** savvas has joined #openstack-ansible | 02:54 | |
| *** savvas has quit IRC | 02:59 | |
| cloudnull | o/ tasker | 03:18 |
| cloudnull | do you have at least one infra node running a repo container on xenial and trusty ? | 03:19 |
| cloudnull | if so you will need to run repo-build to ensure you have the required bits | 03:20 |
| *** flemingo has joined #openstack-ansible | 03:42 | |
| *** flemingo has quit IRC | 03:46 | |
| *** flemingo has joined #openstack-ansible | 03:51 | |
| *** flemingo has quit IRC | 03:56 | |
| *** udesale has joined #openstack-ansible | 04:00 | |
| *** poopcat has quit IRC | 04:08 | |
| *** gkadam has joined #openstack-ansible | 04:30 | |
| *** chyka has joined #openstack-ansible | 04:32 | |
| openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible master: Integrate Congress with OSA. https://review.openstack.org/503971 | 04:33 |
| *** chyka has quit IRC | 04:36 | |
| *** flemingo has joined #openstack-ansible | 04:40 | |
| *** flemingo has quit IRC | 04:44 | |
| *** flemingo has joined #openstack-ansible | 04:54 | |
| *** flemingo has quit IRC | 04:59 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add container journal linking https://review.openstack.org/544518 | 05:04 |
| *** flemingo has joined #openstack-ansible | 05:06 | |
| *** flemingo has quit IRC | 05:11 | |
| *** flemingo has joined #openstack-ansible | 05:22 | |
| *** dariko has joined #openstack-ansible | 05:23 | |
| *** flemingo has quit IRC | 05:26 | |
| *** flemingo has joined #openstack-ansible | 05:33 | |
| *** gkadam_ has joined #openstack-ansible | 05:34 | |
| *** gkadam has quit IRC | 05:36 | |
| *** flemingo has quit IRC | 05:38 | |
| *** flemingo has joined #openstack-ansible | 05:42 | |
| *** hw_wutianwei has quit IRC | 05:45 | |
| *** hw_wutianwei has joined #openstack-ansible | 05:45 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Correct cgroup access on older kernels https://review.openstack.org/553682 | 05:46 |
| *** flemingo has quit IRC | 05:47 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: common-tasks: Container Start/Stop conditions https://review.openstack.org/553094 | 05:53 |
| *** dariko has quit IRC | 06:03 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Correct is_container when deploying containers https://review.openstack.org/553683 | 06:10 |
| *** flemingo has joined #openstack-ansible | 06:10 | |
| *** flemingo has quit IRC | 06:15 | |
| *** flemingo has joined #openstack-ansible | 06:20 | |
| *** flemingo has quit IRC | 06:25 | |
| Taseer | evrardjp: Good morning ! | 06:52 |
| jafeha | good morning everybody | 06:52 |
| Taseer | jafeha: good morning | 06:52 |
| Taseer | evrardjp: can you help me with this => http://logs.openstack.org/71/503971/42/check/openstack-ansible-deploy-congress-ubuntu-xenial/f54e963/job-output.txt.gz#_2018-03-16_05_59_15_421802 | 06:53 |
| jafeha | good morning Taseer | 06:53 |
| *** Sha0000 has joined #openstack-ansible | 06:56 | |
| cloudnull | odyssey4me qq for you - or anyone for that matter - on the pypi server, it looks like all of our requests go through a 404 to local 200 dance, http://paste.openstack.org/show/702880, is that expected? | 07:00 |
| *** jbadiapa has joined #openstack-ansible | 07:02 | |
| *** jbadiapa has quit IRC | 07:02 | |
| *** jbadiapa has joined #openstack-ansible | 07:02 | |
| *** jwitko_ has quit IRC | 07:11 | |
| *** flemingo has joined #openstack-ansible | 07:11 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_trove master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553557 | 07:12 |
| openstackgerrit | Merged openstack/openstack-ansible-os_magnum master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553541 | 07:12 |
| openstackgerrit | Merged openstack/openstack-ansible-os_monasca-ui master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553545 | 07:12 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553524 | 07:12 |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553547 | 07:12 |
| openstackgerrit | Merged openstack/openstack-ansible-os_barbican master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553528 | 07:12 |
| openstackgerrit | Merged openstack/openstack-ansible-nspawn_container_create master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553520 | 07:14 |
| openstackgerrit | Merged openstack/openstack-ansible-os_tacker master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553555 | 07:14 |
| openstackgerrit | Merged openstack/openstack-ansible-specs master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553567 | 07:14 |
| openstackgerrit | Merged openstack/openstack-ansible-galera_server master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553515 | 07:14 |
| openstackgerrit | Merged openstack/openstack-ansible-os_watcher master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553558 | 07:14 |
| openstackgerrit | Merged openstack/openstack-ansible-openstack_hosts master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553522 | 07:14 |
| openstackgerrit | Merged openstack/openstack-ansible-os_ironic master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553537 | 07:14 |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553554 | 07:14 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553548 | 07:14 |
| openstackgerrit | Merged openstack/openstack-ansible-os_congress master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553532 | 07:15 |
| openstackgerrit | Merged openstack/openstack-ansible-galera_client master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553514 | 07:15 |
| openstackgerrit | Merged openstack/openstack-ansible-ceph_client master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553513 | 07:15 |
| openstackgerrit | Merged openstack/openstack-ansible-os_monasca-agent master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553544 | 07:15 |
| openstackgerrit | Merged openstack/openstack-ansible-os_sahara master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553552 | 07:15 |
| openstackgerrit | Merged openstack/openstack-ansible-os_horizon master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553536 | 07:15 |
| openstackgerrit | Merged openstack/openstack-ansible-os_tempest master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553556 | 07:15 |
| openstackgerrit | Merged openstack/openstack-ansible-os_gnocchi master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553534 | 07:15 |
| openstackgerrit | Merged openstack/ansible-hardening master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553511 | 07:16 |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553569 | 07:16 |
| *** flemingo has quit IRC | 07:16 | |
| *** sar has joined #openstack-ansible | 07:16 | |
| openstackgerrit | Merged openstack/openstack-ansible-openstack_openrc master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553523 | 07:16 |
| *** yolanda_ is now known as yolanda | 07:16 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_glance master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553510 | 07:17 |
| openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553562 | 07:17 |
| openstackgerrit | Merged openstack/openstack-ansible-os_cloudkitty master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553531 | 07:17 |
| openstackgerrit | Merged openstack/openstack-ansible-os_almanach master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553525 | 07:18 |
| *** Sha0000 has quit IRC | 07:20 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553563 | 07:20 |
| openstackgerrit | Merged openstack/openstack-ansible-rsyslog_client master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553565 | 07:20 |
| openstackgerrit | Merged openstack/openstack-ansible-os_keystone master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553538 | 07:21 |
| openstackgerrit | Merged openstack/openstack-ansible-os_aodh master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553526 | 07:23 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553518 | 07:23 |
| openstackgerrit | Merged openstack/openstack-ansible-apt_package_pinning master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553512 | 07:23 |
| openstackgerrit | Merged openstack/openstack-ansible-rsyslog_server master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553566 | 07:23 |
| openstackgerrit | Merged openstack/openstack-ansible-os_molteniron master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553542 | 07:24 |
| openstackgerrit | Merged openstack/openstack-ansible-os_zaqar master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553559 | 07:24 |
| openstackgerrit | Merged openstack/openstack-ansible-os_rally master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553551 | 07:25 |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553530 | 07:25 |
| *** indistylo has joined #openstack-ansible | 07:25 | |
| openstackgerrit | Merged openstack/openstack-ansible-plugins master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553561 | 07:26 |
| openstackgerrit | Merged openstack/openstack-ansible-os_panko master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553550 | 07:26 |
| openstackgerrit | Merged openstack/openstack-ansible-os_monasca master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553543 | 07:26 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Add container journal linking https://review.openstack.org/553707 | 07:27 |
| openstackgerrit | Merged openstack/openstack-ansible-memcached_server master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553519 | 07:27 |
| openstackgerrit | Merged openstack/openstack-ansible-os_ceilometer master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553529 | 07:27 |
| openstackgerrit | Merged openstack/openstack-ansible-repo_server master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553564 | 07:28 |
| openstackgerrit | Merged openstack/openstack-ansible-pip_install master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553560 | 07:28 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553517 | 07:28 |
| openstackgerrit | Merged openstack/openstack-ansible-os_searchlight master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553553 | 07:28 |
| openstackgerrit | Merged openstack/openstack-ansible-os_designate master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553533 | 07:29 |
| openstackgerrit | Merged openstack/openstack-ansible-haproxy_server master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553516 | 07:29 |
| openstackgerrit | Merged openstack/openstack-ansible-nspawn_hosts master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553521 | 07:31 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Add container journal linking https://review.openstack.org/553707 | 07:32 |
| *** flemingo has joined #openstack-ansible | 07:33 | |
| *** armaan has joined #openstack-ansible | 07:35 | |
| *** flemingo has quit IRC | 07:37 | |
| hwoarang | odyssey4me: good morning. so far trying to override the cinder end point doesn't work for me. --os-endpoint $internalurl doesn't seem to work but maybe I use it in the wrong way. Do you have any suggestion on how the command line arguments should look like? | 07:43 |
| *** flemingo has joined #openstack-ansible | 07:43 | |
| *** holser__ has joined #openstack-ansible | 07:47 | |
| *** flemingo has quit IRC | 07:48 | |
| *** epalper has joined #openstack-ansible | 08:02 | |
| *** flemingo has joined #openstack-ansible | 08:07 | |
| evrardjp | osnaya -- don't run lb external and internal lb vip on the same address. It makes no sense | 08:10 |
| evrardjp | hwoarang: jafeha Taseer good morning | 08:10 |
| hwoarang | good morning evrardjp | 08:11 |
| evrardjp | Taseer: does congress abide requirements process? | 08:11 |
| *** flemingo has quit IRC | 08:11 | |
| evrardjp | Taseer: could you compare if the requirements to install congress are matching what you can find in upper constraints in github.com/openstack/requirements ? | 08:12 |
| evrardjp | hwoarang: is there nothing like --interface or something like that? | 08:13 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: [DOC] contributor: Add initial documentation for distribution support https://review.openstack.org/549225 | 08:13 |
| hwoarang | evrardjp: can't find anything | 08:13 |
| evrardjp | hwoarang: --os-interface internal | 08:14 |
| hwoarang | what's the meaning of that | 08:15 |
| evrardjp | hwoarang: oh it's maybe worth checking you don't have any env variables set | 08:15 |
| evrardjp | it tries to use the internal endpoint instead of the external endpoint | 08:15 |
| hwoarang | evrardjp: there are a few variables set because of openrc | 08:15 |
| hwoarang | but i dont know which one may affect stuff | 08:15 |
| evrardjp | hwoarang: crash that, if you're using openstack cli, you can directly use your clouds.yaml | 08:15 |
| hwoarang | --help says that cmdline overrides the environment | 08:15 |
| evrardjp | you're on openstack cli right? | 08:16 |
| hwoarang | evrardjp: the role is using the 'cinder' cmd directly | 08:16 |
| evrardjp | oh my bad | 08:16 |
| evrardjp | and it's for cinder things only right? | 08:16 |
| hwoarang | yes | 08:16 |
| *** radeks has joined #openstack-ansible | 08:16 | |
| evrardjp | for backends | 08:16 |
| evrardjp | could you point me to the task please? | 08:16 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_cinder master: tasks: Do not check the load balancer uri for API status https://review.openstack.org/552889 | 08:17 |
| hwoarang | evrardjp: ^ | 08:17 |
| hwoarang | that's the change i am testing but doesn't work | 08:17 |
| evrardjp | thanks | 08:17 |
| hwoarang | i will experiment again later today | 08:18 |
| hwoarang | maybe ask internally some cinder people | 08:18 |
| *** electrofelix has joined #openstack-ansible | 08:18 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_heat master: tox.ini: Expose USER environment variable to execution environment https://review.openstack.org/553535 | 08:20 |
| *** flemingo has joined #openstack-ansible | 08:21 | |
| evrardjp | hwoarang: so | 08:21 |
| evrardjp | these features are available from openstack cli | 08:21 |
| hwoarang | right | 08:22 |
| evrardjp | but by default openstack cli isn't installed on cinder deploys | 08:22 |
| evrardjp | so you'd have to add it | 08:22 |
| hwoarang | do you know if there is a reason we dont have openstack cli everywhere? | 08:22 |
| hwoarang | i mean, in the venvs | 08:23 |
| evrardjp | yeah understood | 08:23 |
| evrardjp | good question -- and openstack is converging into that | 08:23 |
| evrardjp | I guess it was not necessary | 08:23 |
| evrardjp | but my question for you is double | 08:23 |
| evrardjp | do you think it would be possible to use another machine? | 08:23 |
| evrardjp | Like the utility container? | 08:23 |
| evrardjp | I'd prefer NOT do it, but still | 08:24 |
| evrardjp | the question is asked | 08:24 |
| evrardjp | my concern is why reaching ansible_host | 08:24 |
| hwoarang | well, the API LB backend is still down at this point so you can only do stuff from the cinder-api container | 08:24 |
| evrardjp | oh yeah | 08:24 |
| evrardjp | the lb is dead | 08:24 |
| evrardjp | not sure it's gonna work anyway | 08:24 |
| evrardjp | wouldn't keystone point to the LB? | 08:25 |
| hwoarang | yes it does | 08:25 |
| hwoarang | :( | 08:25 |
| evrardjp | I think with openstack client you have something for forcing the cinder endpoint | 08:25 |
| evrardjp | just a sec | 08:25 |
| * hwoarang is deploying HA again so it will take a while | 08:26 | |
| *** flemingo has quit IRC | 08:26 | |
| odyssey4me | cloudnull That should only happen during the repo build because it's not yet got all the wheels, so it tries the pypiserver first - then goes to pypi. | 08:27 |
| evrardjp | hwoarang: I guess here it's worth asking cinder ppl. I think the os-endpoint will take different endpoints depending on your type of token | 08:27 |
| evrardjp | odyssey4me: I thought of something | 08:28 |
| evrardjp | maybe it's old thinking, and it's already solved | 08:28 |
| evrardjp | but if we don't put many roles into a-r-r, the lookup of py_pkgs will have less input, and therefore less things to build. | 08:29 |
| evrardjp | I guess we now are more selective than before, but it just simplifies things "naturally" | 08:29 |
| evrardjp | odyssey4me: I think we should discuss the offline story a little more | 08:31 |
| evrardjp | I know it's not in your path | 08:31 |
| odyssey4me | evrardjp offline can only be fixed once the python simplification and staging is done - if we don't fix those first, then stuff gets much more complicated | 08:31 |
| odyssey4me | one thing that can help the offline story now is to vendor in all the keys | 08:32 |
| evrardjp | all the keys? | 08:32 |
| evrardjp | the gpg keys? | 08:32 |
| evrardjp | for repo sources? | 08:32 |
| evrardjp | anything goes. I think it's worth doing all the things for that story -- ppl are complaining everyday about these things | 08:33 |
| evrardjp | we need to be more reliable | 08:33 |
| odyssey4me | yes | 08:33 |
| evrardjp | because network isn't apparently :) | 08:33 |
| odyssey4me | evrardjp hwoarang could we get this through the door to close the gap in ocata too https://review.openstack.org/553468 | 08:38 |
| odyssey4me | also, I think it's time to remove this from the check pipeline https://review.openstack.org/553590 | 08:38 |
| *** flemingo has joined #openstack-ansible | 08:40 | |
| hwoarang | odyssey4me: lgtm but why isn't the ocata thing a backport? | 08:41 |
| odyssey4me | hwoarang it's actually a forward port from newton | 08:41 |
| hwoarang | ah right | 08:41 |
| odyssey4me | pike onwards is fine - found that bug yesterday | 08:42 |
| odyssey4me | you'll see the change id matches the newton one | 08:43 |
| hwoarang | yep | 08:43 |
| *** flemingo has quit IRC | 08:44 | |
| odyssey4me | would be nice to have another review for https://review.openstack.org/547768 if anyone has a moment to do it | 08:49 |
| *** flemingo has joined #openstack-ansible | 08:49 | |
| *** flemingo has quit IRC | 08:53 | |
| evrardjp | done | 08:56 |
| evrardjp | hey I have a new topic for everyone | 08:56 |
| evrardjp | the Vancouver Forum brainstorming etherpad: | 08:56 |
| evrardjp | Please write down what your users would like to see here: https://etherpad.openstack.org/p/YVR-openstack-ansible-brainstorming | 08:56 |
| *** armaan has quit IRC | 08:59 | |
| evrardjp | odyssey4me: I am thinking it would be nice to relay this into Rackspace ^ | 09:03 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts master: Fix role when running with isolated tags https://review.openstack.org/551032 | 09:04 |
| *** flemingo has joined #openstack-ansible | 09:10 | |
| openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_tacker stable/pike: Modify the etc directory in default/main https://review.openstack.org/553721 | 09:12 |
| odyssey4me | evrardjp did you see https://review.openstack.org/#/c/553683/1 ? | 09:15 |
| *** flemingo has quit IRC | 09:15 | |
| *** savvas has joined #openstack-ansible | 09:16 | |
| evrardjp | oh interesting | 09:19 |
| evrardjp | I didn't know ansible was doing something wrong again. | 09:19 |
| odyssey4me | I wonder if it's stale facts or something. | 09:20 |
| *** savvas has quit IRC | 09:20 | |
| odyssey4me | hm, I'm seing the same behaviour in my env - so it's not just the gate | 09:21 |
| *** flemingo has joined #openstack-ansible | 09:21 | |
| odyssey4me | while cloudnull's approach is to use a hammer - it looks like this is the best option for now | 09:21 |
| odyssey4me | check experimental? | 09:22 |
| odyssey4me | that patch is on the integrated gate - what're you hoping to test? | 09:22 |
| evrardjp | darn. | 09:22 |
| evrardjp | morning. | 09:22 |
| odyssey4me | lol | 09:23 |
| evrardjp | I thought this was openstack_hosts :) | 09:23 |
| evrardjp | haha we are good then | 09:23 |
| evrardjp | don't say anyone. | 09:23 |
| evrardjp | hwoarang: could you vote on some docs patches? This way I can continue my docs crusage | 09:24 |
| evrardjp | crusade* | 09:24 |
| evrardjp | https://review.openstack.org/#/q/status:open+project:openstack/openstack-ansible+branch:stable/queens+topic:bp/docs-improvements-stable/queens | 09:24 |
| evrardjp | at least the first ones that aren't in merge conflict | 09:24 |
| evrardjp | so that the other ones can be rechecked I guess or at least fixed | 09:24 |
| odyssey4me | evrardjp I have to admit, my votes there are rubber stamps. Those patches are impossible to review. | 09:25 |
| openstackgerrit | James Portman proposed openstack/openstack-ansible master: Stop inventory constantly giving containers new IP https://review.openstack.org/553726 | 09:25 |
| evrardjp | odyssey4me: everything is so intertwined :( | 09:25 |
| evrardjp | maybe I should have done a mega patch | 09:25 |
| *** flemingo has quit IRC | 09:26 | |
| jamespo | hey evrardjp: https://review.openstack.org/#/c/553726/ - let me know if I've done something terribly wrong there | 09:26 |
| evrardjp | nah I think all of them can be reviewed separately, but you can only get the big picture when all are landed I guess... That's all about the spec :) | 09:26 |
| evrardjp | jamespo: welcome new contributor! :D | 09:26 |
| evrardjp | thanks for the patch, a new contributor tackling the inventory as first commit, that's one big thing. | 09:27 |
| *** ibmko has joined #openstack-ansible | 09:27 | |
| jamespo | No problem! Hopefully I have understood it correctly, and it's useful | 09:27 |
| odyssey4me | jamespo aha, now it makes sense why the ip's were changing | 09:28 |
| odyssey4me | it had to do with the change in data structure under it | 09:28 |
| evrardjp | jamespo: I think we should move this to a more state machine like inventory | 09:28 |
| odyssey4me | thanks for writing up the bug and working out afix! | 09:28 |
| *** flemingo has joined #openstack-ansible | 09:29 | |
| evrardjp | the cyclomatic complexity of these things is insane | 09:29 |
| odyssey4me | just like those of your sentences ;) | 09:29 |
| jamespo | Yeah there is a lot of complication, the triple assignment is still there and makes me feel slightly bad inside too | 09:30 |
| jamespo | https://github.com/openstack/openstack-ansible/blob/master/osa_toolkit/generate.py#L629 | 09:30 |
| jamespo | I think it's lucky/obscure that network ends up being a pointer to networks[old_address] | 09:31 |
| evrardjp | also a lot of things due to old things not deprecated correctly | 09:31 |
| odyssey4me | jamespo it's hopelessly too complicated - if you can figure out a way to simplify it then we're happy to see patches | 09:31 |
| evrardjp | jamespo: yes | 09:31 |
| evrardjp | exactly what I meant | 09:31 |
| evrardjp | I went to that route earlier this year or last year, and I said "oh really wtf" | 09:31 |
| jamespo | Cool, I might have another look but didn't want to start ripping it all apart just yet | 09:31 |
| jamespo | just wanted to settle the strange behaviour down first | 09:31 |
| evrardjp | jamespo: if you want to help I am working on this on part time. I'd definitely need your help if you have spare cycles. | 09:32 |
| evrardjp | like just a few insights on a regular basis would be good | 09:32 |
| jamespo | Yeah sure, happy to try and help out :) | 09:32 |
| evrardjp | great. | 09:32 |
| *** flemingo has quit IRC | 09:33 | |
| *** aruns__ has joined #openstack-ansible | 09:40 | |
| odyssey4me | evrardjp hwoarang just a little more clean-up when you have a min to review: https://review.openstack.org/551753 | 09:41 |
| *** indistylo has quit IRC | 09:42 | |
| *** esberglu has joined #openstack-ansible | 09:43 | |
| *** esberglu has quit IRC | 09:47 | |
| *** Sha0000 has joined #openstack-ansible | 09:48 | |
| *** flemingo has joined #openstack-ansible | 09:50 | |
| adrianreza | Hi OSA folks, we want to deploy OSA with 1 infra node and 2 compute nodes at first because we're on tight budget | 09:53 |
| adrianreza | but next year we're planning to scale with 3 HA infra node and more compute nodes | 09:53 |
| adrianreza | I have config like this http://paste.openstack.org/show/702937/ https://paste.ubuntu.com/p/9YSTJ2xNnb/ | 09:54 |
| *** aruns__ has quit IRC | 09:54 | |
| *** flemingo has quit IRC | 09:55 | |
| adrianreza | I got this error when setup-openstack.yml occurred http://paste.openstack.org/show/701491/ | 09:56 |
| *** aruns__ has joined #openstack-ansible | 09:56 | |
| adrianreza | we need some advice how to plan OSA deployment with our current situation | 09:57 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/ocata: Switch upgrade test to non-voting https://review.openstack.org/553735 | 09:59 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/ocata: Use constraints for all pip package installations https://review.openstack.org/553468 | 09:59 |
| odyssey4me | evrardjp unfortunately newton->ocata's nova upgrade test is suffering the same issue as pike->queens, so we need to do https://review.openstack.org/553735 | 10:00 |
| *** flemingo has joined #openstack-ansible | 10:01 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift master: Switch pypy functional test to experimental https://review.openstack.org/553590 | 10:01 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_swift stable/queens: Switch pypy functional test to experimental https://review.openstack.org/553737 | 10:01 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_swift stable/pike: Switch pypy functional test to experimental https://review.openstack.org/553740 | 10:04 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_swift stable/ocata: Switch pypy functional test to experimental https://review.openstack.org/553741 | 10:04 |
| *** flemingo has quit IRC | 10:05 | |
| ibmko | odyssey4me, ehm, can I have one more repo related question ? :) | 10:06 |
| *** savvas has joined #openstack-ansible | 10:10 | |
| *** armaan has joined #openstack-ansible | 10:10 | |
| *** savvas has quit IRC | 10:15 | |
| *** armaan_ has joined #openstack-ansible | 10:18 | |
| *** flemingo has joined #openstack-ansible | 10:19 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible stable/queens: Fix BOOTSTRAP_OPTS https://review.openstack.org/553744 | 10:19 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible stable/pike: Fix BOOTSTRAP_OPTS https://review.openstack.org/553745 | 10:20 |
| *** armaan has quit IRC | 10:22 | |
| *** flemingo has quit IRC | 10:23 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Updated from global requirements https://review.openstack.org/549552 | 10:23 |
| *** andymccr_ is now known as andymccr | 10:27 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_keystone stable/queens: Only try to stop services that exist https://review.openstack.org/553746 | 10:27 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_keystone stable/pike: Only try to stop services that exist https://review.openstack.org/553747 | 10:28 |
| *** shardy has joined #openstack-ansible | 10:28 | |
| *** flemingo has joined #openstack-ansible | 10:29 | |
| evrardjp | odyssey4me: did you read the openstack-dev ML's email about requirements process change? | 10:32 |
| *** flemingo has quit IRC | 10:33 | |
| *** armaan has joined #openstack-ansible | 10:39 | |
| *** armaan_ has quit IRC | 10:41 | |
| odyssey4me | evrardjp nope - what's the deal? | 10:45 |
| evrardjp | odyssey4me: well I am not 100% sure. | 10:45 |
| evrardjp | it's still unclear to me | 10:45 |
| evrardjp | http://lists.openstack.org/pipermail/openstack-dev/2018-March/128352.html | 10:45 |
| odyssey4me | odd, it looks like we're still getting disk space errors on rax nodes sometimes - looking into it | 10:46 |
| odyssey4me | looking at ML post | 10:47 |
| evrardjp | I sent you the link | 10:47 |
| evrardjp | this way you don't have to look it up | 10:47 |
| *** flemingo has joined #openstack-ansible | 10:48 | |
| *** udesale has quit IRC | 10:50 | |
| *** flemingo has quit IRC | 10:53 | |
| odyssey4me | evrardjp ok, no effect on us - we only use the u-c anyway | 10:54 |
| odyssey4me | the changes involve lower bounds only, and the synchronising of lower bounds in requirements.txt - so all this means to us is that we'll no longer get the global requirements proposal bot patches... or we'll get them less often | 10:54 |
| evrardjp | yeah, that's the part I understood. But we are using that to bump of versions of sphinx and all right? | 10:56 |
| odyssey4me | sphinx is also in u-c: https://github.com/openstack/requirements/blob/master/upper-constraints.txt#L354 | 10:57 |
| *** Sha0000 has quit IRC | 10:57 | |
| *** flemingo has joined #openstack-ansible | 10:57 | |
| odyssey4me | ans all envs inherit the use of u-c: https://github.com/openstack/openstack-ansible/blob/master/tox.ini#L11 | 10:58 |
| *** ansmith has quit IRC | 10:58 | |
| odyssey4me | so the minimum version never matters | 10:58 |
| odyssey4me | when pip finds a constraint for the package it's installing, it ignores the requirement version spec | 10:58 |
| Taseer | evrardjp: the requirements are not exactly the same in the two. So do I need to update the congress requirements ? | 11:00 |
| *** armaan has quit IRC | 11:01 | |
| evrardjp | Taseer: it could be normal, are they conflicting? | 11:02 |
| *** flemingo has quit IRC | 11:02 | |
| Taseer | I can not say conflicting, e.g, in congress openstackdocstheme===1.19.0 and in requirements it is openstackdocstheme>=1.18.1 # Apache-2.0 | 11:03 |
| *** savvas has joined #openstack-ansible | 11:04 | |
| Taseer | similar are the requirements for sphinx and coverage packages | 11:04 |
| *** flemingo has joined #openstack-ansible | 11:04 | |
| *** savvas has quit IRC | 11:04 | |
| *** savvas has joined #openstack-ansible | 11:05 | |
| *** savvas has quit IRC | 11:05 | |
| *** savvas has joined #openstack-ansible | 11:05 | |
| *** armaan has joined #openstack-ansible | 11:06 | |
| *** bhujay has joined #openstack-ansible | 11:08 | |
| *** flemingo has quit IRC | 11:08 | |
| odyssey4me | evrardjp Taseer that's likely because the os_congress repo is not included in openstack/requirements/projects.txt | 11:15 |
| odyssey4me | evrardjp all the new repositories need to be added there, otherwise they're not included in the g-r process | 11:15 |
| evrardjp | ouch. probably a few ones are missing. | 11:17 |
| Taseer | odyssey4me: is that a complex process ? | 11:17 |
| odyssey4me | Taseer nope, just submitting a patch | 11:17 |
| Taseer | you may delegate it to me.... | 11:17 |
| *** armaan has quit IRC | 11:19 | |
| hwoarang | odyssey4me: in https://review.openstack.org/#/c/553721/ , due you know why only ubuntu-xenial job was executed? | 11:25 |
| hwoarang | *do you know | 11:25 |
| *** flemingo has joined #openstack-ansible | 11:25 | |
| Taseer | odyssey4me: which repository contains "openstack/requirements/projects.txt" file ? | 11:26 |
| Taseer | I cannot find in OSA | 11:26 |
| odyssey4me | hwoarang curious | 11:26 |
| hwoarang | :/ | 11:26 |
| odyssey4me | Taseer that is the repository name: 'openstack/requirements' | 11:26 |
| hwoarang | ah sorry | 11:26 |
| hwoarang | that's tacker not neutron | 11:27 |
| hwoarang | we dont have multidistro jobs there | 11:27 |
| hwoarang | my bad | 11:27 |
| Taseer | oh okay | 11:27 |
| odyssey4me | hwoarang pike doesn't have multi-distro, yeah: https://github.com/openstack/openstack-ansible-os_tacker/blob/stable/pike/zuul.d/project.yaml | 11:27 |
| hwoarang | yeah my bad sorry | 11:27 |
| odyssey4me | Taseer https://github.com/openstack/requirements/blob/master/projects.txt | 11:28 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: zuul: Make openSUSE a voting job https://review.openstack.org/552977 | 11:28 |
| *** flemingo has quit IRC | 11:30 | |
| Taseer | odyssey4me: thanks | 11:31 |
| *** admin0 has joined #openstack-ansible | 11:33 | |
| *** admin0 has quit IRC | 11:33 | |
| *** admin0 has joined #openstack-ansible | 11:34 | |
| evrardjp | I am not sure we re talking about the same thing though | 11:34 |
| evrardjp | it was a failure to install congress right? | 11:35 |
| *** flemingo has joined #openstack-ansible | 11:35 | |
| Taseer | evrardjp: it was on installing tempest plugin requirements | 11:37 |
| Taseer | http://logs.openstack.org/71/503971/42/check/openstack-ansible-deploy-congress-ubuntu-xenial/f54e963/job-output.txt.gz#_2018-03-16_05_59_15_421802 | 11:37 |
| evrardjp | oh yes, my bad | 11:37 |
| odyssey4me | Taseer log? | 11:37 |
| Taseer | odyssey4me: ^ | 11:37 |
| niraj_singh | i am getting below error: | 11:39 |
| niraj_singh | failed: [aio1_masakari_container-bbaa5113] (item={'service_key': 'masakari-api', 'service_name': 'masakari-api', 'service_en': True, 'config_options': '--config-file /etc/masakari/masakari.conf --config-file /etc/masakari/api-paste.ini --log-file=/var/log/masakari/masakari-api.log', 'group': 'masakari_api'}) => {"attempts": 5, "changed": false, "item": {"config_options": "--config-file /etc/masakari/masakari.conf --config-file | 11:39 |
| niraj_singh | /etc/masakari/api-paste.ini --log-file=/var/log/masakari/masakari-api.log", "group": "masakari_api", "service_en": true, "service_key": "masakari-api", "service_name": "masakari-api"}, "msg": "Could not find the requested service masakari-api: host"} | 11:39 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-plugins master: Utilise OrderedDict to ensure no random changes https://review.openstack.org/553757 | 11:39 |
| odyssey4me | Taseer commented on the patch | 11:39 |
| odyssey4me | not sure if it will work, but it's worth a try | 11:39 |
| *** flemingo has quit IRC | 11:40 | |
| Taseer | odyssey4me: let me see | 11:40 |
| *** savvas_ has joined #openstack-ansible | 11:41 | |
| *** savvas has quit IRC | 11:44 | |
| *** Sha000000 has joined #openstack-ansible | 11:45 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_tacker stable/pike: Modify the etc directory in default/main https://review.openstack.org/553721 | 11:48 |
| *** sar has quit IRC | 11:55 | |
| *** odyssey4me has quit IRC | 12:04 | |
| *** odyssey4me has joined #openstack-ansible | 12:04 | |
| *** bhujay has quit IRC | 12:05 | |
| *** sawblade6 has joined #openstack-ansible | 12:15 | |
| *** sawblade6 has quit IRC | 12:19 | |
| *** sawblade6 has joined #openstack-ansible | 12:20 | |
| *** sar has joined #openstack-ansible | 12:27 | |
| admin0 | anyone know how "working" is vitrage with osa ? | 12:29 |
| *** flemingo has joined #openstack-ansible | 12:29 | |
| *** hcericnoan has joined #openstack-ansible | 12:31 | |
| *** gkadam__ has joined #openstack-ansible | 12:32 | |
| *** shardy is now known as shardy_lunch | 12:32 | |
| *** flemingo has quit IRC | 12:34 | |
| *** gkadam_ has quit IRC | 12:34 | |
| *** Sha000000 has quit IRC | 12:42 | |
| *** ansmith has joined #openstack-ansible | 12:42 | |
| *** flemingo has joined #openstack-ansible | 12:45 | |
| *** flemingo has quit IRC | 12:49 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Correct is_container when deploying containers https://review.openstack.org/553683 | 12:51 |
| hwoarang | jmccrory: hello could you ping me when you are around to discuss the cinder/HA problem? :/ | 12:55 |
| hwoarang | thank you :) | 12:55 |
| *** Sha0000 has joined #openstack-ansible | 12:59 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/queens: Correct is_container when deploying containers https://review.openstack.org/553777 | 12:59 |
| *** flemingo has joined #openstack-ansible | 12:59 | |
| *** dave-mccowan has joined #openstack-ansible | 12:59 | |
| *** admin0 has quit IRC | 12:59 | |
| *** admin0 has joined #openstack-ansible | 13:00 | |
| *** gyankum has joined #openstack-ansible | 13:01 | |
| *** throwsb1 has joined #openstack-ansible | 13:02 | |
| *** shardy_lunch is now known as shardy | 13:03 | |
| *** indistylo has joined #openstack-ansible | 13:03 | |
| *** dave-mcc_ has joined #openstack-ansible | 13:04 | |
| *** flemingo has quit IRC | 13:04 | |
| *** dave-mccowan has quit IRC | 13:04 | |
| *** aruns__ has quit IRC | 13:05 | |
| *** aruns__ has joined #openstack-ansible | 13:10 | |
| *** indistylo has quit IRC | 13:10 | |
| *** flemingo has joined #openstack-ansible | 13:11 | |
| *** flemingo has quit IRC | 13:15 | |
| *** esberglu has joined #openstack-ansible | 13:17 | |
| *** dariko has joined #openstack-ansible | 13:20 | |
| *** tasker has quit IRC | 13:26 | |
| *** flemingo has joined #openstack-ansible | 13:28 | |
| *** udesale has joined #openstack-ansible | 13:31 | |
| *** flemingo has quit IRC | 13:32 | |
| *** flemingo has joined #openstack-ansible | 13:37 | |
| evrardjp | good catch odyssey4me | 13:37 |
| evrardjp | on the repo name | 13:37 |
| evrardjp | acl * | 13:37 |
| odyssey4me | evrardjp :) | 13:39 |
| *** flemingo has quit IRC | 13:42 | |
| *** udesale has quit IRC | 13:46 | |
| *** flemingo has joined #openstack-ansible | 13:47 | |
| *** flemingo has quit IRC | 13:51 | |
| admin0 | anyone here will be in vancouver ? | 13:52 |
| *** jwitko has joined #openstack-ansible | 13:52 | |
| admin0 | summit | 13:52 |
| admin0 | or too early to tell | 13:52 |
| evrardjp | admin0: maybe, depends on the travel support :( | 13:52 |
| admin0 | if my vote counts, i will vote for you | 13:52 |
| evrardjp | haha thanks buddy! | 13:52 |
| admin0 | i am not training 2 people on osa .. third one from monday .. after that, we would be able to test/contribute more | 13:53 |
| admin0 | virtage, monitoring/dashboard, SSO etc in scope in OSA to be deployed | 13:54 |
| admin0 | and lots of nfv/epa stuff | 13:55 |
| admin0 | i am refining a heat script that generates a multi node setup with all the mgmt, storage etc networks and vms creaated .. so that test can be done again and gain | 13:56 |
| admin0 | will put it somewhere after refinement | 13:56 |
| admin0 | that way, its a bit more than AIO where the test infra (DTA) can be built using heat | 13:57 |
| admin0 | port security is disabled on the controller ,so the lxc ips work nicely | 13:58 |
| *** aruns__ has quit IRC | 14:03 | |
| *** indistylo has joined #openstack-ansible | 14:03 | |
| *** flemingo has joined #openstack-ansible | 14:10 | |
| *** dariko has quit IRC | 14:11 | |
| *** dariko has joined #openstack-ansible | 14:12 | |
| *** flemingo has quit IRC | 14:15 | |
| mnaser | so really how have people felt about queens | 14:16 |
| mnaser | queens osa release that is | 14:16 |
| mnaser | any big showstoppers or no one rolled it out yet? | 14:16 |
| hwoarang | best release EVER | 14:17 |
| hwoarang | -_- | 14:17 |
| mnaser | :D | 14:20 |
| admin0 | i am rolling out 2 dev environments to test queens | 14:21 |
| admin0 | aim to start testing it from today | 14:21 |
| admin0 | can have more input on monday | 14:21 |
| odyssey4me | mnaser well, we managed to speed up the deploy somewhat for Queens vs Pike - so that's nice | 14:23 |
| mnaser | yeah, we're kinda in a bit of a limbo | 14:23 |
| odyssey4me | I'd expect there are a few loose ends to tie up here and there, so obviously we'd like feedback from as many test deployments as possible. | 14:23 |
| mnaser | new private cloud customer.. not sure if we go pike and upgrade .. or queens and be on top of our stuff | 14:23 |
| odyssey4me | Pike is pretty mature by now, so it's low risk. I guess it depends on your timings. | 14:24 |
| admin0 | mnaser, i am hoping that if by next week, queens is released, i will go queens and help here to iron out bugs | 14:24 |
| odyssey4me | It'll take a bit more time to iron out the upgrade story. I think our CI shows it broken right now. I think evrardjp would love more attention on all that. | 14:25 |
| mnaser | it is my experience that, well, devs are happy to iron out things in newer releases :p | 14:25 |
| admin0 | i promised no SLAs to business, so i have some wiggle room there | 14:25 |
| mnaser | and at this point i cant imagine us running into show stoppers | 14:25 |
| mnaser | the infra stuff is stable | 14:25 |
| mnaser | and openstack is openstack. | 14:25 |
| admin0 | yes..plus i like to have the multi attach cinder and wysiwyg heat stuff | 14:25 |
| mnaser | yeah multiattach cinder is not a thing with rbd | 14:25 |
| mnaser | so boo | 14:25 |
| *** flemingo has joined #openstack-ansible | 14:26 | |
| admin0 | export rbd over nfs :D ? | 14:26 |
| *** Sha0000 has quit IRC | 14:28 | |
| admin0 | because if move with pike now, the thing to test/upgrade will always be with me .. | 14:29 |
| *** gyankum has quit IRC | 14:29 | |
| admin0 | better go queens, help here to iron our bugs and be on latest | 14:29 |
| *** kstev has joined #openstack-ansible | 14:30 | |
| *** flemingo has quit IRC | 14:30 | |
| *** geb has quit IRC | 14:31 | |
| *** geb has joined #openstack-ansible | 14:32 | |
| *** kstev1 has joined #openstack-ansible | 14:32 | |
| *** kstev has quit IRC | 14:35 | |
| cloudnull | odyssey4me: regarding pypi server - https://i.snag.gy/T6GJLk.jpg - this was from a fully built aio this morning. I installed requests on the host and can see that the pypi server is doing the same interaction where it seems to be making two requests all the time | 14:36 |
| cloudnull | mornings all btw | 14:36 |
| *** flemingo has joined #openstack-ansible | 14:41 | |
| spotz | hola cloudnull, como estas? | 14:42 |
| *** flaviosr has joined #openstack-ansible | 14:42 | |
| evrardjp | two requests? | 14:43 |
| evrardjp | you mean the 404? | 14:43 |
| cloudnull | evrardjp: yes | 14:44 |
| evrardjp | cloudnull: do you have requests on your cache? That's not shown on your screenshot | 14:44 |
| evrardjp | cloudnull: mmm | 14:44 |
| cloudnull | it seems to be doing two requests | 14:44 |
| cloudnull | I could try another package, one sec | 14:44 |
| evrardjp | I don't like when you arrive like this in the morning on friday :D | 14:44 |
| evrardjp | cloudnull: well I meant which package do you have in cache already? | 14:44 |
| evrardjp | cloudnull: I am taking the example of chardet for example | 14:45 |
| cloudnull | spotz: todo bien, es Vierness y soy listo para la fin de semana :) | 14:45 |
| *** flemingo has quit IRC | 14:45 | |
| evrardjp | cloudnull: :) | 14:46 |
| evrardjp | that's so nice that spanish is a readable language :) | 14:46 |
| spotz | hehehe, yo tambien | 14:46 |
| spotz | I'm doing a bi-lingual talk at summit. I will be sticking to the English so as not to offend:) | 14:47 |
| evrardjp | nice | 14:48 |
| *** gkadam__ has quit IRC | 14:48 | |
| evrardjp | which language? | 14:48 |
| evrardjp | spanish/english? | 14:48 |
| cloudnull | evrardjp: https://snag.gy/Exb4uk.jpg | 14:48 |
| cloudnull | so that test was installing cachetools and the specific version I had in cache | 14:48 |
| evrardjp | cloudnull: let me check if I understand the log wells for cachetools | 14:48 |
| * cloudnull is a little meta :) | 14:48 | |
| spotz | evrardjp: Yep | 14:48 |
| evrardjp | cloudnull: I see the request, I see the 200. | 14:49 |
| cloudnull | and the 404 | 14:49 |
| evrardjp | but weirdly I also see 127.0.0.1 GET /simple/cachetools as 404 | 14:49 |
| cloudnull | idk if this is expected , maybe its just "the way"-tm | 14:50 |
| evrardjp | could it be a listen/vhost issue? | 14:50 |
| cloudnull | it was just odd imo | 14:50 |
| evrardjp | but it looks like it's serving it from cache later right? | 14:50 |
| evrardjp | only one way to find out: put a firewall on your repo server, do it again :) | 14:50 |
| evrardjp | I mean firewall towards the outside | 14:50 |
| evrardjp | that or the tcpdump | 14:51 |
| cloudnull | it seems to be serving from cache. let me block the world and try again | 14:52 |
| *** gkadam has joined #openstack-ansible | 14:52 | |
| evrardjp | well if you tell it's doing the right thing, it's just about log filtering. Let me spawn up a new instance on my side. | 14:52 |
| evrardjp | I need to do a new instance for logan anyway :) | 14:53 |
| evrardjp | cloudnull: the format for those line is also different | 14:53 |
| evrardjp | I am now intrigued! | 14:54 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-lxc_hosts master: tasks: Ensure machinectl quota value is reasonable https://review.openstack.org/552842 | 14:55 |
| *** flemingo has joined #openstack-ansible | 14:56 | |
| cloudnull | evrardjp: that broke some things https://snag.gy/sN17nG.jpg | 14:57 |
| cloudnull | evrardjp: -A OUTPUT -o eth0 -j REJECT --reject-with icmp-port-unreachable within the repo container | 14:58 |
| cloudnull | seems when it couldn't reachout it died | 14:58 |
| openstackgerrit | Merged openstack/openstack-ansible master: Fix typo and misleading task name for nova-consoleauth https://review.openstack.org/551678 | 14:59 |
| evrardjp | and osloutils is on your cache too? | 14:59 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: [Docs] Introduce next steps https://review.openstack.org/553440 | 14:59 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: [Docs] Less scary landing page https://review.openstack.org/553438 | 14:59 |
| openstackgerrit | Merged openstack/openstack-ansible master: [Docs] Simplify and fix docs configuration https://review.openstack.org/553418 | 14:59 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: Replace host_need_pip with standard override mechanism https://review.openstack.org/547768 | 14:59 |
| cloudnull | it is | 14:59 |
| evrardjp | cloudnull: maybe we are just reverse proxying then :( | 14:59 |
| evrardjp | same version and all? | 14:59 |
| evrardjp | darn | 14:59 |
| evrardjp | that's not cool | 15:00 |
| cloudnull | i ran an `ls` on the package at the top | 15:00 |
| cloudnull | its strange for sure | 15:00 |
| *** flemingo has quit IRC | 15:01 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/queens: Fix typo and misleading task name for nova-consoleauth https://review.openstack.org/553811 | 15:01 |
| evrardjp | oh yes sorry I missed the first 2 lines | 15:02 |
| cloudnull | also, interesting capability of the systemd-networkd role https://review.openstack.org/#/c/552721 - we can generate a production network setup: bonds, bridges, vlans, oh-my! | 15:02 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Fix typo and misleading task name for nova-consoleauth https://review.openstack.org/553812 | 15:02 |
| cloudnull | anyone want to give https://review.openstack.org/#/c/553682/ and https://review.openstack.org/#/c/553707/ a review ? | 15:03 |
| evrardjp | cloudnull: could you check mtu and routes too on https://review.openstack.org/#/c/552721 ? | 15:04 |
| evrardjp | cloudnull: I mean without a shell script | 15:05 |
| evrardjp | if not we can mark it somewhere we need to write that execstart/stop | 15:05 |
| evrardjp | (btw I didn't know execstart/stop, which makes me willing to converge to networkd everywhere) | 15:06 |
| evrardjp | does it break ubuntu 16? | 15:06 |
| evrardjp | cloudnull: what is https://review.openstack.org/#/c/552721/8/tests/roles/bootstrap-host/tasks/prepare_networking.yml L94 for? | 15:08 |
| evrardjp | L94-L98 | 15:08 |
| cloudnull | evrardjp: http://paste.openstack.org/show/703031/ | 15:08 |
| cloudnull | ? | 15:08 |
| cloudnull | the double address ? | 15:09 |
| cloudnull | that's so we can have more than one address on a given interface | 15:09 |
| evrardjp | no I meant | 15:09 |
| cloudnull | on br-vlan we have a .100 address and a .1 which is used as a gateway for flat networking | 15:09 |
| evrardjp | for everything else you're using interface address netmask | 15:09 |
| cloudnull | i've not seen any breakage on ubuntu 16.04 btw | 15:10 |
| evrardjp | I'm surpised this is completely different: config_overrides: Network: Address: list of two ips with cidr notation | 15:10 |
| evrardjp | at least feature wise it's on par with what we are doing in AIO. | 15:11 |
| evrardjp | lgtm | 15:11 |
| evrardjp | just suprised of readability | 15:11 |
| cloudnull | the config overrides is using the direct syntax in the network file which writes addresses in cidr notation. | 15:11 |
| cloudnull | to keep things compatible with our inventory natively the template will join the address and netmask together to create cidr notation. | 15:12 |
| *** electrofelix has quit IRC | 15:12 | |
| *** electrofelix has joined #openstack-ansible | 15:12 | |
| cloudnull | from the template https://github.com/cloudnull/ansible-systemd_networkd/blob/master/templates/systemd-network.j2#L26-L27 | 15:13 |
| evrardjp | I am very confused by networkd sometimes. | 15:13 |
| *** flemingo has joined #openstack-ansible | 15:14 | |
| evrardjp | so question: | 15:14 |
| evrardjp | maybe worth asking on different chan | 15:14 |
| evrardjp | sorry | 15:14 |
| *** electrofelix has quit IRC | 15:14 | |
| *** niraj_singh has quit IRC | 15:14 | |
| *** udesale has joined #openstack-ansible | 15:16 | |
| cloudnull | networkd is confusing as hell coming from traditional network interface setup. however now that i've gotten my head around it I like the layout and the specific small file approach | 15:17 |
| *** flemingo has quit IRC | 15:18 | |
| *** epalper has quit IRC | 15:21 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-plugins master: Utilise sorted to ensure no random changes https://review.openstack.org/553757 | 15:22 |
| *** Sha0000 has joined #openstack-ansible | 15:23 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/ocata: Fix typo and misleading task name for nova-consoleauth https://review.openstack.org/553823 | 15:23 |
| *** flemingo has joined #openstack-ansible | 15:24 | |
| *** sawblade_ has joined #openstack-ansible | 15:25 | |
| *** sawblade6 has quit IRC | 15:28 | |
| *** flemingo has quit IRC | 15:29 | |
| *** sar has quit IRC | 15:30 | |
| *** nattanon has joined #openstack-ansible | 15:30 | |
| *** indistylo has quit IRC | 15:30 | |
| nattanon | Hello everbody really need a help!!! | 15:31 |
| nattanon | Do someone ever face through problem about l3 agent ha have more than one agent active ?? | 15:32 |
| nattanon | I'm implement neutron with hybrid type DVR also normal HA mode. | 15:35 |
| *** dave-mcc_ is now known as dave-mccowan | 15:40 | |
| *** flemingo has joined #openstack-ansible | 15:41 | |
| cloudnull | nattanon: I've had the l3 agents scheduled to more than one active node which was causing flapping, to fix it i simply had to unshedule it from one of them. | 15:42 |
| cloudnull | that was with the linux bridge agent | 15:42 |
| cloudnull | idk about DVR though | 15:42 |
| *** sar has joined #openstack-ansible | 15:43 | |
| *** Sha0000 has quit IRC | 15:44 | |
| *** udesale has quit IRC | 15:44 | |
| nattanon | cloudnull: Could you explain more about step to unschedule it. | 15:45 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-plugins master: Utilise sorted to ensure no random changes https://review.openstack.org/553757 | 15:45 |
| *** flemingo has quit IRC | 15:45 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-plugins master: Utilise sorted to ensure no random changes https://review.openstack.org/553757 | 15:46 |
| *** chyka has joined #openstack-ansible | 15:47 | |
| *** armaan has joined #openstack-ansible | 15:47 | |
| cloudnull | nattanon: i simply had to identify the routers that were double scheduled and then use `neutron l3-agent-router-add/remove` | 15:48 |
| cloudnull | here's a good blog post covering that https://www.openstackfaq.com/openstack-migrate-routers-and-dhcp | 15:49 |
| cloudnull | well the basic steps anyway | 15:49 |
| *** flemingo has joined #openstack-ansible | 15:51 | |
| *** ianychoi_ has joined #openstack-ansible | 15:53 | |
| nattanon | cloudnull: Okay thank you so much i will try. Oh my production T_T. | 15:53 |
| *** flemingo has quit IRC | 15:55 | |
| *** v1k0d3n_ has joined #openstack-ansible | 15:55 | |
| *** wlmbasson_ has joined #openstack-ansible | 15:55 | |
| cloudnull | the interaction should be the same | 15:57 |
| cloudnull | nattanon: I but don't have any quality experience with DVR | 15:57 |
| *** promethe1nfire has joined #openstack-ansible | 15:58 | |
| cloudnull | maybe someone else here might? though it's friday and I suspect a lot of folks are getting ready for the weekend :P | 15:58 |
| cloudnull | nattanon: best of luck and let us know how it goes or if we can help | 15:58 |
| *** prometheanfire has quit IRC | 15:59 | |
| *** promethe1nfire is now known as prometheanfire | 15:59 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Add container journal linking https://review.openstack.org/553707 | 15:59 |
| cloudnull | anyone around to lend a review to https://review.openstack.org/#/c/550609/ | 16:01 |
| nattanon | Actually, For DVR is just passed test still not allow for anyone yet. | 16:01 |
| prometheanfire | cloudnull: commit message should say WHY it's needed | 16:01 |
| *** szaher has quit IRC | 16:02 | |
| prometheanfire | cloudnull: why why do you need both tasks? wouldn't just the tempfiles one work? | 16:02 |
| evrardjp | prometheanfire: good catch! | 16:02 |
| evrardjp | removing my vote! | 16:02 |
| *** jillr has quit IRC | 16:03 | |
| *** ianychoi has quit IRC | 16:03 | |
| *** bradm has quit IRC | 16:03 | |
| *** mattoliverau has quit IRC | 16:03 | |
| *** speedmann has quit IRC | 16:03 | |
| *** wlmbasson has quit IRC | 16:03 | |
| *** v1k0d3n has quit IRC | 16:03 | |
| *** redondo-mk has quit IRC | 16:03 | |
| *** wlmbasson_ is now known as wlmbasson | 16:03 | |
| *** v1k0d3n_ is now known as v1k0d3n | 16:03 | |
| *** speedmann has joined #openstack-ansible | 16:03 | |
| evrardjp | there is something that changed on ansible website, I don't know what | 16:03 |
| evrardjp | font? | 16:03 |
| nattanon | cloudnull: One more things if my problem related with only l3-ha | 16:04 |
| evrardjp | prometheanfire: he lined to the bug | 16:04 |
| evrardjp | linked* | 16:04 |
| prometheanfire | ah, still personal preference | 16:04 |
| evrardjp | :) | 16:04 |
| cloudnull | nattanon: in my prod clouds we're not normally using l3 ha due to flapping issues. we enable it but don't make it the default | 16:05 |
| prometheanfire | and still not sure why just the tmpfiles file isn't enough | 16:05 |
| evrardjp | could anyone vote on those non-merge conflicts backports please? https://review.openstack.org/#/q/topic:bp/docs-improvements-stable/queens+(status:open) | 16:05 |
| cloudnull | evrardjp: done. | 16:07 |
| evrardjp | thanks | 16:07 |
| nattanon | couldnull: If i connect VM direct to the public it should without router it should be worked right ? | 16:07 |
| *** geb has quit IRC | 16:08 | |
| *** redondo-mk has joined #openstack-ansible | 16:10 | |
| pabelanger | cloudnull: so, I'm having trouble getting your ssh connection plugin working locally. No matter where I set physical_host=foo, self._play_context doesn't seem to contain it. Do you know if 2.3.3.0 is supported? | 16:11 |
| cloudnull | yes, we've been using it since 2.1 | 16:11 |
| logan- | is the strategy plugin also present | 16:11 |
| cloudnull | mind sharing your inventory ? | 16:11 |
| logan- | it is still required iirc | 16:11 |
| cloudnull | ++ | 16:11 |
| cloudnull | that's a thing | 16:11 |
| openstackgerrit | Merged openstack/openstack-ansible master: Add missing Ubuntu packages https://review.openstack.org/553394 | 16:12 |
| *** flemingo has joined #openstack-ansible | 16:12 | |
| evrardjp | pabelanger: have a look at what logan- said : | 16:13 |
| evrardjp | :) | 16:13 |
| *** geb has joined #openstack-ansible | 16:13 | |
| cloudnull | i think that requirement was removed in 2.5 | 16:13 |
| pabelanger | http://paste.openstack.org/show/703045/ | 16:13 |
| pabelanger | oh, I'm not using a strategy plugin | 16:14 |
| cloudnull | we no longer need to push things into the connection plugin, it just magically have them | 16:14 |
| pabelanger | just the connection_plugin | 16:14 |
| pabelanger | let me see what that is | 16:14 |
| logan- | https://github.com/openstack/openstack-ansible-plugins/blob/master/strategy/linear.py <-- this plugin will need to be used alongside the connection plugin | 16:14 |
| pabelanger | kk | 16:14 |
| pabelanger | that is likely my issue | 16:15 |
| pabelanger | let me add it | 16:15 |
| pabelanger | hey, that is better | 16:17 |
| *** flemingo has quit IRC | 16:17 | |
| pabelanger | cloudnull: evrardjp: logan-: thanks! | 16:17 |
| logan- | :) | 16:17 |
| cloudnull | anytime | 16:17 |
| *** DanyC has joined #openstack-ansible | 16:17 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/queens: Add missing Ubuntu packages https://review.openstack.org/553844 | 16:18 |
| odyssey4me | ooh, would you look at that - an opensuse pass in check: https://review.openstack.org/#/c/553394/ | 16:18 |
| *** jillr has joined #openstack-ansible | 16:18 | |
| *** bradm has joined #openstack-ansible | 16:18 | |
| *** mattoliverau has joined #openstack-ansible | 16:18 | |
| *** admin0 has quit IRC | 16:19 | |
| idlemind | when in doubt rebuild the container | 16:20 |
| idlemind | #magic | 16:20 |
| *** bhujay has joined #openstack-ansible | 16:22 | |
| pabelanger | boo, physical_host={{ ansible_host}} or physical_host={{ ansible_hostname }} doesn't work. | 16:22 |
| pabelanger | SSH Error: data could not be sent to remote host \"{{ ansible_hostname }}\". Make sure this host can be reached over ssh", "unreachable": true | 16:23 |
| evrardjp | physical_host: "{{ ansible_hostname }}" ? | 16:23 |
| evrardjp | Well I guess it all depends on your file type | 16:23 |
| odyssey4me | pabelanger I may be wrong, but ansible_host/ansible_hostname is the name of the target (container), not the host | 16:23 |
| evrardjp | oh yeah | 16:24 |
| evrardjp | pabelanger: pabelanger https://github.com/openstack/openstack-ansible-os_keystone/blob/master/tests/group_vars/all_containers.yml | 16:24 |
| *** bhujay has quit IRC | 16:24 | |
| evrardjp | and https://github.com/openstack/openstack-ansible-os_keystone/blob/master/tests/host_vars/infra1.yml#L16 | 16:24 |
| evrardjp | these are for creating containers on AIO nodes for our testing | 16:25 |
| evrardjp | I can assure you it works for ages now :) | 16:25 |
| *** bhujay has joined #openstack-ansible | 16:25 | |
| pabelanger | so, I need to dynamically set physical_host at runtime, because I don't know the IP of the remote system | 16:25 |
| *** flemingo has joined #openstack-ansible | 16:25 | |
| *** radeks has quit IRC | 16:25 | |
| odyssey4me | the container need a host property called 'physical_host' which contains a value specifying the host it's on | 16:25 |
| pabelanger | so, figured I could just set it to ansible_host value, which is setup in inventory | 16:25 |
| odyssey4me | ansible_host in the inventory is typically the ansible target (in this case the container) | 16:26 |
| pabelanger | okay, so maybe I am not understanding the plugin. Let me recap what I'd like to happen | 16:27 |
| odyssey4me | a way you can build a dynamic form of this is to connect to the host, then use add_host to add a host to the inventory which includes all containers on the host - out dynamic inventory does that for us, but if you don't have a dynamic inventory for that you'd have to use add_host to do it | 16:27 |
| odyssey4me | *our | 16:27 |
| pabelanger | right now, ansible_host=38.145.33.34 is the remote node I can connect to and run tasks | 16:28 |
| evrardjp | ok | 16:28 |
| pabelanger | as I understand, if I want containers on that host, I would set physical_host=38.145.33.34 | 16:28 |
| evrardjp | and do you have an inventory where you can say container1 ansible_host=container1 physical_host=38.145.33.34 ... but physical_host need to be at run time, that's what you meant? | 16:28 |
| pabelanger | then, the connection_plugin would try to run commands via lxc on that remote node | 16:28 |
| pabelanger | http://paste.openstack.org/show/703045/ | 16:29 |
| pabelanger | is the inventory file I have now | 16:29 |
| odyssey4me | pabelanger yes, although the container has to be on the host already though - and the container's host properties must have that physical_host attribute | 16:29 |
| pabelanger | which, I think is wrong | 16:29 |
| evrardjp | gear01 is your metal node right now | 16:29 |
| odyssey4me | ansible_host should be the container's address | 16:29 |
| *** flemingo has quit IRC | 16:30 | |
| pabelanger | Ah, okay. So then this isn't going to work as I think | 16:30 |
| evrardjp | add another node gear01-container1 ansible_host=container1 physical_host=38.145.33.34 | 16:30 |
| pabelanger | because ansible_host is always going to be the baremetal address | 16:30 |
| pabelanger | because, that is what get setup from zuul / nodepool | 16:30 |
| evrardjp | why ? | 16:31 |
| pabelanger | hardcoded | 16:31 |
| evrardjp | but you can't have extra inventory item? | 16:31 |
| pabelanger | I can, but I cannot change ansible_host | 16:31 |
| evrardjp | like you said, you can't add_host ok... | 16:31 |
| pabelanger | right | 16:31 |
| evrardjp | mmm | 16:31 |
| evrardjp | that's a pickle indeed, because you don't have something to say "I want to reach that container" | 16:32 |
| pabelanger | I mean, I could hack up the connection_plugin for this case | 16:32 |
| evrardjp | what you want is reversing the thing | 16:32 |
| pabelanger | right | 16:32 |
| evrardjp | you're not the only one to talk about this | 16:32 |
| evrardjp | but how would you then select what to run your thing on? | 16:33 |
| evrardjp | if you don't mark it in an inventory of some sort? | 16:33 |
| pabelanger | yah, if I could use the default ansible_host then some how say, i also want to use container commands for that host, it should work | 16:33 |
| pabelanger | yah | 16:33 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-plugins master: Utilise sorted to ensure no random changes https://review.openstack.org/553757 | 16:33 |
| pabelanger | so, does lxc-attach use an ip address? | 16:34 |
| evrardjp | pabelanger: I think your best solution, if you want to do container things with only a limited, unmodifiable set of inventory, you're good at writing something of your own | 16:34 |
| evrardjp | and always reuse that module | 16:35 |
| evrardjp | which will be tedious | 16:35 |
| evrardjp | because you can't delegate to a non inventory thing etc. | 16:35 |
| evrardjp | pabelanger: no | 16:35 |
| evrardjp | it just enter the namespace | 16:35 |
| evrardjp | enters* | 16:35 |
| odyssey4me | pabelanger it doesn't, but if the physical_host and ansible_host addresses match, we assume that it's a physical host | 16:35 |
| evrardjp | nsenter is doing the same almost | 16:35 |
| jmccrory | hwoarang hi is https://review.openstack.org/#/c/552889/ not working in your testing? | 16:36 |
| pabelanger | odyssey4me: Ah, I see. And I want the opposite | 16:36 |
| evrardjp | pabelanger: so to summarize: in a plugin you can say I want to target node x (container or physical host), if x in inventory has physical_host == ansible_host we know it's a physical host, and connect with ssh. If in the inventory it's different, we jump on physical_host through ssh, and attach to namespace of ansible_host | 16:37 |
| pabelanger | so, I could invert that check for a quick test I suspect | 16:37 |
| pabelanger | ack | 16:37 |
| hwoarang | jmccrory: hello :( --os-endpoint has no effect. the LB cinder endpoint is still being returned | 16:37 |
| evrardjp | in a playbook* not in a plugin | 16:37 |
| hwoarang | so the cinder command simply fails | 16:37 |
| pabelanger | evrardjp: where does the physical_host == ansible_host check happen? | 16:37 |
| pabelanger | can you link code path | 16:37 |
| hwoarang | jmccrory: maybe --os--endpoint needs something else or we also need --os-endpoint-type but so far I couldn't find someone internally to discuss that. and i haven't asked in #openstack-cinder yet. | 16:38 |
| evrardjp | https://github.com/openstack/openstack-ansible-plugins/blob/stable/queens/connection/ssh.py#L94 | 16:38 |
| evrardjp | https://github.com/openstack/openstack-ansible-plugins/blob/stable/queens/connection/ssh.py#L149-L165 | 16:39 |
| evrardjp | https://github.com/openstack/openstack-ansible-plugins/blob/stable/queens/connection/ssh.py#L156-L161 | 16:39 |
| evrardjp | pabelanger: ^ | 16:39 |
| *** flemingo has joined #openstack-ansible | 16:39 | |
| jmccrory | hwoarang hmm other place i know of that OSA does something similar is in keystone with 'ignore_catalog' to our plugin | 16:40 |
| jmccrory | https://github.com/openstack/openstack-ansible-os_keystone/blob/master/tasks/keystone_service_setup.yml#L89 | 16:40 |
| jmccrory | https://github.com/openstack/openstack-ansible-plugins/blob/master/library/keystone#L609-L610 | 16:40 |
| evrardjp | jmccrory: not sure if that can be done from cinder cli :( | 16:41 |
| *** savvas_ has quit IRC | 16:41 | |
| evrardjp | logan-: okay I finally got time to look at your issue. Friday. Like 1 week after the initial conversation ! \o/ | 16:41 |
| pabelanger | evrardjp: Hmm, are you sure? because when I set ansible_host == physical_host, I do see lxc-attach --clear-env --name gear01 command | 16:41 |
| pabelanger | evrardjp: or is it some other process later on that will fail | 16:41 |
| evrardjp | mmm | 16:42 |
| logan- | evrardjp: woot | 16:42 |
| pabelanger | http://paste.openstack.org/show/703051/ | 16:42 |
| evrardjp | pabelanger: mmm maybe we are using container_name and physical_host | 16:42 |
| evrardjp | let me double check | 16:42 |
| pabelanger | I mean, that looks like it is working, if I had lxc setup | 16:43 |
| *** savvas has joined #openstack-ansible | 16:43 | |
| evrardjp | tech debt! | 16:43 |
| hwoarang | jmccrory: the cinder cmdline doesn't offer such option | 16:43 |
| hwoarang | so we are sort of stuck :/ | 16:43 |
| evrardjp | pabelanger: haha my bad | 16:43 |
| evrardjp | we are indeed using container_name now | 16:43 |
| pabelanger | WOOT | 16:43 |
| evrardjp | and container_name is in our inventory | 16:43 |
| pabelanger | okay, so this might actually work :) | 16:43 |
| evrardjp | well, how come? | 16:44 |
| evrardjp | you have to multiply your hosts? | 16:44 |
| jmccrory | hwoarang don't think i've tried anything like this before, taking a look at one of regions here to test things out | 16:44 |
| evrardjp | well I guess it would workd :) | 16:44 |
| *** flemingo has quit IRC | 16:44 | |
| evrardjp | it's very unfriendly though! | 16:44 |
| evrardjp | I guess you should try pabelanger :) | 16:44 |
| evrardjp | logan-: haha | 16:44 |
| pabelanger | I think I just need to add some code to get physical_host={{ ansible_host }} to run via jinja2 parser in the connection_plugin | 16:45 |
| pabelanger | I'll hack on that after some food | 16:45 |
| spotz | evrardjp: You put in for on-boarding and project update for summit right? | 16:45 |
| evrardjp | yes I did | 16:45 |
| evrardjp | want to be in? | 16:45 |
| evrardjp | logan-: ok | 16:45 |
| evrardjp | so | 16:46 |
| evrardjp | found out the issue | 16:46 |
| evrardjp | well | 16:46 |
| evrardjp | not really | 16:46 |
| evrardjp | but the issue is definitely on the git lcone | 16:46 |
| evrardjp | let me dig deeper | 16:46 |
| spotz | evrardjp: If you want me there sure, I was just making sure:) | 16:48 |
| *** admin0 has joined #openstack-ansible | 16:48 | |
| lbragstad | o/ does osa support setting up k2k? or just service provider federation? | 16:49 |
| odyssey4me | lbragstad k2k I think was implemented too | 16:50 |
| lbragstad | cool | 16:50 |
| lbragstad | i have a container i'd like to try and run it against | 16:50 |
| odyssey4me | https://docs.openstack.org/openstack-ansible-os_keystone/latest/configure-federation-sp.html | 16:51 |
| cloudnull | idlemind: ++ i typically spend about 5 min trying to work on a problem within a container, then nuke it and rebuild it | 16:51 |
| jmccrory | hwoarang yeah...seems broken. bypass-url, os-endpoint, and CINDER_ENDPOINT...all still use catalog with the client | 16:51 |
| cloudnull | idlemind: one of the neat features of the zfs, btrfs, machinectl backends is that you can clone containers instantly. so if i have to rebuild one, i just clone from a functional one, then rerun the playbooks to setup that target correctly. | 16:53 |
| *** armaan has quit IRC | 16:53 | |
| idlemind | cloudnull ya i'll have to figure out a monitoring strategy ... or if it's worth it or not | 16:53 |
| idlemind | cloudnull that's pretty sweet actually. i've heard good things about brtfs | 16:53 |
| cloudnull | it works well enough. | 16:54 |
| idlemind | i'm still slummin' in lvm land | 16:54 |
| cloudnull | i've been beating on it at home w/ software raid and several levels of a special kind of hell and it's not failed me yet. | 16:54 |
| idlemind | nice nice | 16:54 |
| *** yolanda has quit IRC | 16:55 | |
| hwoarang | jmccrory: right :/ | 16:55 |
| *** flemingo has joined #openstack-ansible | 16:55 | |
| cloudnull | I even put BTRFS on my laptop as root, running opensuse tumbleweed. so far so good | 16:55 |
| idlemind | i just got some cheapo 8 tb sata drives for my home lab maybe i'll migrate to it when i get around to installing them | 16:55 |
| spotz | hwoarang: jmccrory - I know some stuff shows up in the client help for cinder that isn't actually implemented | 16:55 |
| cloudnull | idlemind: as for monitoring, have a look at https://github.com/openstack/openstack-ansible-ops/tree/master/cluster_metrics | 16:56 |
| cloudnull | or netdata | 16:56 |
| jmccrory | spotz it's strange because it mentions one thing as being deprecated and suggests to use something else, but neither of them work | 16:56 |
| cloudnull | it's more metrics than alert monitoring | 16:56 |
| lbragstad | odyssey4me: is the k2k stuff tested yet, or run in the osa gate? | 16:56 |
| cloudnull | but gives a lot of great insight into the clouds operation | 16:56 |
| idlemind | cool i'll take a peek there | 16:56 |
| *** nattanon has quit IRC | 16:56 | |
| odyssey4me | lbragstad nope, not at all | 16:57 |
| lbragstad | ok - just curious | 16:57 |
| odyssey4me | lbragstad it's not had any feedback or changes since kilo, so it's likely out of date :/ | 16:57 |
| spotz | jmccrory: I had talked to them about --project and it's never been implemented, so I wouldn't be surprised about other stuff | 16:57 |
| idlemind | i've been thinking about my horizon ssl issue ... instead of disabling ssl verification i wonder if it would be better to tell horizon to trust the self-signed certs it should expect to see | 16:57 |
| evrardjp | logan-: env.d/haproxy.yml rename haproxy_container with haproxy | 16:57 |
| lbragstad | odyssey4me: ack | 16:57 |
| odyssey4me | lbragstad most of our federation consumers trickle feedback for keystone as an SP in over time, and that covers their use-case | 16:57 |
| lbragstad | odyssey4me: right - not many people using keystone as an actual identity provider | 16:58 |
| odyssey4me | one would hope not ;) | 16:58 |
| evrardjp | logan-: I will check on the root cause but at least I found out the problem | 16:58 |
| evrardjp | well | 16:58 |
| *** flemingo has quit IRC | 16:59 | |
| evrardjp | logan-: "the dynamic inventory is all broken" | 17:00 |
| logan- | evrardjp: ouch :/ | 17:00 |
| admin0 | idlemind, its also better to just have aSSL cert .. | 17:00 |
| logan- | lol | 17:00 |
| idlemind | admin0 oh i don't disagree | 17:01 |
| evrardjp | logan-: but for your case, it would be enough to solve your "issue" | 17:01 |
| evrardjp | moving to haproxy or haproxy_containers | 17:01 |
| logan- | gotcha | 17:01 |
| admin0 | its like less than 5$ per year | 17:01 |
| evrardjp | instead of haproxy_container | 17:01 |
| idlemind | admin0 but generically; ppl come to osa to "trial" openstack and they up with an install that can't upload images in horizon or connect to consoles (spice/novnc) | 17:01 |
| evrardjp | logan-: could you check on what was done before? | 17:01 |
| evrardjp | just in case it's not a typo | 17:01 |
| idlemind | admin0 so the idea is to create "sensible" defaults that create a more "functional" install with less customization | 17:02 |
| idlemind | if we generate and use self-signed certificates they should be used sensibly in the rest of the environment | 17:02 |
| admin0 | but we do generate a self signed cert as well | 17:02 |
| logan- | evrardjp: pretty sure I copypasta'd that file directly from one that hasnt changed since liberty or mitaka, checking | 17:02 |
| idlemind | admin0 https://bugs.launchpad.net/openstack-ansible/+bug/1747684 | 17:02 |
| openstack | Launchpad bug 1747684 in openstack-ansible "Default Values Do Not Allow Image Uploads to Glance from Horizon" [Medium,Confirmed] | 17:02 |
| idlemind | ^^ that's bug i wrote up on it | 17:02 |
| evrardjp | logan-: this is an extract of my default inventory: http://paste.openstack.org/show/703052/ | 17:03 |
| evrardjp | so I'd say applying stuff to haproxy_containers would make sense | 17:03 |
| evrardjp | no | 17:03 |
| evrardjp | it doesn't | 17:03 |
| evrardjp | scratch what I just said. | 17:03 |
| *** esberglu has quit IRC | 17:04 | |
| logan- | the cinder-volume is applied the same way, but that one's not breaking stuff? | 17:04 |
| idlemind | admin0 so 1 option, disabling cert verification works, a better option to keep cert validation in place would be to teach horizon to trust the certs used in the external endpoints | 17:04 |
| idlemind | only when using the generated self-signed certs | 17:04 |
| idlemind | (me thinks) | 17:04 |
| logan- | also yea confirmed this exact method is what ive been using for haproxy for a long time heh | 17:04 |
| evrardjp | logan-: meh ! | 17:04 |
| *** holser__ has quit IRC | 17:04 | |
| evrardjp | I don't say I have an idea right now of what's going on. | 17:05 |
| admin0 | idlemind, in production you will use real SSL .. for test, can do without SSL as well .. why go to horizon and agian have it trust some arbritary self signed certs | 17:05 |
| evrardjp | but I expect something in our defaults env.d | 17:05 |
| logan- | taking a glance at https://github.com/openstack/openstack-ansible/blob/stable/queens/inventory/env.d/haproxy.yml it looks like 'haproxy_container' is the correct place to apply it | 17:05 |
| logan- | weird | 17:05 |
| evrardjp | just giving you hints on what's the cause and what's an appropriate fix for your env | 17:05 |
| logan- | yeah | 17:05 |
| logan- | really strange | 17:05 |
| idlemind | admin0 because in production i've also seen self-signed certs; look at any customers esx environment | 17:05 |
| logan- | it seems like that fix shouldn't even work | 17:06 |
| idlemind | admin0 esx/vsphere | 17:06 |
| evrardjp | logan-: did you try it ? | 17:06 |
| evrardjp | Just to see if I'm insane or not? | 17:06 |
| evrardjp | :p | 17:06 |
| logan- | nope not yet gonna bootstrap it now | 17:06 |
| admin0 | well, then its not an openstack problem is it ? | 17:06 |
| logan- | and see | 17:06 |
| idlemind | admin0 right; it's about sensible defaults to provide a good customer experience | 17:06 |
| evrardjp | yeah, change your env.d BEFORE generating the inventory | 17:06 |
| admin0 | if someone choses to use self signed in prod, then well.. they will get errors here and there | 17:06 |
| logan- | right | 17:06 |
| evrardjp | or remove your openstack_inventory.json | 17:07 |
| *** udesale has joined #openstack-ansible | 17:07 | |
| *** holser__ has joined #openstack-ansible | 17:08 | |
| *** holser__ has quit IRC | 17:08 | |
| admin0 | idlemind, i agree defaults should be to provide a good experience, and i think the current defaults allows me good usage in both both prod and dev .. you can't have lets use self-signed-ssl but also lets made changes in code so that we don't get errors when using self-signed code | 17:08 |
| admin0 | cert* | 17:08 |
| openstackgerrit | Merged openstack/openstack-ansible master: scripts: scripts-library.sh: Fix typo https://review.openstack.org/552050 | 17:09 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: [Docs] Flatten out maintenance task https://review.openstack.org/553445 | 17:09 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: [Docs] Link to the AIO https://review.openstack.org/553437 | 17:09 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: [Docs] Simplify structure of operations guide https://review.openstack.org/553442 | 17:09 |
| admin0 | well. just my view .. | 17:09 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: [Docs] Update manage networks location https://review.openstack.org/553455 | 17:09 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: [Docs] Add haproxy verification https://review.openstack.org/553456 | 17:09 |
| admin0 | i might be wrong :) | 17:09 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: [Docs] Promote scale the environment https://review.openstack.org/553457 | 17:09 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: [Docs] Promote Backup and Restore https://review.openstack.org/553458 | 17:09 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: [Docs] Move the upgrade reference into reference https://review.openstack.org/553448 | 17:09 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: [Docs] Migrate CLI reference to reference guide https://review.openstack.org/553479 | 17:09 |
| openstackgerrit | Merged openstack/openstack-ansible stable/queens: [Docs] Simplify advanced config https://review.openstack.org/553480 | 17:09 |
| idlemind | admin0 ya i've been investigating it and it seems so far i'm able to tell if a self-signed certificate is being used so i'd image i could teach horizon to inject and trust that certificate | 17:09 |
| admin0 | even if we fix horizon, with self signed certs, curl will complain, api will complain | 17:10 |
| idlemind | admin0 right at that point i trust the person is intelligent enough to know how to trust the cert | 17:10 |
| admin0 | which is beyond us .. so evne if we fix, we are not addressing the issue | 17:10 |
| *** flemingo has joined #openstack-ansible | 17:10 | |
| idlemind | admin0 w/o the horizon fix it's nestled fairly deep in developer tools of chrome to find out that horizon isn't trusting it | 17:10 |
| idlemind | and not the local machine which trusts it as soon as you hit the "accept and continue" button in the browser | 17:11 |
| admin0 | by default logic, it should not right .. to trust self signed certs | 17:11 |
| idlemind | admin0 again agreed there | 17:11 |
| idlemind | admin0 but a fledging enterprise customer might go "meh, gargabe just as i expected, can't even install right" | 17:12 |
| idlemind | because all the user gets is a random unmeaning error message in horizon | 17:12 |
| idlemind | i think it is literally "undefined error" | 17:12 |
| idlemind | but ya i'll keep looking at a fix to propose | 17:13 |
| idlemind | maybe we'll find a happy medium | 17:13 |
| admin0 | so isn't a good fix to fix the documentatin saying if using for demo/test, use SSL, if prod use real SSL certs else these are the issue you face ? | 17:13 |
| evrardjp | logan-: the plot thickens -- reusing your env.d and generating the inventory looks like it produces a valid inventory, with no cycles. | 17:13 |
| admin0 | don't use SSL* | 17:13 |
| admin0 | real enterprize customes will have money and every (throw every known compliance and iso requirment ) to not go prod without a vendor | 17:14 |
| *** flemingo has quit IRC | 17:14 | |
| idlemind | admin0 definitely an option; doc's blurb about it and a few config blurbs to paste in to fix the problem | 17:15 |
| idlemind | which could very well be keep self-signed but disable the cert verification | 17:15 |
| idlemind | which does owrk | 17:15 |
| logan- | evrardjp: maybe https://github.com/continuouspw/openstack-ci-cloud/blob/28a2652bc5abd836dcf19a1ef1e7f6f1d785c321/openstack_deploy/env.d/haproxy.yml#L19 should be "False"? | 17:16 |
| *** udesale has quit IRC | 17:16 | |
| hwoarang | jmccrory: i think i am running out of ideas. i can't even get it to work with the openstack client tool | 17:17 |
| *** pabelanger has quit IRC | 17:17 | |
| *** pabelanger has joined #openstack-ansible | 17:17 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible stable/queens: Correct is_container when deploying containers https://review.openstack.org/553777 | 17:17 |
| jmccrory | hwoarang think i managed to narrow down a bit. https://github.com/openstack/python-cinderclient/blob/master/cinderclient/client.py#L682-L719 | 17:17 |
| jmccrory | it'll only make use of the bypass_url if session is None, still need to figure out how to make that happen though | 17:18 |
| evrardjp | logan-: checking the generated inventory I see that you have haproxy_all: that contains both hosts and children | 17:18 |
| hwoarang | i see | 17:19 |
| evrardjp | logan-: that wouldn't hurt. | 17:19 |
| evrardjp | that's technically what it hsould be | 17:19 |
| evrardjp | let me try | 17:19 |
| logan- | evrardjp: yep thats a standard quirk with dynamic inventory when the host group and service group match names | 17:19 |
| logan- | I just tried it, no luck | 17:19 |
| logan- | testing your fix now | 17:19 |
| evrardjp | logan-: don't work better | 17:20 |
| evrardjp | logan-: my fix isn't a fix | 17:20 |
| logan- | yeah | 17:20 |
| logan- | gotcha | 17:20 |
| evrardjp | your inventory still gonna get busted | 17:20 |
| logan- | it just populates 'haproxy' group with the metal hosts | 17:20 |
| evrardjp | or not doing what you expecte | 17:20 |
| logan- | basically the same as not having that env override | 17:20 |
| evrardjp | what I am telling here is that we have something with haproxy override. | 17:21 |
| evrardjp | Not so sure why, but it looks like the whole thing is wrongly generated | 17:21 |
| logan- | yep agreed | 17:21 |
| logan- | crazy that the haproxy override wrecks it but the cinder one doesn't | 17:21 |
| *** rpittau has quit IRC | 17:22 | |
| odyssey4me | evrardjp logan- is this perhaps a differential in the env.d definitions? | 17:22 |
| evrardjp | logan-: I think it's a problem in the basic env.d provided by OSA | 17:22 |
| evrardjp | and our dynamic inventory | 17:22 |
| evrardjp | odyssey4me: maybe | 17:23 |
| evrardjp | let me check | 17:23 |
| *** ianychoi__ has joined #openstack-ansible | 17:24 | |
| evrardjp | I just think I did shit on the first env.d/haproxy.yaml it was long ago | 17:24 |
| *** admin0 has quit IRC | 17:26 | |
| *** ianychoi_ has quit IRC | 17:27 | |
| evrardjp | I guess it's really time for that inventory migration tool :p | 17:28 |
| *** DanyC has quit IRC | 17:30 | |
| odyssey4me | evrardjp that time was yesterday, two cycles back ;) | 17:30 |
| evrardjp | haha. | 17:31 |
| evrardjp | well... It still can be independant! | 17:31 |
| evrardjp | (I hear the WOOOOOOOOOOO and "did he dare?" ) | 17:32 |
| mnaser | what have i created on twitter | 17:32 |
| * mnaser hides from the foundation | 17:32 | |
| odyssey4me | mnaser what did you do? | 17:32 |
| odyssey4me | lol | 17:32 |
| mnaser | apparently created the "talk rejects club" | 17:32 |
| *** savvas_ has joined #openstack-ansible | 17:33 | |
| mnaser | im sure we can work something out like an evening event but not sure we'd have a lot of time for talks | 17:33 |
| evrardjp | mnaser: lol | 17:33 |
| evrardjp | You are the cause of all of this! | 17:34 |
| evrardjp | hahaha | 17:34 |
| mnaser | cause if we can get some sort of evening event | 17:34 |
| mnaser | it would be cool but i dont think we'd have much more time than like | 17:34 |
| mnaser | 2 talks | 17:34 |
| cloudnull | mnaser: leading a revolution ! | 17:35 |
| cloudnull | :) | 17:35 |
| odyssey4me | I'm out for the w/end folks. See you next week (or maybe on the weekend if I get insomnia, which is likely). | 17:36 |
| *** savvas has quit IRC | 17:36 | |
| mnaser | later | 17:39 |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia master: Fixes mgmt net fact setting https://review.openstack.org/553866 | 17:41 |
| evrardjp | later odyssey4me | 17:43 |
| jmccrory | hwoarang this seems to pretty much force a session unfortunately, https://github.com/openstack/python-cinderclient/blob/master/cinderclient/shell.py#L693-L694 https://github.com/openstack/python-cinderclient/blob/master/cinderclient/shell.py#L889-L912 | 17:44 |
| evrardjp | jmccrory: can we use the openstack client for this? | 17:48 |
| jmccrory | not sure, i'll look into that later today | 17:49 |
| Bico_Fino | Hello, trying to rebuild a container and I'm getting this http://paste.openstack.org/show/703061/ | 17:52 |
| cloudnull | Bico_Fino: include hosts in your limit | 17:53 |
| cloudnull | when I rebuild a container i run the command like so: `openstack-ansible lxc-container-destroy.yml lxc-container-create.yml --limit 'hosts:infra1_galera_container-b350d733'` | 17:54 |
| Bico_Fino | cloudnull I was following https://docs.openstack.org/openstack-ansible/pike/admin/maintenance-tasks/galera.html#galera-cluster-recovery | 17:54 |
| cloudnull | hum. IDK if limit twice works. | 17:55 |
| cloudnull | I think you have to combine it with ":" or a "," | 17:55 |
| Bico_Fino | yeap that worked | 17:56 |
| Bico_Fino | thanks cloudnull | 17:56 |
| *** flemingo has joined #openstack-ansible | 18:02 | |
| hwoarang | jmccrory: i see :( | 18:02 |
| hwoarang | evrardjp: i had a look with the openstack cli but i couldn't get it to work either | 18:02 |
| hwoarang | but maybe i wasn't doing the right thing | 18:02 |
| pabelanger | cloudnull: evrardjp: odyssey4me: so far, it seems only set_fact is the only way physical_host: "{{ ansible_host }}" parses correctly. All other methods of inventory / host / group variable don't work. | 18:06 |
| pabelanger | http://paste.openstack.org/show/703062/ | 18:06 |
| pabelanger | so, I should be able to use that for now, but need to better understand internals of ansible to see where the template parsing is invoked from | 18:06 |
| *** flemingo has quit IRC | 18:07 | |
| cloudnull | its all done in a module called templar (or used to be). | 18:07 |
| pabelanger | k | 18:08 |
| cloudnull | we use the group_vars in our role tests. like so https://github.com/openstack/openstack-ansible-os_keystone/blob/master/tests/group_vars/all_containers.yml#L24 | 18:08 |
| pabelanger | yah, I cannot get that to work with an ansible variable | 18:09 |
| cloudnull | you should also be able to get it to work using a static inventory like so http://paste.openstack.org/show/703064/ | 18:12 |
| *** bhujay has quit IRC | 18:13 | |
| *** flemingo has joined #openstack-ansible | 18:15 | |
| *** rootd00d has joined #openstack-ansible | 18:17 | |
| rootd00d | hey folks! anyone using the ansible-hardening role from openstack? | 18:17 |
| cloudnull | yes. we use it by default | 18:18 |
| cloudnull | if you deploy osa you have that role run on your hosts by default. | 18:18 |
| cloudnull | you can disable that interaction if you want but we recommend running it | 18:18 |
| pabelanger | cloudnull: yah, that does work. I'm going to try creating an lxc container now and interact with it | 18:18 |
| *** flemingo has quit IRC | 18:19 | |
| cloudnull | pabelanger: the same interaction can be done with an nspawn container too | 18:19 |
| cloudnull | not that you want to dive into that quite yet, but the plugin supports both | 18:19 |
| pabelanger | cloudnull: any reason to use that over lxc? Or just the new and shiny? | 18:21 |
| evrardjp | we have a page for that that explain the strength of each | 18:21 |
| evrardjp | or at least we should have! | 18:21 |
| evrardjp | basically one is provided with systemd, so shipped everywhere | 18:21 |
| cloudnull | ^ | 18:22 |
| pabelanger | yah | 18:22 |
| evrardjp | oh no this doc patch hasn't merged yet. | 18:22 |
| evrardjp | it's on my list! | 18:22 |
| cloudnull | its actually not all that new or shinny, its just ubiquitous. | 18:22 |
| cloudnull | the interaction between the two is the same | 18:22 |
| cloudnull | so is the management | 18:22 |
| *** gkadam has quit IRC | 18:23 | |
| cloudnull | IMHO it's nice because we can support it the same on all distros without having to rely on third party repos, packages, etc. | 18:23 |
| cloudnull | I also like all of the built in resource management tools | 18:24 |
| cloudnull | all that said, for osa, lxc is the tried and true method | 18:24 |
| *** savvas_ has quit IRC | 18:24 | |
| pabelanger | Yah, can see that | 18:25 |
| *** savvas has joined #openstack-ansible | 18:25 | |
| cloudnull | in the coming cycles I'd like to make nspawn a first class option. | 18:26 |
| cloudnull | for now i'll settle for it being tested in the gate :) | 18:27 |
| *** flaviosr has quit IRC | 18:28 | |
| *** flemingo has joined #openstack-ansible | 18:30 | |
| *** flemingo has quit IRC | 18:34 | |
| *** deadnull has joined #openstack-ansible | 18:35 | |
| *** flemingo has joined #openstack-ansible | 18:50 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Convert role to use a common systemd role(s) https://review.openstack.org/552047 | 18:51 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Converge distro interfaces to systemd-networkd https://review.openstack.org/552721 | 18:52 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Use production like interfaces in the gate https://review.openstack.org/553636 | 18:52 |
| *** flemingo has quit IRC | 18:54 | |
| *** flemingo has joined #openstack-ansible | 19:00 | |
| *** Sha0000 has joined #openstack-ansible | 19:04 | |
| *** flemingo has quit IRC | 19:04 | |
| *** shardy has quit IRC | 19:11 | |
| idlemind | haproxy_use_keepalived: False wouldn't that be True for a keepalived driven haproxy not false? https://docs.openstack.org/openstack-ansible-haproxy_server/latest/configure-haproxy.html#making-haproxy-highly-available | 19:12 |
| idlemind | or am i just reading the blurb above it incorrectly? | 19:12 |
| *** deadnull has quit IRC | 19:13 | |
| *** chyka has quit IRC | 19:14 | |
| *** flemingo has joined #openstack-ansible | 19:14 | |
| openstackgerrit | Olivier Bourdon proposed openstack/openstack-ansible master: Add more infos into error message https://review.openstack.org/553881 | 19:16 |
| *** flemingo has quit IRC | 19:19 | |
| *** kstev1 has quit IRC | 19:21 | |
| *** Sha0000 has quit IRC | 19:21 | |
| *** ibmko has quit IRC | 19:22 | |
| *** flemingo has joined #openstack-ansible | 19:23 | |
| *** flemingo has quit IRC | 19:27 | |
| evrardjp | logan-: I am sorry I got side tracked to many other things. It will have to wait for next week. | 19:32 |
| logan- | evrardjp: no probs.. you narrowed it down a lot. it seems like its just something where ansible 2.4 doesn't like the inventory json... and since you narrowed it down to that specific part, im hacking on the inventory json some to see if i can figure it out | 19:33 |
| *** ansmith has quit IRC | 19:33 | |
| logan- | and at least you scoped the problem.. I was worried we'd see this coming up for a lot of people trying to deploy queens, but it seems like the impact will be fairly limited now | 19:34 |
| idlemind | evrardjp is the keepalived role that is used by us the same as: https://github.com/evrardjp/ansible-keepalived/ | 19:35 |
| evrardjp | logan-: sorry I could only spend a total of 10 minutes on that | 19:35 |
| evrardjp | :( | 19:35 |
| *** flemingo has joined #openstack-ansible | 19:35 | |
| evrardjp | Yes I am concerned about the test gap logan- | 19:35 |
| evrardjp | idlemind: yes | 19:36 |
| idlemind | cool thanks | 19:36 |
| evrardjp | https://github.com/openstack/openstack-ansible/blob/master/ansible-role-requirements.yml#L31 | 19:36 |
| evrardjp | you can use keepalived role for many things. | 19:36 |
| *** admin0 has joined #openstack-ansible | 19:36 | |
| spotz | evrardjp: portal is fixed | 19:37 |
| *** throwsb1 has quit IRC | 19:38 | |
| evrardjp | spotz: oh great | 19:39 |
| evrardjp | what did we have to do? | 19:39 |
| spotz | evrardjp: Jimmy fixed it, I'm asking if there's anything we can do in the future | 19:40 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible stable/queens: Correct is_container when deploying containers https://review.openstack.org/553777 | 19:40 |
| *** flemingo has quit IRC | 19:40 | |
| *** kstev has joined #openstack-ansible | 19:40 | |
| idlemind | evrardjp thx where / how is it called by osa? doesn't seem to be in -haproxy_server | 19:41 |
| idlemind | (and ya am interested in ipv6 support) | 19:41 |
| evrardjp | idlemind: it's called in the playbook for haproxy | 19:41 |
| evrardjp | idlemind: ipv6 works | 19:41 |
| evrardjp | or at least it used to | 19:41 |
| evrardjp | spotz: great | 19:42 |
| evrardjp | idlemind: https://github.com/openstack/openstack-ansible/blob/master/playbooks/haproxy-install.yml#L32-L33 | 19:42 |
| evrardjp | I am off for today | 19:42 |
| evrardjp | have a good week-end everyone | 19:42 |
| idlemind | thx | 19:42 |
| idlemind | you too | 19:42 |
| spotz | seeya evrardjp | 19:42 |
| evrardjp | thanks spotz on the project-navigator! | 19:42 |
| evrardjp | great job there | 19:43 |
| evrardjp | ! | 19:43 |
| evrardjp | cya ! | 19:43 |
| *** kstev has quit IRC | 19:44 | |
| *** kstev has joined #openstack-ansible | 19:46 | |
| *** flemingo has joined #openstack-ansible | 19:51 | |
| *** flemingo has quit IRC | 19:55 | |
| openstackgerrit | Olivier Bourdon proposed openstack/openstack-ansible master: Add more infos into error message https://review.openstack.org/553881 | 19:58 |
| pabelanger | success, just deployed first lxc container and use the connection_plugin to echo hello work inside it :D | 19:59 |
| cloudnull | pabelanger: nice! | 20:20 |
| cloudnull | next step, conquer the world! | 20:20 |
| *** savvas has quit IRC | 20:23 | |
| *** savvas has joined #openstack-ansible | 20:24 | |
| *** ansmith has joined #openstack-ansible | 20:25 | |
| *** savvas has quit IRC | 20:28 | |
| *** hw_wutianwei1 has joined #openstack-ansible | 20:32 | |
| *** hw_wutianwei has quit IRC | 20:33 | |
| *** hw_wutianwei1 is now known as hw_wutianwei | 20:33 | |
| *** savvas has joined #openstack-ansible | 20:34 | |
| *** flemingo has joined #openstack-ansible | 20:34 | |
| *** chyka has joined #openstack-ansible | 20:35 | |
| *** chyka has quit IRC | 20:36 | |
| *** chyka has joined #openstack-ansible | 20:37 | |
| *** savvas has quit IRC | 20:38 | |
| *** savvas has joined #openstack-ansible | 20:38 | |
| *** flemingo has quit IRC | 20:39 | |
| *** dave-mccowan has quit IRC | 21:01 | |
| *** jwitko has quit IRC | 21:22 | |
| *** sawblade6 has joined #openstack-ansible | 21:28 | |
| *** flemingo has joined #openstack-ansible | 21:29 | |
| *** sawblade_ has quit IRC | 21:30 | |
| *** flemingo has quit IRC | 21:33 | |
| *** ibmko has joined #openstack-ansible | 21:44 | |
| *** savvas has quit IRC | 21:50 | |
| *** savvas has joined #openstack-ansible | 21:51 | |
| *** savvas has quit IRC | 21:58 | |
| *** savvas has joined #openstack-ansible | 22:03 | |
| *** savvas has quit IRC | 22:03 | |
| *** savvas has joined #openstack-ansible | 22:04 | |
| *** savvas has quit IRC | 22:05 | |
| *** savvas has joined #openstack-ansible | 22:05 | |
| *** sawblade6 has quit IRC | 22:07 | |
| *** chyka has quit IRC | 22:07 | |
| *** sep has quit IRC | 22:07 | |
| *** rickflare has quit IRC | 22:07 | |
| *** DimGR has quit IRC | 22:07 | |
| *** coolj has quit IRC | 22:07 | |
| *** olivierbourdon38 has quit IRC | 22:07 | |
| *** sawblade6 has joined #openstack-ansible | 22:07 | |
| *** chyka has joined #openstack-ansible | 22:07 | |
| *** sep has joined #openstack-ansible | 22:07 | |
| *** rickflare has joined #openstack-ansible | 22:07 | |
| *** DimGR has joined #openstack-ansible | 22:07 | |
| *** coolj has joined #openstack-ansible | 22:07 | |
| *** olivierbourdon38 has joined #openstack-ansible | 22:07 | |
| *** savvas has quit IRC | 22:09 | |
| *** chyka has quit IRC | 22:13 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Use hostnamectl to set the container hostname https://review.openstack.org/553911 | 22:20 |
| *** kstev has quit IRC | 22:22 | |
| logan- | cloudnull: testing out the networkd changes in lxc_container_create.. nice work. one question on the eth0 lxc net interface.. ifquery --status shows that ifupdown is managing eth0, and eth0 if setup for dhcp in /etc/network/interfaces. but there is also networkd config present for eth0 and networkd seems to think it is managing it also | 22:24 |
| logan- | same with lo | 22:25 |
| logan- | shouldn't we remove /etc/network/interfaces? | 22:26 |
| logan- | or delete everything from it | 22:26 |
| logan- | networkctl shows eth0 and eth1 stuck in configuring state. when i delete all of the contents of /etc/network/interfaces, now it shows eth0/eth1 as "configured" | 22:27 |
| cloudnull | we could migrate the eth0/1/lo interfaces into networkd, but the idea was to first allow them to coexist on the test hosts. | 22:31 |
| cloudnull | at least as it pertains to the bootstrap-hosts role | 22:32 |
| cloudnull | I think it'd be something worthwhile to work on though to migrate existing network config into networkd. | 22:32 |
| logan- | not looking at the host interfaces though.. im talking about inside the containers | 22:33 |
| logan- | eth1 is not configured in ifupdown anymore | 22:33 |
| *** flemingo has joined #openstack-ansible | 22:33 | |
| logan- | eth0 is configured in both networkd and ifupdown | 22:33 |
| logan- | just not sure if that will cause issues | 22:33 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_hosts master: Truncate all default sys config https://review.openstack.org/553912 | 22:35 |
| cloudnull | ah yes. | 22:36 |
| logan- | it also seems like eth0 being configured in ifupdown prevents BOTH eth0 and eth1 from showing "configured" state in networkctl | 22:36 |
| cloudnull | we should remove those interfaces | 22:36 |
| logan- | they stay in "configuring" forever | 22:36 |
| logan- | gotcha | 22:36 |
| cloudnull | I guess we need to create a lo interface file using networkd in the lxc container create role | 22:37 |
| *** flemingo has quit IRC | 22:37 | |
| cloudnull | and just nuke the ENI files | 22:37 |
| logan- | interestingly even with lo removed from /etc/network/interfaces, ifquery --state still shows it configured lo | 22:38 |
| logan- | so we might not be able to move lo to networkd, but at least eth0 we can get rid of | 22:38 |
| cloudnull | logan-: http://0pointer.de/blog/projects/the-new-configuration-files | 22:41 |
| cloudnull | looks like it's just something that happens | 22:41 |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia master: Adds certificate generation https://review.openstack.org/553630 | 22:42 |
| cloudnull | so we could likely nuke the eni file | 22:42 |
| logan- | cool | 22:42 |
| cloudnull | and any other files in /etc/network/interfaces.d | 22:42 |
| cloudnull | i wonder if it'd all still work with state absent on /etc/network/interfaces.d and /etc/network/interfaces ? | 22:43 |
| *** flemingo has joined #openstack-ansible | 22:45 | |
| *** chyka has joined #openstack-ansible | 22:46 | |
| *** hw_wutianwei has quit IRC | 22:46 | |
| *** hw_wutianwei has joined #openstack-ansible | 22:46 | |
| pabelanger | giving openstack-ansible-lxc_container_create a try locally, but think I am hitting a bug | 22:47 |
| pabelanger | http://paste.openstack.org/show/703091/ | 22:47 |
| pabelanger | using stable/queens branch | 22:47 |
| logan- | yes it does work cloudnull | 22:48 |
| *** admin0 has quit IRC | 22:49 | |
| logan- | pabelanger: have to go afk for 30 mins... but with a delegation issue like that it'd be interesting to see your group vars for the container/physical hosts, and also your inventory | 22:49 |
| *** chyka has quit IRC | 22:49 | |
| pabelanger | sure, I can paste, but I'm just following http://git.openstack.org/cgit/openstack/openstack-ansible-lxc_container_create/tree/examples/playbook.yml for my playbook. I'm likely missing group_vars | 22:50 |
| *** flemingo has quit IRC | 22:50 | |
| pabelanger | oh, I might see the issue | 22:52 |
| pabelanger | properties is a dict, which filters don't work with by default | 22:53 |
| pabelanger | so, adding it to group_vars to see if fixes | 22:53 |
| pabelanger | okay, that fixed it. So, a minor bug. I can push up a patch here later this evening to fix the default filter with dictionaries | 22:59 |
| pabelanger | thanks! | 23:00 |
| *** chyka has joined #openstack-ansible | 23:02 | |
| *** flemingo has joined #openstack-ansible | 23:11 | |
| *** flemingo has quit IRC | 23:17 | |
| *** savvas has joined #openstack-ansible | 23:24 | |
| *** savvas has quit IRC | 23:29 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: common-tasks: Container Start/Stop conditions https://review.openstack.org/553094 | 23:33 |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible master: Increase DIB disk space to 3GB https://review.openstack.org/496832 | 23:34 |
| *** openstackgerrit has quit IRC | 23:48 | |
| *** chyka has quit IRC | 23:49 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
