| *** gyee has quit IRC | 00:01 | |
| *** hjohnson1 has left #openstack-ansible | 00:22 | |
| *** Weifan has quit IRC | 00:41 | |
| *** Weifan has joined #openstack-ansible | 00:42 | |
| *** Weifan has quit IRC | 00:47 | |
| *** hjohnson1 has joined #openstack-ansible | 00:57 | |
| *** hjohnson1 has left #openstack-ansible | 00:58 | |
| *** hwoarang has quit IRC | 01:01 | |
| *** hwoarang has joined #openstack-ansible | 01:03 | |
| *** kplant has joined #openstack-ansible | 01:44 | |
| *** Weifan has joined #openstack-ansible | 02:04 | |
| *** Weifan has quit IRC | 02:07 | |
| *** dasp has joined #openstack-ansible | 02:11 | |
| *** kplant has quit IRC | 02:29 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: [wip] Add support for Fedora 28 https://review.opendev.org/658242 | 02:39 |
|---|---|---|
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: [wip] Add support for Fedora 28 https://review.opendev.org/658242 | 02:41 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: [wip] Add support for Fedora 28 https://review.opendev.org/658242 | 02:43 |
| *** udesale has joined #openstack-ansible | 03:03 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-openstack_hosts master: [redhat] Update variables to be version agnostic https://review.opendev.org/658250 | 03:39 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-repo_server master: [redhat] Install rsync-daemon for newer releases https://review.opendev.org/658251 | 03:49 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-galera_client master: [redhat] Update variables to be version agnostic https://review.opendev.org/658252 | 03:53 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-galera_server master: [redhat] Update variables to be version agnostic https://review.opendev.org/658253 | 03:56 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: [wip] Add support for Fedora 28 https://review.opendev.org/658242 | 03:59 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_keystone master: [redhat] Update variables to be version agnostic https://review.opendev.org/658254 | 04:04 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: [wip] Add support for Fedora 28 https://review.opendev.org/658242 | 04:04 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_placement master: [redhat] Update variables to be version agnostic https://review.opendev.org/658255 | 04:07 |
| *** dasp has quit IRC | 04:08 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_glance master: [redhat] Update variables to be version agnostic https://review.opendev.org/658256 | 04:10 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_cinder master: [redhat] Update variables to be version agnostic https://review.opendev.org/658257 | 04:15 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_cinder master: [redhat] Update variables to be version agnostic https://review.opendev.org/658257 | 04:25 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-repo_server master: [redhat] Install rsync-daemon for newer releases https://review.opendev.org/658251 | 04:25 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: [wip] Add support for Fedora 28 https://review.opendev.org/658242 | 04:28 |
| *** raukadah is now known as chandankumar | 04:28 | |
| *** dasp has joined #openstack-ansible | 04:30 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_nova master: [redhat] Update variables to be version agnostic https://review.opendev.org/658259 | 04:31 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_nova master: [redhat] Update variables to be version agnostic https://review.opendev.org/658259 | 04:37 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_nova master: [redhat] Update variables to be version agnostic https://review.opendev.org/658259 | 04:40 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: [wip] Add support for Fedora 28 https://review.opendev.org/658242 | 04:44 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_neutron master: [redhat] Update variables to be version agnostic https://review.opendev.org/658261 | 04:46 |
| *** shyamb has joined #openstack-ansible | 04:49 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_tempest master: [redhat] Update variables to be version agnostic https://review.opendev.org/658262 | 04:50 |
| *** redrobot has quit IRC | 04:53 | |
| *** dxiri has joined #openstack-ansible | 05:02 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-openstack_hosts master: [redhat] Update variables to be version agnostic https://review.opendev.org/658250 | 05:05 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: [wip] Add support for Fedora 28 https://review.opendev.org/658242 | 05:06 |
| *** dxiri has quit IRC | 05:07 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_placement master: zuul: drop functional jobs https://review.opendev.org/658265 | 05:07 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-os_placement master: [redhat] Update variables to be version agnostic https://review.opendev.org/658255 | 05:07 |
| *** radeks_ has joined #openstack-ansible | 05:12 | |
| *** ivve has quit IRC | 05:14 | |
| *** udesale has quit IRC | 05:44 | |
| *** udesale has joined #openstack-ansible | 05:45 | |
| *** DanyC has joined #openstack-ansible | 05:53 | |
| *** miloa has joined #openstack-ansible | 06:04 | |
| miloa | Morning | 06:04 |
| *** slin has joined #openstack-ansible | 06:14 | |
| *** ivve has joined #openstack-ansible | 06:14 | |
| *** slin has quit IRC | 06:16 | |
| *** slin has joined #openstack-ansible | 06:17 | |
| *** DanyC has quit IRC | 06:18 | |
| *** DanyC has joined #openstack-ansible | 06:19 | |
| *** DanyC has quit IRC | 06:20 | |
| *** kopecmartin|off is now known as kopecmartin | 06:26 | |
| *** shardy has joined #openstack-ansible | 06:27 | |
| fnpanic | goodmorning | 06:35 |
| slin | what is the supportde way to extend deployed aio with features? enable in conf.d, rerun setup-infrastructure + deploy_openstack playbooks? | 06:35 |
| slin | gm | 06:35 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible master: Imported Translations from Zanata https://review.opendev.org/658274 | 06:36 |
| fnpanic | slin: hi | 06:37 |
| fnpanic | i would do it that way | 06:37 |
| fnpanic | not sure about aio if you also need to create containers for the new services | 06:38 |
| slin | hmmm. then setup-hosts as first | 06:38 |
| fnpanic | so also the setup-hosts | 06:38 |
| fnpanic | yeah | 06:38 |
| fnpanic | or just the container-deploy | 06:39 |
| fnpanic | if it is no a service on metal | 06:39 |
| fnpanic | does anyone know why the cpu allocation ratio is not honored in the horizon dashboard admin view? | 06:40 |
| fnpanic | i set it to nova_cpu_allocation_ratio: | 06:40 |
| fnpanic | 4.0 | 06:40 |
| fnpanic | it is correctly appied to nova | 06:40 |
| slin | dunno what is the default, i just wondering on unbound/designate is not part of default aio however the network service sets the router as dns => no dns in the tenant vms | 06:40 |
| fnpanic | but in horizon only the cores are counted | 06:41 |
| fnpanic | the hypervisor only reports 48 (2x12 + HT) | 06:43 |
| slin | fnpanic: cpu overbooking ratio is just an information for the scheduler afaik | 06:43 |
| slin | btw horizon... should i see octavia/designate/whatever config/deployments over horizon? | 06:44 |
| fnpanic | for octavia depends on the driver. you should see the amphore instances | 06:46 |
| fnpanic | but not sure if there is an "admin" panel, only a user panel as i last tried it | 06:46 |
| fnpanic | rest is cli | 06:46 |
| fnpanic | slin: ok, not perfect but why does the nova agent not report the ammount of vcpus left? does not make sense to me | 06:47 |
| *** hwoarang has quit IRC | 06:48 | |
| *** mcarden has quit IRC | 06:48 | |
| *** hwoarang has joined #openstack-ansible | 06:49 | |
| fnpanic | https://bugs.launchpad.net/horizon/+bug/1202965 | 06:50 |
| openstack | Launchpad bug 1202965 in OpenStack cinder-backup charm "The name VCPUs (total) of Hypervisors is confusing" [Undecided,Incomplete] | 06:50 |
| fnpanic | this thing is open forever :-) | 06:52 |
| slin | fnpanic: quality opensource product :) | 06:55 |
| fnpanic | :-P | 06:56 |
| slin | fnpanic: well, let's see :) users will configure via terraform, so the horizon feedback is just about able to check/confirm what they did | 06:56 |
| fnpanic | ;-) | 06:57 |
| fnpanic | it is in the admin view only anyway | 06:58 |
| fnpanic | so nice grafana dashbaord would be suffiecent | 06:58 |
| *** jbadiapa has quit IRC | 06:59 | |
| slin | hmmm... is possible that ubount and designate conflicts? | 07:03 |
| fnpanic | not sure | 07:04 |
| slin | container is started but _infra can't deploy/restart the unbound service - will try to survive withou | 07:07 |
| slin | t | 07:07 |
| *** starborn has joined #openstack-ansible | 07:07 | |
| fnpanic | has anyone a working example for glance cors settings when horizon image upload method is direct? | 07:09 |
| fnpanic | CeeMac: :-) | 07:09 |
| slin | fnpanic: conflicts but with sytemd-resolved... lets try to fix | 07:26 |
| slin | unbound_listen_interface: "{{ ansible_eth1.ipv4.address }}" fixes | 07:29 |
| *** DanyC has joined #openstack-ansible | 07:30 | |
| *** DanyC has quit IRC | 07:35 | |
| *** shyamb has quit IRC | 07:38 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Imported Translations from Zanata https://review.opendev.org/658274 | 07:47 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/stein: Temporarily set bionic job to non-voting https://review.opendev.org/658285 | 07:50 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/stein: Temporarily set bionic job to non-voting https://review.opendev.org/658285 | 07:50 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/stein: Remove install_test_requirements flag https://review.opendev.org/658286 | 07:51 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests stable/stein: Fix clone list for opendev changes https://review.opendev.org/658287 | 07:53 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Set CentOS 7 distro job to non-voting https://review.opendev.org/658289 | 07:57 |
| *** udesale has quit IRC | 07:57 | |
| *** udesale has joined #openstack-ansible | 07:58 | |
| *** udesale has quit IRC | 08:05 | |
| *** udesale has joined #openstack-ansible | 08:08 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests stable/stein: Replace git.openstack.org with opendev.org https://review.opendev.org/658295 | 08:17 |
| *** udesale has quit IRC | 08:22 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Use opendev links https://review.opendev.org/658298 | 08:28 |
| *** shyamb has joined #openstack-ansible | 08:39 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Replace git.openstack.org with opendev.org https://review.opendev.org/658299 | 08:42 |
| odyssey4me | cores - could I get reviews for https://review.opendev.org/658287 please to help get stein passing gates | 08:47 |
| CeeMac | morning fnpanic , sadly I haven't looked at CORS for upload method yet | 08:48 |
| noonedeadpunk | mornings | 08:51 |
| slin | hmmm... | 08:53 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/stein: Temporarily set bionic & centos_distro jobs to non-voting https://review.opendev.org/658285 | 08:54 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/stein: Remove install_test_requirements flag https://review.opendev.org/658286 | 08:54 |
| CeeMac | morning noonedeadpunk | 08:55 |
| noonedeadpunk | odyssey4me: isn't https://review.opendev.org/#/c/658289/1 enough? | 08:56 |
| odyssey4me | noonedeadpunk in master that patch merged before the tempest thing broke | 08:56 |
| odyssey4me | for stein the bionic and centos distro jobs are broken, so we need https://review.opendev.org/658286 to merge before anything will merge in the integrated repo | 08:57 |
| noonedeadpunk | oh, yep, just to ublock things it's good | 08:57 |
| noonedeadpunk | you're right | 08:58 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Set CentOS 7 distro job to non-voting https://review.opendev.org/658289 | 08:58 |
| odyssey4me | updated to reflect the dependency | 08:58 |
| slin | how about "Invalid input for operation: physical_network 'lbaas' unknown for flat provider network." during octavia install? (aio) | 08:58 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/stein: Revert "Temporarily set bionic & centos_distro jobs to non-voting" https://review.opendev.org/658313 | 09:01 |
| CeeMac | slin, does the 'lbaas' provider network exist? | 09:01 |
| CeeMac | slin, did you have the provider network config set up in openstack_user_config.yml? | 09:03 |
| CeeMac | slin, https://docs.openstack.org/openstack-ansible-os_octavia/rocky/configure-octavia.html#setup-a-neutron-network-for-use-by-octavia | 09:04 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Use opendev links https://review.opendev.org/658298 | 09:06 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Replace git.openstack.org with opendev.org https://review.opendev.org/658299 | 09:06 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Re-add the upgrade playbook removing pip.conf https://review.opendev.org/658145 | 09:07 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Ensure that pip/setuptools/wheel python wheels are built https://review.opendev.org/655521 | 09:07 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Make utility install playbook idempotent https://review.opendev.org/655486 | 09:07 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Use hashes for the upgrade playbook markers https://review.opendev.org/658146 | 09:07 |
| odyssey4me | ok, I think everything's got the right deps and are rebased on top of the right patches - now just to wait and see if everything passes | 09:10 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests stable/stein: Replace git.openstack.org with opendev.org https://review.opendev.org/658295 | 09:13 |
| odyssey4me | missed that one, that should do it | 09:15 |
| slin | CeeMac:its the default aio config - will check | 09:15 |
| *** DanyC has joined #openstack-ansible | 09:16 | |
| *** DanyC has quit IRC | 09:17 | |
| *** DanyC has joined #openstack-ansible | 09:17 | |
| slin | CeeMac: can i point to my existing external network with octavia? | 09:24 |
| *** udesale has joined #openstack-ansible | 09:27 | |
| *** SmearedBeard has joined #openstack-ansible | 09:31 | |
| slin | to be more precise, i would like to use the floating ip range with | 09:34 |
| *** vnogin has joined #openstack-ansible | 09:37 | |
| *** aedc has joined #openstack-ansible | 09:39 | |
| *** hamzaachi has joined #openstack-ansible | 09:40 | |
| *** electrofelix has joined #openstack-ansible | 09:47 | |
| CeeMac | slin, you should be able to override the variables for the cidr | 09:49 |
| CeeMac | although the variables all refer to the management network | 09:50 |
| CeeMac | octavia_management_net_subnet_cidr is set to 172.29.232.0/22 by default | 09:50 |
| slin | or br-lbaas + octavia queue is only the control plane? | 09:50 |
| CeeMac | slin, are you talking about attaching the amphora / LB VM to our external network ? | 09:50 |
| CeeMac | *your | 09:51 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Bump requirement for master https://review.opendev.org/656598 | 09:51 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Bump global-requirement-pins before release https://review.opendev.org/656596 | 09:52 |
| CeeMac | slin, default variables are detailed here: https://docs.openstack.org/openstack-ansible-os_octavia/rocky/ | 09:53 |
| slin | CeeMac:ok, i see.... | 09:56 |
| openstackgerrit | Merged openstack/openstack-ansible-tests stable/stein: Fix clone list for opendev changes https://review.opendev.org/658287 | 10:01 |
| *** vnogin has quit IRC | 10:03 | |
| *** Smeared_Beard has joined #openstack-ansible | 10:04 | |
| *** SmearedBeard has quit IRC | 10:06 | |
| *** shyamb has quit IRC | 10:07 | |
| *** shyamb has joined #openstack-ansible | 10:07 | |
| noonedeadpunk | odyssey4me: seems that we either need to disable centos7 or backport https://review.opendev.org/#/c/657386/ as well | 10:07 |
| odyssey4me | noonedeadpunk for which branch? that patch is in stein | 10:08 |
| noonedeadpunk | oh, it was Connection failure, disregard then | 10:09 |
| * noonedeadpunk went to make some coffee | 10:10 | |
| noonedeadpunk | so it doesn't seem like temp fix now... probably it's even worth staying? | 10:12 |
| odyssey4me | yeah, it seems like providing a general option to use aria or use get_url is sensible given that suse also has trouble | 10:13 |
| odyssey4me | or maybe just use get_url to be simpler | 10:13 |
| odyssey4me | it will be slower, but simpler | 10:14 |
| odyssey4me | hmm, looks like https://review.opendev.org/658285 is failing on tempestconf - I need another pair of eyes to figure out why http://logs.openstack.org/85/658285/3/check/openstack-ansible-functional-tempestconf-centos-7/5e5e079/ | 10:15 |
| noonedeadpunk | but your suggestion about retries make sense. get_url is not as reliable as aria (except such packaging stuff, which may occur with anything) | 10:15 |
| odyssey4me | well, if get_url is used then it should not be async and should implement retries on the task | 10:15 |
| noonedeadpunk | don't we need https://review.opendev.org/#/c/656891/ to fix it? | 10:19 |
| odyssey4me | I don't *think* so, because stable/stein of nova still has placement... but there is a chance that our tests are installing master - busy checking that now | 10:20 |
| odyssey4me | yep - all u-c enstries are coming from master here: http://logs.openstack.org/85/658285/3/check/openstack-ansible-functional-tempestconf-centos-7/5e5e079/logs/ara-report/result/f971ebeb-b61c-482e-b0ea-94edb5e329f1/ | 10:21 |
| *** aedc has quit IRC | 10:21 | |
| odyssey4me | ok, I guess we have a bug in the way that's all setup to work... | 10:22 |
| odyssey4me | hmm, ok it seems like it should be right - but isn't somehow | 10:24 |
| odyssey4me | well, I guess for now we can port back the placement pin and use it to unblock the legacy role tests while I try to figure out the issue | 10:25 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests stable/stein: Pin nova_git_install_branch to before placement removal https://review.opendev.org/658327 | 10:27 |
| odyssey4me | cores - https://review.opendev.org/658327 will unblock the legacy role tests for stable/stein, please review asap | 10:29 |
| *** shyamb has quit IRC | 10:30 | |
| *** shyamb has joined #openstack-ansible | 10:38 | |
| evrardjp | odyssey4me: is TESTING_BRANCH env var properly set ? | 10:40 |
| odyssey4me | evrardjp busy checking that - in code it seems right | 10:41 |
| odyssey4me | working it backwards: https://github.com/openstack/openstack-ansible-tests/blob/stable/stein/test-vars.yml#L21 -> https://github.com/openstack/openstack-ansible-os_tempest/blob/stable/stein/run_tests.sh#L42-L47 -> https://github.com/openstack/openstack-ansible-os_tempest/blob/stable/stein/tox.ini#L22 | 10:42 |
| odyssey4me | I'm testing with a VM to validate | 10:42 |
| evrardjp | should be okay for osa-tests and for all the things that releases has branched | 10:43 |
| odyssey4me | I'm guessing that the var isn't being passed through properly - it's just to figure out where | 10:43 |
| evrardjp | do you have a role where you see it failing? | 10:44 |
| odyssey4me | yes, os_tempest stable/stein - as noted earlier, it's clearly using the master branch u-c, not stable/stein | 10:44 |
| odyssey4me | see http://logs.openstack.org/85/658285/3/check/openstack-ansible-functional-tempestconf-centos-7/5e5e079/logs/ara-report/result/f971ebeb-b61c-482e-b0ea-94edb5e329f1/ from https://review.opendev.org/658285 | 10:45 |
| evrardjp | ok let me check that | 10:45 |
| evrardjp | with you I mean | 10:46 |
| odyssey4me | confirmed - TESTING_BRANCH is empty when running the functional test script, so ansible will default it back to master | 10:47 |
| odyssey4me | right, so that means that tox isn't passing it through which is likely a config issue in tox.ini - figuring it out now | 10:47 |
| evrardjp | yes that's where I am at | 10:48 |
| evrardjp | odyssey4me: do you think for those we need to have a passenv from [testenv]passenv ? | 10:48 |
| evrardjp | or something like that? | 10:48 |
| openstackgerrit | melissaml proposed openstack/ansible-role-systemd_networkd master: Replace git.openstack.org URLs with opendev.org URLs https://review.opendev.org/658328 | 10:48 |
| evrardjp | not too familiar with tox | 10:48 |
| odyssey4me | will test that now, although I didn't think so | 10:49 |
| evrardjp | checking the base report, see how it goes | 10:50 |
| odyssey4me | nope, that doesn't work either | 10:50 |
| evrardjp | odyssey4me: same job: http://logs.openstack.org/85/658285/3/check/openstack-ansible-functional-tempestconf-centos-7/5e5e079/ara-report/ | 10:52 |
| evrardjp | sorry | 10:52 |
| evrardjp | http://logs.openstack.org/85/658285/3/check/openstack-ansible-functional-tempestconf-centos-7/5e5e079/ara-report/result/916e8762-c012-4e44-aab6-ed9580d82960/ | 10:52 |
| odyssey4me | evrardjp what about that output should I look at? | 10:52 |
| evrardjp | because you are taking the reverse path, I took the forward path | 10:53 |
| evrardjp | the functional tempestconf centos 7 is running run_tests.sh tempestconf | 10:53 |
| evrardjp | as you can see in this link, the TESTING_BRANCH is correctly set | 10:53 |
| odyssey4me | yes, which invokes tox | 10:53 |
| odyssey4me | it is set correctly in run_tests.sh, but not passed through tox to test-ansible-functional.sh | 10:54 |
| evrardjp | correct | 10:54 |
| evrardjp | We should fail if unset, I thought we had this | 10:55 |
| evrardjp | instead of defaulting to master | 10:55 |
| evrardjp | (side note) | 10:55 |
| *** Smeared_Beard has quit IRC | 10:58 | |
| evrardjp | odyssey4me: could you try to wrap the ANSIBLE_PARAMETERS | 10:58 |
| evrardjp | with double quotes? | 10:58 |
| evrardjp | that confuses the heck out of me | 10:58 |
| evrardjp | in tox.ini tempestconf | 10:58 |
| odyssey4me | oh no, that breaks other things | 10:58 |
| odyssey4me | also, I tried an env without that setting - no difference | 10:58 |
| evrardjp | question | 10:59 |
| evrardjp | does that problem apply to not only tempestconf ? | 10:59 |
| *** SmearedBeard has joined #openstack-ansible | 10:59 | |
| odyssey4me | yes, it seems so given that I made a fresh env and it has the same issue | 10:59 |
| evrardjp | and only for role only testing (the integrated works fine?) | 11:00 |
| evrardjp | oh great we have to think about tempest plugins too | 11:00 |
| evrardjp | now that I am thinking about it | 11:01 |
| odyssey4me | well, I don't think tempest plugins have stable branches... but yes, for stable branches we should switch to using the pypi packages instead of git | 11:02 |
| evrardjp | yes | 11:02 |
| odyssey4me | ok, so that passenv is definitely not working - lemme try and figure out why | 11:03 |
| evrardjp | odyssey4me: I pinged you something to try | 11:06 |
| evrardjp | but I doubt this is it | 11:06 |
| evrardjp | odyssey4me: https://tox.readthedocs.io/en/latest/config.html#conf-passenv | 11:07 |
| evrardjp | alternatively we can use that TOX_TESTENV_PASSENV | 11:08 |
| odyssey4me | well, this doesn't work either: passenv = COMMON_TESTS_PATH HOME TESTING_BRANCH TESTING_HOME USER | 11:08 |
| evrardjp | oh so you mean using spaces isn't working great either? | 11:09 |
| evrardjp | But we used to have this working | 11:09 |
| evrardjp | what changed? | 11:09 |
| evrardjp | tox version? | 11:09 |
| evrardjp | (this is unconstrained iirc) | 11:09 |
| odyssey4me | I don't know if it ever worked | 11:09 |
| evrardjp | role tests were working with branches in the past | 11:10 |
| odyssey4me | yes, but this is the first time we're using the TESTING_BRANCH variable | 11:10 |
| evrardjp | got it | 11:10 |
| odyssey4me | I think we may have tried before, but ended up just hard coding it in test-vars rather than figuring out the source of the issue. | 11:11 |
| evrardjp | could you try what I said? | 11:12 |
| evrardjp | TOX_TESTENV_PASSENV? | 11:12 |
| odyssey4me | trying it now | 11:14 |
| evrardjp | oh I got an idea | 11:15 |
| odyssey4me | COMMON_TESTS_PATH doesn't appear to work either | 11:15 |
| evrardjp | let me check something real quick | 11:16 |
| evrardjp | I think none of them are passed :p | 11:16 |
| odyssey4me | yep, none of them at all | 11:16 |
| *** radeks has joined #openstack-ansible | 11:17 | |
| evrardjp | mmm | 11:21 |
| evrardjp | so my little poc shos that tox.ini is good | 11:21 |
| evrardjp | shows* | 11:22 |
| evrardjp | I mean the env passing is correct | 11:22 |
| evrardjp | where is the common/test-ansible-functional.sh cloned in this case? | 11:23 |
| evrardjp | (sorry to traceback here) | 11:25 |
| evrardjp | it's the run_tests that does it before invoking the tox | 11:25 |
| evrardjp | (sorry to talk out loud) | 11:25 |
| odyssey4me | interesting - even https://stackoverflow.com/a/32252679 isn't working for me | 11:26 |
| *** shyamb has quit IRC | 11:26 | |
| odyssey4me | yeah, so run_tests.sh sets those vars, then sources tests/common/run_tests_common.sh, which then invokes tox | 11:26 |
| odyssey4me | run_tests_common will fail if those vars aren't set | 11:26 |
| odyssey4me | so it would seem that this is somehow broken in the invocation of tox from a script | 11:27 |
| evrardjp | yes, those are set to the correct value, as seen in the log I pasted above | 11:27 |
| odyssey4me | oh, duh - I see it | 11:27 |
| odyssey4me | we aren't exporting the vars | 11:27 |
| evrardjp | so I did change bash -c "env" | 11:27 |
| evrardjp | and I see the vars there | 11:27 |
| *** hamzaachi has quit IRC | 11:28 | |
| evrardjp | oh interesting | 11:28 |
| evrardjp | I did an export indeed :p | 11:28 |
| evrardjp | DUH | 11:28 |
| *** aedc has joined #openstack-ansible | 11:28 | |
| odyssey4me | yep, if we export the vars then it works just fine | 11:28 |
| evrardjp | ofc :p | 11:28 |
| odyssey4me | well, so - to prevent having to do this in run_tests.sh and result in a propogation of the change to all repositories... I think I'll just put it into run_tests_common | 11:29 |
| evrardjp | Yes | 11:29 |
| evrardjp | I would have put it there | 11:29 |
| evrardjp | but it's weird though | 11:29 |
| evrardjp | because the tox call should still have that var | 11:30 |
| evrardjp | can you do a patch directly on the stein branch for PoC? | 11:30 |
| odyssey4me | well, that won't work because depends-on is broken until the current chain of patches is merged | 11:31 |
| evrardjp | I would also totally display those | 11:31 |
| odyssey4me | but yes, I'll work up a patch to make it fail if those aren't present, and to display them | 11:31 |
| evrardjp | but this one is for run_tests.sh though | 11:31 |
| evrardjp | the make it fail if unset | 11:31 |
| odyssey4me | run_tests_common already does fail if it's not set | 11:32 |
| odyssey4me | I'll make test-ansible-functional fail if it's unset | 11:32 |
| evrardjp | k | 11:33 |
| evrardjp | odyssey4me: very weird -- I have a machine now | 11:36 |
| evrardjp | and it tells me the thing is set | 11:36 |
| evrardjp | when doing bash -c "env" in the tox.ini | 11:36 |
| evrardjp | Double DUH | 11:39 |
| evrardjp | just forgot to unset it first. | 11:39 |
| *** samueldmq has joined #openstack-ansible | 11:39 | |
| odyssey4me | well, I just found that we have a duplication of the branch determination and many other things in https://github.com/openstack/openstack-ansible-tests/blob/master/test-ansible-env-prep.sh#L51-L56 | 11:40 |
| evrardjp | yes | 11:40 |
| odyssey4me | it's perhaps just best to make that an export and leave everything else alone, given that this whole mechanism is due to be retired | 11:40 |
| evrardjp | odyssey4me: I believe we should export it in the first place we parse it, which should be in run_tests | 11:41 |
| evrardjp | is it really due to be retired though? | 11:41 |
| odyssey4me | yep | 11:41 |
| evrardjp | how will we do role tests? | 11:41 |
| odyssey4me | once stein is out, that's my next target | 11:41 |
| odyssey4me | all role tests are being replaced with integrated repo-based tests | 11:42 |
| evrardjp | odyssey4me: if you could remove the part where it's set in test-ansible-env-prep.sh it would be great | 11:42 |
| evrardjp | or export it there again | 11:42 |
| *** shyamb has joined #openstack-ansible | 11:42 | |
| *** hjohnson1 has joined #openstack-ansible | 11:43 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests master: Ensure that TESTING_BRANCH is exported https://review.opendev.org/658339 | 11:43 |
| *** kplant has joined #openstack-ansible | 11:43 | |
| evrardjp | odyssey4me: that alone won't fix the world | 11:44 |
| odyssey4me | it should - what gap are you seeing | 11:45 |
| evrardjp | because run_tests sources the run_tests_common with the variable set, but tox can run in a different env where test-ansible-env-prep doesnt' run | 11:45 |
| odyssey4me | well, yes - but TESTING_BRANCH is only used by ansible, and all ansible checks use test-ansible-env-prep | 11:45 |
| evrardjp | lint for example, will not be using the right branch | 11:46 |
| evrardjp | not that it's a big deal | 11:46 |
| odyssey4me | however, you are correct - the best place to export them properly is in run_tests.sh right from the start | 11:46 |
| evrardjp | oh lint is sourcing it | 11:46 |
| evrardjp | so it's good | 11:46 |
| evrardjp | but yeah it's more about a race condition that I am afraid of | 11:46 |
| odyssey4me | so I'll push up a patch now on top of this one to only set them in one place, but we can wait before merging it until we have all our gates working again after the opendev patches and stein is released | 11:47 |
| evrardjp | sounds good to me | 11:47 |
| evrardjp | we should clean up all of those and fail if unset | 11:47 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Ensure that pip/setuptools/wheel python wheels are built https://review.opendev.org/655521 | 11:47 |
| evrardjp | removing all those setting of vars | 11:47 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Re-add the upgrade playbook removing pip.conf https://review.opendev.org/658145 | 11:47 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Make utility install playbook idempotent https://review.opendev.org/655486 | 11:48 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Bump global-requirement-pins before release https://review.opendev.org/656596 | 11:48 |
| evrardjp | noonedeadpunk: could you vote on https://review.opendev.org/#/c/658339/1 ? | 11:48 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/stein: Use hashes for the upgrade playbook markers https://review.opendev.org/658146 | 11:48 |
| evrardjp | it is a "good enough" stop gap | 11:48 |
| odyssey4me | ok, those are rebased on the opendev patches which are passing - so they should pass too and will go into gate once the required patches are through | 11:49 |
| noonedeadpunk | evrardjp: looking | 11:49 |
| *** redrobot has joined #openstack-ansible | 11:50 | |
| noonedeadpunk | yeah.... | 11:50 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests master: Ensure that TESTING_BRANCH is exported https://review.opendev.org/658339 | 11:56 |
| odyssey4me | evrardjp better commit msg? cc noonedeadpunk | 11:56 |
| evrardjp | odyssey4me: wow thanks. I just wrote it in the review for the record, but that's even better | 11:59 |
| evrardjp | noonedeadpunk: could you revote? | 11:59 |
| noonedeadpunk | yep, already did:) | 12:00 |
| noonedeadpunk | just one typo, but I think we can leave it | 12:01 |
| odyssey4me | heh, yeah - sorry :p | 12:01 |
| *** radeks has quit IRC | 12:04 | |
| odyssey4me | ok, I'm taking a break for lunch - bbl | 12:04 |
| *** strattao has joined #openstack-ansible | 12:07 | |
| *** radeks has joined #openstack-ansible | 12:10 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Fix Fwaas Dashboard URL https://review.opendev.org/658341 | 12:12 |
| evrardjp | while we are at fixing things: ^ | 12:12 |
| openstackgerrit | Merged openstack/openstack-ansible-tests stable/stein: Pin nova_git_install_branch to before placement removal https://review.opendev.org/658327 | 12:13 |
| *** nowster has quit IRC | 12:20 | |
| *** nowster has joined #openstack-ansible | 12:24 | |
| evrardjp | noonedeadpunk: odyssey4me: if we bring https://review.opendev.org/#/c/658341/ , then we'll be able to resume releasing for master | 12:29 |
| *** CeeMac_ has joined #openstack-ansible | 12:30 | |
| odyssey4me | we have a few stein patches ready for review.- https://review.opendev.org/658295, https://review.opendev.org/658298, https://review.opendev.org/658299 | 12:31 |
| *** CeeMac has quit IRC | 12:33 | |
| *** CeeMac_ is now known as CeeMac | 12:33 | |
| noonedeadpunk | I think I shouldn't vote on some of them... | 12:34 |
| odyssey4me | noonedeadpunk it's ok as long as you don't own the patch - if it's a cherry-pick from your merged patch to master, then you can verify it by voting | 12:38 |
| noonedeadpunk | odyssey4me: didn't know that | 12:40 |
| fnpanic | hi | 12:44 |
| *** radeks has quit IRC | 12:44 | |
| fnpanic | has anyone a working example for glance cors settings when horizon image upload method is direct? | 12:45 |
| noonedeadpunk | fnpanic: probably I have:0 | 12:49 |
| fnpanic | :-) | 12:49 |
| fnpanic | can you share it? | 12:49 |
| noonedeadpunk | fnpanic: I think it's kinda default one... | 12:50 |
| noonedeadpunk | /etc/openstack_deploy/user_variables.yml:glance_cors_allowed_origin: "{{ (glance_show_multiple_locations | bool) | ternary(openstack_service_publicuri_proto | default('http') + '://' + external_lb_vip_address, None) }}" | 12:50 |
| fnpanic | that is the default from the role | 12:51 |
| noonedeadpunk | however it's working for me... And I'm on 18.1.5 | 12:52 |
| fnpanic | mhhh ok | 12:52 |
| fnpanic | strange | 12:52 |
| *** hamzaachi has joined #openstack-ansible | 12:52 | |
| noonedeadpunk | But not really the default... | 12:53 |
| fnpanic | ? | 12:53 |
| noonedeadpunk | only since https://opendev.org/openstack/openstack-ansible-os_glance/commit/242b0ac2d1c5925d5f6f4870c17934c331580de2 | 12:53 |
| noonedeadpunk | and 18.1.5 is before this patch | 12:54 |
| noonedeadpunk | so I had to overwrite it in user_variables | 12:54 |
| fnpanic | ok. we are still on 18.1.4 | 12:55 |
| fnpanic | that explains it | 12:55 |
| fnpanic | :-) | 12:55 |
| fnpanic | i will give it a try | 12:55 |
| fnpanic | thanks a lot! | 13:03 |
| noonedeadpunk | welcome:) | 13:04 |
| *** shyamb has quit IRC | 13:14 | |
| *** shyamb has joined #openstack-ansible | 13:22 | |
| noonedeadpunk | again http://logs.openstack.org/86/658286/2/check/openstack-ansible-functional-tempestconf-centos-7/c33103c/logs/ara-report/result/0e51e90f-ca0b-4200-a3e7-71e2cd5ba447/ | 13:26 |
| *** vnogin has joined #openstack-ansible | 13:29 | |
| evrardjp | can we reverse proxy those files? | 13:31 |
| evrardjp | reverse proxy cache | 13:31 |
| *** vnogin has quit IRC | 13:31 | |
| odyssey4me | :/ I'd really prefer to just switch to using d-i-b for base images, but haven't had the time to get that done | 13:32 |
| odyssey4me | I did manage to get an ubuntu set done as part of dib | 13:33 |
| odyssey4me | it'd be nice if someone could work on a suse & centos version of these: https://review.opendev.org/#/q/topic:systemd-container-elements+(status:open+OR+status:merged) | 13:34 |
| evrardjp | the job names we have in OSA are currently very inconsistent :p | 13:38 |
| evrardjp | I realise that I know what these do because I am just used to them | 13:38 |
| evrardjp | I will change that when everything will be more stable | 13:39 |
| *** radeks_ has quit IRC | 13:41 | |
| evrardjp | odyssey4me: do you know why, when we moved to use integrated jobs for everything, some roles didn't get that? like the galera_server/galera_client | 13:46 |
| evrardjp | I mean if we really want to tests nothing breaks the integrated, it makes sense to add those there too | 13:46 |
| odyssey4me | evrardjp we started with the openstack service roles first, those are next - but to replace them we need to implement a system for increasing affinity | 13:46 |
| odyssey4me | that's something I'm keen to get to asap, so that we can switch all the roles over | 13:47 |
| evrardjp | oh I see what you mean -- -but it's a different subject | 13:47 |
| evrardjp | I mean testing it's not blocking gates for integrated | 13:47 |
| evrardjp | not that the full coverage of role tests goes into that | 13:47 |
| odyssey4me | currently the rabbitmq/galera_server roles actually test a cluster - if wwe switch to integrated, we actually *decrease* our test coverage | 13:47 |
| evrardjp | which is why I said not switch | 13:48 |
| evrardjp | add jobs | 13:48 |
| evrardjp | I want to make the opensuse experience consistent | 13:48 |
| evrardjp | so I want to add the integrated jobs everywhere, and make that very explicit we work on leap 15 | 13:48 |
| odyssey4me | well sure, that can be done | 13:48 |
| evrardjp | ok | 13:48 |
| evrardjp | let me work on that, and we'll improve coverage later | 13:49 |
| evrardjp | (well it's kinda improving coverage too) | 13:49 |
| evrardjp | (as long as I don't remove jobs) | 13:49 |
| odyssey4me | Seeing as we're now waiting for a bunch of things to pass tests and merge, I'll get to work on my idea to bring in the increased affinity. I have a plan. | 13:49 |
| evrardjp | yes but this can be done on the side I guess | 13:50 |
| *** SmearedBeard has quit IRC | 13:50 | |
| evrardjp | I just want to go through this very quickly | 13:50 |
| odyssey4me | evrardjp mind adding your review here? https://review.opendev.org/#/c/658285/ | 13:51 |
| evrardjp | sure thing | 13:53 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-galera_server master: Add integrated coverage https://review.opendev.org/658368 | 13:53 |
| evrardjp | overriding only for tempest odyssey4me? | 13:54 |
| *** shyamb has quit IRC | 13:54 | |
| odyssey4me | I started using an ansible playbook to create and prep my AIO VM's... I honestly have no idea why I didn't do this before. | 13:54 |
| odyssey4me | evrardjp yep, see there's a revert patch later in that chain | 13:54 |
| evrardjp | odyssey4me: like setting up libvirt and stuff? | 13:54 |
| evrardjp | you should have told me yesterday for throwback thursday | 13:55 |
| odyssey4me | nah, just creating it - then copying over my vimrc, installing/removing packages and checking out the OSA repo for me | 13:55 |
| odyssey4me | the cloud I'm using uses a volume-booted disk, so doing that via CLI is a PITA | 13:56 |
| odyssey4me | with ansible, it's a single task :) | 13:56 |
| odyssey4me | I did try and build a volume yesterday using btrfs as a base, so I could stap and revert... but converting it didn't work, so I'll have to try another strategy | 13:56 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-haproxy_server master: Add integrated coverage https://review.opendev.org/658373 | 13:58 |
| *** vnogin has joined #openstack-ansible | 13:59 | |
| *** ianychoi_ is now known as ianychoi | 14:00 | |
| odyssey4me | evrardjp perhaps good to use the topic 'use-integrated-repo-for-all-tests' to follow on from the previous changes for that purpose | 14:00 |
| odyssey4me | https://review.opendev.org/#/q/topic:use-integrated-repo-for-all-tests+(status:open+OR+status:merged) | 14:00 |
| evrardjp | sounds good | 14:00 |
| odyssey4me | ok, brb - relocating | 14:01 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-openstack_hosts master: Add integrated coverage https://review.opendev.org/658374 | 14:01 |
| *** SmearedBeard has joined #openstack-ansible | 14:05 | |
| *** aedc has quit IRC | 14:06 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_aodh master: Add integrated coverage https://review.opendev.org/658376 | 14:08 |
| evrardjp | odyssey4me: ^ that might look weird, but please bare with me :) | 14:12 |
| evrardjp | I just want consistent experience | 14:12 |
| noonedeadpunk | Don't we get it with https://opendev.org/openstack/openstack-ansible/src/branch/master/zuul.d/project-templates.yaml ? | 14:13 |
| evrardjp | correct I am using those templates | 14:13 |
| evrardjp | I want to use a template for all the distro/distro+pkgs ones | 14:14 |
| evrardjp | everywhere | 14:14 |
| evrardjp | then we can adapt to be smarter | 14:14 |
| evrardjp | sorry | 14:14 |
| openstackgerrit | Antony Messerli proposed openstack/openstack-ansible-os_keystone stable/stein: Remove keystone directory on source installs https://review.opendev.org/658377 | 14:14 |
| *** aedc has joined #openstack-ansible | 14:14 | |
| evrardjp | I meant metal // metal + distro pkgs | 14:14 |
| slin | CeeMac: that part fixed, next part is octavia does not trust haproxy cert :) | 14:15 |
| *** strattao has quit IRC | 14:16 | |
| CeeMac | slin, great news. progress is good! | 14:18 |
| CeeMac | however i'm not sure what to do about the untruster cert | 14:18 |
| noonedeadpunk | yeah ok. I don't see the full picture right now, but will be looking at the progress | 14:18 |
| CeeMac | you probably want to speak with jrosser, hes been doing a lot with the self-signed certs iirc | 14:19 |
| slin | could be :) | 14:20 |
| *** strattao has joined #openstack-ansible | 14:20 | |
| slin | "TASK [os_octavia : Setup the service project quota] *****************************************************************************************************************************************" | 14:20 |
| slin | fatal: [aio1_octavia_server_container-37535d12 -> localhost]: FAILED! => {"changed": false, "extra_data": null, "msg": "SSL exception connecting to https://100.67.0.245:5000/tenants: HTTPSConnectionPool(host='100.67.0.245', port=5000): Max retries exceeded with url: /tenants (Caused by SSLError(SSLError(\"bad handshake: Error([('SSL routines', 'tls_process_server_certificate', 'certificate verify | 14:20 |
| slin | failed')],)\",),))"} | 14:20 |
| slin | (still aio) | 14:20 |
| *** SmearedBeard has quit IRC | 14:21 | |
| *** SmearedBeard has joined #openstack-ansible | 14:24 | |
| evrardjp | noonedeadpunk: yes I guess the idea is that I want to simplify things and have a consistent interface | 14:29 |
| evrardjp | there are jobs that might impact a few roles, not only their own role | 14:29 |
| evrardjp | in those cases, you still want a single job to be testing across the board to validate that scenario | 14:30 |
| evrardjp | in this case, why not adding it to a single template, instead of multiplying the templates, and you then have to apply to every single role | 14:31 |
| evrardjp | you can envision this as "official" testing | 14:31 |
| noonedeadpunk | but actually telemetry scenario covers aodh, ceilometer, panko and gnocchi. | 14:32 |
| evrardjp | good example | 14:32 |
| evrardjp | it's a single job, right? | 14:32 |
| noonedeadpunk | yep | 14:32 |
| evrardjp | why not making this use case part of the current testing? | 14:33 |
| evrardjp | (and make sure all the roles are inline with this) | 14:33 |
| evrardjp | else there is a role that can change and break your use case | 14:33 |
| noonedeadpunk | it's part - openstack-ansible-deploy-aio_telemetry_metal-jobs | 14:33 |
| evrardjp | I understand that | 14:34 |
| evrardjp | see what I just said above :) | 14:34 |
| odyssey4me | evrardjp if we end up adding all uses cases everywhere, we're just wasting test resources | 14:34 |
| evrardjp | I was told it's not that bad. | 14:35 |
| odyssey4me | a change in os_aodh does nothing to an aio job, so all we're doing is increasing the chance of unrelated failures due to a bad node | 14:35 |
| noonedeadpunk | so, you mean to add telemetry scenario to every build of every role? | 14:35 |
| evrardjp | noonedeadpunk: correct | 14:35 |
| evrardjp | odyssey4me: when did we really have a bad node? | 14:36 |
| odyssey4me | evrardjp we get them all the time | 14:36 |
| evrardjp | I mean think about the consistency here, it's very interesting for new devs | 14:36 |
| noonedeadpunk | so yeah, than test resources is also smth that worries me... and time needed for tests to finish | 14:36 |
| evrardjp | here I can't point to someone and say: Have a look at this job, it's the one that matters to you | 14:37 |
| evrardjp | noonedeadpunk: when we decided to go for the integrated, we waved this | 14:37 |
| odyssey4me | I don't think that consistency to that degree is important. It's over the top IMO. | 14:37 |
| evrardjp | noonedeadpunk: this is why I made, in the past, those jobs experimental instead of voting | 14:37 |
| evrardjp | odyssey4me: for me it's unnecessary. | 14:38 |
| odyssey4me | right now the aio tests the basic use case, and any repo adds the applicable service(es) to make that service work on top of the basic use-case | 14:38 |
| evrardjp | for other ppl in my company that's different | 14:38 |
| evrardjp | odyssey4me: I agree with that point | 14:39 |
| evrardjp | I just want the basic use case to be covered everywhere | 14:39 |
| evrardjp | and we'll carefully increase those basic use cases | 14:39 |
| odyssey4me | evrardjp and it is - the infra roles still need it, but the roles like os_horizon and the telemetry roles already do it - even if the job name is different | 14:39 |
| evrardjp | if telemetry is NOT a basic use case, then it deserves a different project template, and then added to the roles appropriately | 14:40 |
| CeeMac | slin, I can see where the octavia role is generating certs for its own use, but can't see where you could specify to trust the haproxy certs that are generated | 14:40 |
| evrardjp | odyssey4me: that's my point -- we can't point those to the people | 14:40 |
| evrardjp | if we don't agree on that, I better should not edit those, and ignore all those roles for suse support | 14:41 |
| odyssey4me | evrardjp I guess I'm missing the point then - the jobs have been defined with a particular pattern to make it intuitive to understand what it tests. | 14:41 |
| evrardjp | so we become opinionated | 14:41 |
| evrardjp | I am fine with that too | 14:41 |
| *** ivve has quit IRC | 14:41 | |
| evrardjp | I think the problem is that I got mixed messages -- either we want SUSE support to ALL roles that are integrated, and then for that I would like to have simple jobs everywhere | 14:42 |
| odyssey4me | anything using 'aio' in the name, tests the basic use-case - anything with 'metal' tests without container, etc | 14:42 |
| evrardjp | odyssey4me: It's barely consistent everywhere | 14:42 |
| evrardjp | why are some not having aio_ in the name | 14:42 |
| odyssey4me | evrardjp all os_roles are tested using that for functional tests | 14:43 |
| odyssey4me | those with no aio_ in the name are not testing glance/cinder/nova/neutron because they don't need to | 14:43 |
| evrardjp | let me rephrase the problem. I want to point ppl to 1 single job to monitor for SUSE. Then increase coverage of said jobs. | 14:43 |
| evrardjp | here there are blind spots. | 14:44 |
| evrardjp | AND | 14:44 |
| evrardjp | mnaser: asked to add support for every single role in osa | 14:44 |
| evrardjp | so we have conflicting requirements here | 14:45 |
| odyssey4me | evrardjp well, that's not how these jobs are designed - using a regex would be better: ^openstack-ansible-.*-opensuse.*$ | 14:45 |
| evrardjp | or we make jobs per distro | 14:45 |
| evrardjp | so let's take a step back | 14:47 |
| *** aludwar has joined #openstack-ansible | 14:48 | |
| openstackgerrit | Antony Messerli proposed openstack/openstack-ansible-ops master: Skip neutron rebalance if ha-tool is disabled https://review.opendev.org/652205 | 14:48 |
| odyssey4me | potentially another way to do this would be to make zuul introspect the additional services based on the repo being tested | 14:48 |
| evrardjp | I see the integrated jobs as user stories | 14:48 |
| odyssey4me | for example, if the repo is aodh - then add '_telemetry' to the scenario env var | 14:49 |
| evrardjp | that sounds nice and scalable | 14:49 |
| odyssey4me | then we could have a single job definition, and zuul does the magic - but it's still easy enough to test manually | 14:50 |
| evrardjp | so basically somewhere in the integrated we have a map role:scenario | 14:50 |
| odyssey4me | well, lemme take a look - I think it could be quite simple to do | 14:51 |
| evrardjp | please note this can be an extra job or injected jobs | 14:51 |
| odyssey4me | I don't think it needs to be | 14:51 |
| evrardjp | what I mean by that is that we can say, if something spawn sub jobs | 14:51 |
| odyssey4me | this might also be a nice way to have periodics for all OSA os_ repositories to continually test them | 14:52 |
| odyssey4me | spawn sub jobs? sounds more complicated than I'm thinking | 14:52 |
| evrardjp | it's complementary to your thinking :) | 14:52 |
| evrardjp | your idea tells what to run, my idea on top then how to run | 14:53 |
| evrardjp | for example: | 14:53 |
| evrardjp | (base jobs + extra_role_job) are part of a template | 14:53 |
| odyssey4me | is there a zuul var for the repo that initiated the change? | 14:53 |
| evrardjp | yes | 14:53 |
| evrardjp | which one, I can't say | 14:54 |
| odyssey4me | lol, ok - I'll look in the docs | 14:54 |
| evrardjp | but let me finish my idea, to tell me if I am weird or not :) | 14:54 |
| evrardjp | (no whiteboard, damn :p ) | 14:54 |
| evrardjp | so basically you are saying if <repo initiating change> == <repo in scenarioA>, then add scenarioA to run, skip base jobs. | 14:55 |
| evrardjp | s/you are saying/I am saying to zuul/ | 14:55 |
| odyssey4me | not quite | 14:55 |
| evrardjp | because adding scenarioA to run basically covers base | 14:56 |
| evrardjp | (not 100% true, but you get the idea) | 14:56 |
| odyssey4me | I'm saying that for the aio job, check the repo name for the job originating the change and, if applicable, add the extra service name to the scenario variable | 14:56 |
| evrardjp | yes, I agree on that | 14:56 |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: Ensure that TESTING_BRANCH is exported https://review.opendev.org/658339 | 14:57 |
| evrardjp | I just thought this deserved a different jobname | 14:57 |
| odyssey4me | we already have the special name logic here: https://opendev.org/openstack/openstack-ansible/src/branch/master/tests/roles/bootstrap-host/vars/main.yml#L28-L54 | 14:57 |
| evrardjp | and that this jobname would be able to skip jobs where things already run | 14:57 |
| evrardjp | I think I am not clear :) | 14:58 |
| evrardjp | my idea totally needs your idea to work :p | 14:58 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests stable/stein: Ensure that TESTING_BRANCH is exported https://review.opendev.org/658385 | 15:00 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests stable/stein: Revert "Pin nova_git_install_branch to before placement removal" https://review.opendev.org/658386 | 15:00 |
| *** hjohnson1 has quit IRC | 15:00 | |
| *** dxiri has joined #openstack-ansible | 15:00 | |
| evrardjp | odyssey4me: https://zuul-ci.org/docs/zuul/user/jobs.html?highlight=zuul_return#skipping-child-jobs is what I meant | 15:01 |
| *** DanyC_ has joined #openstack-ansible | 15:01 | |
| evrardjp | so basically, if you are from a certain repo, don't run the default integrated | 15:02 |
| evrardjp | (that is, if we want to have different jobs names) | 15:03 |
| evrardjp | I would still prefer a single job name | 15:03 |
| odyssey4me | yep, lemme do a quick poc for havig a single job name that introspects based on the job's originating repo | 15:03 |
| jrosser | slin: i think that the problem you are seeing is with this task https://github.com/openstack/openstack-ansible-os_octavia/blob/ea21d76875336d2296ac0f667a70689488efc763/tasks/octavia_security_group.yml#L27-L38 | 15:03 |
| jrosser | slin: it should not be using the external endpoint to do the service setup imho so this is likley a bug | 15:04 |
| odyssey4me | looks like I can use https://zuul-ci.org/docs/zuul/user/jobs.html#var-zuul.project.short_name | 15:04 |
| *** DanyC has quit IRC | 15:04 | |
| jrosser | slin: see that the next task specifies "interface: admin" but that one that fails for you does not specify and interface, and the default is 'public' | 15:05 |
| evrardjp | odyssey4me: yes that's the one I was thinking about! | 15:06 |
| slin | jrosser: will change the task locally, let's see | 15:06 |
| evrardjp | question is, should I continue with a standard approach? | 15:07 |
| evrardjp | I have the impression it's better and simpler | 15:07 |
| slin | jrosser: confirmed, changing the role fixed the issue on this level | 15:14 |
| jrosser | slin: it would be great if you could submit a patch, or create a bug on launchpad for that, then we can make sure it gets fixed | 15:16 |
| slin | jrosser: will do | 15:17 |
| *** gyee has joined #openstack-ansible | 15:17 | |
| *** samueldmq has quit IRC | 15:20 | |
| *** starborn has quit IRC | 15:26 | |
| evrardjp | continuing the crusade of the uniform jobs | 15:27 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_barbican master: Add integrated coverage https://review.opendev.org/658396 | 15:27 |
| slin | jrosser: #1828599 | 15:28 |
| *** hamzaachi has quit IRC | 15:28 | |
| *** hamzaachi has joined #openstack-ansible | 15:29 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Add services for testing based on test repo name https://review.opendev.org/658398 | 15:29 |
| odyssey4me | evrardjp ^ let's see if that works | 15:29 |
| odyssey4me | it won't work for everything just yet - just the os_ roles for now... but it's a start | 15:30 |
| odyssey4me | lemme push up a test patch in another role to test it out - unless I can change one of yours? | 15:30 |
| evrardjp | wow everything in one line :) | 15:31 |
| evrardjp | you can change one of mine if you like | 15:31 |
| *** macza has joined #openstack-ansible | 15:31 | |
| odyssey4me | hmm, let me refine that a bit | 15:32 |
| openstackgerrit | Merged openstack/openstack-ansible master: Bump requirement for master https://review.opendev.org/656598 | 15:34 |
| openstackgerrit | Merged openstack/openstack-ansible-os_tempest stable/stein: Temporarily set bionic & centos_distro jobs to non-voting https://review.opendev.org/658285 | 15:34 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Add services for testing based on test repo name https://review.opendev.org/658398 | 15:37 |
| evrardjp | logan-: hahaha | 15:37 |
| evrardjp | we totally think alike :) | 15:38 |
| *** BjoernT has joined #openstack-ansible | 15:40 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Add services for testing based on test repo name https://review.opendev.org/658398 | 15:41 |
| *** BjoernT has quit IRC | 15:43 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_blazar master: Add integrated coverage https://review.opendev.org/658402 | 15:46 |
| *** ivve has joined #openstack-ansible | 15:47 | |
| odyssey4me | evrardjp FYI, we switched to 'all' in meta to reduce unnecessary code churn - the vars file now longer changes, so if a new version comes out, we may need to change nothing for a new version to work | 15:47 |
| *** sreejithp has joined #openstack-ansible | 15:50 | |
| *** sreejithp has quit IRC | 15:51 | |
| *** sreejithp has joined #openstack-ansible | 15:51 | |
| odyssey4me | hmm, I wonder just how much we can reduce the job definitions by doing some more dynamic introspection :) | 15:52 |
| *** hjohnson1 has joined #openstack-ansible | 15:53 | |
| *** hamzaachi has quit IRC | 15:55 | |
| logan- | evrardjp: lol | 15:56 |
| evrardjp | odyssey4me: reducing everything to one job | 15:57 |
| evrardjp | Launch all the "you had one job" memes now | 15:57 |
| odyssey4me | evrardjp I don't see how that's possible, given that we have to specify a node type in a job | 15:57 |
| evrardjp | odyssey4me: that was a joke? | 15:57 |
| odyssey4me | ah yes | 15:58 |
| evrardjp | sorry, I mean: that was a joke! ?! | 15:58 |
| odyssey4me | I keep thinking I should make a job generator - specify a matrix you want, then generate the zuul job yaml file | 15:58 |
| evrardjp | it's harder to maintain than saying "Run with these parameters, and we'll test your use case" | 15:59 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_aodh master: [TEST] Use the dynamic aio scenario https://review.opendev.org/658405 | 15:59 |
| evrardjp | and for the meta/ -all I am now definitely against, because it breaks expectations (and some code I have) :p | 16:00 |
| evrardjp | odyssey4me: where do you set dynamic? | 16:00 |
| evrardjp | I think you forgot to add that to the zuul scenario | 16:01 |
| logan- | odyssey4me: yeah the dynamic job pattern is really handy. your setup there is similar to what i do with DIB builds: http://paste.openstack.org/show/751232/ :) it ends up being pretty similar to what we used to have with zuulv2 job templates | 16:01 |
| evrardjp | zuul job* | 16:01 |
| evrardjp | logan-: I see :) | 16:01 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_barbican master: [TEST] Use the dynamic aio scenario https://review.opendev.org/658406 | 16:02 |
| *** chandankumar is now known as raukadah | 16:02 | |
| odyssey4me | evrardjp what do you mean? it just uses the standard aio job template | 16:02 |
| evrardjp | the template doesn't have dynamic_scenario defined | 16:03 |
| evrardjp | does it? | 16:03 |
| odyssey4me | it doesn't need it | 16:03 |
| odyssey4me | dynamic_scenario can be overridden in any job, but it's assumed to be on | 16:04 |
| evrardjp | it's not (True) | 16:04 |
| odyssey4me | oops | 16:04 |
| evrardjp | :) | 16:04 |
| evrardjp | (dynamic_scenario | default(True)) | bool | 16:04 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Add services for testing based on test repo name https://review.opendev.org/658398 | 16:04 |
| evrardjp | should do | 16:04 |
| odyssey4me | I have *no idea* what I was thinking with negating that. | 16:05 |
| odyssey4me | friday afternoon :p | 16:05 |
| odyssey4me | logan- in your job definition, don't you need to specify the node type? | 16:06 |
| logan- | all of those run on a bionic node | 16:06 |
| logan- | so no its not dynamic that way | 16:07 |
| odyssey4me | oh I see - you're just giving it the target image name | 16:07 |
| evrardjp | it's about what we do, when we think about it :) | 16:07 |
| logan- | yeah but i think you could build your scenario string this way sort of | 16:07 |
| evrardjp | we just have legacy things in bash scripts because we have "user interfaces" | 16:07 |
| logan- | i haven't read your patches yet i was on the phone when i wrote that comment but figured its similar enough i'd share | 16:07 |
| logan- | yep | 16:08 |
| odyssey4me | yeah, that is similar to what I'm doing | 16:08 |
| openstackgerrit | Nicolas Bock proposed openstack/openstack-ansible-lxc_container_create master: Update LXC container config options https://review.opendev.org/657785 | 16:12 |
| *** udesale has quit IRC | 16:12 | |
| *** udesale has joined #openstack-ansible | 16:12 | |
| *** vnogin has quit IRC | 16:13 | |
| *** SmearedBeard has quit IRC | 16:16 | |
| *** strattao has quit IRC | 16:17 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: [WIP] Add services for testing based on test repo name https://review.opendev.org/658398 | 16:18 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: [wip] Add support for Fedora 28 https://review.opendev.org/658242 | 16:21 |
| mnaser | odyssey4me: https://review.opendev.org/#/c/658251/ do you have an idea how I can clean up this failure? | 16:22 |
| mnaser | it looks like it tries to install "an empty string" :X | 16:22 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: [WIP] Add services for testing based on test repo name https://review.opendev.org/658398 | 16:24 |
| mnaser | No package matching '' found available, installed or updated | 16:25 |
| odyssey4me | mnaser yep, we have a pattern to solve that one - although the fix should really be made in the ansible module upstream | 16:25 |
| odyssey4me | lemme find the workaround | 16:25 |
| logan- | mnaser: under the repo server install distro packages task | 16:25 |
| logan- | use this: name: "{{ repo_server_distro_packages | reject('equalto', '') }}" | 16:25 |
| mnaser | odyssey4me: ^ I guess thats it | 16:25 |
| odyssey4me | yep, that's it | 16:25 |
| mnaser | shenanigans | 16:25 |
| odyssey4me | it's only yum that fails on an empty item | 16:26 |
| odyssey4me | apt and zypper don't quote the package names, but for some reason the yum module does | 16:26 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-repo_server master: [redhat] Install rsync-daemon for newer releases https://review.opendev.org/658251 | 16:26 |
| mnaser | so just fyi fedora 28 is the base for RHEL 8 | 16:26 |
| mnaser | so by testing against it we can be _kinda_ sure that we support CentOS 8 when its out | 16:27 |
| logan- | hmm i think older versions of ansible used to have that problem with apt. i know i have idiotic workarounds in my user_variables template for like galera server package filtering where its like "{% if whatever %}package{% else %}bash{% endif %}" | 16:27 |
| logan- | or galera client package filtering rather | 16:27 |
| mnaser | btw, thoughts on dropping security hardening from our playbooks? | 16:27 |
| *** aedc has quit IRC | 16:27 | |
| *** aedc has joined #openstack-ansible | 16:28 | |
| mnaser | (we can move it to a user story "how to harden meh machines" | 16:28 |
| mnaser | given that no one is maintaining Ansible-hardening and it's a bit out of scope | 16:28 |
| *** slin has quit IRC | 16:28 | |
| logan- | i agree but i actually like ansible-hardening and i feel like it would reduce the test coverage a lot. also its a one liner to disable it | 16:28 |
| odyssey4me | could do - and could refer them to ansible-lockdown as a maintained alternative | 16:29 |
| odyssey4me | we could perhaps just deprecate it now, and indicate that we will remove it in U | 16:29 |
| logan- | maybe we could turn it off by default and have a separate job that leaves it enabled? | 16:29 |
| logan- | or disable by default and leave it enabled in the gate? | 16:29 |
| mnaser | logan-: yeah.. but it is also very invasive too in some way | 16:29 |
| odyssey4me | bbiab | 16:29 |
| mnaser | like you want to deploy openstack and then boom all of a sudden your sshd files are changed | 16:30 |
| mnaser | things are disabled/enabled | 16:30 |
| mnaser | https://ansiblelockdown.io | 16:30 |
| mnaser | this is interesting | 16:30 |
| mnaser | so my goal is to simplify our stuff much more and be far less opinionated in how we deploy boxes, leaving the hardening/etc as an exercise for the user | 16:31 |
| evrardjp | mnaser: I agree | 16:31 |
| evrardjp | if we don't have maintainers anymore, we should drop it | 16:31 |
| evrardjp | with a cycle for deprecation | 16:32 |
| mnaser | like I think as the project originated in something like rax where they needed tooling for hardening, sure | 16:32 |
| evrardjp | however, I believe to really know this, we should ask the questions on the ML, and point to ansiblelockdown | 16:32 |
| mnaser | but as the project goes 'mainstream' .. our focus goes onto deploying openstack | 16:32 |
| evrardjp | mnaser: it is still use by more people than just osa | 16:32 |
| evrardjp | used* | 16:32 |
| mnaser | because I think to say that we provide adequate hardening in OSA is a promise we can't necessarily deliver on | 16:32 |
| evrardjp | recently some guy from fujitsu asked me about it | 16:32 |
| *** aedc has quit IRC | 16:33 | |
| evrardjp | mnaser: correct | 16:33 |
| mnaser | I'm totally okay with keeping the role active and around. I'm just not sure if it has a place inside our integrated stuff | 16:33 |
| logan- | mnaser: i don't think we're saying that. imo what we're saying is OSA doesn't break when you use ansible-hardening because we test that. | 16:33 |
| evrardjp | logan-: it's under our governance | 16:33 |
| evrardjp | if it was outside governance, I would totally agree | 16:34 |
| evrardjp | but at the same time, we say in our docs its a community project | 16:34 |
| noonedeadpunk | btw, ansiblelockdown doesn't cover ubuntu and suse at all | 16:34 |
| evrardjp | yes, lockdown is not as good as hardening in terms of OS coverage, but I never got the chance to merge the communities | 16:35 |
| mnaser | I dunno, running ansible-hardening on all of our jobs *but* not providing it by default seems like a bit of a flaw. maybe running integrated jobs with hardening inside the role only rather than all of our jobs | 16:35 |
| evrardjp | we can also not run it at all in OSA, and do it on the side with molecule | 16:35 |
| mnaser | id be okay with running Ansible-hardening manually *then* run integrated in the ansible-hardening role so we can test if any changes break | 16:36 |
| evrardjp | so basically hardening and config_template are more like libraries we use | 16:36 |
| mnaser | evrardjp: right but logan- seems to want to make sure ansible-hardening + OSA work together | 16:36 |
| logan- | we test lots of features that are compatible but opt-in, ie metal, proxy, ceph | 16:36 |
| noonedeadpunk | actually I also like hardending, as it provides at least some basic stuff whic is must-have tbh for each system.. | 16:36 |
| evrardjp | logan-: that's true too | 16:36 |
| evrardjp | mnaser: it seems we already have a few ppl out there using it, why deprecating it? :p | 16:37 |
| *** DanyC_ has quit IRC | 16:37 | |
| mnaser | id be okay with a single job then. I'm aiming to go towards something that's not very invasive | 16:37 |
| evrardjp | just ask on the ML for help on the next stig, and if nobody wants to do it, we'll say we are feature complete, and won't handle the new cases? | 16:37 |
| *** DanyC has joined #openstack-ansible | 16:37 | |
| mnaser | aka "I wanted openstack but I ended up with a modified sshd config" | 16:37 |
| mnaser | "and now kerberos is disabled" | 16:38 |
| evrardjp | I believe it should be opt-in too | 16:38 |
| mnaser | "and ntp was replaced by crony" | 16:38 |
| evrardjp | not opt-out | 16:38 |
| mnaser | chrony ** | 16:38 |
| logan- | yeah changing it to opt-in makes sense i think | 16:38 |
| logan- | reducing it to a single job might not, because we only test 1 OS then | 16:38 |
| logan- | although we still have full coverage in the hardening role itself | 16:38 |
| mnaser | yeah all of our one-offs are tested on 1 OS (except for metal) | 16:38 |
| logan- | so maybe that's fine? | 16:38 |
| logan- | idk | 16:38 |
| *** vnogin has joined #openstack-ansible | 16:38 | |
| mnaser | proxy is bionic only, Ceph is bionic only right now | 16:38 |
| logan- | yeah | 16:39 |
| mnaser | if our jobs were stable as heck (metal are pretty good) I'd be inclined to put all 4 but then we start covering a lot of combinations | 16:40 |
| noonedeadpunk | but probably moving it out of scope is not a bad idea. I'm wondering if we can get extra users/contributors by moving it and not relating so much with osa itself | 16:40 |
| mnaser | btw | 16:40 |
| *** vnogin has quit IRC | 16:40 | |
| mnaser | https://linux-system-roles.github.io | 16:40 |
| mnaser | could be interesting | 16:40 |
| mnaser | to work with | 16:41 |
| logan- | yeah i wouldn't like to add separate hardening jobs. i'm split between off by default, and on by default in integrated gate jobs. or off by default, and on in a separate single job | 16:41 |
| odyssey4me | evrardjp no sense in working on the next stig in that role, because ansible-lockdown are actively working on all that and doing it in our role would duplicate effort | 16:41 |
| odyssey4me | I'd rather see any interest move to their roles. | 16:41 |
| logan- | odyssey4me: are they actively working on multi OS support also? | 16:42 |
| mnaser | I never understood how 'rhel 7' stig applies on 'ubuntu' :) | 16:42 |
| odyssey4me | logan- well, yes - although they need hands to get it done, so perhaps we should ML post that we're not actively maintaining it and that ansible-lockdown is seeking contributors | 16:42 |
| odyssey4me | yep, ubuntu now has a stig apparently | 16:42 |
| odyssey4me | back then it did not | 16:42 |
| logan- | mnaser: because common sense hardening practices are only distro specific in implementation, not conceptually :) | 16:42 |
| mnaser | gotcha | 16:43 |
| *** miloa has quit IRC | 16:44 | |
| logan- | odyssey4me: ++ | 16:44 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: [WIP] Add services for testing based on test repo name https://review.opendev.org/658398 | 16:45 |
| odyssey4me | reviews please: https://review.opendev.org/658385 | 16:45 |
| *** cmart has joined #openstack-ansible | 16:46 | |
| odyssey4me | and https://review.opendev.org/658386 | 16:46 |
| *** cmart has quit IRC | 16:47 | |
| logan- | odyssey4me: wow, how many releases back are affected by that? | 16:47 |
| odyssey4me | logan- just stein, I think - I think we hacked around it in rocky and didn't have that mechanism before then | 16:48 |
| logan- | ah | 16:48 |
| odyssey4me | once we have stein back to normal, I'll take a peek at the older branches | 16:48 |
| logan- | shouldn't be too hard to find out. all we have to do is run a nova job. :) | 16:48 |
| odyssey4me | well, any role test which includes nova will break | 16:49 |
| logan- | yeah quite some interesting fallout from the placement split. | 16:49 |
| *** udesale has quit IRC | 16:49 | |
| logan- | thanks for patching it | 16:50 |
| odyssey4me | mnaser https://review.opendev.org/658289 won't work until the dependent change is merged (depends-on is broken until the later opendev patches merge) | 16:50 |
| openstackgerrit | Merged openstack/openstack-ansible-os_tempest stable/stein: Remove install_test_requirements flag https://review.opendev.org/658286 | 16:50 |
| mnaser | yeah I noticed after clicking 'recheck' | 16:50 |
| logan- | yeah i stacked some things wrong yesterday. i need to go restack changes today | 16:51 |
| logan- | i think it all works (including https://review.opendev.org/#/q/topic:fix-credentials-scoping), but i'll work on getting it all passing CI in a bit | 16:51 |
| *** SmearedBeard has joined #openstack-ansible | 16:52 | |
| mnaser | fatal: [aio1]: FAILED! => {"attempts": 5, "changed": false, "msg": "No package matching '<generator object select_or_reject at 0x7efd4dc61f00>' is available"} | 16:55 |
| mnaser | logan-: did I need to add |list or something after? | 16:55 |
| logan- | yeah | 16:55 |
| logan- | guess so | 16:55 |
| logan- | sorry, I didn't look at any references for that | 16:56 |
| logan- | I know we have similar stuff implemented and i'll assume it is using the reject filter also :P | 16:56 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-repo_server master: [redhat] Install rsync-daemon for newer releases https://review.opendev.org/658251 | 16:56 |
| mnaser | ya all good | 16:56 |
| *** strattao has joined #openstack-ansible | 16:57 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: [wip] Add support for Fedora 28 https://review.opendev.org/658242 | 16:57 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Skip neutron rebalance if ha-tool is disabled https://review.opendev.org/652205 | 17:00 |
| *** kopecmartin is now known as kopecmartin|off | 17:04 | |
| *** DanyC_ has joined #openstack-ansible | 17:06 | |
| *** DanyC has quit IRC | 17:10 | |
| *** DanyC_ has quit IRC | 17:10 | |
| *** DanyC has joined #openstack-ansible | 17:11 | |
| *** SmearedBeard has quit IRC | 17:15 | |
| *** SmearedBeard has joined #openstack-ansible | 17:17 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts master: Returning of the Suse LXC image https://review.opendev.org/655851 | 17:19 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Add services for testing based on test repo name https://review.opendev.org/658398 | 17:19 |
| *** SmearedBeard has quit IRC | 17:23 | |
| *** SmearedBeard has joined #openstack-ansible | 17:27 | |
| *** electrofelix has quit IRC | 17:30 | |
| openstackgerrit | Merged openstack/openstack-ansible-tests stable/stein: Ensure that TESTING_BRANCH is exported https://review.opendev.org/658385 | 17:30 |
| odyssey4me | evrardjp yay! https://review.opendev.org/658398 is working :) | 17:36 |
| odyssey4me | logan- does that ability to disable the dynamic scenario work for you? | 17:37 |
| odyssey4me | mnaser what do you think of https://review.opendev.org/658398 ? there are tester patches https://review.opendev.org/658406, https://review.opendev.org/658405 | 17:40 |
| *** Weifan has joined #openstack-ansible | 17:40 | |
| logan- | odyssey4me: yeah that looks nice! | 17:41 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_barbican master: Use the dynamic aio scenario https://review.opendev.org/658406 | 17:42 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_aodh master: Use the dynamic aio scenario https://review.opendev.org/658405 | 17:42 |
| openstackgerrit | Merged openstack/openstack-ansible-tests stable/stein: Revert "Pin nova_git_install_branch to before placement removal" https://review.opendev.org/658386 | 17:42 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_blazar master: Use the dynamic aio scenario https://review.opendev.org/658430 | 17:47 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ceilometer master: Use the dynamic aio scenario https://review.opendev.org/658431 | 17:47 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_congress master: Use the dynamic aio scenario https://review.opendev.org/658432 | 17:50 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_designate master: Use the dynamic aio scenario https://review.opendev.org/658433 | 17:52 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_gnocchi master: Use the dynamic aio scenario https://review.opendev.org/658434 | 17:53 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_heat master: Use the dynamic aio scenario https://review.opendev.org/658435 | 17:54 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Add services for testing based on test repo name https://review.opendev.org/658398 | 18:02 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_horizon master: Use the dynamic aio scenario https://review.opendev.org/658436 | 18:02 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Add services for testing based on test repo name https://review.opendev.org/658398 | 18:07 |
| *** strattao has quit IRC | 18:13 | |
| mnaser | odyssey4me: I like that. it'll be a lot cleaner, tho the only thing people might be confused on why a thing is not being tested | 18:14 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Add services for testing based on test repo name https://review.opendev.org/658398 | 18:14 |
| mnaser | odyssey4me: the other thing is there are some dependencies like ironic needs swift for example | 18:14 |
| odyssey4me | mnaser we deal with that in the aio bootstrap | 18:15 |
| mnaser | yeah so we can just refactor that in there indeed | 18:15 |
| odyssey4me | these are only scenario key entries, the interpretation of the full service list happens in the aio bootstrap | 18:15 |
| logan- | ye[ | 18:15 |
| mnaser | we still have a few functional things we need to kill at some point i.e. galera/keystone etc | 18:15 |
| logan- | yep* | 18:15 |
| mnaser | multinode testing anyone ha | 18:15 |
| odyssey4me | yep, working on that after this - then will work on moving the lint tests to the integrated repo | 18:16 |
| odyssey4me | any other feedback on https://review.opendev.org/658398 ? I think I have it now | 18:16 |
| *** hamzy_ has quit IRC | 18:17 | |
| logan- | lgtm | 18:20 |
| *** DanyC has quit IRC | 18:21 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ironic master: Use the dynamic aio scenario https://review.opendev.org/658442 | 18:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_magnum master: Use the dynamic aio scenario https://review.opendev.org/658443 | 18:22 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_masakari master: Use the dynamic aio scenario https://review.opendev.org/658445 | 18:24 |
| *** hamzy has joined #openstack-ansible | 18:24 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_mistral master: Use the dynamic aio scenario https://review.opendev.org/658448 | 18:25 |
| odyssey4me | that's it for tonight, I'll keep an eye on the patches in flight to unbreak stein over the w/end | 18:26 |
| mnaser | thanks odyssey4me | 18:27 |
| mnaser | anyone has any ideas why wsrep_evs_state is not going into 'OPERATIONAL' ? | 18:28 |
| odyssey4me | have a great w/end all! | 18:28 |
| logan- | anything useful in /var/log/mysql_logs/galera_whatever.log | 18:28 |
| logan- | usually that holds the answers when there's wsrep problems | 18:28 |
| mnaser | that doesn't even log | 18:29 |
| mnaser | but it looks like include /etc/my.cnf.d isn't happening | 18:29 |
| logan- | :/ | 18:29 |
| mnaser | so its dropping configs in `/etc/my.cnf.d/` | 18:30 |
| mnaser | but but `galera_etc_include_dir` is `/etc/mysql/conf.d` | 18:30 |
| mnaser | ah but that path exist stoo | 18:31 |
| *** strattao has joined #openstack-ansible | 18:31 | |
| mnaser | `!includedir /etc/mysql/conf.d/` is inside the `[mysqldump]` section.. does that matter? | 18:31 |
| *** jamesdenton has quit IRC | 18:37 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Add services for testing based on test repo name https://review.opendev.org/658398 | 18:39 |
| *** slin has joined #openstack-ansible | 18:41 | |
| logan- | reviews on https://review.opendev.org/#/q/topic:calico-tests+status:open appreciated | 18:42 |
| mnaser | https://www.irccloud.com/pastebin/0jMUwkck/ | 18:42 |
| mnaser | so its not parsing the config file at all | 18:42 |
| logan- | weird | 18:42 |
| mnaser | `mysqld --verbose --help` shows the right settings | 18:43 |
| mnaser | so maybe this is a bootstrap setting | 18:44 |
| *** Weifan has quit IRC | 18:46 | |
| *** Weifan has joined #openstack-ansible | 18:50 | |
| mnaser | im super stuck :\ | 19:03 |
| spotz | mnaser: double check permissions? And what happens if you manually do it(if possible) | 19:12 |
| mnaser | perms are ok, no errors for perms, nothing being parsed | 19:12 |
| spotz | just trying to think of why it can't parse the conf:( | 19:13 |
| jrosser | for some os there is a good deal of messing with the location of the config dirs / symlinks | 19:13 |
| spotz | I know I ran into some errors on something else with bionic needing a mysql user vs using root but that doesn't seem to be the issue | 19:14 |
| mnaser | jrosser: yeah I suspect this might have to do with it but I mean its the same rpms | 19:17 |
| *** radeks_ has joined #openstack-ansible | 19:18 | |
| mnaser | im quite stuck honestly, I have no clue. | 19:19 |
| mnaser | I can pop in someone's ssh keys if they want to dig | 19:19 |
| jrosser | theres no selinux interfereing is there? | 19:24 |
| mnaser | jrosser: nope | 19:26 |
| mnaser | SELinux is permissive | 19:26 |
| jrosser | i'd probably be reaching for strace on open() syscalls next to see what it actually opens/trys to open | 19:28 |
| spotz | mnaser silly thought but all the instances can reach other and started in order? | 19:28 |
| mnaser | spotz: its just an aio | 19:28 |
| spotz | grr:( RUnning out of ideas | 19:29 |
| mnaser | https://www.irccloud.com/pastebin/jgjoSbJp/ | 19:33 |
| mnaser | manually starting it up work | 19:33 |
| mnaser | but with systems it .. doesn't | 19:33 |
| mnaser | ? | 19:33 |
| spotz | so permissions most likely somewhere with the user or the startup file? | 19:35 |
| mnaser | I think it might have to do with the systemd unit | 19:37 |
| jrosser | https://mariadb.com/kb/en/library/configuring-mariadb-with-option-files/ | 19:38 |
| *** shardy has quit IRC | 19:39 | |
| mnaser | I think somehow when it starts with systemd it can't see that file | 19:42 |
| mnaser | mysqld --print-defaults shows the right things | 19:44 |
| jrosser | can you paste the unit file? | 19:45 |
| mnaser | jrosser: grrr | 19:46 |
| mnaser | jrosser: sudo -u mysql stat /etc/mysql/my.cnf | 19:46 |
| mnaser | stat: cannot stat '/etc/mysql/my.cnf': Permission denied | 19:46 |
| jrosser | errr? thought the permissions looked ok? | 19:47 |
| mnaser | I didn't check them *as* mysql user | 19:47 |
| mnaser | :< | 19:47 |
| mnaser | looks like we create all those folders | 19:49 |
| mnaser | but never set permissions | 19:49 |
| mnaser | I wonder how it ever worked before? | 19:49 |
| mnaser | /etc/mysql perms are 755 on cenetos | 19:51 |
| openstackgerrit | Merged openstack/openstack-ansible-repo_server master: [redhat] Install rsync-daemon for newer releases https://review.opendev.org/658251 | 19:51 |
| mnaser | we should probably lock that down to the mysql user only. | 19:51 |
| mnaser | that fixed it. | 19:52 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-galera_server master: [redhat] Update variables to be version agnostic https://review.opendev.org/658253 | 19:54 |
| *** slin has quit IRC | 19:55 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: ci: use local connection inside AIO https://review.opendev.org/658460 | 20:01 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-rabbitmq_server master: config: setup correct owner and group https://review.opendev.org/658463 | 20:06 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: [wip] Add support for Fedora 28 https://review.opendev.org/658242 | 20:07 |
| *** chalupaul has joined #openstack-ansible | 20:12 | |
| chalupaul | howdy! I was wondering if anyone had a rough idea when we can expect a 19.0.0 tag. | 20:13 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible master: Use fast SSH cipher in the gate https://review.opendev.org/658473 | 20:26 |
| *** Weifan has quit IRC | 20:35 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-galera_server master: [redhat] Update variables to be version agnostic https://review.opendev.org/658253 | 20:35 |
| *** Smeared_Beard has joined #openstack-ansible | 20:37 | |
| *** SmearedBeard has quit IRC | 20:40 | |
| *** radeks_ has quit IRC | 20:42 | |
| logan- | mnaser: that's weird.. ansible is creating /etc/mysql with 0755 by default right since we weren't providing a mode? what bits does the mysql package set normally? | 20:51 |
| mnaser | logan-: im gonna guess maybe the umask under /etc changed from centos 7 to fedora 28? | 20:51 |
| logan- | yeah thats what i'm guessing, just curious what mode it was set to? | 20:52 |
| logan- | rather than setting 755, another idea is we could just move that task to the bottom of the install file and inherit the mode from the mariadb package since it will create /etc/mysql | 20:53 |
| *** hjohnson1 has left #openstack-ansible | 20:54 | |
| logan- | just trying to understand why it was broken in the first place. http://paste.openstack.org/raw/751241/ | 20:56 |
| *** sreejithp has quit IRC | 20:58 | |
| spotz | Ha I was right!:) | 21:08 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible master: Use fast SSH cipher in the gate https://review.opendev.org/658473 | 21:19 |
| *** kplant has quit IRC | 21:38 | |
| *** Smeared_Beard has quit IRC | 22:01 | |
| *** aludwar has quit IRC | 22:04 | |
| *** hamzaachi has joined #openstack-ansible | 22:14 | |
| *** kplant has joined #openstack-ansible | 22:23 | |
| *** strattao has quit IRC | 22:31 | |
| *** Weifan has joined #openstack-ansible | 22:37 | |
| *** Weifan has quit IRC | 22:38 | |
| *** Weifan has joined #openstack-ansible | 22:38 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible master: Import placement auth vars to global scope https://review.opendev.org/658176 | 22:43 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible master: Migrate to placement_service_password https://review.opendev.org/658217 | 22:43 |
| *** macza has quit IRC | 22:45 | |
| *** Weifan has quit IRC | 22:46 | |
| *** macza has joined #openstack-ansible | 22:47 | |
| *** Weifan has joined #openstack-ansible | 22:47 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-tests master: Clean up placement test vars https://review.opendev.org/658502 | 22:53 |
| *** Weifan has quit IRC | 22:53 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible master: Migrate to placement_service_password https://review.opendev.org/658217 | 23:00 |
| *** hamzaachi has quit IRC | 23:09 | |
| *** macza has quit IRC | 23:11 | |
| *** Weifan has joined #openstack-ansible | 23:12 | |
| *** hjohnson1 has joined #openstack-ansible | 23:17 | |
| *** hjohnson1 has left #openstack-ansible | 23:17 | |
| *** Weifan has quit IRC | 23:18 | |
| *** Weifan has joined #openstack-ansible | 23:19 | |
| *** gyee has quit IRC | 23:26 | |
| *** Weifan has quit IRC | 23:42 | |
| *** aedc has joined #openstack-ansible | 23:53 | |
| *** Weifan has joined #openstack-ansible | 23:58 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!