| *** brosky has quit IRC | 01:12 | |
| *** evrardjp has quit IRC | 04:35 | |
| *** evrardjp has joined #openstack-ansible | 04:35 | |
| *** schwicht has quit IRC | 04:36 | |
| *** schwicht has joined #openstack-ansible | 04:42 | |
| *** pcaruana has joined #openstack-ansible | 06:35 | |
| *** andrea15 has joined #openstack-ansible | 06:39 | |
| *** cshen_ has quit IRC | 08:45 | |
| *** cshen has joined #openstack-ansible | 08:49 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Update hacking for Python3 https://review.opendev.org/718997 | 09:16 |
|---|---|---|
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Ensure procps is present in debian container image https://review.opendev.org/710251 | 09:17 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Move all tox environments to python3 https://review.opendev.org/710253 | 09:17 |
| *** tosky has joined #openstack-ansible | 09:36 | |
| *** cshen has quit IRC | 10:08 | |
| *** cshen has joined #openstack-ansible | 10:43 | |
| *** cshen has quit IRC | 10:48 | |
| *** cshen has joined #openstack-ansible | 11:23 | |
| *** cshen has quit IRC | 11:28 | |
| openstackgerrit | Andreas Jaeger proposed openstack/ansible-hardening master: Cleanup py27 support https://review.opendev.org/723084 | 13:07 |
| openstackgerrit | Merged openstack/openstack-ansible-os_keystone master: Cleanup py2 https://review.opendev.org/720546 | 13:13 |
| openstackgerrit | Andreas Jaeger proposed openstack/ansible-role-python_venv_build master: Cleanup py27 support https://review.opendev.org/723088 | 13:17 |
| openstackgerrit | Andreas Jaeger proposed openstack/ansible-role-systemd_mount master: Cleanup py27 support https://review.opendev.org/723092 | 13:47 |
| *** AJaeger has joined #openstack-ansible | 13:54 | |
| AJaeger | is there are reason that pip_install is still used in tests or can it be removed like https://review.opendev.org/#/c/723089/2/tests/ansible-role-requirements.yml ? | 13:55 |
| openstackgerrit | Andreas Jaeger proposed openstack/ansible-role-systemd_networkd master: Cleanup py27 support https://review.opendev.org/723094 | 14:01 |
| openstackgerrit | Andreas Jaeger proposed openstack/ansible-role-systemd_service master: Cleanup py27 support https://review.opendev.org/723095 | 14:04 |
| -openstackstatus- NOTICE: Zuul is currently failing some jobs with MERGER_FAILURE, this needs investigation by OpenDev team. Please refrain from rechecking until we give the all-clear. | 14:16 | |
| *** vesper11 has quit IRC | 15:02 | |
| *** vesper11 has joined #openstack-ansible | 15:03 | |
| *** nwonknu has quit IRC | 15:06 | |
| *** jroll has quit IRC | 15:06 | |
| *** nwonknu has joined #openstack-ansible | 15:06 | |
| *** jroll has joined #openstack-ansible | 15:07 | |
| *** irclogbot_0 has quit IRC | 15:55 | |
| *** irclogbot_0 has joined #openstack-ansible | 15:58 | |
| *** idlemind has quit IRC | 15:59 | |
| *** idlemind has joined #openstack-ansible | 16:00 | |
| *** irclogbot_0 has quit IRC | 16:26 | |
| *** irclogbot_2 has joined #openstack-ansible | 16:28 | |
| *** evrardjp has quit IRC | 16:35 | |
| *** evrardjp has joined #openstack-ansible | 16:35 | |
| *** pcaruana has quit IRC | 18:33 | |
| *** cshen has joined #openstack-ansible | 18:40 | |
| *** cshen has quit IRC | 18:45 | |
| fridtjof[m] | can someone fill me in on how DNS/hostname resolution works on the deployment host when executing playbooks? | 18:59 |
| fridtjof[m] | I'm going through setup-hosts on train right now, and it's failing in the task "Create container(s)" in containers-lxc-create.yml because it fails to resolve the container host names | 18:59 |
| fridtjof[m] | is it directly connecting to the containers, or does it go through the physical host they're on? | 19:01 |
| fridtjof[m] | hm, when looking at openstack_inventory.json, it seems they're not being allocated IPs | 19:02 |
| fridtjof[m] | ansible_host and container_address are both null | 19:02 |
| fridtjof[m] | okay, i noticed that the containers on my infra1 hosts exist, but don't have any network interfaces except for lo | 19:20 |
| fridtjof[m] | am i missing something? | 19:20 |
| fridtjof[m] | in the preceding step, "Write default container config" i can see it's not adding any includes for network interfaces, compared to the config file the containers have in my existing rocky deployment | 19:27 |
| fridtjof[m] | okay, i'm understanding that containers don't have IPs from the start? | 19:51 |
| fridtjof[m] | that's what https://docs.openstack.org/openstack-ansible/train/reference/inventory/manage-inventory.html implies, at least | 19:51 |
| *** cshen has joined #openstack-ansible | 20:00 | |
| *** cshen has quit IRC | 20:05 | |
| fridtjof[m] | so from what I gathered, "Gather container facts" (the task that fails) happens before any interfaces for containers are ever configured | 20:52 |
| fridtjof[m] | but that task makes the deployment host try to connect to the containers through SSH | 20:52 |
| fridtjof[m] | - the container's can't be reached from the deployment host at this point, though | 20:53 |
| fridtjof[m] | - is the deployment host going to know their addresses if they did have one? | 20:53 |
| fridtjof[m] | for reference, I'm on the 20.1.0 branch with all involved hosts on ubuntu 18.04 | 20:54 |
| jrosser | the deploy host *always* connects to the containers via the physical host they are on, never directly by ssh | 20:58 |
| jrosser | if they are missing addresses in the inventory json then it is most likley you have an error in openstack_user_config relating perhaps to the management network | 20:59 |
| jrosser | building a train AIO in a vm would give you something quick to compare against | 21:02 |
| fridtjof[m] | is that what delegate_to: {{ physical_host }} does in the role's playbooks? | 21:02 |
| fridtjof[m] | (connecting through the physical host) | 21:02 |
| jrosser | no, the connection plugin does that | 21:02 |
| jrosser | every time you run the dynamic inventory script the inventory json is updated, that’s the source of the ip addresses | 21:04 |
| jrosser | it’s not worth proceeding any further until your opentack_user_config is fixed to make the containers be assumed addresses, I think | 21:05 |
| jrosser | *assigned addresses | 21:05 |
| fridtjof[m] | okay, so they should definitely not have their addresses = null, right? | 21:06 |
| fridtjof[m] | (in the json file, that is) | 21:06 |
| jrosser | that does sound wrong | 21:08 |
| jrosser | iirc the address assignment will all happen up front in the dynamic inventory python script | 21:08 |
| fridtjof[m] | so, the errors i'm getting on that task look like this: | 21:09 |
| fridtjof[m] | fatal: [infra1_neutron_server_container-c8a11a51]: UNREACHABLE! => {"changed": false, "msg": "Data could not be sent to remote host \"infra1_neutron_server_container-c8a11a51\". Make sure this host can be reached over ssh: ssh: Could not resolve hostname infra1_neutron_server_container-c8a11a51: Temporary failure in name resolution\r\n", "unreachable": true} | 21:09 |
| jrosser | scripts/inventory_manage.py can pretty-print the inventory data for you | 21:09 |
| fridtjof[m] | but i'll take a quick look at the dynamic inventory stuff | 21:09 |
| fridtjof[m] | that's a python script in inventory/, right? | 21:09 |
| fridtjof[m] | ah, okay | 21:09 |
| fridtjof[m] | http://paste.openstack.org/show/792718/ | 21:11 |
| fridtjof[m] | this looks.... wrong | 21:11 |
| fridtjof[m] | http://paste.openstack.org/show/792719/ | 21:12 |
| fridtjof[m] | this is my openstack_user_config | 21:13 |
| fridtjof[m] | br-mgmt should be exactly as the example says | 21:14 |
| fridtjof[m] | (compute1 is on unexpected downtime for the time being, don't mind it being commented out) | 21:14 |
| fridtjof[m] | running inventory/dynamic-inventory.py --check gives me a "configuration ok!" :/ | 21:15 |
| jrosser | i think that the indention in provider_networks might be wrong | 21:16 |
| jrosser | see here https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/openstack_user_config.yml.prod.example#L23-L25 | 21:16 |
| fridtjof[m] | from my experience with yaml, that shouldn't matter | 21:20 |
| fridtjof[m] | i'll try it though | 21:20 |
| fridtjof[m] | OH | 21:20 |
| fridtjof[m] | you're right, that should actually make a difference | 21:20 |
| fridtjof[m] | the way it is right now, network: is just another key besides all the other ones | 21:20 |
| fridtjof[m] | YUP | 21:23 |
| jrosser | awesome :) | 21:23 |
| fridtjof[m] | thank you so much | 21:23 |
| fridtjof[m] | they all have IPs now | 21:24 |
| fridtjof[m] | sometimes it's stupid things like this | 21:24 |
| fridtjof[m] | at least I learned a little more about ansible and OSA tooling in the process today ^^ | 21:24 |
| jrosser | no worries.... ideally the code should error out, becasue 'None' is nonsense | 21:24 |
| fridtjof[m] | i'd gladly make a PR for that, if i had the time :( | 21:25 |
| *** tosky has quit IRC | 23:32 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!