| jrosser | morning | 08:32 |
|---|---|---|
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Deploy step-ca when 'stepca' is part of the deployment scenario. https://review.opendev.org/c/openstack/openstack-ansible/+/876637 | 08:41 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Add a /etc/hosts entry for the external IP of an AIO https://review.opendev.org/c/openstack/openstack-ansible/+/876638 | 08:41 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Use certbot to generate SSL cert for the external VIP in 'stepca' scenario https://review.opendev.org/c/openstack/openstack-ansible/+/876693 | 08:41 |
| noonedeadpunk | o/ | 09:09 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-rabbitmq_server master: Do not use 'always' tag in inappropriate places https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/875971 | 09:10 |
| noonedeadpunk | these patches are interesting... | 09:11 |
| noonedeadpunk | I thought we'd need integration with pki role for smth like that? | 09:11 |
| jrosser | noonedeadpunk: well - these patches just test the external VIP getting a cert with the haproxy acme/LE code | 09:13 |
| noonedeadpunk | aha | 09:13 |
| jrosser | and the step-ca installation is there in place of using actual LE endpoint | 09:13 |
| jrosser | i think i'm particularly interested in what you think of 876638 | 09:14 |
| jrosser | this is all a bit "first attempt" btw so if you can see any improvements then we can do that | 09:14 |
| noonedeadpunk | Well. I can recall having ipv6 only providers for opendev CI | 09:15 |
| noonedeadpunk | So default_ipv4 might not be a thing | 09:15 |
| noonedeadpunk | I'm not sure if they're still present, but smth to take into consideration | 09:15 |
| jrosser | thats just totally copy/pasted from the way `external_lb_vip_address` is defined | 09:17 |
| jrosser | https://opendev.org/openstack/openstack-ansible/src/branch/master/etc/openstack_deploy/openstack_user_config.yml.aio#L37 | 09:18 |
| noonedeadpunk | oh, huh | 09:31 |
| jrosser | i need to think about that a bit more | 09:43 |
| jrosser | we need to support the old way with with an IP, also `external.openstack.local` for testing ACME as certbot refuses to issue for an IP, and also you might have a legitimate `aio.example.org` proper DNS entry that you want to use | 09:45 |
| noonedeadpunk | well. same would apply for internal then | 09:46 |
| noonedeadpunk | and we don't have anything to cover it anyway. | 09:46 |
| noonedeadpunk | So I won't relate legit FQDN for aio with test of step-ca - these are different things we should address | 09:47 |
| jrosser | it affects how the variables are set up i think | 09:47 |
| jrosser | need the right point to be able to override your own FQDN and i think i've not allowed that right now in these patches | 09:48 |
| jrosser | but not sure tbh | 09:48 |
| jrosser | still concerned about horizon and the new haproxy setup tbh which is why i started this | 09:49 |
| jrosser | default_backend: 'horizon' -> this fails if the horizon backend is not defined | 09:49 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-rabbitmq_server stable/wallaby: Bump erlang versions https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/876708 | 09:52 |
| noonedeadpunk | Yeah, I'd need to update things with your findings to pass only specific variables to haproxy rather then adding it to group | 09:53 |
| noonedeadpunk | I'm just trying to finilize role for generic resource creation, that everyone said is quite easy and useless to have in plugins... But I've already spent a day and it covers only half of what's needed... | 09:54 |
| noonedeadpunk | But I think since TripleO is deprecated, we can now easily drop crap out of tempest role :D | 09:56 |
| jrosser | oh yes i also thought of that | 09:56 |
| jrosser | i am also looking at what to do about compute.example.com / dashboard.example.com approach for haproxy | 09:57 |
| jrosser | because even though it seems like more complexity, actually i think an implementation that supports that will bring further simplification to the work damiandabrowski has done already | 09:57 |
| jrosser | in particular for this tricky horizon case | 09:57 |
| jrosser | i will try to do something that moves the ACL definitions we currently have on the haproxy frontend out to a map file rather than have them directly in the config | 09:59 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_tempest master: Remove tripleo CI jobs https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/876711 | 10:06 |
| Losraio | Hello | 13:04 |
| Losraio | Are you seeing this? | 13:04 |
| mgariepy | Losraio, yes.. | 13:06 |
| Losraio | Great | 13:06 |
| Losraio | I would like some help regarding my openstack-ansible deployment | 13:07 |
| Losraio | When I try to run the setup-hosts.yml I get the following error: | 13:09 |
| Losraio | [WARNING]: Unable to parse /etc/openstack_deploy/inventory.ini as an inventory source | 13:09 |
| Losraio | Which is a problem because at some point later during the run I get this error: | 13:09 |
| Losraio | TASK [lxc_container_create : Gather container facts] ******************************************************************* [WARNING]: Unhandled error in Python interpreter discovery for host hua-openstack-controller_glance_container-42a67e6c: Failed to connect to the host via ssh: ssh: Could not resolve hostname inventory_hostname: Temporary failure in name resolution [WARNING]: Unhandled error in Python interpreter discovery for host hua-opensta | 13:09 |
| Losraio | And it prints for each container | 13:09 |
| noonedeadpunk | Losraio: inventory.ini is red herring, you can ignore that as it should not affect deployment anyhow | 13:12 |
| noonedeadpunk | Also for pastes we use https://paste.openstack.org/ as it gets trimmed | 13:12 |
| Losraio | oh | 13:13 |
| noonedeadpunk | But I assume you've posted paste in a bug report yestarday? | 13:13 |
| Losraio | Yes | 13:13 |
| Losraio | Let me paste it again | 13:13 |
| Losraio | https://paste.openstack.org/show/bB83GCum5mN0thj8aLvr/ | 13:13 |
| Losraio | FYI, the deployment host is also the the controller node of the whole deployment | 13:14 |
| noonedeadpunk | Eventually pasting your configuration would be helpful here | 13:14 |
| Losraio | Sure | 13:15 |
| noonedeadpunk | Yeah, deployment host == controller should not cause any troubles | 13:15 |
| Losraio | Here you go: | 13:16 |
| Losraio | https://paste.openstack.org/show/b2XcmStbbD3STqes1uoW/ | 13:16 |
| noonedeadpunk | So first thing - internal_lb_vip_address and external_lb_vip_address can not be exact same. They can be from the same network, but different | 13:17 |
| Losraio | Alright | 13:17 |
| Losraio | Let me change that real quick | 13:18 |
| noonedeadpunk | But it's not that causes this failure | 13:18 |
| noonedeadpunk | *the issue | 13:18 |
| Losraio | I see... | 13:18 |
| opendevreview | Merged openstack/openstack-ansible-galera_server stable/yoga: Allow maridbcheck socket to FreeBind https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/874733 | 13:18 |
| noonedeadpunk | It will fail in setup-infrastructure with that ) | 13:19 |
| Losraio | OK | 13:20 |
| Losraio | Fixed | 13:20 |
| Losraio | What else could be the culprit then | 13:20 |
| jrosser | i wonder if there is a limit on hostname length being exceeded somewhere | 13:23 |
| jrosser | Losraio: does the contents of /etc/hosts look sensible for those hostnames you have, like `hua-openstack-controller_memcached_container- | 13:25 |
| jrosser | ab2ed959` | 13:25 |
| Losraio | Yeah | 13:26 |
| Losraio | Here's another paste of it | 13:26 |
| Losraio | https://paste.openstack.org/show/bqJGBCkG02oemhjWaYZR/ | 13:27 |
| jrosser | oh well there we are! see all the 'None' in there | 13:27 |
| Losraio | Aha | 13:28 |
| noonedeadpunk | jrosser: eventually we should verify length with dynamic_inventory script | 13:28 |
| jrosser | iirc there is some comment in the code about length | 13:28 |
| Losraio | Should I perhaps delete the openstack entries from the hosts file? | 13:29 |
| Losraio | And trying to run the deployment again? | 13:29 |
| noonedeadpunk | Losraio: do you have bridges created on the host? | 13:29 |
| Losraio | Let me check | 13:29 |
| noonedeadpunk | As there should be br-mgmt at very least with IP on it | 13:29 |
| Losraio | I only have an lxcbr0 bridge | 13:29 |
| noonedeadpunk | Ok, that's the issue | 13:30 |
| Losraio | But I thought these bridges were created automatically? | 13:30 |
| opendevreview | Merged openstack/openstack-ansible-os_nova stable/yoga: Add authentication for [cinder] section of nova.conf https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/873570 | 13:30 |
| noonedeadpunk | So until very lately on master we don't manage network configuration and don't have option to do that with osa | 13:30 |
| noonedeadpunk | So we leave net config to deployers | 13:30 |
| Losraio | Right | 13:30 |
| noonedeadpunk | I'm not even sure if it was merged on master or not | 13:31 |
| noonedeadpunk | The only thing that is managed is lxcbr0 but it can be disabled as well | 13:31 |
| jrosser | also i think that having all `cidr_networks` be the same CIDR is not going to end well | 13:31 |
| jrosser | i can't say how, but multi-homing several interfaces in the same subnet usually avoided | 13:32 |
| noonedeadpunk | Yeah, if you're going to use same interface for storage/mgmt/tenent - you can do that but then you'll need to drop these extra interfaces and cidrs | 13:32 |
| Losraio | Yes, I know that its suboptimal but I'm just testing right now | 13:32 |
| jrosser | Losraio: have you tried an all-on-one deployment? | 13:33 |
| noonedeadpunk | as it can live just fine with just br-mgmt | 13:33 |
| Losraio | Yes I have. It works but I need a multi-node deployment | 13:33 |
| noonedeadpunk | Yeah, that's where impression about interfaces come from :D | 13:33 |
| noonedeadpunk | As in AIO we indeed create them for quite some tome | 13:33 |
| noonedeadpunk | *time | 13:33 |
| Losraio | Oh | 13:33 |
| jrosser | there is an advantage to keeping the setup kind of similar between the AIO and your multinode | 13:33 |
| Losraio | I see | 13:34 |
| jrosser | then you can compare one vs. the other when something is broken | 13:34 |
| Losraio | So, I should setup the bridges according the deployment guide? | 13:34 |
| noonedeadpunk | yeah | 13:34 |
| jrosser | Losraio: https://docs.openstack.org/project-deploy-guide/openstack-ansible/latest/targethosts.html | 13:34 |
| Losraio | And just attach them to my NIC | 13:34 |
| noonedeadpunk | You can have only 1 bridge as well if you want | 13:34 |
| noonedeadpunk | Eventually you can have pretty much any setup you want, guides describe jsut most common options | 13:35 |
| Losraio | But the documentation says otherwise | 13:35 |
| Losraio | I thought these bridges were mandatory | 13:35 |
| jrosser | the documentation describes a set of "sensible defaults" | 13:35 |
| jrosser | but you can make pretty much any valid deployment that you like | 13:35 |
| Losraio | I see | 13:36 |
| jrosser | openstack-ansible is not a tool that has "only one way" - everything is customisable, but that means there is not "one right answer" | 13:36 |
| noonedeadpunk | For example - you might want to deploy on bare metal without lxc at all - then you don't need any bridges at all | 13:36 |
| jrosser | the sensible defaults described in the documentation are a great starting point for something that will cover 95% of use cases | 13:37 |
| Losraio | Yeah I was looking for the option to deploy osa on bare metal without containers | 13:38 |
| Losraio | But I couldn't find it | 13:38 |
| jrosser | https://docs.openstack.org/openstack-ansible/latest/reference/inventory/configure-inventory.html#deploying-directly-on-hosts | 13:39 |
| jrosser | again depending on your use case that might, or might not be what you want to do | 13:39 |
| jrosser | pros/cons with all of these things | 13:39 |
| Losraio | Yes I know, I just want to set it up without containers for the sake of simplicity | 13:40 |
| jrosser | right - sure | 13:40 |
| Losraio | I will give it a shot | 13:41 |
| jrosser | you can look at simplicity multiple ways too, so without containers you have less things involved | 13:41 |
| jrosser | but then you might say "how to a completely delete / reprovision the galera database on one infra node" | 13:41 |
| jrosser | thats very easy with containers, less so without | 13:41 |
| Losraio | So without containers, I don't need to declare networks in the user_config.yml? | 13:41 |
| jrosser | so it's not always a simple judgement | 13:42 |
| noonedeadpunk | I think you can define `no_containers` just in global_overrides? | 13:42 |
| Losraio | I didn't see such option | 13:43 |
| jrosser | like this https://github.com/openstack/openstack-ansible/blob/17a37653e69282112eccc8416112f1253d7cf3d2/etc/openstack_deploy/openstack_user_config.yml.aio.j2#L46 | 13:43 |
| noonedeadpunk | so your openstack_user_config for bare metal may look like that: https://paste.openstack.org/show/byPvye0l04ZeNzWeYfk8/ | 13:44 |
| Losraio | Right | 13:44 |
| noonedeadpunk | You can fully skip defining provider_networks, but then you need to define neutron_provider_networks in user_variables, which is way more clear IMO https://opendev.org/openstack/openstack-ansible-os_neutron/src/branch/master/defaults/main.yml#L386-L397 | 13:46 |
| Losraio | Let's see how that goes | 13:46 |
| noonedeadpunk | Just don't forget to drop /etc/openstack_deploy/openstack_inventory.json if you're switching to bare metal from containers | 13:46 |
| noonedeadpunk | But don't drop it if containers are already created | 13:47 |
| Losraio | Well, now that I have messed up so much I might as well clean install Ubuntu on my VMs and try a bare metal deployment | 13:47 |
| noonedeadpunk | :D | 13:47 |
| Losraio | xD | 13:47 |
| noonedeadpunk | I don't think you've messed to much | 13:48 |
| noonedeadpunk | /etc/hosts will be cleaned if you run setup-hosts.yml playbook | 13:48 |
| noonedeadpunk | As it's dynamicaly generated based on the inventory | 13:48 |
| Losraio | Now I'm running the bare metal deployment config that you recommened | 13:49 |
| Losraio | let's see how long it will take until it crashes | 13:49 |
| Losraio | Huh | 13:52 |
| Losraio | Nope, it didn't work | 14:06 |
| Losraio | Oh well, I will definitely reset these Vms | 14:06 |
| Losraio | Could someone please paste the page where the bare metal option is described because I can't find it? | 14:07 |
| jrosser | Losraio: https://docs.openstack.org/openstack-ansible/latest/reference/inventory/configure-inventory.html#deploying-directly-on-hosts | 14:13 |
| Losraio | Thanks a lot | 14:13 |
| jrosser | kind of a good example of when there is not "one right answer" | 14:14 |
| jrosser | `no_containers` is an ansible variable, and you can apply that to one host, or the whole deployment depending what you want to happen | 14:15 |
| jrosser | that doc shows applying it at the level of a host or a component on a host | 14:15 |
| jrosser | but equally you can put it in `global_overrides` of your `openstack_user_config.yml` to apply to the whole deployment | 14:16 |
| Losraio | I see | 14:20 |
| Losraio | So, if I do that then I don't need to declare a network in that file? | 14:20 |
| noonedeadpunk | Yes | 14:27 |
| Losraio | Hmm, I see | 14:30 |
| Losraio | It seems like I have to do a lot more reading | 14:31 |
| Losraio | Thank you both so much though | 14:31 |
| noonedeadpunk | no worries | 14:38 |
| opendevreview | Merged openstack/ansible-role-python_venv_build stable/yoga: Drop empty elements from constraint/requirement files https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/874735 | 14:40 |
| noonedeadpunk | #startmeeting openstack_ansible_meeting | 15:00 |
| opendevmeet | Meeting started Tue Mar 7 15:00:19 2023 UTC and is due to finish in 60 minutes. The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot. | 15:00 |
| opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 15:00 |
| opendevmeet | The meeting name has been set to 'openstack_ansible_meeting' | 15:00 |
| noonedeadpunk | #topic rollcall | 15:00 |
| noonedeadpunk | o/ | 15:00 |
| ebbex | o/ | 15:02 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-rabbitmq_server stable/xena: Restore integrated jobs https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/876748 | 15:03 |
| noonedeadpunk | #topic office hours | 15:04 |
| noonedeadpunk | We don't have any new reported bugs for discussion | 15:04 |
| NeilHanlon | o/ heyo | 15:04 |
| noonedeadpunk | But we're having broken stables branches due to rabbit and changes to the neutron tempest plugin | 15:05 |
| noonedeadpunk | Also this becomes harder to sort out due to upgrade jobs that require to fix prior branches as well | 15:05 |
| noonedeadpunk | So we can't merge Xena fix due to Wallaby that is currently under EM | 15:05 |
| noonedeadpunk | I'm going to push changes to W now, but I assume that prior ones are also broken. | 15:06 |
| ebbex | probably | 15:06 |
| noonedeadpunk | With that I do have a proposal to remove upgrade jobs from releases that are already on EM | 15:06 |
| noonedeadpunk | Which would mean we should stop testing from Wallaby to Xena and all prior ones as of today | 15:07 |
| noonedeadpunk | As EM basically means community support, so it should not blocking actively supported ones from my perspective | 15:08 |
| ebbex | Yep, is victoria in em aswell? | 15:08 |
| noonedeadpunk | Yep | 15:08 |
| noonedeadpunk | V, U, W.... | 15:09 |
| ebbex | ah, then disable upgrade-jobs from EM seems fair | 15:09 |
| noonedeadpunk | lol, sorry, U, V, W ofc | 15:10 |
| noonedeadpunk | but you got the gist :) | 15:10 |
| ebbex | yup | 15:10 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server master: Add support for haproxy map files https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/876749 | 15:11 |
| noonedeadpunk | At the same time I will try to fix what is possible to fix with little blood | 15:11 |
| jrosser | o/ hello | 15:11 |
| noonedeadpunk | Next thing - we should pick up time for the PTG. | 15:14 |
| noonedeadpunk | I wonder how much time we need. As it feels we have plenty of things going on, but not sure if we have much to discuss on them rather then keep working on implementation | 15:15 |
| jrosser | getting these haproxy changed landed is enough maybe | 15:15 |
| noonedeadpunk | Well. How about breaking smth last minute? :D | 15:16 |
| noonedeadpunk | But eventually I'm also thinking about covering systemd bug when services are not restarted | 15:17 |
| noonedeadpunk | And new mariadb LTS | 15:17 |
| noonedeadpunk | And quorum queues.... | 15:17 |
| noonedeadpunk | But I'm fine cutting smth out | 15:18 |
| noonedeadpunk | So maybe we should be fine withing 3 hours? Alss it seems there's no operator hours this year | 15:19 |
| noonedeadpunk | So I'm thinking 15-17 UTC? | 15:20 |
| jrosser | feels like none of the things you list are totally unknown | 15:21 |
| jrosser | so thats good | 15:21 |
| noonedeadpunk | We have quite uknown which is modular libvirt... | 15:22 |
| noonedeadpunk | But things seems to be working as of today, so maybe there's no rush... | 15:22 |
| noonedeadpunk | Ah, lol, and PKI to use pipes. But realistically - there's no time for that | 15:23 |
| jrosser | tbh i am the most concerned about the haproxy changes | 15:24 |
| jrosser | and not making a huge mess | 15:24 |
| noonedeadpunk | ++ | 15:24 |
| jrosser | i would like to merge the stepca stuff soon so we have some coverage in CI for all certbot stuff | 15:24 |
| jrosser | and then 876749 is the starting point for a bunch more tidyness | 15:25 |
| noonedeadpunk | Ok, improving coverage is always good thing to do | 15:27 |
| noonedeadpunk | So, should we vote on vPTG time ? :) | 15:28 |
| noonedeadpunk | March 28, 15-17 UTC? | 15:29 |
| noonedeadpunk | or any alternative proposals? | 15:29 |
| jrosser | thats ok for me | 15:29 |
| ebbex | same | 15:30 |
| noonedeadpunk | ok, great | 15:30 |
| noonedeadpunk | I'll try to invest some more time into haproxy stuff later this week or early next one | 15:31 |
| noonedeadpunk | We're also in shortage of reviewers right now, so all cores are welcome to do some | 15:32 |
| noonedeadpunk | #link http://bit.ly/osa-review-board-v4_1 | 15:33 |
| Elnaz | running `ansible-playbook site.yml $USER_VARS`, It encounters with these errors: | 15:43 |
| Elnaz | `ansible.errors.AnsibleUndefinedVariable: 'ansible_memtotal_mb' is undefined` | 15:43 |
| Elnaz | `ansible.errors.AnsibleUndefinedVariable: {{ (((ansible_memtotal_mb | int) * 0.35) // 1) | int }}: 'ansible_memtotal_mb' is undefined` | 15:44 |
| Elnaz | that seems something has not been loaded! | 15:44 |
| jrosser | Elnaz: that might mean that facts have not been gathered | 15:47 |
| noonedeadpunk | #endmeeting | 15:49 |
| opendevmeet | Meeting ended Tue Mar 7 15:49:31 2023 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 15:49 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/openstack_ansible_meeting/2023/openstack_ansible_meeting.2023-03-07-15.00.html | 15:49 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/openstack_ansible_meeting/2023/openstack_ansible_meeting.2023-03-07-15.00.txt | 15:49 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/openstack_ansible_meeting/2023/openstack_ansible_meeting.2023-03-07-15.00.log.html | 15:49 |
| Elnaz | (I was disconnected for seconds) | 15:54 |
| Elnaz | I'm running it as described in the README file: | 15:54 |
| Elnaz | `cd /opt/openstack-ansible-ops/elk_metrics_7x` | 15:54 |
| Elnaz | `ansible-playbook site.yml $USER_VARS` | 15:55 |
| Elnaz | "If required add `-e@/opt/openstack-ansible/inventory/group_vars/all/all.yml`" | 15:55 |
| Elnaz | jrosser: Am I missing something? | 15:55 |
| Elnaz | full log: https://paste.ubuntu.ir/xgsre | 15:58 |
| Elnaz | Are the conversations here on IRC logged somewhere? | 15:59 |
| noonedeadpunk | Elnaz: they are - https://meetings.opendev.org/irclogs/%23openstack-ansible/ | 16:00 |
| Elnaz | Great | 16:01 |
| jrosser | the error is "ansible_memtotal_mb" | 16:03 |
| jrosser | is undefined | 16:03 |
| jrosser | and thats probably because of "Refresh minimal facts" | 16:04 |
| jrosser | ^ not including that | 16:04 |
| jrosser | Elnaz: can you edit this line https://opendev.org/openstack/openstack-ansible-ops/src/branch/master/elk_metrics_7x/roles/elastic_data_hosts/tasks/main.yml#L18 | 16:06 |
| jrosser | and add "hardware" to that list | 16:06 |
| jrosser | and try again | 16:06 |
| Elnaz | yes | 16:14 |
| Elnaz | Passed: | 16:18 |
| Elnaz | `TASK [elasticsearch : Drop jvm conf file(s)] ***********************************` | 16:18 |
| Elnaz | `changed: [logging1_elastic-logstash_container-8c1f4468] => (item={'src': 'jvm.options.j2', 'dest': '/etc/elasticsearch/jvm.options'})` | 16:18 |
| Elnaz | jrosser: ^ | 16:19 |
| jrosser | Elnaz: would you like to make a patch to fix that? | 16:20 |
| Elnaz | Do you mean like sending a PR on GitHub? | 16:23 |
| Elnaz | jrosser: It's odd! I can't find any login button in https://opendev.org/ | 16:27 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Add releasenote for ANSIBLE_INJECT_FACT_VARS defaulting to false https://review.opendev.org/c/openstack/openstack-ansible/+/876764 | 16:28 |
| jrosser | Elnaz: it's a bit like github, but also a bit different - there is documentation here https://docs.openstack.org/contributors/code-and-documentation/quick-start.html | 16:29 |
| noonedeadpunk | Elnaz: but in short login button is here https://review.opendev.org/ | 16:36 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Use certbot to generate SSL cert for the external VIP in 'stepca' scenario https://review.opendev.org/c/openstack/openstack-ansible/+/876639 | 16:36 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Deploy step-ca when 'stepca' is part of the deployment scenario. https://review.opendev.org/c/openstack/openstack-ansible/+/876637 | 16:59 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Add a /etc/hosts entry for the external IP of an AIO https://review.opendev.org/c/openstack/openstack-ansible/+/876638 | 16:59 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Use certbot to generate SSL cert for the external VIP in 'stepca' scenario https://review.opendev.org/c/openstack/openstack-ansible/+/876639 | 16:59 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible stable/zed: Fix comment typo in nova install playbook https://review.opendev.org/c/openstack/openstack-ansible/+/876678 | 17:08 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible stable/yoga: Fix comment typo in nova install playbook https://review.opendev.org/c/openstack/openstack-ansible/+/876679 | 17:08 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible stable/xena: Fix comment typo in nova install playbook https://review.opendev.org/c/openstack/openstack-ansible/+/876680 | 17:08 |
| jrosser | noonedeadpunk: have you done "add compute node" recently - like this https://docs.openstack.org/openstack-ansible/latest/admin/scale-environment.html#add-a-compute-host | 17:09 |
| noonedeadpunk | Nope, but I'm right now re-orginizing our internal stuff to make use of it | 17:09 |
| jrosser | getting a bit of this https://paste.opendev.org/show/bqFfHPztJ5CrA3oFD6Ql/ | 17:10 |
| noonedeadpunk | hm. Is `ansible_local` smth potentially related to us disabling everything except ansible_facts? | 17:12 |
| noonedeadpunk | Not sure if we merged that or no... | 17:12 |
| jrosser | this is on Z | 17:14 |
| noonedeadpunk | So I assume it's failing somewhere here? https://opendev.org/openstack/openstack-ansible/src/branch/master/playbooks/os-nova-install.yml#L52-L56 | 17:17 |
| jrosser | yeah right there | 17:17 |
| jrosser | i find that code quite obscure | 17:18 |
| noonedeadpunk | It hasn't changed for ages | 17:18 |
| noonedeadpunk | And we leverage it to proceed with migrations iirc | 17:19 |
| noonedeadpunk | So I assume by that time we should have facts for `groups['nova_all']` | 17:20 |
| Elnaz | jrosser: I created an account but It takes me a while to get familiar with this system. If possible, please edit it yourself this time. | 17:20 |
| noonedeadpunk | Obviously, that new compute should not have that.... | 17:20 |
| jrosser | what i'm surprised by is that ansible_local appears in the facts cache for the new compute node | 17:20 |
| jrosser | not for localhost | 17:20 |
| noonedeadpunk | and compute1a01 is exactly the new compute? | 17:20 |
| jrosser | so i think i misunderstand how this is supposed to work | 17:20 |
| jrosser | yes thats right | 17:21 |
| jrosser | oh | 17:21 |
| jrosser | so it delegates to `localhost` but accesses the hostvars of everything else from there | 17:21 |
| jrosser | i wonder if there is an interaction with --limit here | 17:23 |
| noonedeadpunk | eventually, I've checked what our ops are doing, and they're doing exactly the same | 17:23 |
| jrosser | regarding implicit localhost | 17:23 |
| noonedeadpunk | but we're still on X | 17:23 |
| noonedeadpunk | openstack-ansible playbooks/setup-hosts.yml --limit "localhost,$limit_hosts" && openstack-ansible playbooks/setup-openstack.yml --limit "localhost,$limit_hosts" | 17:24 |
| opendevreview | Merged openstack/ansible-role-systemd_service master: Fix tags usage for included tasks https://review.opendev.org/c/openstack/ansible-role-systemd_service/+/876321 | 17:25 |
| noonedeadpunk | So I might assume smth has changed in the meanwhile | 17:26 |
| noonedeadpunk | with ansible-core versions. But I'm not sure why it's not happy, as we have both localhost and compute in play. And that works without limits... | 17:26 |
| noonedeadpunk | (I assume) | 17:27 |
| jrosser | ok, will dig some more | 17:27 |
| noonedeadpunk | Unless... It tires to resolve that for everything that is not part of limit | 17:28 |
| noonedeadpunk | And fail there.... | 17:28 |
| jrosser | we've not done this for a very long time, like just add one host | 17:28 |
| jrosser | right `Implicit localhost does not appear in the hostvars magic variable unless demanded, such as by "{{ hostvars['localhost'] }}"` | 17:28 |
| jrosser | according to https://docs.ansible.com/ansible/latest/inventory/implicit_localhost.html | 17:28 |
| noonedeadpunk | We did added couple of computes like that week ago or so, but again on X | 17:28 |
| noonedeadpunk | Yes, but we're checking not localhost hostvars | 17:29 |
| noonedeadpunk | But we're trying to fetch hostvars from each member of groups['nova_all'] | 17:29 |
| noonedeadpunk | I wonder if refreshing facts for nova_all will just help... | 17:29 |
| noonedeadpunk | As this reffer to `/etc/ansible/facts.d/openstack_ansible.fact` content on computes | 17:31 |
| noonedeadpunk | or well, everywhere | 17:31 |
| jrosser | yeah | 17:31 |
| noonedeadpunk | Oh, huh, so by this time every compute should have that | 17:31 |
| jrosser | it does :) | 17:31 |
| noonedeadpunk | even new one | 17:31 |
| jrosser | thats the wierd thing | 17:31 |
| noonedeadpunk | and we collect facts right before as well | 17:32 |
| noonedeadpunk | Yeah, no idea | 17:32 |
| jrosser | no worries - will investigate | 17:33 |
| prometheanfire | ya, may have to | 17:35 |
| opendevreview | Merged openstack/openstack-ansible-haproxy_server master: Fix tags usage for letsencrypt setup https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/875772 | 17:40 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server stable/zed: Fix tags usage for letsencrypt setup https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/876681 | 17:42 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server stable/yoga: Fix tags usage for letsencrypt setup https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/876682 | 17:42 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server stable/xena: Fix tags usage for letsencrypt setup https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/876683 | 17:42 |
| opendevreview | Merged openstack/openstack-ansible master: Remove support for calico ml2 driver. https://review.opendev.org/c/openstack/openstack-ansible/+/866119 | 18:03 |
| opendevreview | Merged openstack/openstack-ansible-haproxy_server master: Serialise initial issuing of LetsEncrypt certificates https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/875781 | 18:12 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server stable/zed: Serialise initial issuing of LetsEncrypt certificates https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/876684 | 18:12 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server stable/yoga: Serialise initial issuing of LetsEncrypt certificates https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/876685 | 18:16 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server stable/xena: Serialise initial issuing of LetsEncrypt certificates https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/876686 | 18:16 |
| opendevreview | Merged openstack/openstack-ansible master: Disable fact variables https://review.opendev.org/c/openstack/openstack-ansible/+/778396 | 18:16 |
| opendevreview | Merged openstack/openstack-ansible master: Add validation zuul job for hosts setup https://review.opendev.org/c/openstack/openstack-ansible/+/843002 | 18:16 |
| opendevreview | Merged openstack/openstack-ansible master: Bump OpenStack-Ansible master https://review.opendev.org/c/openstack/openstack-ansible/+/876043 | 18:26 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Deploy step-ca when 'stepca' is part of the deployment scenario. https://review.opendev.org/c/openstack/openstack-ansible/+/876637 | 19:08 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Add a /etc/hosts entry for the external IP of an AIO https://review.opendev.org/c/openstack/openstack-ansible/+/876638 | 19:09 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Use certbot to generate SSL cert for the external VIP in 'stepca' scenario https://review.opendev.org/c/openstack/openstack-ansible/+/876639 | 19:09 |
| opendevreview | Merged openstack/openstack-ansible-haproxy_server stable/yoga: Use let's encrypt standalone flag only for http-01 https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/874620 | 20:25 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!