| *** Daemoen is now known as demone | 00:00 | |
| *** demone is now known as mmercer | 00:00 | |
| *** macza has quit IRC | 00:14 | |
| *** hwoarang has quit IRC | 00:15 | |
| *** hwoarang has joined #openstack-ansible | 00:17 | |
| *** spatel has joined #openstack-ansible | 00:17 | |
| *** spatel has quit IRC | 00:22 | |
| *** rodolof has quit IRC | 00:43 | |
| *** rodolof has joined #openstack-ansible | 00:44 | |
| *** tosky has quit IRC | 00:55 | |
| *** rodolof has quit IRC | 01:04 | |
| *** rodolof has joined #openstack-ansible | 01:04 | |
| *** gyee has quit IRC | 01:08 | |
| *** nurdie has joined #openstack-ansible | 01:11 | |
| *** vnogin has joined #openstack-ansible | 01:26 | |
| *** vnogin has quit IRC | 01:30 | |
| *** rodolof has quit IRC | 01:30 | |
| *** tinwood has quit IRC | 02:10 | |
| *** tinwood has joined #openstack-ansible | 02:11 | |
| *** cshen has joined #openstack-ansible | 02:36 | |
| *** cshen has quit IRC | 02:40 | |
| *** DanyC has joined #openstack-ansible | 02:40 | |
| *** DanyC has quit IRC | 02:45 | |
| *** maharg101 has quit IRC | 02:49 | |
| *** nurdie has quit IRC | 02:50 | |
| *** nurdie has joined #openstack-ansible | 02:52 | |
| *** nurdie_ has joined #openstack-ansible | 02:53 | |
| *** spatel has joined #openstack-ansible | 02:56 | |
| *** nurdie has quit IRC | 02:56 | |
| *** hw_wutianwei_ has joined #openstack-ansible | 03:17 | |
| *** fnpanic has quit IRC | 03:28 | |
| *** spatel has quit IRC | 03:39 | |
| *** nurdie_ has quit IRC | 04:47 | |
| *** nurdie has joined #openstack-ansible | 04:48 | |
| *** udesale has joined #openstack-ansible | 04:49 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/rocky: Update all SHAs for 18.1.2 https://review.openstack.org/624407 | 05:37 |
|---|---|---|
| *** cshen has joined #openstack-ansible | 05:51 | |
| *** cshen has quit IRC | 05:56 | |
| *** nurdie_ has joined #openstack-ansible | 06:51 | |
| *** nurdie__ has joined #openstack-ansible | 06:53 | |
| *** nurdie has quit IRC | 06:54 | |
| *** nurdie_ has quit IRC | 06:55 | |
| *** pcaruana has joined #openstack-ansible | 07:12 | |
| *** nurdie has joined #openstack-ansible | 07:15 | |
| *** nurdie has quit IRC | 07:17 | |
| *** nurdie has joined #openstack-ansible | 07:18 | |
| *** nurdie__ has quit IRC | 07:18 | |
| *** nurdie has quit IRC | 07:22 | |
| *** cshen has joined #openstack-ansible | 07:22 | |
| *** kopecmartin|off is now known as kopecmartin | 07:35 | |
| *** shardy has joined #openstack-ansible | 07:55 | |
| *** maharg101 has joined #openstack-ansible | 08:33 | |
| *** nurdie has joined #openstack-ansible | 08:48 | |
| *** nurdie has quit IRC | 08:54 | |
| *** cshen has quit IRC | 09:06 | |
| *** cshen has joined #openstack-ansible | 09:06 | |
| *** trident has quit IRC | 09:16 | |
| *** DanyC has joined #openstack-ansible | 09:17 | |
| *** trident has joined #openstack-ansible | 09:19 | |
| *** DanyC has quit IRC | 09:21 | |
| *** dcdamien has joined #openstack-ansible | 09:29 | |
| *** DanyC has joined #openstack-ansible | 09:33 | |
| *** tosky has joined #openstack-ansible | 09:37 | |
| *** markvoelker has joined #openstack-ansible | 09:46 | |
| noonedeadpunk | spotz: yeah, API has been removed, but it still collects data and sends it to gnocchi. I've filled in a bug https://bugs.launchpad.net/ceilometer/+bug/1801348 and placed a patch for it before Berlin summit, and try to ping in IRC at least once a week, but never saw anyone there | 09:56 |
| openstack | Launchpad bug 1801348 in Ceilometer "notification agent fails with KeyError on volume create/delete" [Undecided,In progress] - Assigned to Dmitriy Rabotjagov (noonedeadpunk) | 09:56 |
| *** rodolof has joined #openstack-ansible | 09:58 | |
| *** rgogunskiy has quit IRC | 10:03 | |
| noonedeadpunk | Actually, it's smth, that might be fixed by OSA like with https://review.openstack.org/#/c/623239/ but if the project is really dead, it'll look like project maintaining. | 10:04 |
| *** electrofelix has joined #openstack-ansible | 10:04 | |
| *** trident has quit IRC | 10:15 | |
| *** trident has joined #openstack-ansible | 10:15 | |
| *** ppetit has joined #openstack-ansible | 10:20 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Chage cpu model to allow for nested virt to work https://review.openstack.org/624780 | 10:34 |
| *** ppetit has quit IRC | 10:41 | |
| Miouge | good morning folks | 10:46 |
| *** nurdie has joined #openstack-ansible | 10:50 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_hosts stable/pike: prep: remove old machinectl workarounds https://review.openstack.org/624674 | 10:51 |
| odyssey4me | o/ Miouge how're things going? | 10:52 |
| *** DanyC has quit IRC | 10:53 | |
| *** DanyC has joined #openstack-ansible | 10:54 | |
| Miouge | odyssey4me: All good, I’ve been testing an OSA Dockerfile :D | 10:54 |
| *** DanyC has joined #openstack-ansible | 10:55 | |
| *** nurdie has quit IRC | 10:55 | |
| *** gisak has joined #openstack-ansible | 11:17 | |
| openstackgerrit | Maxime Guyot proposed openstack/openstack-ansible master: [WIP] Add doc for Alpine deployment host https://review.openstack.org/621127 | 11:27 |
| openstackgerrit | Maxime Guyot proposed openstack/openstack-ansible master: [WIP] Add doc for Alpine deployment host https://review.openstack.org/621127 | 11:28 |
| openstackgerrit | Maxime Guyot proposed openstack/openstack-ansible master: Add doc for Docker Alpine deployment host https://review.openstack.org/621127 | 11:33 |
| *** ansmith has joined #openstack-ansible | 11:43 | |
| odyssey4me | Miouge I'm curious why you didn't just use an Ubuntu/CentOS container base, and chose to use Alpine? | 11:47 |
| Miouge | odyssey4me: If you use Ubuntu as a core image, it’s exactly the same procedure as a standard deploy host, so nothing to document there. | 11:49 |
| odyssey4me | Miouge ah, fair enough | 11:50 |
| *** DanyC has quit IRC | 11:59 | |
| odyssey4me | noonedeadpunk I'd suggest trying to make contact with the PTL for telemetry to find out what's going on - mnaser can likely help with that. | 11:59 |
| *** DanyC has joined #openstack-ansible | 12:00 | |
| *** DanyC has joined #openstack-ansible | 12:00 | |
| noonedeadpunk | odyssey4me: I'll try to, thanks! | 12:02 |
| noonedeadpunk | btw, should I place patch for removing git retrievment for ceilometer, and then add https://review.openstack.org/#/c/623239/ , or place such patch and rebase that on on top of it? | 12:03 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Ensure that control plane hosts are also updated correctly https://review.openstack.org/624965 | 12:04 |
| odyssey4me | noonedeadpunk that one will also backport to queens right? if so - leave as-is and put the one to change it to use the venv on top of that | 12:04 |
| *** django_ is now known as django | 12:05 | |
| odyssey4me | jrosser this is a glaring bug in the upgrade process which will need porting back to rocky: https://review.openstack.org/624965 | 12:05 |
| * jrosser looks | 12:06 | |
| * odyssey4me is looking at http://zuul.openstack.org/builds?project=OpenStack%2Fopenstack-ansible&branch=stable%2Frocky&pipeline=periodic | 12:09 | |
| noonedeadpunk | odyssey4me: I think it shouldn't be backported (even if I'd wish to), as it introduces override and configuration for 2 extra files, so it's kinda new functional. | 12:10 |
| odyssey4me | noonedeadpunk yep, I agree | 12:12 |
| jrosser | odyssey4me: i think somehow i dodged that during my Q->R upgrade | 12:12 |
| jrosser | i have dedicated neutron nodes | 12:12 |
| odyssey4me | jrosser ah, yes, that would have covered it | 12:16 |
| odyssey4me | I think, overall, we really need a better method | 12:17 |
| jrosser | i also did it manually | 12:17 |
| jrosser | too much trust required to just let the script rip and sit back and watch | 12:17 |
| odyssey4me | it's ideal to be able to do a fan-out for all the deployment changes, but then only serialise the restarts - that'll be much, uch faster | 12:17 |
| *** ansmith has quit IRC | 12:17 | |
| odyssey4me | yeah, I agree - I prefer to execute manually, then inspect things after each step to confirm everything is as expected | 12:18 |
| odyssey4me | but meh, some don't | 12:18 |
| jrosser | otoh the script (or equivalent) is needed for CI | 12:18 |
| odyssey4me | yep, that was the original reason it was created | 12:19 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Ensure that control plane hosts are also updated correctly https://review.openstack.org/624965 | 12:21 |
| *** markvoelker has quit IRC | 12:24 | |
| *** udesale has quit IRC | 12:25 | |
| *** udesale has joined #openstack-ansible | 12:26 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server master: upgrade: start service before applying policies https://review.openstack.org/620378 | 12:43 |
| *** mkuf_ has joined #openstack-ansible | 12:45 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_keystone stable/pike: Ensure that LDAP config is deployed on all keystone hosts https://review.openstack.org/624434 | 12:45 |
| gisak | hi guys, one question: why would the public-network's bridge (in my case brq88038979-47) not bridge any interfaces at all ? | 12:49 |
| *** mkuf has quit IRC | 12:49 | |
| *** nurdie has joined #openstack-ansible | 12:51 | |
| *** rodolof has quit IRC | 12:54 | |
| *** rodolof has joined #openstack-ansible | 12:54 | |
| *** nurdie has quit IRC | 12:56 | |
| jamesdenton | gisak you can check linuxbridge_agent.ini to see what the mapping should be. physical_interface_mappings is the option | 12:58 |
| gisak | thnx, in my case I have eno1 as internet access interface and it is bridged with br-vlan, how to manage the correct bridging topology? | 13:02 |
| jamesdenton | how many interfaces are you working with? | 13:02 |
| gisak | 4 on network node | 13:02 |
| jamesdenton | and what did that mapping look like | 13:02 |
| gisak | eno1 is br-vlan, eno2 br-mgmt eno3 br-vxlan and eno4 br-storage | 13:03 |
| gisak | should i delete br-vlan and assign eno1 to physical_interface_mappings from linuxbridge_agent.ini? | 13:04 |
| jamesdenton | ok. what is the current physical_interface_mappings value? | 13:04 |
| spotz | off to dog related stuff back next week! | 13:07 |
| openstackgerrit | Christian Zunker proposed openstack/openstack-ansible-os_neutron master: Set admin url endpoints insecure flag for nova client https://review.openstack.org/624987 | 13:09 |
| gisak | flat:eno1,vlan:br-vlan | 13:10 |
| *** markvoelker has joined #openstack-ansible | 13:10 | |
| gisak | I use flat, in this case i should remove vlan:br-vlan and then delete br-vlan brdige from node and try again ? | 13:11 |
| jamesdenton | well, to get this working now you can set vlan:eno1 and remove eno1 from the br-vlan bridge | 13:11 |
| jamesdenton | but to do it right, you'll need to update the provider networks in openstack_user_config.yml and add a 'host_bind_override: eno1' to the network block: https://docs.openstack.org/project-deploy-guide/openstack-ansible/draft/app-config-prod.html | 13:13 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create master: cleanup: remove rocky-only upgrade code https://review.openstack.org/620343 | 13:18 |
| gisak | in this case, if I set 'host_bind_override: eno1' in openstack_user_config.yml and rerun playbooks, will i still need the br-vlan ? | 13:18 |
| gisak | or br-vlan becomes useless in this case ? | 13:18 |
| jamesdenton | it would be used on infra/network nodes, but not computes | 13:19 |
| jamesdenton | i would recommend leaving itfor now | 13:19 |
| jamesdenton | as in, not deleting it | 13:19 |
| odyssey4me | jamesdenton have you managed to figure out how to get the neutron agents to install on the same network_host when doing the p->q upgrade? | 13:29 |
| jamesdenton | no | 13:29 |
| odyssey4me | I wonder if it wouldn't be as trivial as using the extra var 'on_metal=yes' when running the os-neutron-install playbook | 13:30 |
| jamesdenton | i can try | 13:31 |
| odyssey4me | otherwise it might have to be more subtle - like adding the hosts into the right group | 13:32 |
| *** hamzaachi has joined #openstack-ansible | 13:33 | |
| jamesdenton | yeah, i've noticed that the inventory post-upgrade looks like this: http://paste.openstack.org/show/737214/ | 13:33 |
| jamesdenton | the Network01 host is one i added after the upgrade, and it was added to the neutron_agent group as baremetal. | 13:34 |
| odyssey4me | jamesdenton and neutron_agent is the group that matters here, from a play/task/template targeting standpoint, right? | 13:36 |
| gisak | thank you jamesdenton | 13:37 |
| *** gillesMo has joined #openstack-ansible | 13:37 | |
| gisak | the gateway finally is up, but floating IP's now dont work ( | 13:38 |
| gisak | are unreachable | 13:38 |
| gisak | but the instance has internet | 13:38 |
| *** ansmith has joined #openstack-ansible | 13:38 | |
| jamesdenton | odyssey4me i think so? | 13:40 |
| gisak | btw, in openstack_user_config.yml the host_bind_override is set to eno1 ) | 13:41 |
| jamesdenton | for the vlan block as well as flat? | 13:42 |
| jamesdenton | odyssey4me the bare metal node would also need to be part of the other agent groups (neutron_linuxbridge_agent, ovs, etc) | 13:43 |
| gisak | this is the openstack_user_config.yml I run the playbooks: http://paste.openstack.org/show/737217/ | 13:43 |
| gillesMo | gisak: sorry if I missed something, I'm just arriving ! But, are you sure that you just don't allow traffic via your security groups ? | 13:43 |
| *** fnpanic has joined #openstack-ansible | 13:44 | |
| fnpanic | hi | 13:44 |
| fnpanic | quick question | 13:45 |
| fnpanic | is there any downside of disabling nat for the lxc containers? | 13:45 |
| odyssey4me | jamesdenton I have an idea - going to try it out now. | 13:45 |
| jamesdenton | odyssey4me so it looks like you can use inventory-manage.py -r to remove the container, then run the dynamic inventory script to update the inventory to include the baremetal infra node in the proper groups. But like you said, you would lose access to that container | 13:45 |
| *** masterpe has joined #openstack-ansible | 13:45 | |
| odyssey4me | fnpanic yes, they would lose internet access - unless you've made provision for that another way | 13:46 |
| gillesMo | Has someone tried to "host migrate" a compute node in Horizon (Queens) ? I've tried yesterday and my instances stay in migrating state since then (24h...). I see a ssh connection between two compute nodes, but there is WARNING in my nova-compute logs : | 13:46 |
| jamesdenton | well you don't lose ssh access per say | 13:46 |
| gisak | ok my bad here with sec groups :) neither ping nor ssh access was opened ) | 13:46 |
| odyssey4me | jamesdenton yeah, but we lose the ability to target it via ansible directly as a host in the inventory | 13:46 |
| fnpanic | odyssey4me: thanks. we will use a quid in the mgmt network | 13:46 |
| jamesdenton | odyssey4me but at that point, does it matter? | 13:46 |
| odyssey4me | jamesdenton it does, because we still need to delete it later | 13:47 |
| jamesdenton | lxc-delete. hehe | 13:47 |
| odyssey4me | I mean, we could hack it away - but I think there's possibly a more elegant option - simply have a temporary ini file which adds network_hosts as a child to each of those other groups. | 13:48 |
| gillesMo | Pb migrating instance while evacuate a host (migrate host in Horizon after disabling a compute node)... nova-compute logs shows "Instance not resizing, skipping migration" | 13:48 |
| odyssey4me | this would be something we drop into user space for during the upgrade, then remove after we've removed the containers | 13:49 |
| jamesdenton | so the ini will supplement, or be merged, against the existing inventory? | 13:49 |
| odyssey4me | jamesdenton ansible will merge the two - but the dynamic inventory will know nothing of it | 13:49 |
| jamesdenton | ahh ok | 13:49 |
| odyssey4me | gillesMo it's been a long time since I did any of that - but typically live/cold migrations will not happen if the base image the instances was created from was deleted - because it cannot pull it down to the target host... so you have to copy it over manually into the base cache | 13:52 |
| jamesdenton | odyssey4me alright, just so i understand. The approach could be to see which groups contain existing neutron agent containers, add the network_hosts group to those groups, install the bits, then remove the members of neutron_agents_container group from inventory? | 13:53 |
| odyssey4me | jamesdenton yep, that's basically my thinking | 13:53 |
| odyssey4me | before removing those members, we ensure that the agents are disabled, everything's migrated over, etc | 13:54 |
| *** stuartgr has joined #openstack-ansible | 13:54 | |
| jamesdenton | So you want to automate that, too? | 13:54 |
| odyssey4me | yep, it seems automatable to me | 13:54 |
| gillesMo | odyssey4me: One additional info : I use Ceph everywhere. One of my instance is booted from a volume, the other is from image, which is still here (in Ceph)... | 13:54 |
| jamesdenton | It's a bold strategy, let | 13:55 |
| odyssey4me | whether anyone uses the automation is up to them - but the playbook would act as a guide for the procedure anyway | 13:55 |
| jamesdenton | let's see if it pays off | 13:55 |
| jamesdenton | kk it's optional, then | 13:55 |
| odyssey4me | gillesMo ah ok, then I dunno - live migrations always worked for me then... evacuation wasn't possible back when I operated a cloud ;) | 13:55 |
| *** tosky has quit IRC | 13:56 | |
| odyssey4me | jamesdenton it would be included in run-upgrade, but every part of run-upgrade can be run in the individual steps too - as preferred by the operator | 13:56 |
| jamesdenton | #yolo | 13:57 |
| odyssey4me | :) | 13:58 |
| fnpanic | so aio is broken currently? | 13:58 |
| fnpanic | ? | 13:58 |
| odyssey4me | fnpanic what do you mean? | 13:58 |
| jamesdenton | so in the inventory... children and hosts... children simply refers to groups while hosts is individual machines? | 13:59 |
| odyssey4me | fnpanic the AIO is used daily to merge patches, so no - it's not broken | 13:59 |
| odyssey4me | jamesdenton yep | 14:00 |
| *** tosky has joined #openstack-ansible | 14:00 | |
| *** irclogbot_0 has quit IRC | 14:00 | |
| *** markvoelker has quit IRC | 14:03 | |
| fnpanic | odyssey4me: ok | 14:05 |
| fnpanic | i was following the guide for quick aio | 14:06 |
| fnpanic | and it was not working | 14:06 |
| fnpanic | strange | 14:06 |
| odyssey4me | fnpanic you'll have to be much more specific - what distro, where did it break? | 14:06 |
| *** irclogbot_0 has joined #openstack-ansible | 14:07 | |
| odyssey4me | also, which specific guide are you following (master/latest, or rocky, or older?) | 14:08 |
| jamesdenton | odyssey4me New update to https://bugs.launchpad.net/openstack-ansible/+bug/1804770 confirms what we saw re: not migration to BM during upgrade. And another bug: https://bugs.launchpad.net/openstack-ansible/+bug/1785592. Just FYI | 14:09 |
| openstack | Launchpad bug 1804770 in openstack-ansible "Pike -> Queens Upgrade: Documentation for OVS setup and neutron agent rebalancing missing" [High,Confirmed] - Assigned to James Denton (james-denton) | 14:09 |
| openstack | Launchpad bug 1785592 in openstack-ansible "dynamic inventory doesn't handle is_metal: false->true change" [High,Confirmed] - Assigned to Kevin Carter (kevin-carter) | 14:09 |
| *** mkuf has joined #openstack-ansible | 14:10 | |
| fnpanic | odyssey4me: ubuntu 1804 with rocky | 14:10 |
| odyssey4me | jamesdenton ok, I've assigned that other one to me and will have a go at it to see if it works | 14:11 |
| jamesdenton | right on | 14:11 |
| odyssey4me | otherwise the other option is to remove the containers from the inventory as we discussed, but I suspect that'll be more hacky and prone to failure | 14:11 |
| odyssey4me | also possibly more disruptive | 14:12 |
| *** mkuf_ has quit IRC | 14:13 | |
| *** irclogbot_0 has quit IRC | 14:14 | |
| jamesdenton | one might require a lot less work :D | 14:16 |
| *** irclogbot_0 has joined #openstack-ansible | 14:23 | |
| *** thuydang has joined #openstack-ansible | 14:23 | |
| *** rodolof has quit IRC | 14:32 | |
| odyssey4me | jamesdenton bother - the only imaged build of pike I have is RPC-O, which already has the env.d file from queens as an override, so I can't test this | 14:32 |
| *** rodolof has joined #openstack-ansible | 14:32 | |
| jamesdenton | are there some changes you want me to look at? I can redeploy Pike today | 14:33 |
| odyssey4me | jamesdenton oh no - if that's your plan, then I can just do an AIO now | 14:33 |
| *** nurdie has joined #openstack-ansible | 14:33 | |
| jamesdenton | Will an AIO tell the whole story, though? Won't the AIO already be baremetal since it runs compute? | 14:34 |
| odyssey4me | well, I need to validate just ansible and the inventory behaviour before I go ahead and build a patch | 14:34 |
| jamesdenton | k | 14:34 |
| odyssey4me | I'll fire up a fresh MNAIO of pike too for testing the patch later. | 14:35 |
| *** markvoelker has joined #openstack-ansible | 14:36 | |
| *** nurdie has quit IRC | 14:39 | |
| *** hamzaachi has quit IRC | 14:39 | |
| ansmith | odyssey4me: if you have a minute, could you take glance at https://review.openstack.org/#/c/624184/ | 14:40 |
| ansmith | odyssey4me: I started a review from the top for the qdrouterd integration, in the experimental job, no hosts matched for creating the qdrouterd container | 14:40 |
| ansmith | odyssey4me: http://logs.openstack.org/84/624184/4/experimental/openstack-ansible-deploy-aio_rpc-qdrouterd-ubuntu-bionic/d07c9c3/job-output.txt.gz | 14:41 |
| ansmith | odyssey4me: not quite sure what I am missing | 14:41 |
| *** priteau has joined #openstack-ansible | 14:42 | |
| odyssey4me | ansmith ok, let me finish a thread here and I'll take a look | 14:43 |
| ansmith | odyssey4me: thanks | 14:44 |
| *** cshen has quit IRC | 14:51 | |
| fnpanic | so the aio guide is missing something? | 15:12 |
| fnpanic | i am trying it on ubuntu 1804 with rocky | 15:12 |
| jrosser | fnpanic: is your question about the AIO also related to wanting to disable nat on the containers? | 15:14 |
| odyssey4me | fnpanic you still haven't said where it's failing for you or what broke when you tried to follow the intructions there | 15:15 |
| odyssey4me | fnpanic if you could share a transcript of any error output or the commands you ran or *anything* informative in a paste service or gist or something, that'd be helpful | 15:16 |
| *** ThiagoCMC has joined #openstack-ansible | 15:18 | |
| ThiagoCMC | cloudnull, hey man, are you around? | 15:19 |
| odyssey4me | ThiagoCMC cloudnull is on holiday until next year afaik | 15:19 |
| ThiagoCMC | Oh no!! LOL | 15:19 |
| ThiagoCMC | I'm seeing a problem with Ansible, when running it against Ubuntu LXD containers and, after googling about it, I can see that cloudnull faced this problem too but, I don't know how to fix it... | 15:20 |
| ThiagoCMC | The thing is... | 15:20 |
| ThiagoCMC | When running Ansible against an Ubuntu LXD or nspawn container, the Ansible APT module fails to install packages, it shows state "ok:" but, it does nothing! | 15:21 |
| ThiagoCMC | Are you guys aware of this? | 15:21 |
| ThiagoCMC | I tried Ansible 2.5 from Ubuntu 18.04, and 2.7 from their PPA. | 15:21 |
| ThiagoCMC | Both fails... | 15:21 |
| jrosser | i do ansible + lxd for non osa stuff and it works ok | 15:21 |
| ThiagoCMC | But which OS inside of the container? | 15:22 |
| ThiagoCMC | Ubuntu? | 15:22 |
| jrosser | yes | 15:22 |
| ThiagoCMC | Ok, cool! | 15:22 |
| ThiagoCMC | I have no idea about what's happening here... :-( | 15:22 |
| jrosser | i set the lxd containers up to look just like hosts, no lxdbr nat or anything, they get a static IP when they're created | 15:22 |
| jrosser | and regular ansible + ssh | 15:22 |
| odyssey4me | jrosser mnaser happy with https://review.openstack.org/624965 ? I'd like to get that ported back asap. | 15:23 |
| ThiagoCMC | I have the very same RAM image, that if I boot on KVM, Ansible against it works, if I boot it with nspawn, it doesn't. | 15:23 |
| ThiagoCMC | jrosser, mtacvlan? | 15:23 |
| jrosser | no, bridge | 15:23 |
| ThiagoCMC | I mean, macvlan | 15:23 |
| ThiagoCMC | ok | 15:23 |
| ThiagoCMC | sorry, auto corrector... =P | 15:23 |
| ThiagoCMC | I'll test it again today. | 15:24 |
| ThiagoCMC | Thank you! | 15:24 |
| mnaser | odyssey4me: +2 | 15:25 |
| jrosser | i just kept it super simple, bridges and no lxd nat, static IP | 15:25 |
| *** weezS has joined #openstack-ansible | 15:26 | |
| ThiagoCMC | yes, I'm trying to keep it KISS. | 15:27 |
| ThiagoCMC | But macvlan instead of bridge, container have IP from my LAN, even ipv6. | 15:28 |
| ThiagoCMC | I can ssh into container, run apt update and etc | 15:28 |
| ThiagoCMC | it's fine | 15:28 |
| ThiagoCMC | Just Ansible APT isn't working. | 15:28 |
| ThiagoCMC | Other tasks works. | 15:28 |
| ThiagoCMC | very weird. | 15:28 |
| odyssey4me | ThiagoCMC in that case, why bother with the lxd connection plugin - just treat it as a host | 15:28 |
| ThiagoCMC | I'm using it as if it was a regular server, via SSH, not lxd connection thing. | 15:29 |
| odyssey4me | ok, perhaps you're hitting an issue relating to ansible_tmp - I seem to recall something along those lines being an issue | 15:30 |
| ThiagoCMC | There is an issue, definitely! =P | 15:30 |
| ThiagoCMC | I remember about ansible_tmp as well! | 15:31 |
| ThiagoCMC | It isn't the first time that I'm seeing this as well. | 15:31 |
| ThiagoCMC | Last time was in 2016 | 15:31 |
| ThiagoCMC | I thought that it wasn't an issue anymore... | 15:31 |
| ThiagoCMC | I'll try to change the ansible_tmp! | 15:31 |
| *** cshen has joined #openstack-ansible | 15:32 | |
| openstackgerrit | weizj proposed openstack/openstack-ansible-os_heat stable/queens: Replace Chinese punctuation with English punctuation https://review.openstack.org/625027 | 15:34 |
| openstackgerrit | weizj proposed openstack/openstack-ansible-ceph_client stable/queens: Replace Chinese punctuation with English punctuation https://review.openstack.org/625028 | 15:34 |
| openstackgerrit | weizj proposed openstack/openstack-ansible-rsyslog_client stable/queens: Replace Chinese punctuation with English punctuation https://review.openstack.org/625029 | 15:34 |
| jamesdenton | mnaser you still out doing the Kube thing? | 15:34 |
| openstackgerrit | weizj proposed openstack/openstack-ansible-repo_build stable/queens: Replace Chinese punctuation with English punctuation https://review.openstack.org/625030 | 15:34 |
| *** cshen has quit IRC | 15:36 | |
| *** cshen has joined #openstack-ansible | 15:37 | |
| *** markvoelker has quit IRC | 15:38 | |
| mgariepy | anyone here knows someone that works on ubuntu kernel team ? | 15:43 |
| mgariepy | i'm having quite a fun issue with the 4.15.0 kernel.. | 15:44 |
| odyssey4me | mgariepy not personally, but tinwood or jamespage might be able to point you in the right direction | 15:44 |
| odyssey4me | otherwise pop into #ubuntu/#ubuntu-devel and ask | 15:45 |
| mgariepy | i'm in #ubuntu-kernel and was in -virt before that, but seems to be more kernel related than virt.. | 15:46 |
| odyssey4me | ah ok :/ | 15:46 |
| *** hamzaachi has joined #openstack-ansible | 15:46 | |
| mnaser | jamesdenton: yes, I’m heading back today at night (it’s 7 am here) but can I help with anything ? | 15:47 |
| jamesdenton | naw, it's just been awful quiet in here :) | 15:47 |
| ioni | mgariepy, what problem do you have with 4.15? | 15:53 |
| ioni | i'm just curious because i'm on that as well on most of compute nodes | 15:54 |
| mgariepy | i have pci passthrough for some gamer pci video card | 15:55 |
| mgariepy | and i spawn 120G vms with 16 core and 4 titanx card . | 15:56 |
| mgariepy | when i boot the vm the ram is pre-allocated and it goes fast for about 50% then is slows down | 15:56 |
| mgariepy | on the 4.13 kernel i don't have any issues | 15:57 |
| ioni | does the qemu version changes? | 15:57 |
| ioni | or only the kernel | 15:57 |
| mgariepy | only changing the kernel | 15:58 |
| ioni | maybe it's due to l1tf | 15:58 |
| ioni | did you tried to disable the protection? | 15:58 |
| mgariepy | i did try an older version of qemu and libvirt but with the 4.15 kernel i have the issue. | 16:00 |
| mgariepy | looking into it. | 16:00 |
| pabelanger | Yay, stable/rocky deployed from zuul: https://object-storage-ca-ymq-1.vexxhost.net/swift/v1/a0b4156a37f9453eb4ec7db5422272df/logs/4c/4c91f44a55cd37a5c80ddf7e147ba0252465195c/post/packet-ci-cloud-deploy/0aa4dbb/logs/ara-report/ | 16:02 |
| odyssey4me | jamesdenton it looks like this inventory.ini puts all the network_hosts into the right groups - allowing us to use automation to get everything done. :) https://gist.github.com/odyssey4me/5e9df7ebce1eff5e9321bfbaf6959a3f | 16:02 |
| odyssey4me | pabelanger :) yay! | 16:02 |
| ioni | mgariepy, https://www.kernel.org/doc/html/latest/admin-guide/l1tf.html | 16:03 |
| jamesdenton | odyssey4me genius | 16:05 |
| ioni | hmm | 16:05 |
| ioni | this sounds something i need for P->Q | 16:06 |
| odyssey4me | jamesdenton ok, so now I'll wait for my pike deploy to complete - then image it - and get on with building out the actual automation, but I think this has a fighting chance of working | 16:07 |
| odyssey4me | ioni right now it's something that has to be done by hand, but thanks to some inspiration from jamesdenton and two bug reports, I think we can automate it | 16:08 |
| odyssey4me | FYI jamesdenton - I updated the gist with more complete instructions | 16:09 |
| jamesdenton | odyssey4me let me know how i can help. i will reimage multinode w/ Pike and upgrade when you have the first set of patches | 16:09 |
| mgariepy | ioni, seems to help. | 16:09 |
| odyssey4me | jamesdenton yep, once I have something that works for me - more testing and ironing out kinks for other deployment setups would be nice - I'll be working with linuxbridge only, but you're able to try more configs out and provide feedback | 16:10 |
| jamesdenton | ok yeah, i've been doing OVS | 16:11 |
| jamesdenton | which lines up with the recent biug report | 16:11 |
| odyssey4me | excellent, so we'll be able to confirm it for at least those two options | 16:11 |
| odyssey4me | it seems to me that the procedure will be the same regardless, but it's better when confirmed | 16:11 |
| ioni | odyssey4me, do you mind giving me the bug reports? | 16:12 |
| gillesMo | Does someone knows how to get rid of messages like "Instance 3689d4c2-20b1-438e-ab0d-429c1f352485 has allocations against this compute host but is not found in the database". After failed migrations... | 16:17 |
| mgariepy | ioni, well. not so much after the reboot.. | 16:17 |
| mgariepy | that's weird.. | 16:17 |
| ioni | mgariepy, i don't have ideas. sorry | 16:19 |
| ioni | it's the last version of package for 4.13? because all versions have fixes hardware bugs in procs | 16:20 |
| odyssey4me | ioni I've noted them in the gist | 16:20 |
| ioni | odyssey4me, thanks | 16:20 |
| odyssey4me | ok, let me take a peek at ansmith's things before I dive into that | 16:25 |
| *** nurdie has joined #openstack-ansible | 16:31 | |
| mgariepy | ioni, latest 4.13 works fine without the l1tf setting | 16:36 |
| mgariepy | ho, there is no l1tf patch in 4.13 it seems. | 16:37 |
| *** udesale has quit IRC | 16:41 | |
| *** tosky has quit IRC | 16:43 | |
| *** tosky has joined #openstack-ansible | 16:44 | |
| mnaser | jrosser: https://review.openstack.org/#/c/620378/2 if you got a few sec, i potato'd and almost +2d my own thing | 16:48 |
| mnaser | :p | 16:48 |
| * jrosser looks | 16:52 | |
| jrosser | lgtm | 16:54 |
| *** cshen has quit IRC | 17:01 | |
| *** cshen has joined #openstack-ansible | 17:02 | |
| *** hamzy_ has quit IRC | 17:10 | |
| *** hamzy_ has joined #openstack-ansible | 17:10 | |
| ioni | mgariepy, interesting | 17:11 |
| odyssey4me | ansmith ok, reviewed and made some suggestions to change it up to ensure it's opt-in, but that qdrouterd is preferred if the right inventory entries are found | 17:11 |
| ioni | maybe you didn't disable all the workarounds? | 17:11 |
| odyssey4me | ansmith it'll also make it simpler to configure | 17:11 |
| ansmith | odyssey4me: thanks for checking it out, i will take action on your suggestions | 17:12 |
| odyssey4me | ansmith the test result was success, so I think we're now down to refining it and finalising it :) | 17:12 |
| odyssey4me | ansmith I also reviewed the depends-on patches | 17:12 |
| mgariepy | ioni, : noibrs noibpb nospectre_v2 nopti | 17:13 |
| *** gisak has quit IRC | 17:26 | |
| odyssey4me | mnaser I'm not sure what to say about http://logs.openstack.org/65/624965/2/gate/openstack-ansible-deploy-aio_lxc-centos-7/6b66364/logs/ara-report/result/3c5c77ea-9d8d-429c-9b84-047aaa343bbb/ :/ | 17:28 |
| odyssey4me | which is from https://review.openstack.org/#/c/624965/ | 17:28 |
| odyssey4me | that is a very odd error | 17:29 |
| *** gillesMo has quit IRC | 17:30 | |
| mnaser | odyssey4me: i wonder if that was another change that broke it | 17:33 |
| mnaser | the one that used loops in the incluhde_role rather than looping the include_role | 17:34 |
| *** arxcruz is now known as arxcruz|off|next | 17:34 | |
| mnaser | odyssey4me: https://review.openstack.org/#/c/607814/ this? | 17:34 |
| *** arxcruz|off|next is now known as arxcruz|next_yea | 17:34 | |
| openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server master: upgrade: start service before applying policies https://review.openstack.org/620378 | 17:35 |
| *** arxcruz|next_yea is now known as arxcruz|next_yr | 17:35 | |
| odyssey4me | mnaser then why do all other platforms work, and centos does not :p | 17:37 |
| mnaser | valid question | 17:37 |
| mnaser | log digging | 17:37 |
| odyssey4me | of course all other platforms might be silently ignoring things | 17:37 |
| odyssey4me | however, that very same patch passed check earlier today | 17:37 |
| mnaser | Dec 13 16:51:07 localhost cinder-volume: 2018-12-13 16:51:07.821 4729 ERROR cinder.cmd.volume [-] Configuration for cinder-volume does not specify "enabled_backends". Using DEFAULT section to configure drivers is not supported since Ocata.#033[00m | 17:38 |
| odyssey4me | how nice :) | 17:38 |
| mnaser | "cinder-rtstool is not installed correctly: OSError: [Errno 2] No such file or directory" | 17:39 |
| mnaser | thats what is breaking it | 17:39 |
| odyssey4me | I thought that was fixed some time ago. | 17:40 |
| mnaser | what the hell | 17:40 |
| mnaser | https://review.openstack.org/#/c/607814/ | 17:41 |
| mnaser | http://logs.openstack.org/14/607814/5/check/openstack-ansible-functional-centos-7/bfdf67e/logs/openstack/keystone/stestr_results.html | 17:41 |
| mnaser | that failed | 17:41 |
| mnaser | but reported as passed | 17:41 |
| mnaser | we fixed it but that patch i remember wasnt working or needed some work to make it work with the env variables | 17:41 |
| mnaser | and i thought it was fixed because job passed | 17:41 |
| mnaser | but tempest failed | 17:41 |
| *** ianychoi has quit IRC | 17:42 | |
| odyssey4me | odd though, that merged two days ago, and centos has been passing master patches since | 17:42 |
| mnaser | "Fail if tempest tests did not succeed" | 17:42 |
| mnaser | "All assertions passed" | 17:42 |
| mnaser | i think we messed up tempest | 17:42 |
| *** dcdamien has quit IRC | 17:42 | |
| mnaser | and even failing tempest shows up as passing | 17:42 |
| mnaser | http://logs.openstack.org/14/607814/5/check/openstack-ansible-functional-centos-7/bfdf67e/job-output.txt.gz | 17:42 |
| mnaser | same job doesnt fail but failed tempest | 17:42 |
| odyssey4me | ah, must be https://github.com/openstack/openstack-ansible-os_tempest/commit/ea8ae41f6146d762bfe4b146a71dbda6740dcf08 | 17:43 |
| odyssey4me | the 'always:' should not have been removed there | 17:44 |
| odyssey4me | lemme push that up | 17:44 |
| mnaser | odyssey4me: darn, a bit of a poor yaml spacing made it seem confusing | 17:45 |
| mnaser | we should maybe run a yaml linter in the future so the spacing is more visible | 17:46 |
| mnaser | odyssey4me: are you sure that is the issue though? it looks like the "Fail if tempest tests did not succeed" task ran anyways | 17:47 |
| mnaser | and asserted true, so it did return 0 | 17:47 |
| odyssey4me | yeah, I'm actually inclined to revert, because it seems to me that tempest returns differently with the subunit output | 17:48 |
| mnaser | https://github.com/openstack/tempest/blob/f9650269a32800fdcb873ff63f366b7bc914b3d7/tempest/cmd/run.py#L174-L183 | 17:49 |
| mnaser | doesnt look like it would affect too much | 17:49 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest master: Revert "Use tempest run for generating subunit results" https://review.openstack.org/625070 | 17:49 |
| mnaser | lemme ping mtreinish for a sec | 17:50 |
| odyssey4me | yeah, that's pretty odd - that all seems right | 17:50 |
| odyssey4me | however, perhaps the subunit generation is overriding the test result return code | 17:51 |
| mnaser | odyssey4me: that is my guess | 17:52 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest master: Return the 'always' block https://review.openstack.org/625073 | 17:52 |
| odyssey4me | if that is the case, then we found a bug in tempest which could affect everyone - so we did good. :) | 17:52 |
| openstackgerrit | Merged openstack/openstack-ansible-galera_server stable/queens: Fix SSL support https://review.openstack.org/624633 | 17:52 |
| openstackgerrit | Merged openstack/openstack-ansible-galera_server stable/queens: Fix Galera self-signed SSL functionality https://review.openstack.org/624482 | 17:52 |
| mnaser | i pinged mtreinish who is behing stestr but i dunno how much before a response | 17:53 |
| mnaser | we wont be able to reach arxcruz|next_yr until next year =) | 17:53 |
| odyssey4me | well, he's already approved the revert :p | 17:53 |
| mnaser | oh :p | 17:54 |
| odyssey4me | we should probably get that merged asap, and we may have to deal with fallout in other roles as a result of it | 17:56 |
| pabelanger | mnaser: | 17:59 |
| pabelanger | mnaser: sorry, did you see we are now CDing openstack-ansible from zuul, in ansible-network | 18:00 |
| pabelanger | https://object-storage-ca-ymq-1.vexxhost.net/v1/a0b4156a37f9453eb4ec7db5422272df/logs/4c/4c91f44a55cd37a5c80ddf7e147ba0252465195c/post/packet-ci-cloud-deploy/0aa4dbb/ | 18:00 |
| pabelanger | logs/ara-report for osa bits | 18:00 |
| *** kopecmartin is now known as kopecmartin|off | 18:01 | |
| pabelanger | the next step for me, is to turn https://github.com/ansible-network/packet-ci-cloud/blob/master/openstack_deploy/openstack_inventory.json into a static inventory file | 18:01 |
| pabelanger | but, haven't figured out that step yet | 18:01 |
| *** gyee has joined #openstack-ansible | 18:09 | |
| *** sep has quit IRC | 18:23 | |
| *** sep has joined #openstack-ansible | 18:24 | |
| *** ansmith has quit IRC | 18:30 | |
| *** ansmith has joined #openstack-ansible | 18:30 | |
| *** electrofelix has quit IRC | 18:34 | |
| *** ansmith has quit IRC | 18:35 | |
| *** electrofelix has joined #openstack-ansible | 18:44 | |
| *** dcdamien has joined #openstack-ansible | 18:45 | |
| *** shardy has quit IRC | 18:46 | |
| *** electrofelix has quit IRC | 18:51 | |
| goldenfri | Does anyone happen to have an example network interfaces file using netplan for OSA? Not looking forward to translating all of this. | 19:14 |
| noonedeadpunk | I had, but reverted everything back, as netplan does not support infiniband, and it's a case for me | 19:15 |
| noonedeadpunk | I'd said, it's not rreally complicated | 19:15 |
| goldenfri | thanks, thats good to know | 19:15 |
| goldenfri | yea not complicated but annoying | 19:15 |
| goldenfri | also not sure how the eth12 stuff will translate | 19:16 |
| noonedeadpunk | I have example only for computing node though http://paste.openstack.org/show/737250/ | 19:18 |
| noonedeadpunk | but it's kinda container related, while you should just configure right bridges with netplan | 19:18 |
| goldenfri | Ok thanks, yea it looks like you aren't creating the veth pair for eth12 | 19:20 |
| jamesdenton | you don't need to. that happens automatically | 19:20 |
| noonedeadpunk | lxc should do it for you, based on it's interfaces config files | 19:21 |
| goldenfri | oh hrm | 19:22 |
| jamesdenton | goldenfri here is a two interface xample: http://paste.openstack.org/show/737252/ | 19:22 |
| *** ansmith has joined #openstack-ansible | 19:22 | |
| *** DanyC_ has joined #openstack-ansible | 19:24 | |
| *** DanyC has quit IRC | 19:24 | |
| goldenfri | I use eth12 for my provider network | 19:24 |
| jamesdenton | eth12 is usually only used on an infra node in a neutron_agent container | 19:25 |
| jamesdenton | but i have seen some users recently create a veth with one end named eth12, and the other end in br-vlan. Kinda lines up with the docs, but not the best approach | 19:26 |
| *** weezS has quit IRC | 19:27 | |
| noonedeadpunk | I'm not using eth12 at all, as it's required only by linuxbridge-agent (in my case) wich is bare-metal | 19:28 |
| noonedeadpunk | so yeah, every setup differs:) | 19:30 |
| *** vnogin has joined #openstack-ansible | 19:30 | |
| *** DanyC_ has quit IRC | 19:33 | |
| *** fnpanic_ has joined #openstack-ansible | 19:33 | |
| fnpanic_ | hi | 19:33 |
| fnpanic_ | i am still trying quickstart aio on ubuntu 18.04 | 19:34 |
| fnpanic_ | without luck | 19:34 |
| fnpanic_ | it looks like the /etc/openstack_deploy is not populated | 19:34 |
| fnpanic_ | i also needed to install ansible from universe prior running the bootstrap.sh | 19:35 |
| *** vnogin has quit IRC | 19:35 | |
| goldenfri | jamesdenton: yes that is how I am doing it. Was the only way I could get it to work at the time based the docs etc | 19:35 |
| jamesdenton | fnpanic_ you ran scripts/bootstrap-ansible.sh first? | 19:35 |
| fnpanic_ | yes | 19:36 |
| fnpanic_ | but it fails because ansible is not there | 19:36 |
| *** rodolof has quit IRC | 19:36 | |
| jamesdenton | goldenfri when you define the provider networks in openstack_user_config.yml, specify 'host_bind_override: <interface>' to use a physical interface of some kind. Then you bypass that eth12 veth and br-vlan | 19:36 |
| fnpanic_ | here is the logput from setup-hosts | 19:36 |
| fnpanic_ | http://paste.openstack.org/show/737257/ | 19:36 |
| *** rodolof has joined #openstack-ansible | 19:37 | |
| fnpanic_ | and /etc/openstack_deploy only hase ansible_facts dir | 19:37 |
| fnpanic_ | so why is the bootstrap.sh failing on 18.04 | 19:37 |
| fnpanic_ | it was freshly installed :-) | 19:38 |
| fnpanic_ | does ist produce a log? | 19:38 |
| jrosser | have you run bootstrap-aio.sh? | 19:38 |
| fnpanic_ | YES | 19:39 |
| goldenfri | jamesdenton: right but if I remember why I did it this way, my interfaces are not consistent across hosts so I use eth12 | 19:39 |
| fnpanic_ | yes | 19:39 |
| fnpanic_ | sorry :-) caps was a fault | 19:39 |
| jrosser | if you don't have a populated /etc/openstack_deploy then something went wrong at the bootstrap-aio point | 19:40 |
| fnpanic_ | i guess so | 19:40 |
| jamesdenton | goldenfri You can override those interfaces on a per-host basis if you want to. Or, the method you're using is fine for now. I'm guessing you're asking about netplan because you want something akin to a 'post-up' to configure that veth? | 19:40 |
| fnpanic_ | the playbook finished without error | 19:41 |
| fnpanic_ | i am re-running it | 19:41 |
| fnpanic_ | localhost : ok=131 changed=70 unreachable=0 failed=0 | 19:43 |
| *** cshen has quit IRC | 19:43 | |
| goldenfri | jamesdenton: I'm just asking about netplan because I am starting to test out rocky and 18.04 and need to translate my queens config | 19:43 |
| fnpanic_ | now the files are there | 19:44 |
| fnpanic_ | wtf? | 19:44 |
| jamesdenton | fair enough. i ask because running post-up equivalent commands with netplan isn't really intuitive. | 19:44 |
| fnpanic_ | this is crazy | 19:44 |
| fnpanic_ | the first run was also without errors | 19:44 |
| jamesdenton | anyway, that link i sent earlier has an example of setting up the bridges and whatnot | 19:44 |
| fnpanic_ | and now the files are there | 19:44 |
| fnpanic_ | i am not sure what it is but i had a similar "problem" when i first cloned the git and run bootstrap-aio.sh it was not able to find the sshd role | 19:46 |
| fnpanic_ | then i wiped the /opt/openstack-ansible on the same machine and it worked.... | 19:47 |
| *** pcaruana has quit IRC | 19:47 | |
| fnpanic_ | any idea? | 19:47 |
| jrosser | i think that was a bug which got fixed | 19:47 |
| fnpanic_ | ??? | 19:47 |
| fnpanic_ | what bug? | 19:47 |
| fnpanic_ | i just re-run the script | 19:47 |
| mgariepy | ioni, transparent_hugepage=never seems to help. | 19:48 |
| goldenfri | Thanks, jamesdenton that will get me started, but yea I will need to figure out that pre-up/post-down stuff sometime soon | 19:48 |
| *** cshen has joined #openstack-ansible | 19:49 | |
| jamesdenton | goldenfri i ended up creating a service file to do it. Ghetto, but works.: http://paste.openstack.org/show/737258/ | 19:49 |
| jrosser | fnpanic_: i had to dig a bit but it shouldnt affect you with a recent branch, ansible-sshd messed with it's repo and broke some older osa stuff | 19:53 |
| jrosser | anyway - the aio build process should be really solid because it is what's used for all the CI job gating | 19:54 |
| fnpanic_ | jrosser: so fixed in master? | 19:54 |
| jrosser | no, becasue it only affects old releases | 19:54 |
| jrosser | hence not something to worry about | 19:54 |
| fnpanic_ | i am confused.... old? | 19:54 |
| jrosser | pike/ocata | 19:54 |
| fnpanic_ | mhhh but this was rocky i was using | 19:54 |
| jrosser | anyway the point is the aio should work out of the box | 19:55 |
| jrosser | it's run many times a day as part of the osa CI | 19:55 |
| fnpanic_ | for me it is strange that when i have no ansible installed and run bootstrap-aio.sh it fails saying ansible-playbook command not found | 19:55 |
| fnpanic_ | then i install it and it does not find the sshd role | 19:55 |
| jamesdenton | did you run them in the right order? | 19:55 |
| jrosser | did you first do bootstrap-ansible.sh ? | 19:56 |
| odyssey4me | fnpanic_ I'm confused - there is no bootstrap.sh... what documentation are you reading - can you provide a URL? | 19:56 |
| fnpanic_ | then i wipe the /opt/openstack-ansible | 19:56 |
| fnpanic_ | and clone the git again | 19:56 |
| *** tosky has quit IRC | 19:56 | |
| fnpanic_ | it works but does not copy the config to /etc/openstack_deploy | 19:56 |
| fnpanic_ | when i run it a second time it works | 19:56 |
| fnpanic_ | this is reproducable | 19:56 |
| *** tosky has joined #openstack-ansible | 19:57 | |
| odyssey4me | fnpanic_ cloning the git repo does not create anything other than a copy of the git repo - you appear to be missing steps | 19:57 |
| fnpanic_ | tried it on two different vms and hosts (kvm and vbox) | 19:57 |
| odyssey4me | either that or your over summarising | 19:57 |
| jrosser | fnpanic_: just hold on - you shoudnt have to install ansible yourself | 19:57 |
| fnpanic_ | bootstrap-aio.sh | 19:57 |
| fnpanic_ | yes | 19:57 |
| odyssey4me | fnpanic_ did you do bootstrap-ansible before that? | 19:57 |
| jamesdenton | fnpanic_ https://docs.openstack.org/openstack-ansible/rocky/user/aio/quickstart.html | 19:58 |
| fnpanic_ | no | 19:58 |
| fnpanic_ | bootstrap-aio.sh | 19:58 |
| odyssey4me | clone repo - bootstrap-ansible- - bootstrap-aio... the docs are pretty clear on that | 19:58 |
| odyssey4me | right, so the issue is that you have not done the ansible bootstrap as specified by the docs | 19:58 |
| fnpanic_ | ohhhhhhhhh | 19:58 |
| fnpanic_ | i see | 19:58 |
| fnpanic_ | damn | 19:58 |
| fnpanic_ | i am stupid and cannot read | 19:58 |
| fnpanic_ | :-( | 19:58 |
| odyssey4me | the ansible bootstrap puts ansible down, and puts the roles in place | 19:58 |
| fnpanic_ | sorry | 19:58 |
| fnpanic_ | my brain was only reading bootstrap and did autocomplete the rest | 19:59 |
| fnpanic_ | sorry | 19:59 |
| jamesdenton | it's all good | 19:59 |
| odyssey4me | the deploy guide also makes it clear: https://docs.openstack.org/project-deploy-guide/openstack-ansible/rocky/deploymenthost.html#install-the-source-and-dependencies | 19:59 |
| fnpanic_ | you are absolutly right.... | 20:00 |
| fnpanic_ | also the quickstart AIO | 20:00 |
| odyssey4me | :) IT gives us a curse - we skim read, instead of reading thoroughly | 20:00 |
| fnpanic_ | i earned a channel ban for the next 5 days.... | 20:00 |
| odyssey4me | there is always too much waffle | 20:00 |
| odyssey4me | *but* thankfully we figured out what was going wrong, and you're a step forward :) | 20:01 |
| fnpanic_ | you are very kind.... | 20:01 |
| fnpanic_ | even helping the dumpest people | 20:02 |
| odyssey4me | lol, I am probably the dumbest of all :p | 20:02 |
| fnpanic_ | btw the unbound role, does it install dns in lxc and then points the containers to the ip? | 20:02 |
| fnpanic_ | dns resolver | 20:03 |
| fnpanic_ | i mean | 20:03 |
| *** hamzaachi has quit IRC | 20:03 | |
| fnpanic_ | where does it pick the upstream dns servers from? the infra hosts? | 20:04 |
| odyssey4me | fnpanic_ the default will modify /etc/hosts on the hosts to allow dns resolution everywhere... however if you add unbound hosts then I think it sets up unbound on the hosts and makes all containers use it... although logan- would have to explain better given he set it up and uses it | 20:07 |
| odyssey4me | I suspect there may be some bugs for queens+ for that setup, because it's not gate tested. | 20:08 |
| odyssey4me | anyway, I'm out for the night - will be back online tomorrow to work out a way to finalise that P2Q upgrade stuff for neutron for jamesdenton and ioni :) | 20:09 |
| jamesdenton | thanks odyssey4me! have a great night | 20:09 |
| fnpanic_ | ok | 20:11 |
| fnpanic_ | thanks | 20:11 |
| fnpanic_ | have a great night | 20:11 |
| fnpanic_ | we will stick with host files :-) | 20:11 |
| prometheanfire | odyssey4me: I think https://review.openstack.org/#/c/623240/ already merged :P | 20:13 |
| *** priteau has quit IRC | 20:15 | |
| *** fnpanic_ has quit IRC | 20:18 | |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-haproxy_server master: Force force-tlsv12 only https://review.openstack.org/622411 | 20:20 |
| *** nurdie has quit IRC | 20:25 | |
| *** nurdie has joined #openstack-ansible | 20:25 | |
| *** nurdie has quit IRC | 20:30 | |
| *** vnogin has joined #openstack-ansible | 20:31 | |
| *** vnogin has quit IRC | 20:35 | |
| *** ianychoi has joined #openstack-ansible | 20:59 | |
| *** cshen has quit IRC | 21:02 | |
| redkrieg | Has anyone experienced memcache import errors when trying to set up the IdP for a federated setup in openstack-ansible? http://paste.openstack.org/show/737263/ | 21:03 |
| *** macza has joined #openstack-ansible | 21:09 | |
| *** markvoelker has joined #openstack-ansible | 21:11 | |
| *** cshen has joined #openstack-ansible | 21:24 | |
| *** DanyC has joined #openstack-ansible | 21:26 | |
| *** DanyC has quit IRC | 21:31 | |
| *** vnogin has joined #openstack-ansible | 21:32 | |
| *** ansmith has quit IRC | 21:33 | |
| *** vnogin has quit IRC | 21:37 | |
| *** DanyC has joined #openstack-ansible | 21:47 | |
| *** DanyC has quit IRC | 21:51 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server master: Allow backend and backup node rise/fall parameters to be set https://review.openstack.org/625124 | 21:55 |
| *** DanyC has joined #openstack-ansible | 21:59 | |
| *** markvoelker has quit IRC | 22:06 | |
| *** vnogin has joined #openstack-ansible | 22:33 | |
| *** vnogin has quit IRC | 22:37 | |
| *** lbragstad has quit IRC | 22:37 | |
| redkrieg | I was able to complete the federation setup by installing the following packages in the keystone container: python-keystoneclient python-memcache python-crypto | 22:39 |
| redkrieg | Is there an accepted way to list additional packages for install in a specific container? | 22:39 |
| *** dcdamien has quit IRC | 22:45 | |
| *** cshen has quit IRC | 22:57 | |
| *** DanyC has quit IRC | 22:58 | |
| *** cshen has joined #openstack-ansible | 23:00 | |
| *** noonedeadpunk has quit IRC | 23:02 | |
| *** noonedeadpunk has joined #openstack-ansible | 23:03 | |
| *** cshen has quit IRC | 23:08 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_ceilometer master: gnocchi_resources override fixed https://review.openstack.org/622956 | 23:16 |
| openstackgerrit | Merged openstack/openstack-ansible-os_ceilometer master: Add ability to override meters.yaml and event_definitions.yaml https://review.openstack.org/623239 | 23:16 |
| redkrieg | I've hit another issue that I' | 23:23 |
| redkrieg | m really not sure how best to work around. keystone-manage doesn't seem to have saml2 support accessible: http://paste.openstack.org/show/737273/ | 23:24 |
| *** nurdie has joined #openstack-ansible | 23:26 | |
| *** tosky has quit IRC | 23:27 | |
| redkrieg | I've also noticed that keystone isn't even running on port 5000 on the SP. The IdP seems fine. My user_federation.yml from my SP: http://paste.openstack.org/show/737274/ | 23:28 |
| *** nurdie has quit IRC | 23:31 | |
| *** vnogin has joined #openstack-ansible | 23:33 | |
| *** vnogin has quit IRC | 23:38 | |
| *** thuydang has quit IRC | 23:56 | |
| *** thuydang has joined #openstack-ansible | 23:56 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!