| *** tosky has quit IRC | 00:00 | |
| *** macz_ has quit IRC | 01:11 | |
| *** Underknowledge has quit IRC | 01:23 | |
| *** Underknowledge has joined #openstack-ansible | 01:24 | |
| *** macz_ has joined #openstack-ansible | 01:26 | |
| *** Carcer has quit IRC | 01:26 | |
| *** Carcer has joined #openstack-ansible | 01:27 | |
| *** macz_ has quit IRC | 01:31 | |
| *** cp- has quit IRC | 01:58 | |
| *** cp- has joined #openstack-ansible | 02:04 | |
| *** macz_ has joined #openstack-ansible | 02:07 | |
| *** macz_ has quit IRC | 02:11 | |
| *** LowKey has quit IRC | 02:20 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Switch back to willshersystems ansible-sshd role https://review.opendev.org/c/openstack/openstack-ansible/+/776965 | 02:45 |
|---|---|---|
| *** cp- has quit IRC | 02:45 | |
| *** cp- has joined #openstack-ansible | 02:46 | |
| *** spatel has joined #openstack-ansible | 03:35 | |
| *** gyee has quit IRC | 05:20 | |
| *** macz_ has joined #openstack-ansible | 05:27 | |
| *** macz_ has quit IRC | 05:32 | |
| *** evrardjp has quit IRC | 05:33 | |
| *** evrardjp has joined #openstack-ansible | 05:33 | |
| *** luksky has joined #openstack-ansible | 06:13 | |
| *** spatel has quit IRC | 06:14 | |
| *** ierdem has joined #openstack-ansible | 06:28 | |
| *** ioni has joined #openstack-ansible | 06:56 | |
| *** macz_ has joined #openstack-ansible | 07:08 | |
| *** macz_ has quit IRC | 07:13 | |
| *** pto has joined #openstack-ansible | 07:16 | |
| *** miloa has joined #openstack-ansible | 07:21 | |
| *** pto has quit IRC | 07:22 | |
| *** macz_ has joined #openstack-ansible | 07:24 | |
| *** macz_ has quit IRC | 07:29 | |
| *** miloa has quit IRC | 07:34 | |
| *** rpittau|afk is now known as rpittau | 07:57 | |
| ierdem | Hi everyone, I am facing "EHOSTUNREACH" error in my nova container journals. When I checked the network with tcpdump I saw this log http://paste.openstack.org/show/802989/. The IP addresses in this log are not used but they used for containers before. Now, I cleaned my inventory and re-installed OpenStack, I do not know where these IPs comes from, | 08:06 |
| ierdem | how my containers find these IPs? Can you help me please? | 08:06 |
| *** andrewbonney has joined #openstack-ansible | 08:13 | |
| *** maharg101 has joined #openstack-ansible | 08:28 | |
| *** tosky has joined #openstack-ansible | 08:38 | |
| *** mathlin_ has quit IRC | 08:41 | |
| kleini | ierdem: can you just grep for those IPs in all files of the nova container? especially /etc/nova, /etc/hosts | 08:50 |
| ierdem | I grep all files, there is nothing about those in any files. I also search other containers, servers etc | 08:52 |
| *** pto has joined #openstack-ansible | 08:56 | |
| *** pto has joined #openstack-ansible | 08:57 | |
| *** pto has quit IRC | 08:58 | |
| *** macz_ has joined #openstack-ansible | 08:58 | |
| *** macz_ has quit IRC | 09:03 | |
| kleini | oh, you have that error message in the container journals. from which process are these errors? | 09:08 |
| *** Underknowledge has quit IRC | 09:37 | |
| *** Underknowledge1 has joined #openstack-ansible | 09:37 | |
| *** Underknowledge1 is now known as Underknowledge | 09:37 | |
| *** macz_ has joined #openstack-ansible | 09:46 | |
| jrosser | noonedeadpunk: do you think we have a good reason for setting "ANSIBLE_CACHE_PLUGIN_TIMEOUT": "86400" | 09:47 |
| jrosser | it means that i come to the AIO i was using > 24hrs ago and wierd things break | 09:47 |
| jrosser | like run os-nova-install and it breaks on python_venv_build because the facts for the repo server are now out of date | 09:48 |
| noonedeadpunk | I think in all playbook we've added cache gathering on launch | 09:49 |
| jrosser | yes, but only for the hosts in the play | 09:49 |
| jrosser | so i have fresh facts for nova_all, but not for repo_all | 09:49 |
| noonedeadpunk | ah... | 09:49 |
| jrosser | leads to really wierd breakage because the files all look good in /etc/openstack_deploy/ansible_facts | 09:50 |
| noonedeadpunk | it's default value actually https://docs.ansible.com/ansible/latest/reference_appendices/config.html#cache-plugin-timeout | 09:50 |
| *** macz_ has quit IRC | 09:51 | |
| jrosser | oh! interesting | 09:51 |
| admin0 | ierdem, still having the same issues ? | 09:51 |
| admin0 | \o | 09:51 |
| jrosser | wonder how this is not causing more issues already then | 09:51 |
| noonedeadpunk | well, I just run all facts gathering before running smth with tags | 09:51 |
| noonedeadpunk | as eventually it's only tags that are breaking I guess | 09:51 |
| noonedeadpunk | since then gather_facts is not performed on role startup | 09:52 |
| noonedeadpunk | or we should have always here https://opendev.org/openstack/openstack-ansible/src/branch/master/playbooks/os-nova-install.yml#L42-L43 (and everywhere in playbooks) | 09:52 |
| jrosser | well i think that still won't make old repo_all facts become valid running the nova playbook will it? | 09:56 |
| *** jhesketh_ has joined #openstack-ansible | 09:57 | |
| jrosser | os-nova-install.yml has a dependancy on repo_all facts because it uses python_venv_build | 09:57 |
| noonedeadpunk | so it was kind of working with ansible_ but not with ansible_facts, right? | 09:59 |
| jrosser | well no i'm not sure it can | 09:59 |
| *** brad- has joined #openstack-ansible | 10:00 | |
| jrosser | ok: [localhost] => { | 10:00 |
| jrosser | "hostvars['aio1_repo_container-d8de7ef9']['ansible_facts']": {} | 10:00 |
| *** arxcruz has quit IRC | 10:00 | |
| *** arxcruz has joined #openstack-ansible | 10:01 | |
| *** macz_ has joined #openstack-ansible | 10:02 | |
| *** Carcer has quit IRC | 10:03 | |
| *** jhesketh has quit IRC | 10:03 | |
| *** brad[] has quit IRC | 10:03 | |
| *** kleini has quit IRC | 10:03 | |
| *** irclogbot_2 has quit IRC | 10:03 | |
| *** mcarden has quit IRC | 10:03 | |
| jrosser | noonedeadpunk: http://paste.openstack.org/show/802991/ | 10:03 |
| noonedeadpunk | hm | 10:04 |
| noonedeadpunk | How it has been even working indeed? | 10:05 |
| jrosser | i guess its always been less than 24 hours since running something that gathers the needed facts | 10:05 |
| *** macz_ has quit IRC | 10:06 | |
| jrosser | this has just come up i think because i pretty much didnt touch my AIO for a whole day then came back to it half way through the deployment | 10:06 |
| noonedeadpunk | and I was running only nova-config tags...? | 10:06 |
| *** irclogbot_2 has joined #openstack-ansible | 10:06 | |
| jrosser | if that doesnt do any of the venv stuff it's probably not going to occur | 10:06 |
| noonedeadpunk | well my aio on U working for half a year or so... And I think I can just re-run playbooks and it should be working... | 10:07 |
| jrosser | it was just accident i saw the timeout value in a dump of hostvars | 10:07 |
| jrosser | setting the timeout to 0 seems to do what i'd expect http://paste.openstack.org/show/802992/ | 10:09 |
| noonedeadpunk | I re-run glance and it failed but because lxc-dnsmasq has been killed by oom, so it couldn't update apt. But it haven't failed http://paste.openstack.org/show/802993/ | 10:11 |
| noonedeadpunk | oh, hm https://opendev.org/openstack/ansible-role-python_venv_build/src/branch/master/tasks/python_venv_wheel_build.yml#L16-L21 | 10:12 |
| noonedeadpunk | maye we should move it to main? | 10:13 |
| jrosser | heres mine at a similar point for nova http://paste.openstack.org/show/802994/ | 10:14 |
| noonedeadpunk | I think setting cache to 0 is also an option, it would just influence os upgrades that you need to go and clean up facts manually | 10:14 |
| noonedeadpunk | maybe we've breake it with https://opendev.org/openstack/ansible-role-python_venv_build/commit/ab68e944256648cb6d714e02dbe88d228a2c04a5 ? | 10:15 |
| noonedeadpunk | as I don't have it yet in my aio | 10:16 |
| jrosser | i think that patch is legitimate | 10:17 |
| jrosser | better fix is to set the cache time to 0, as I think we do gather facts correctly in each playbook when things are deployed | 10:18 |
| jrosser | it's more about when one thing wants facts from somewhere out of the scope of that playbook that we have a problem | 10:18 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/ansible-role-python_venv_build master: Gather facts for repo containers https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/777559 | 10:20 |
| noonedeadpunk | well we can disable facts cache timeout. But even then the only way to get the updated is run repo_all which is not really intuitive as well | 10:22 |
| jrosser | i'm going to revisit a lot of this with the ansible_facts[] stuff | 10:23 |
| noonedeadpunk | and I think we might miss proper facts gathering somewhere in terms of tags usage | 10:24 |
| jrosser | specifically for nova i think we should add repo_all to the group here https://github.com/openstack/openstack-ansible/blob/master/playbooks/common-playbooks/nova.yml#L16-L20 | 10:25 |
| jrosser | yeah, gathering facts for hosts: "{{ nova_hosts }},repo_all" in common_playbooks/nova.yml is the proper way and robust | 10:30 |
| jrosser | just checked and that works fine | 10:31 |
| ierdem | admin0 yes same problem .. :/ | 10:43 |
| masterpe | Hi, I try to upgrade from Train to Ussuri, I'm following https://docs.openstack.org/openstack-ansible/ussuri/admin/upgrades/major-upgrades.html for this. But as I understand Ceph also needs to upgraded from Nautilus to Octopus. For this I need to run openstack-ansible /opt/openstack-ansible/playbooks/ceph-install.yml with the rolling_update=true option. | 10:52 |
| masterpe | But I hit the following issue: https://github.com/ceph/ceph-ansible/issues/5400 | 10:52 |
| masterpe | I'm I correct that a ceph upgrade is required? | 10:53 |
| *** pto has joined #openstack-ansible | 10:59 | |
| *** pto has quit IRC | 11:04 | |
| noonedeadpunk | well, I'd say that doing that in python_venv_build is more proper because that's the role that failing it... But I can agree that this will mean multiplication on facts gathering for each role. But from other saide, each playbook would have this included anyway... | 11:06 |
| noonedeadpunk | It's jsut less work to patch pythn_venv_build like this https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/777559 | 11:06 |
| *** sshnaidm|afk is now known as sshnaidm|pto | 11:22 | |
| noonedeadpunk | masterpe: yes, we use ceph-ansible 5 in Ususri | 11:25 |
| noonedeadpunk | so generally you need a ceph update or just omit ceph-install playbook and leave ceph as-is and upgrade it later. Eventually openstack can work nicely with Nautilus as well | 11:26 |
| noonedeadpunk | but yes, it's better to upgrade anyway | 11:27 |
| noonedeadpunk | jrosser: is ansible_facts['env'] some object so `ansible_facts['env']['HOME']` just doesn't work? | 11:48 |
| noonedeadpunk | (works for me) | 11:50 |
| jrosser | noonedeadpunk: works for me too - did i miss changing a .HOME somewhere? | 11:52 |
| CeeMac | ierdem: did you try tcpdump against the rabbit port to see what nova is trying to talk to? Regarding your tcpdump, is this a dedicated network to your OSA install, or do you have none OSA things running on it to? The container could just be seeing arp requests from other devices on the same subnet | 11:52 |
| noonedeadpunk | yeah is several roles in several places:) So wasn't sure it works:) | 11:53 |
| jrosser | oh yes i see the emails now | 11:53 |
| jrosser | this can all be fixed | 11:54 |
| jrosser | i just added the spaces to make it easier to see what the code was meaning | 11:54 |
| jrosser | i had to space it out here to understand what the nesting was to change it, which felt like somehow the expression was a bit too complex | 11:55 |
| noonedeadpunk | ok, got it. just didn't get used to spaces inside list barckets | 11:56 |
| noonedeadpunk | maybe it's good to have them... | 11:56 |
| noonedeadpunk | I have comments for https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/777179 that probably relevant for all roles | 12:05 |
| *** macz_ has joined #openstack-ansible | 12:05 | |
| noonedeadpunk | what do you think about it? Shouldn't we clean up a bit once we're revising that? | 12:05 |
| jonher | hi, i'm trying to deploy a AIO on focal with stable/victoria, default AIO setup with the addition of the adjutant role but i'm getting this: http://paste.openstack.org/show/802997/ any ideas on what i could check to help troubleshoot this? it seems like the ci jobs deploy this fine against master: https://review.opendev.org/c/openstack/openstack-ansible-os_adjutant/+/772020 | 12:06 |
| *** macz_ has quit IRC | 12:09 | |
| noonedeadpunk | oops, looks like some bug | 12:12 |
| noonedeadpunk | yeah, we use new resolver on master | 12:12 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/openstack-ansible-lxc_container_create/+/776958 | 12:15 |
| noonedeadpunk | I have a meting right now, will be able to check right after that | 12:15 |
| jonher | tyvm :) | 12:16 |
| *** macz_ has joined #openstack-ansible | 12:21 | |
| openstackgerrit | Jonathan Rosser proposed openstack/ansible-role-python_venv_build master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/777068 | 12:24 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-openstack_openrc master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/openstack-ansible-openstack_openrc/+/777070 | 12:25 |
| *** macz_ has quit IRC | 12:25 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-openstack_hosts master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/776960 | 12:26 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-galera_server master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/777060 | 12:27 |
| jrosser | noonedeadpunk: would you prefer me to combine cleanup stuff into the ansible_facts[] patches? | 12:27 |
| jrosser | i'e tried not to change anything else, like the centos-8 ternary stuff or suse/zypper | 12:28 |
| jrosser | was going to take another pass over the whole thing for that | 12:28 |
| noonedeadpunk | well, that would be awesome to have another pass, but it's jsut about time and if you have it... As I'd also loved to have such pass but not sure when will be able to get to it | 12:41 |
| noonedeadpunk | so if we don't need to use ternary somewhere at all - why we even replace it? | 12:42 |
| noonedeadpunk | not sure though... | 12:42 |
| noonedeadpunk | It's not strong opinion anyway | 12:42 |
| noonedeadpunk | was just asking how do you think. I'm ok to merge this as is actually | 12:43 |
| jrosser | i think atm i prefer to change just one thing | 12:43 |
| noonedeadpunk | ok then | 12:43 |
| jrosser | trying out vscode rather than vim and these things are much easier across * repos | 12:43 |
| jrosser | vscode remote into AIO = win | 12:44 |
| noonedeadpunk | yeah, vscode is super nice :) | 12:44 |
| *** macz_ has joined #openstack-ansible | 12:52 | |
| *** macz_ has quit IRC | 12:56 | |
| openstackgerrit | Merged openstack/ansible-role-uwsgi master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/ansible-role-uwsgi/+/777386 | 13:04 |
| openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/777071 | 13:14 |
| openstackgerrit | Merged openstack/openstack-ansible-memcached_server master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/openstack-ansible-memcached_server/+/777069 | 13:15 |
| *** kleini_ has joined #openstack-ansible | 13:18 | |
| openstackgerrit | Merged openstack/ansible-role-systemd_networkd master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/ansible-role-systemd_networkd/+/776955 | 13:30 |
| noonedeadpunk | ok, so, adjutant | 13:36 |
| openstackgerrit | Merged openstack/openstack-ansible-repo_server master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/777073 | 13:40 |
| *** jamesdenton has joined #openstack-ansible | 13:40 | |
| *** gshippey has joined #openstack-ansible | 13:57 | |
| *** spatel has joined #openstack-ansible | 13:59 | |
| noonedeadpunk | jonher: ok, I found it. what if you set `adjutant_venv_tag: "{{ venv_tag | default('untagged') }}"` in you user_variables and re-run it? | 14:04 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_adjutant master: Use global service variables https://review.opendev.org/c/openstack/openstack-ansible-os_adjutant/+/777596 | 14:07 |
| jonher | i put that into /etc/openstack_deploy/user_variables.yml and ran: openstack-ansible os-adjutant-install.yml but it looks like it's hitting the same issue, do i need to set venv rebuild to true as well? | 14:08 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_adjutant master: Use venv_tag https://review.opendev.org/c/openstack/openstack-ansible-os_adjutant/+/777597 | 14:09 |
| noonedeadpunk | can you pls copy that new issue? | 14:11 |
| noonedeadpunk | weird thing is - `You are using pip version 20.2.3; however, version 21.0.1 is available.` so it's pip version that should have old resolver, right? | 14:11 |
| jrosser | i don't think we merged the patch yet to actually change the resolver | 14:11 |
| jonher | http://paste.openstack.org/show/803003/ | 14:11 |
| jrosser | becasue os_neturon is br0k | 14:12 |
| noonedeadpunk | oh, indeed | 14:12 |
| noonedeadpunk | jonher: can I also ask you to check `/var/log/python_wheel_build.log` inside repo_all[0] container? | 14:13 |
| noonedeadpunk | yeah eventually this one does not contain reason why it's failed :( http://paste.openstack.org/show/803005/ | 14:16 |
| noonedeadpunk | so we really need to check that log file | 14:18 |
| jonher | ah sorry, i emptied the log to make a fresh one, re-ran the playbook and then it complained about adjutant-22.0.2.dev7-constraints.txt not exisiting, that is where i'm at right now | 14:19 |
| noonedeadpunk | I thought it should be `adjutant-22.0.2.dev7-source-constraints.txt` hm | 14:25 |
| jonher | adjutant-22.0.2.dev7-global-constraints.txt adjutant-22.0.2.dev7-requirements.txt adjutant-22.0.2.dev7-source-constraints.txt adjutant-untagged-global-constraints.txt adjutant-untagged-requirements.txt adjutant-untagged-source-constraints.txt exists in repo | 14:30 |
| noonedeadpunk | but it claims on `adjutant-22.0.2.dev7-constraints.txt`? | 14:31 |
| jonher | yep: FAILED! => {"changed": false, "msg": "file not found: /var/www/repo/os-releases/22.0.2.dev7/adjutant-22.0.2.dev7-constraints.txt"} | 14:31 |
| noonedeadpunk | you can try passing venv_rebuild=true in case you're on stable/victoria and have this patch applied https://opendev.org/openstack/ansible-role-python_venv_build/commit/9df14bdd0ba8221abd43ce26190f13d48ec8c08f. This should help | 14:35 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: DNM Test workaround octavia centos bug https://review.opendev.org/c/openstack/openstack-ansible/+/777601 | 14:39 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_octavia master: [reno] Stop publishing release notes https://review.opendev.org/c/openstack/openstack-ansible-os_octavia/+/772042 | 14:39 |
| andrewbonney | noonedeadpunk: you might need https://review.opendev.org/c/openstack/openstack-ansible-os_octavia/+/772559 for the octavia workaround too | 14:45 |
| andrewbonney | Otherwise I think it'll fail looking for openstacksdk on the deploy host | 14:45 |
| noonedeadpunk | yeah, you're right | 14:46 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_octavia master: [reno] Stop publishing release notes https://review.opendev.org/c/openstack/openstack-ansible-os_octavia/+/772042 | 14:46 |
| mgariepy | noonedeadpunk, quick one about the sha bump. how is the new tag generated ? | 14:47 |
| noonedeadpunk | there's a tooling but lately I just write it up by hands since it's faster. eventually x.x.Y release is a bugfix, x.Y.x is a feature release and Y.x.x is major release eventually | 14:49 |
| noonedeadpunk | so if there're feature release notes it should be feature release kind of | 14:49 |
| mgariepy | ok, and the tag is added where ? | 14:51 |
| mgariepy | once the bump is merged, you simply add the tag to the repo ? | 14:51 |
| jonher | edited the task removing the line that is in the commit, then tried: openstack-ansible os-adjutant-install.yml --extra-vars "venv_rebuild=true" and got: http://paste.openstack.org/show/803008/ | 14:53 |
| noonedeadpunk | mgariepy: once bump is merged I push commit to the releases repo with change to the deliverables https://opendev.org/openstack/releases/ | 14:58 |
| mgariepy | ha, less black magic :) haha | 14:59 |
| noonedeadpunk | it's actually pretty straightforward :p | 14:59 |
| noonedeadpunk | jonher: | 14:59 |
| noonedeadpunk | huh. | 14:59 |
| noonedeadpunk | do we even need mysqlclient there? | 15:00 |
| *** pto has joined #openstack-ansible | 15:00 | |
| mgariepy | that's what i was thinking but i didn't knew where it was done ;) haha | 15:00 |
| noonedeadpunk | hm https://opendev.org/openstack/adjutant/src/branch/master/requirements.txt#L19 | 15:01 |
| jonher | yeah was just looking at that | 15:01 |
| *** spatel has quit IRC | 15:02 | |
| jonher | it has it's database in galera/mysql unless you run a local sqlite | 15:02 |
| noonedeadpunk | well, every service we run has own mysql (all almost every) | 15:02 |
| noonedeadpunk | but none use mysqlclient I guess | 15:03 |
| noonedeadpunk | ok, let's install it then | 15:05 |
| noonedeadpunk | sec | 15:05 |
| noonedeadpunk | can you add override like that? http://paste.openstack.org/show/803009/ | 15:06 |
| noonedeadpunk | considering it's ubuntu | 15:06 |
| *** pto has quit IRC | 15:06 | |
| jonher | now it gets past the initial error but can't find libmariadb.so.3 when trying to import the module: http://paste.openstack.org/show/803010/ | 15:12 |
| noonedeadpunk | doh | 15:13 |
| noonedeadpunk | I hate mysqlclient.... | 15:13 |
| noonedeadpunk | why not to use PyMySQL like all other projects... | 15:13 |
| jonher | why does this work in ci against master but not on stable/victoria? i'm still lost why that is the case | 15:14 |
| noonedeadpunk | I think because in CI we use metal jobs where libmariadbclient-dev got installed by part of another role | 15:15 |
| noonedeadpunk | or with utility | 15:15 |
| jonher | oh | 15:15 |
| noonedeadpunk | if you manually install that package inside container, does it solve the issue? | 15:15 |
| noonedeadpunk | (inside adjutant container) | 15:15 |
| noonedeadpunk | I will patch this if this helps | 15:16 |
| *** spatel has joined #openstack-ansible | 15:16 | |
| *** macz_ has joined #openstack-ansible | 15:22 | |
| jonher | yep now os-adjutant-install.yml is OK | 15:22 |
| noonedeadpunk | well, we said about experimental support :p | 15:22 |
| noonedeadpunk | btw, I think you would love to have horizon integration as well?:) | 15:23 |
| noonedeadpunk | as I never had time/motivation to made it | 15:23 |
| jonher | hehe yep, that's next step to see how it would be installed | 15:23 |
| jonher | i looked for the same for cloudkitty etc.. and saw some similar references in the horizon role | 15:23 |
| noonedeadpunk | well, at this point it won't | 15:24 |
| noonedeadpunk | for cloudkitty we have half-backed solution that never was integrated properly | 15:24 |
| admin0 | anyone using victoria + AD integration + multi domain auth and has their horizon not broken ? | 15:24 |
| noonedeadpunk | but there should notbe too much work. we just never had interested enough people and for us it' hard to make smth nobody use | 15:25 |
| *** macz_ has quit IRC | 15:26 | |
| jonher | yeah i see, do you know on top of your head a os-ansible project that also publishes dashboards to horizon? | 15:26 |
| jonher | and thanks very much for the help in the adjutant case, i thought i did something wrong there at first or had a broken env | 15:27 |
| noonedeadpunk | I can take a look at horizon and adjutant dashboard tomorrow if you will be ready to test that :) | 15:28 |
| noonedeadpunk | kolla-ansible? | 15:28 |
| noonedeadpunk | well, we added experimental support of adjutant in victoria so it's really new | 15:28 |
| *** rh-jelabarre has joined #openstack-ansible | 15:29 | |
| jonher | yeah sure, would like to deploy adjutant with os-ansible instead of manually on the side like we used to if itwas not too much work, which is why i'm now testing how much of an effort it would be :) | 15:29 |
| *** rh-jelabarre has quit IRC | 15:29 | |
| *** rh-jelabarre has joined #openstack-ansible | 15:30 | |
| noonedeadpunk | well I had a brief look on adjutant dashboard and it should be pretty straigtforward to add to the role | 15:30 |
| noonedeadpunk | but without having a usecase it super hard to understand if you're doing right and eventually there're other things that need to be done asap | 15:30 |
| jonher | we have been using adjutant for 3+ years for inviting users / managing quota / requesting new projects / password resets and it works really well for that :) | 15:31 |
| jonher | i'll be available for testing, i'm not very good at ansible yet but i'll be happy to test | 15:31 |
| noonedeadpunk | yeah, agree. I was just using whmcs for that before and now we don't use even horizon, so... | 15:31 |
| noonedeadpunk | but I always liked the project | 15:32 |
| *** fanfi has joined #openstack-ansible | 15:33 | |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_adjutant master: Install mysql client libraries https://review.opendev.org/c/openstack/openstack-ansible-os_adjutant/+/777607 | 15:38 |
| noonedeadpunk | jonher: ^ that should be it | 15:38 |
| *** macz_ has joined #openstack-ansible | 16:02 | |
| *** spatel has quit IRC | 16:12 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/victoria: Add reno about barbican_simple_crypto_key https://review.opendev.org/c/openstack/openstack-ansible/+/775856 | 16:27 |
| *** tsturm has joined #openstack-ansible | 16:32 | |
| tsturm | When we update openstack via openstack-ansible, do we need to update the conaiiners in the controllers as well? It doesn't look like the update playbook covers updating them. | 16:34 |
| *** spatel has joined #openstack-ansible | 16:51 | |
| noonedeadpunk | well yes, you do need. Evnetually you can look what upgrade script does https://opendev.org/openstack/openstack-ansible/src/branch/master/scripts/run-upgrade.sh#L170-L189 | 16:53 |
| noonedeadpunk | but I think it depends on understanding what update of containers include.. | 16:53 |
| noonedeadpunk | all services on containers should be updated as well as galera/rabbitmq version and etc | 16:54 |
| jrosser | tsturm: not sure if it's clear that they are system containers not application containers, so unlike docker type things you don't need to "update the container" to get a new version of the service installed | 16:57 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_nova master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/777627 | 16:58 |
| tsturm | We are specifically asking about os packages within the system containers. | 17:00 |
| noonedeadpunk | yes, you need to run setup-openstack.yml as a part of upgrade to get new os service venvs/packages and db upgrades | 17:01 |
| *** fresta has quit IRC | 17:13 | |
| *** fresta has joined #openstack-ansible | 17:14 | |
| *** fresta_ has joined #openstack-ansible | 17:21 | |
| *** fresta has quit IRC | 17:21 | |
| *** rpittau is now known as rpittau|afk | 17:34 | |
| *** maharg101 has quit IRC | 17:38 | |
| *** fresta_ has quit IRC | 17:52 | |
| *** fresta has joined #openstack-ansible | 17:54 | |
| *** luksky has quit IRC | 18:01 | |
| *** luksky has joined #openstack-ansible | 18:01 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_nova master: Do not use service_facts https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/777644 | 18:09 |
| djhankb | tsturm: If you are talking about os packages as in "operating system" and not "openstack" then you can use ansible to handle apt updates in the case of debian: ansible all_containers -m apt -a "upgrade=yes update_cache=yes cache_valid_time=86400" --become | 18:14 |
| tsturm | That's what we were looking for. Thanks! | 18:15 |
| djhankb | you would substitute the appropriate group you would like to target, my example would include all containers - you should be careful to upgrade galera mysql packages beforehand as this command could bring down all galera nodes at the same time forcing you to re-bootstrap the cluster | 18:15 |
| mgariepy | we should add support for ansible vault in the pwgent script :D | 18:16 |
| djhankb | you can use this command to list out your groups (run this from the /opt/openstack-ansible directory): ansible localhost -m debug -a 'var=groups' | 18:16 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_neutron master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/777650 | 18:42 |
| spatel | Can i use barbican as a password vault ? | 18:45 |
| spatel | admin0 i am not using AD but i used FreeIPA LDAP. Ldap is very standard so it should work with AD also https://satishdotpatel.github.io//openstack-ldap-integration/ | 18:47 |
| *** andrewbonney has quit IRC | 18:49 | |
| *** ierdem has quit IRC | 18:50 | |
| admin0 | i have it running on ussuri .. i upgraded to vic and horizon crashed | 18:51 |
| admin0 | so trying to figure it out why | 18:51 |
| admin0 | luckily in a lab env | 18:51 |
| spatel | +1 | 18:53 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_horizon master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/777651 | 18:59 |
| *** ioni has quit IRC | 19:01 | |
| *** maharg101 has joined #openstack-ansible | 19:34 | |
| mgariepy | admin0, when horizon fail, usualy it's kinda easier to just wipe a container and start from scratch | 19:39 |
| *** maharg101 has quit IRC | 19:39 | |
| admin0 | mgariepy, i do that for everything except galera :) | 19:40 |
| admin0 | i find it faster to redo than to troubleshoot | 19:40 |
| mgariepy | for upgrades, if i know it can happen, i usualy just delete one at the begining, then when running the upgrade witll deploy a new one. | 19:40 |
| mgariepy | then wipe the remaining one if they fail (can be disabled in haproxy, before deploying) | 19:41 |
| mgariepy | noonedeadpunk, wow. wtf for the 127.0.0.1 in /etc/hosts | 20:01 |
| mgariepy | all that because of the extra nodes in inventory ? | 20:02 |
| mgariepy | hha. nop i was mistaken i guess :D | 20:20 |
| *** spatel has quit IRC | 20:27 | |
| *** ioni has joined #openstack-ansible | 20:28 | |
| *** logan- has quit IRC | 20:35 | |
| *** fresta_ has joined #openstack-ansible | 20:35 | |
| *** tinwood has quit IRC | 20:35 | |
| *** tinwood has joined #openstack-ansible | 20:36 | |
| *** mugsie_ has joined #openstack-ansible | 20:36 | |
| *** persia_ has joined #openstack-ansible | 20:36 | |
| *** bverschueren_ has joined #openstack-ansible | 20:36 | |
| *** grabes has quit IRC | 20:36 | |
| *** bverschueren has quit IRC | 20:36 | |
| *** zigo has quit IRC | 20:36 | |
| *** gary_perkins has quit IRC | 20:36 | |
| *** persia has quit IRC | 20:36 | |
| *** melwitt has quit IRC | 20:36 | |
| *** snadge has quit IRC | 20:36 | |
| *** gregwork has quit IRC | 20:36 | |
| *** mugsie has quit IRC | 20:36 | |
| *** mmercer has quit IRC | 20:36 | |
| *** melwitt has joined #openstack-ansible | 20:36 | |
| *** grabes has joined #openstack-ansible | 20:37 | |
| *** rh-jlabarre has joined #openstack-ansible | 20:37 | |
| *** mmercer has joined #openstack-ansible | 20:37 | |
| *** jmccrory_ has joined #openstack-ansible | 20:37 | |
| *** tosky_ has joined #openstack-ansible | 20:37 | |
| *** rh-jlabarre has quit IRC | 20:37 | |
| *** MrClayPole_ has joined #openstack-ansible | 20:37 | |
| *** fresta_ has quit IRC | 20:37 | |
| *** fresta_ has joined #openstack-ansible | 20:37 | |
| *** rh-jlabarre has joined #openstack-ansible | 20:37 | |
| *** jmccrory has quit IRC | 20:37 | |
| *** jmccrory_ is now known as jmccrory | 20:38 | |
| *** bradm has quit IRC | 20:38 | |
| *** logan- has joined #openstack-ansible | 20:38 | |
| *** evrardjp_ has joined #openstack-ansible | 20:38 | |
| *** heikkine has quit IRC | 20:38 | |
| *** fresta has quit IRC | 20:38 | |
| *** cp- has quit IRC | 20:38 | |
| *** mubix has quit IRC | 20:38 | |
| *** poopcat1 has joined #openstack-ansible | 20:38 | |
| *** aedan1 has joined #openstack-ansible | 20:39 | |
| *** jrosser has quit IRC | 20:40 | |
| *** rh-jelabarre has quit IRC | 20:40 | |
| *** csmart has quit IRC | 20:40 | |
| *** tosky has quit IRC | 20:40 | |
| *** noonedeadpunk has quit IRC | 20:40 | |
| *** trident has quit IRC | 20:40 | |
| *** fanfi has quit IRC | 20:40 | |
| *** evrardjp has quit IRC | 20:40 | |
| *** jrosser has joined #openstack-ansible | 20:40 | |
| *** poopcat has quit IRC | 20:40 | |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Workaround nova bug https://review.opendev.org/c/openstack/openstack-ansible/+/777601 | 20:41 |
| *** tosky_ is now known as tosky | 20:42 | |
| *** mubix has joined #openstack-ansible | 20:42 | |
| *** masterpe has quit IRC | 20:42 | |
| *** manti has quit IRC | 20:42 | |
| *** trident has joined #openstack-ansible | 20:42 | |
| *** fridtjof[m] has quit IRC | 20:42 | |
| *** janno has quit IRC | 20:43 | |
| *** MrClayPole has quit IRC | 20:43 | |
| *** dasp has quit IRC | 20:43 | |
| *** ioni has quit IRC | 20:43 | |
| *** cp- has joined #openstack-ansible | 20:43 | |
| *** ioni has joined #openstack-ansible | 20:43 | |
| *** janno has joined #openstack-ansible | 20:44 | |
| *** dasp has joined #openstack-ansible | 20:45 | |
| mgariepy | hmm. placement issue when upgrading from U to V. | 20:57 |
| mgariepy | anyone seen it ? python_venv_build failed | 20:58 |
| mgariepy | on 2 out of 3 nodes. | 20:58 |
| *** ioni has quit IRC | 21:01 | |
| *** ajg20 has joined #openstack-ansible | 21:02 | |
| *** csmart has joined #openstack-ansible | 21:04 | |
| mgariepy | anyone seens this before? http://paste.openstack.org/show/803022/ | 21:10 |
| mgariepy | one did install correctly but the 2 other did fail with that | 21:10 |
| *** manti has joined #openstack-ansible | 21:24 | |
| jrosser | which host is that log from? | 21:24 |
| mgariepy | i was the placement container on the 3 infra host | 21:25 |
| mgariepy | in the first run only the first infra host passed.. | 21:25 |
| jrosser | it should not be trying to build the placement code on the placement container | 21:26 |
| jrosser | for some reason it’s not got it from the repo server | 21:26 |
| mgariepy | then after trying various things the second one did pass. | 21:26 |
| mgariepy | hmm. | 21:26 |
| jrosser | could be lsync fail on the repo containers | 21:26 |
| jrosser | 1-in-n usually points to loadbalancer induced trouble | 21:27 |
| *** fridtjof[m] has joined #openstack-ansible | 21:29 | |
| mgariepy | restated lsyncd.. | 21:31 |
| mgariepy | hmm. same issue. | 21:31 |
| jrosser | you need to look through more if the venv build log and see it try to get the placement wheel from the repo server, and why that’s not happened | 21:34 |
| mgariepy | after rerunning another time, the downlaod seems to have passed. i'll debug the repo conrtainer tomorrow. | 21:37 |
| mgariepy | thanks for you help jrosser | 21:37 |
| mgariepy | it was working great until that :) haha | 21:37 |
| jrosser | hmm really has the feel of inconsistent content in the repo servers somehow | 21:38 |
| mgariepy | it was running so smoothly until that haha | 21:38 |
| mgariepy | anyway, i'll check it tomorrow | 21:39 |
| mgariepy | and keep you updated if i find something. | 21:39 |
| *** pto has joined #openstack-ansible | 21:46 | |
| mgariepy | hmm. is there a way to force re-download of the file ? | 21:48 |
| mgariepy | -e venv_rebuild=true ? | 21:48 |
| *** pto has quit IRC | 21:50 | |
| *** jbadiapa has quit IRC | 21:52 | |
| openstackgerrit | David Moreau Simard proposed openstack/openstack-ansible master: scripts-library: simplify ara setup https://review.opendev.org/c/openstack/openstack-ansible/+/777696 | 22:26 |
| openstackgerrit | David Moreau Simard proposed openstack/openstack-ansible-tests master: test-ansible-env-prep: simplify ara setup https://review.opendev.org/c/openstack/openstack-ansible-tests/+/777697 | 22:27 |
| *** luksky has quit IRC | 22:35 | |
| *** bradm has joined #openstack-ansible | 22:45 | |
| jonher | mgariepy --extra-vars "venv_rebuild=true" worked for me, me and noonedeadpunk spoke of a issue earlier today on victoria: https://opendev.org/openstack/ansible-role-python_venv_build/commit/9df14bdd0ba8221abd43ce26190f13d48ec8c08f this might be what you are seeing | 22:51 |
| *** gshippey has quit IRC | 23:33 | |
| *** mathlin has joined #openstack-ansible | 23:33 | |
| *** aedan1 is now known as aedan | 23:34 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!