| *** dave-mccowan has quit IRC | 00:08 | |
| *** dotnetted has joined #openstack-ansible | 00:47 | |
| dotnetted | Hey all - Is the "cinder-manage" command available on a Victoria cinder install? If so, where? Thanks | 00:50 |
|---|---|---|
| *** spatel_ has joined #openstack-ansible | 01:24 | |
| *** spatel_ is now known as spatel | 01:24 | |
| poopcat | @dotnetted -- it's on the controller nodes. If your services are in containers, try logging into the cinder-api LXC container and sourcing '/openstack/venvs/cinder-*/bin/activate' | 02:15 |
| poopcat | otherwise, just do that on the baremetal controller | 02:15 |
| *** spatel has quit IRC | 02:15 | |
| openstackgerrit | YuehuiLei proposed openstack/openstack-ansible-haproxy_server master: setup.cfg: Replace dashes with underscores https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/789692 | 02:18 |
| *** spatel_ has joined #openstack-ansible | 02:21 | |
| *** spatel_ is now known as spatel | 02:21 | |
| *** evrardjp has quit IRC | 02:33 | |
| *** evrardjp has joined #openstack-ansible | 02:33 | |
| *** spatel has quit IRC | 02:41 | |
| dotnetted | poopcat:- aha! thanks - was missing the venv activation | 03:04 |
| *** dotnetted has quit IRC | 03:29 | |
| *** ChipOManiac has joined #openstack-ansible | 04:01 | |
| ChipOManiac | Hey all, anyone know what the problem is when "Slurp up constraints file for later re-deployment" in python_venv_build goes "constraints file not found". I've come to believe it's just that the repo is still downloading the file, but it's been a whole day already and it still throws the same error. | 04:03 |
| *** mpjetta has joined #openstack-ansible | 04:13 | |
| *** mpjetta has quit IRC | 04:18 | |
| *** oleksandry has quit IRC | 04:18 | |
| *** shyamb has joined #openstack-ansible | 04:46 | |
| *** miloa has joined #openstack-ansible | 04:59 | |
| *** miloa has quit IRC | 05:21 | |
| *** shyamb has quit IRC | 05:24 | |
| *** shyamb has joined #openstack-ansible | 05:25 | |
| *** shyam89 has joined #openstack-ansible | 05:39 | |
| *** oleksandry has joined #openstack-ansible | 05:39 | |
| *** shyamb has quit IRC | 05:41 | |
| *** poopcat has quit IRC | 06:05 | |
| *** poopcat has joined #openstack-ansible | 06:07 | |
| *** d34dh0r53 has quit IRC | 06:18 | |
| *** d34dh0r53 has joined #openstack-ansible | 06:20 | |
| *** gyee has quit IRC | 06:32 | |
| *** oleksandry has quit IRC | 06:48 | |
| *** oleksandry has joined #openstack-ansible | 06:57 | |
| *** gokhani has joined #openstack-ansible | 07:05 | |
| openstackgerrit | YuehuiLei proposed openstack/openstack-ansible-os_cinder master: setup.cfg: Replace dashes with underscores https://review.opendev.org/c/openstack/openstack-ansible-os_cinder/+/789712 | 07:05 |
| openstackgerrit | YuehuiLei proposed openstack/openstack-ansible-os_cinder master: setup.cfg: Replace dashes with underscores https://review.opendev.org/c/openstack/openstack-ansible-os_cinder/+/789712 | 07:09 |
| *** rpittau|afk is now known as rpittau | 07:12 | |
| *** SiavashSardari has joined #openstack-ansible | 07:12 | |
| openstackgerrit | YuehuiLei proposed openstack/openstack-ansible-os_horizon master: setup.cfg: Replace dashes with underscores https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/789713 | 07:14 |
| openstackgerrit | YuehuiLei proposed openstack/openstack-ansible-os_glance master: setup.cfg: Replace dashes with underscores https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/789714 | 07:16 |
| openstackgerrit | YuehuiLei proposed openstack/openstack-ansible-os_ironic master: setup.cfg: Replace dashes with underscores https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/789715 | 07:20 |
| jrosser | ChipOManiac: the file is grabbed from the repo server as part of building the venv, its not downloaded from the internet | 07:22 |
| jrosser | it's a file that should be created as part of the deployment, not an external file | 07:22 |
| *** andrewbonney has joined #openstack-ansible | 07:22 | |
| openstackgerrit | YuehuiLei proposed openstack/openstack-ansible-os_tempest master: setup.cfg: Replace dashes with underscores https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/789716 | 07:23 |
| *** oleksandry has quit IRC | 07:28 | |
| noonedeadpunk | jrosser: regarding min facts gathering (I think we already merged?) - do you recall how hardware facts are gathered to resolve ansible_facts['processor_vcpus'] ? | 07:29 |
| noonedeadpunk | As it's hardware facts, not min? | 07:29 |
| jrosser | hmm | 07:29 |
| jrosser | iirc i did have to deal with this | 07:30 |
| *** tosky has joined #openstack-ansible | 07:30 | |
| noonedeadpunk | ok-ok, just faced some issues with that on V, and can't recall what we did in regards of that in master | 07:31 |
| jrosser | oh well i think that there should be no changes on V, other than facts older than 24hours maybe leading to a surprise | 07:33 |
| openstackgerrit | Andrew Bonney proposed openstack/openstack-ansible-os_zun master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/openstack-ansible-os_zun/+/780733 | 07:41 |
| ChipOManiac | jrosser which means that something went wrong during deployment for it not to be created? | 07:42 |
| jrosser | ChipOManiac: yes, thats right | 07:42 |
| jrosser | which release are you using? | 07:42 |
| noonedeadpunk | yeah, I think in my case it's just smth simple, just pushed me to some thoughts about current state:) | 07:42 |
| ChipOManiac | jrosser Upgraded from Train to Ussuri, it's a PoC. Would doing a venv_rebuild=True work? | 07:42 |
| jrosser | well, i'm suspecting that folklore advice to use venv_rebuild=True is actually what has made this break | 07:43 |
| ChipOManiac | I barely use it. I usually resort to that when a error of this sorts pops up. | 07:44 |
| jrosser | right, well it leaves these constraints files on the repo server in an inconsistent state | 07:45 |
| ChipOManiac | Which makes it worse? | 07:46 |
| jrosser | yeah well the trouble is for the one time you use venv_rebuild=true it does indeed sometimes 'fix' things for that run | 07:47 |
| jrosser | but it will then leave the state broken for any subsequent runs | 07:47 |
| ChipOManiac | Ohwell. | 07:48 |
| ChipOManiac | Constraints error popped up. Thought it was the repo server, so I reinstalled that. Then it got stuck at creating wheels. So tried it with a venv_rebuild=Tue, and it's still stuck at crating wheels. | 07:49 |
| jrosser | right yes, so this is the symptom of having used venv_rebuild=true | 07:49 |
| jrosser | if you attach to the repo container which it's using for the venv build | 07:50 |
| jrosser | and look in /var/www/repo/os-releases/<version-number> | 07:50 |
| jrosser | i think there you will find that for some of the venv there are missing .txt files, there should be four for each | 07:51 |
| *** shyam89 has quit IRC | 07:53 | |
| *** shyamb has joined #openstack-ansible | 07:58 | |
| jrosser | noonedeadpunk: on V i think this would be it https://github.com/openstack/openstack-ansible/blob/stable/victoria/scripts/openstack-ansible.rc#L38 | 07:59 |
| noonedeadpunk | jrosser: well, just folks reported about having http://paste.openstack.org/show/804968/ while running 22.1.2 | 08:01 |
| noonedeadpunk | which looks super weird for me tbh | 08:01 |
| noonedeadpunk | but yeah, I've already found that | 08:02 |
| jrosser | thats interesting | 08:03 |
| noonedeadpunk | can't reproduce though... | 08:05 |
| jrosser | stale facts would do it | 08:06 |
| noonedeadpunk | whatever though I think. I believe they might mix up cherry-picking with checkout .... | 08:09 |
| *** ianychoi_ has quit IRC | 08:13 | |
| noonedeadpunk | would be great to have second pair of eyes on https://c98a6745a1bc52cbe8a6-6774f2c76ec10220eac79c064c510a87.ssl.cf5.rackcdn.com/789376/7/check/openstack-tox-docs/0fd26ec/docs/admin/upgrades/compatability-matrix.html as I'm not sure how valid these support matrix is... | 08:13 |
| noonedeadpunk | especially in terms of distro path | 08:14 |
| noonedeadpunk | probably centos for P and Q should be also marked in warning (as it was kind of experimental there iirc) | 08:15 |
| ChipOManiac | jrosser How'd i get those files added back? | 08:15 |
| *** shyamb has quit IRC | 08:19 | |
| noonedeadpunk | jrosser: it apeeared to happen on master as there was really weird checkout | 08:19 |
| noonedeadpunk | that should be easy to reproduce | 08:22 |
| *** shyamb has joined #openstack-ansible | 08:36 | |
| *** shyam89 has joined #openstack-ansible | 08:38 | |
| *** shyam89 has quit IRC | 08:40 | |
| *** shyam89 has joined #openstack-ansible | 08:40 | |
| *** shyamb has quit IRC | 08:41 | |
| *** shyam89 has quit IRC | 08:51 | |
| *** ChipOManiac has quit IRC | 09:00 | |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/openstack-ansible-tests/+/781059 | 09:07 |
| openstackgerrit | Xinxin Shen proposed openstack/openstack-ansible-tests master: setup.cfg: Replace dashes with underscores https://review.opendev.org/c/openstack/openstack-ansible-tests/+/789761 | 09:14 |
| noonedeadpunk | jrosser: reproduced http://paste.openstack.org/show/804970/ | 09:20 |
| jrosser | is this an AIO? | 09:21 |
| noonedeadpunk | yep | 09:21 |
| jrosser | i expect that part of bootstrap_host role gathers * facts | 09:21 |
| noonedeadpunk | they do... | 09:21 |
| jrosser | and then subsequently in the playbook we only gather a subset | 09:22 |
| noonedeadpunk | so `!all,min` kind of break things out of ci now | 09:22 |
| noonedeadpunk | https://opendev.org/openstack/openstack-ansible/src/branch/master/tests/bootstrap-aio.yml#L18 | 09:24 |
| jrosser | could you check which facts subset the mounts are in? | 09:26 |
| noonedeadpunk | I think hardware, but checking | 09:26 |
| jrosser | perhaps this all needs a sanity check for the vcpu stuff too | 09:27 |
| noonedeadpunk | yep, hardware, same as vcpu | 09:27 |
| noonedeadpunk | but adding hardware without filtering expands facts file from 6.5k to 36k | 09:28 |
| jrosser | we can use the filter, like i did on the dynamic_address_fact | 09:32 |
| jrosser | but that might need lots and lots of patches | 09:32 |
| noonedeadpunk | well, or just one but huge (in the integrated repo for playbooks) | 09:32 |
| jrosser | oh right yes, it's there isnt it | 09:33 |
| jrosser | https://github.com/openstack/openstack-ansible/blob/master/playbooks/common-tasks/dynamic-address-fact.yml#L19 | 09:34 |
| noonedeadpunk | we also need to clean up facts after boostrap-aio I guess | 09:35 |
| jrosser | hopefully this https://github.com/openstack/openstack-ansible/blob/2aa71dfebcfd618b8ee937cbec428e6f35cb90e4/tests/bootstrap-aio.yml#L88 | 09:36 |
| noonedeadpunk | oh, huh... | 09:36 |
| noonedeadpunk | why then it works.... | 09:36 |
| jrosser | though i am wondering if that is clearing facts for localhost | 09:36 |
| jrosser | and not aio1 | 09:36 |
| jrosser | as the play is targetting localhost | 09:37 |
| noonedeadpunk | yeah, I think it does... Should we just delegate it to all hosts? | 09:37 |
| noonedeadpunk | but we don't have inventory during this play, right | 09:37 |
| jrosser | we don't, but the hostname does get changed during the play | 09:38 |
| jrosser | that would be interesting to see if it has an effect | 09:38 |
| noonedeadpunk | we can actually run `ansible -m meta -a clear_facts all` after this playbook in gate-check-commit.sh | 09:39 |
| jrosser | needs some care, as the inventory will now be present but the hosts are not all there | 09:40 |
| noonedeadpunk | I think it's not running against hosts at all | 09:41 |
| jrosser | maybe instead add a second play at the end if bootstrap-aio.yml with just the clear_facts task, and have it target localhost,aio1 | 09:41 |
| noonedeadpunk | meta is run locally | 09:41 |
| jrosser | oh hmm | 09:42 |
| noonedeadpunk | it cleans up /etc/openstack_deploy/ansible_facts/ | 09:42 |
| jrosser | `causes the gathered facts for the hosts specified in the play's list of hosts to be cleared, including the fact cache` | 09:42 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Clean up gathered facts during AIO bootstrap https://review.opendev.org/c/openstack/openstack-ansible/+/789769 | 09:47 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Clean up gathered facts during AIO bootstrap https://review.opendev.org/c/openstack/openstack-ansible/+/789769 | 09:49 |
| *** shyamb has joined #openstack-ansible | 09:53 | |
| *** shyamb has quit IRC | 09:53 | |
| noonedeadpunk | tbh, doing filters for every playbook and extra facts gathering and maintaining it is kind of.... | 09:57 |
| noonedeadpunk | not sure makes much sense comparing to just gather hardware facts additionally. as mounts add up to filesize of 12k (which is twice from just min) and considering we need also some cpu info... | 10:13 |
| noonedeadpunk | and considering that setup will run regardless of having facts or not, we probably won't gain any performance profit at the end | 10:15 |
| noonedeadpunk | *of having facts cache | 10:15 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: DNM Gather additional required facts to min https://review.opendev.org/c/openstack/openstack-ansible/+/789776 | 10:15 |
| *** sshnaidm|afk is now known as sshnaidm | 10:34 | |
| *** tosky has quit IRC | 10:38 | |
| *** tosky has joined #openstack-ansible | 10:38 | |
| jrosser | hmm so using --tags rabbitmq-config,rabbitmq_server-config rewrites the config file if needed but doesnt restart the service | 10:40 |
| noonedeadpunk | ummmmm | 10:41 |
| noonedeadpunk | we don't handlers here https://opendev.org/openstack/openstack-ansible-rabbitmq_server/src/branch/master/tasks/rabbitmq_post_install.yml#L37 doh | 10:41 |
| noonedeadpunk | that's stupid.... | 10:42 |
| noonedeadpunk | ah, well, we reastart not with handlers https://opendev.org/openstack/openstack-ansible-rabbitmq_server/src/branch/master/tasks/rabbitmq_post_install.yml#L77-L78 | 10:42 |
| *** tosky_ has joined #openstack-ansible | 10:45 | |
| *** tosky has quit IRC | 10:47 | |
| noonedeadpunk | so seems the way more trickier | 10:48 |
| *** tosky_ is now known as tosky | 10:49 | |
| jrosser | oh, isnt it missing tags off here https://opendev.org/openstack/openstack-ansible-rabbitmq_server/src/branch/master/tasks/rabbitmq_post_install.yml#L77-L78 | 10:53 |
| noonedeadpunk | yeah, and actually I'm not sure if we should also add them to https://opendev.org/openstack/openstack-ansible-rabbitmq_server/src/branch/master/tasks/rabbitmq_started.yml | 10:55 |
| noonedeadpunk | as tags doesn't seem to pass through https://opendev.org/openstack/openstack-ansible-rabbitmq_server/src/branch/master/tasks/rabbitmq_restart.yml#L16 as well | 10:56 |
| openstackgerrit | Andrew Bonney proposed openstack/openstack-ansible-os_zun master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/openstack-ansible-os_zun/+/780733 | 10:58 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Gather hardware facts by default https://review.opendev.org/c/openstack/openstack-ansible/+/789784 | 11:07 |
| *** mgariepy has quit IRC | 11:08 | |
| noonedeadpunk | btw, galera tags doesn't work as expected as well | 11:09 |
| jrosser | rabbit is another role with too many small files including each other | 11:12 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-galera_server master: Add galera devel packages installation https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/789786 | 11:22 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-galera_server master: Add galera devel packages installation https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/789786 | 11:22 |
| openstackgerrit | Jonathan Rosser proposed openstack/ansible-role-pki master: WIP - create certificate authorities https://review.opendev.org/c/openstack/ansible-role-pki/+/787404 | 11:26 |
| openstackgerrit | Jonathan Rosser proposed openstack/ansible-role-pki master: WIP - Create server certificates https://review.opendev.org/c/openstack/ansible-role-pki/+/788021 | 11:26 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-rabbitmq_server master: DNM - Test PKI role https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/788032 | 11:27 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-rabbitmq_server master: Fix service restart when using tags https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/789788 | 11:27 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-rabbitmq_server master: Modernise TLS configuration https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/789789 | 11:27 |
| *** premkumarar has joined #openstack-ansible | 11:29 | |
| *** recyclehero has quit IRC | 11:35 | |
| jrosser | noonedeadpunk: do you remember where we had the trouble with rabbitmq/ssl? would i see it with keystone in just an infra AIO? | 11:35 |
| premkumarar | hi all while creating stack i got this issues. | 11:35 |
| premkumarar | Create Failed | 11:35 |
| premkumarar | Resource Create Failed: Resourceinerror: Resources.My Instance: Went To Status Error Due To "Message: No Valid Host Was Found. , Code: 500 | 11:35 |
| noonedeadpunk | jrosser: iirc we saw that only in tempest when instance was spawned and some service was rejecting to connect to rabbit (like nova-conductor) | 11:36 |
| noonedeadpunk | premkumarar: what's in nova-scheduler log? | 11:37 |
| noonedeadpunk | is nova-compute service is up in openstack compute service list? | 11:38 |
| noonedeadpunk | (and present in openstack hypervisor list) | 11:38 |
| premkumarar | noonedeadpunk nova-compute aio1 nova Enabled Up | 11:41 |
| premkumarar | cinder-scheduler aio1-cinder-api-container-845d8e39 nova Enabled Up | 11:41 |
| premkumarar | but cinder volume is down | 11:42 |
| premkumarar | cinder-volume aio1@lvm nova Enabled Down 2 days, 2 hours | 11:42 |
| *** gokhani has quit IRC | 11:42 | |
| *** gokhani has joined #openstack-ansible | 11:56 | |
| *** jbadiapa has joined #openstack-ansible | 12:02 | |
| *** mgariepy has joined #openstack-ansible | 12:13 | |
| premkumarar | Hi i have a basic question. Installed openstack aio using the ansible in Ussuri. Everything is working fine. Once i restart the system. My service are going down. After heat service completely went down and cinder volume also down. | 12:26 |
| premkumarar | Do we have any steps or document what i need to do after the restart | 12:27 |
| *** rh-jelabarre has joined #openstack-ansible | 12:29 | |
| 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 | 12:32 |
| noonedeadpunk | premkumarar: nope, I don't think we have one. This happens, because we don't do persistant loop mounts for aio | 12:33 |
| noonedeadpunk | and we create loop drives for cinder/glance/etc | 12:33 |
| noonedeadpunk | without loop drive being present, lvm group can't activate and thus cinder-volume is down | 12:33 |
| noonedeadpunk | that's smth that worth fixing, but never have time on that, as aio is used mainly for testing and not like prod envs | 12:34 |
| noonedeadpunk | I have devices set like this in aio http://paste.openstack.org/show/804973/ | 12:35 |
| noonedeadpunk | and `/dev/loop3` is used as cinder-volumes volume group on LVM | 12:36 |
| jrosser | noonedeadpunk: maybe we should convert the creation of the loop devices to systemd units | 12:54 |
| jrosser | then they should come back at reboot | 12:54 |
| *** premkumarar has quit IRC | 13:10 | |
| openstackgerrit | Arx Cruz proposed openstack/openstack-ansible-os_tempest master: Moving tripleo train job to non-voting https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/789830 | 13:18 |
| arxcruz | noonedeadpunk: jrosser ^ | 13:19 |
| arxcruz | we are working to fix the train job, meanwile, moving it to nv | 13:19 |
| *** akahat|ruck has quit IRC | 13:21 | |
| *** dpawlik has quit IRC | 13:22 | |
| *** jbadiapa has quit IRC | 13:23 | |
| *** jbadiapa has joined #openstack-ansible | 13:23 | |
| *** akahat has joined #openstack-ansible | 13:33 | |
| *** macz_ has joined #openstack-ansible | 14:05 | |
| *** macz_ has quit IRC | 14:09 | |
| *** dpawlik5 has joined #openstack-ansible | 14:27 | |
| *** dpawlik5 is now known as dpawlik | 14:32 | |
| MrClayPole | Hi All, we are currently running OSA Rocky and have a single infrastructure node running 3 x lxc Rabbit and Galera instances. We deployed 2 additional servers and deploy Galera and rabbit on these so we now have 5 Rabbit and Galera configured. We would now like to scale down the 3 x lxc Rabbitt and Galera instances to 1 on the original server. | 14:42 |
| MrClayPole | We are just wondering on the best/safest way to do this? | 14:42 |
| noonedeadpunk | That's really interesting question :) | 14:43 |
| jrosser | galera is in some way easier as you have the loadbalancer in the way | 14:43 |
| noonedeadpunk | yeah, for rabbit things might be tough, as you need to adjust all services configuration | 14:44 |
| jrosser | i guess for rabbit you'd need to remove the unwanted ones from the inventory but not delete them | 14:44 |
| jrosser | redeploy all the services to get the configs updated | 14:44 |
| noonedeadpunk | well, I'd do with galera exactly the same I guess | 14:44 |
| jrosser | then drop the old containers | 14:44 |
| noonedeadpunk | redeploy in terms of re-running roles | 14:45 |
| jrosser | yeah | 14:45 |
| noonedeadpunk | for rabbit I'd first changed services configs and only after re-configured rabbit | 14:45 |
| jrosser | then i guess also the rabbit/galera roles need re-running last to reduce the size of the clusters in their config | 14:47 |
| noonedeadpunk | oh, but for galera I'd probably run haproxy role first | 14:47 |
| noonedeadpunk | so that "master" didn't appear on dropped node | 14:48 |
| jrosser | MrClayPole: lots of handwaving here ^^^ :) - so i guess the take-away is to get the ordering right | 14:48 |
| noonedeadpunk | jrosser: btw it seems we do gather facts somewhere after bootstrap-aio.... | 14:51 |
| noonedeadpunk | in terms of all facts | 14:51 |
| noonedeadpunk | as otherwise I'd expect this to fail really badly https://review.opendev.org/c/openstack/openstack-ansible/+/789769 | 14:51 |
| *** Premkumarar has joined #openstack-ansible | 14:52 | |
| Premkumarar | noonedeadpunk is there any work around to reslove that issue | 14:52 |
| openstackgerrit | Oleksandr Yeremko proposed openstack/openstack-ansible-specs master: Protecting plaintext configs https://review.opendev.org/c/openstack/openstack-ansible-specs/+/788829 | 14:52 |
| MrClayPole | Thanks guys I had a feel this wouldn't be straight forward | 14:53 |
| noonedeadpunk | Premkumarar: nope, not really. we just were discussing that it would be great to get this fixed, but not high prio for me personally (and I guess for nobody) | 14:53 |
| *** mobuntu has joined #openstack-ansible | 14:54 | |
| jrosser | noonedeadpunk: looks like all facts are present for aio1 https://aa62d61626cb9330e709-81a8be848ef91b58aa974b4cb791a408.ssl.cf2.rackcdn.com/789769/2/check/openstack-ansible-deploy-aio_lxc-ubuntu-focal/01b1208/logs/etc/host/openstack_deploy/ansible_facts/index.html | 14:55 |
| noonedeadpunk | yep | 14:55 |
| jrosser | hmm so maybe we are hiding broken things for multinode | 14:55 |
| jrosser | i was concerned about this tbh | 14:55 |
| noonedeadpunk | and we do this somewhere between bootstrap aio and hardening | 14:56 |
| noonedeadpunk | so should be not that hard to track it down | 14:56 |
| jrosser | when i was messing with this i had a terminal open with "watch ls -l /etc/openstack_deploy/ansible_facts" | 14:56 |
| jrosser | and just ran the plays through one by one | 14:56 |
| noonedeadpunk | sounds not so fun | 14:56 |
| jrosser | it was very obvious when one fo the files went huge | 14:57 |
| *** macz_ has joined #openstack-ansible | 14:57 | |
| noonedeadpunk | yeah, will try to find where we collect things | 14:57 |
| jrosser | wtf is Connection failed: [SSL: WRONG_VERSION_NUMBER] wrong version number | 14:57 |
| noonedeadpunk | do we use smth like tls1.0 or ssl3? | 14:58 |
| noonedeadpunk | I thought actually we even set ciphers.... | 14:58 |
| noonedeadpunk | but we probably used tls1.1 | 14:58 |
| noonedeadpunk | (might be worth switching to smth more descent) | 14:59 |
| mobuntu | Hey guys, noob here. I'm trying to create two targetable host groups for my compute. But I can't seem to get the host groups to be created properly in my inventory. This is how i have them defined in my openstack_user_config.yml https://pastebin.com/qJdiddyd anyone have a working example of this? | 14:59 |
| jrosser | noonedeadpunk: ah for the TLS setup for rabbit previously it was pretty undefined so i made this https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/789789 | 14:59 |
| jrosser | so i am trying to use tls1.2 | 15:00 |
| noonedeadpunk | mobuntu: you alse need to create env.d files. spatel does have and already published that here one day... | 15:00 |
| noonedeadpunk | jrosser: ah! I believe tls1.1 shoudl be valid nowadays? Or it's already not? | 15:01 |
| noonedeadpunk | https://support.umbrella.com/hc/en-us/articles/360033350851-End-of-Life-for-TLS-1-0-1-1- hm | 15:01 |
| noonedeadpunk | seems it's eol as well | 15:01 |
| *** mgariepy has quit IRC | 15:14 | |
| *** SiavashSardari has quit IRC | 15:15 | |
| jrosser | mobuntu: from your paste you've got an identation problem with the yaml | 15:24 |
| jrosser | mobuntu: you should have something like this http://paste.openstack.org/show/804975/ | 15:25 |
| *** mgariepy has joined #openstack-ansible | 15:46 | |
| *** rpittau is now known as rpittau|afk | 15:52 | |
| * jrosser has TLS connection to rabbitmq \o/ | 16:02 | |
| jrosser | helps if you actually connect to port 5671 instead of 5672..... doh | 16:02 |
| noonedeadpunk | haha | 16:06 |
| noonedeadpunk | yeah, that's true) | 16:06 |
| noonedeadpunk | jrosser: should we also filter here? https://opendev.org/openstack/openstack-ansible/src/branch/master/playbooks/openstack-hosts-setup.yml#L49 | 16:17 |
| noonedeadpunk | hm.... interesting... | 16:20 |
| noonedeadpunk | feels like we don't need virtual there o_O | 16:28 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Don't collect virtual facts https://review.opendev.org/c/openstack/openstack-ansible/+/789926 | 16:31 |
| noonedeadpunk | sooooo. what have we decided in terms of hardware facts - gather them all or do a hook for each playbook? | 16:31 |
| *** gyee has joined #openstack-ansible | 16:33 | |
| noonedeadpunk | I think I'd vote for gathering all hardware facts - there're not too much of them I think... well, it's actually 6 times more... but dunno about having all these filters. probably we can add it to the common-tasks and just do simple include | 16:34 |
| jrosser | the filter could also be a var defined in group_vars | 16:38 |
| jrosser | that would make maintainance eaiser | 16:38 |
| noonedeadpunk | well, yeah, agree | 16:38 |
| noonedeadpunk | I just not really like that setup will run each time - despite we have things already cached or not | 16:43 |
| *** oleksandry has joined #openstack-ansible | 16:44 | |
| fridtjof[m] | hmmm, trying to run an OSA deployment right now (osa 22.1.2), and specifically glance and cinder containers are failing to start | 16:46 |
| fridtjof[m] | starting them manually in foreground mode gives me this error over and over: | 16:47 |
| fridtjof[m] | lxc-start: infra1_glance_container-3965f658: cgroups/cgfsng.c: cgfsng_monitor_create: 1264 Failed to create cgroup "(null)" | 16:47 |
| fridtjof[m] | all other containers are running fine | 16:47 |
| fridtjof[m] | weird thing is that this happens on two separate infrastructure hosts (albeit VMs on the same physical machine, but...) for exactly glance and cinder_api containers | 16:48 |
| fridtjof[m] | hosts are all running ubuntu 20.04 | 16:49 |
| *** oleksandry has quit IRC | 16:50 | |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: DNM Gather additional required facts to min https://review.opendev.org/c/openstack/openstack-ansible/+/789776 | 16:56 |
| fridtjof[m] | update: rebooted both hosts, now it works | 17:07 |
| fridtjof[m] | strange | 17:07 |
| *** Premkumarar has quit IRC | 17:15 | |
| fridtjof[m] | ugh it's happening again | 17:25 |
| *** andrewbonney has quit IRC | 17:31 | |
| fridtjof[m] | oh, both cinder-api and glance want connectivity to br-storage? | 17:34 |
| fridtjof[m] | i kind of get that for glance, as I set it up to use cinder as a backing store (which does explain the new requirement), but why does cinder-api have to get directly onto the storage network? | 17:34 |
| jrosser | fridtjof[m]: if it's the same place that cinder-volume would go when that component is not on metal, it may explain it | 17:38 |
| fridtjof[m] | I based my configuration off of this: https://docs.openstack.org/openstack-ansible/victoria/user/prod/example.html | 17:40 |
| fridtjof[m] | it's worked before on stein, but I think I did connect the infra hosts to the storage network too back then | 17:40 |
| fridtjof[m] | (i didn't this time) | 17:40 |
| jrosser | cinder-volume runs on the infra hosts, so that will almost certainly need to be on the storage network | 17:41 |
| jrosser | unless you change the config to make it containerised | 17:42 |
| fridtjof[m] | I do see a discrepancy on that page - the "IP Assignments" table implies you don't have to connect the infra host(s) to br-storage, but the following openstack_user_config then binds cinder-api and glance-api to that network | 17:42 |
| fridtjof[m] | cinder-volume in my case is only running on a separate storage host | 17:43 |
| jrosser | oh right in that example it tells you to make the adjustments in "Environment customizations" to put the cinder-volume service in containers | 17:43 |
| fridtjof[m] | infra hosts are only running cinder-api, which (from what i've read) is supposed to just control cinder-volume through br-mgmt, right? | 17:43 |
| jrosser | the cinder-volume service (as which isnt necessarily where the storage is) is also on the infra hosts | 17:44 |
| jrosser | in the example those services are using the NFS server at 172.29.244.15 | 17:45 |
| *** oleksandry has joined #openstack-ansible | 17:45 | |
| fridtjof[m] | mhm | 17:45 |
| fridtjof[m] | I get why cinder-volume needs to be on br-storage, does cinder-api need to be on there though? | 17:46 |
| fridtjof[m] | (same question for glance-api) | 17:46 |
| jrosser | i think that glance would need to talk to a ceph backend for example to upload images to the storage | 17:47 |
| jrosser | the same would go for NFS i expect | 17:47 |
| fridtjof[m] | ah, yep | 17:47 |
| jrosser | and really i think that the storage network connected to the cinder-api container is a generalisation, it's not strictly necessary | 17:47 |
| fridtjof[m] | previously I only ever ran the file backend (which i realized wasn't the best idea after i reinstalled an infra host lol) for glance | 17:48 |
| jrosser | unless like in the example the cinder-volume service is also in that container | 17:48 |
| jrosser | so rather than overcomplicate the "if this else this...." documentaion it's easier to just connect it to the bridge | 17:48 |
| fridtjof[m] | in any case, seems like i'll have to connect br-storage in any case for glance-api to work with cinder | 17:49 |
| jrosser | glace will use its own space on the NFS i think | 17:49 |
| fridtjof[m] | (not using NFS, just cinder-volume with a local LVM store) | 17:50 |
| jrosser | oh well thats the case when you have to run cinder-volume on the infra host itself | 17:50 |
| jrosser | becasue the iscsi stuff doesnt work containerised iirc | 17:50 |
| jrosser | assuming we're talking about the same thing :) | 17:51 |
| fridtjof[m] | i think i'm not running into _that_ problem :) | 17:52 |
| fridtjof[m] | (i have a single "storage" physical server with a bunch of drives for this deployment, those are one LVM VG, and cinder-volume runs on metal there) | 17:52 |
| fridtjof[m] | but for glance-api to be able to use cinder as a backend, it'll need access to the storage network (which I still have) to be able to store images there, right? | 17:53 |
| fridtjof[m] | or would iscsi not work with glance-api containerised? | 17:55 |
| jrosser | i think it's the server part of cinder-volume that needs to be on the host, as you have it now | 17:57 |
| jrosser | and yes i think you'll need the storage interface in the glance container in order to access it directly | 17:58 |
| fridtjof[m] | alright | 18:01 |
| *** oleksandry has quit IRC | 18:09 | |
| *** gokhani has quit IRC | 18:13 | |
| admin0 | how do I override network interfaces for a flat networking when diff hypervisors have diff network interface names | 18:14 |
| *** MrClayPole_ has joined #openstack-ansible | 18:19 | |
| *** MrClayPole has quit IRC | 18:23 | |
| jrosser | admin0: one simple way round that is to use a veth and a new interface, thats why you see eth12 all over the example configs, to give a well defined interface name to neutron | 18:26 |
| jrosser | alternatively, depending on how many different combinations you have this could be useful too https://docs.openstack.org/openstack-ansible/latest/user/prod/provnet_groups.html | 18:27 |
| admin0 | jrosser, exactly what i needed | 18:28 |
| mgariepy | what are you using for the network configuration ? | 18:28 |
| admin0 | many thansk | 18:28 |
| admin0 | simple bridges | 18:28 |
| admin0 | lb | 18:28 |
| admin0 | not ovs | 18:29 |
| admin0 | mgariepy, or was the question diff and i did not understand | 18:29 |
| mgariepy | i thend to just rename the interface to something like : 25G-1 , 25G-2 etc. | 18:29 |
| admin0 | netplan | 18:29 |
| mgariepy | netplan ? systemd-networkd ? | 18:29 |
| admin0 | focal 20 | 18:29 |
| admin0 | yep | 18:29 |
| admin0 | hmm..so just rename from udev :) | 18:30 |
| admin0 | 3 methods to chose from \o/ | 18:30 |
| mgariepy | http://paste.openstack.org/show/804980/ | 18:31 |
| *** recyclehero has joined #openstack-ansible | 18:32 | |
| mgariepy | netplan can rename them. but sometimes you can have ussues with the file generation (because netplan only do file generation for other backend) the match might not work on some corner case that might be fixed now. | 18:32 |
| *** Jeffrey4l has quit IRC | 18:44 | |
| *** Jeffrey4l has joined #openstack-ansible | 18:47 | |
| *** Adri2000 has joined #openstack-ansible | 18:52 | |
| *** oleksandry has joined #openstack-ansible | 19:01 | |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_glance master: [goal] Deprecate the JSON formatted policy file https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/780749 | 19:13 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_aodh master: [goal] Deprecate the JSON formatted policy file https://review.opendev.org/c/openstack/openstack-ansible-os_aodh/+/780844 | 19:15 |
| *** oleksandry has quit IRC | 19:26 | |
| admin0 | thanks mgariepy | 19:32 |
| openstackgerrit | Merged openstack/openstack-ansible-os_glance master: setup.cfg: Replace dashes with underscores https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/789714 | 19:34 |
| *** oleksandry has joined #openstack-ansible | 19:39 | |
| mgariepy | admin0, you are welcome if you have issues, do not hesitate to ping me, | 19:40 |
| *** oleksandry has quit IRC | 20:26 | |
| *** mobuntu has quit IRC | 21:41 | |
| *** gyee has quit IRC | 22:41 | |
| *** tosky has quit IRC | 23:03 | |
| *** macz_ has quit IRC | 23:06 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!