| *** ysandeep|out is now known as ysandeep | 04:53 | |
| opendevreview | OpenStack Release Bot proposed openstack/openstack-ansible stable/yoga: Update .gitreview for stable/yoga https://review.opendev.org/c/openstack/openstack-ansible/+/847242 | 06:33 |
|---|---|---|
| opendevreview | OpenStack Release Bot proposed openstack/openstack-ansible stable/yoga: Update TOX_CONSTRAINTS_FILE for stable/yoga https://review.opendev.org/c/openstack/openstack-ansible/+/847243 | 06:33 |
| opendevreview | OpenStack Release Bot proposed openstack/openstack-ansible master: Update master for stable/yoga https://review.opendev.org/c/openstack/openstack-ansible/+/847244 | 06:33 |
| *** ysandeep is now known as ysandeep|afk | 07:30 | |
| opendevreview | Merged openstack/openstack-ansible stable/yoga: Update .gitreview for stable/yoga https://review.opendev.org/c/openstack/openstack-ansible/+/847242 | 08:02 |
| opendevreview | Merged openstack/openstack-ansible stable/yoga: Update TOX_CONSTRAINTS_FILE for stable/yoga https://review.opendev.org/c/openstack/openstack-ansible/+/847243 | 08:02 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Continue Zed development https://review.opendev.org/c/openstack/openstack-ansible/+/847272 | 08:12 |
| jrosser_ | should we drop rocky-8 and centos-8-stream straight away? | 08:19 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_manila master: Create backends when running against manila_share https://review.opendev.org/c/openstack/openstack-ansible-os_manila/+/847276 | 08:26 |
| noonedeadpunk | Yeah, I think so. I just wanted to see with my eyes that they fail first :D | 08:41 |
| jrosser_ | hmm i wonder if we are pulling in openstacksdk 0.99 https://zuul.opendev.org/t/openstack/build/41651e569b0a4170994f1bb8ed68fbfa/log/job-output.txt#9532 | 09:08 |
| jrosser_ | yeah https://github.com/openstack/requirements/blob/master/upper-constraints.txt#L456 | 09:11 |
| jrosser_ | thats going to end badly i think | 09:11 |
| *** ysandeep|afk is now known as ysandeep | 09:20 | |
| jrosser_ | noonedeadpunk: i wonder if we now need to add some stuff to globally edit the u-c file we download | 09:29 |
| jrosser_ | because changing the openstacksdk version constraint everywhere is going to be a huge mess otherwise | 09:29 |
| noonedeadpunk | I think it all depends on collections version we are to use tbh | 09:31 |
| noonedeadpunk | As basically we need 0.99 as well as master sdk... | 09:31 |
| jrosser_ | and also anywhere that sdk is used inside services, which might have different version requirements again | 09:32 |
| jrosser_ | perhaps the first thing to do is be more in control of the openstacksdk version that gets used in the ansible-runtime venv | 09:33 |
| jrosser_ | then we can be a bit independant about that compared to what we deploy with the services | 09:33 |
| noonedeadpunk | maybe it's time to try https://review.opendev.org/c/openstack/openstack-ansible/+/836382 again? | 09:35 |
| noonedeadpunk | btw let me bump collections anyway... | 09:35 |
| noonedeadpunk | because tbh, for system scopes in Zed we one way or another should have 0.99.... | 09:36 |
| noonedeadpunk | maybe not not though | 09:36 |
| jrosser_ | omg there are so many moving parts here :( | 09:36 |
| jrosser_ | for tooling we have here, using the published collections and sdk 0.99 just doesnt work | 09:37 |
| jrosser_ | we have to pin to < 0.99 | 09:37 |
| noonedeadpunk | I'd say - there's sooooo much mess in clients/collections/sdk now | 09:38 |
| noonedeadpunk | and eventually on master collections some things were already fixed.... | 09:41 |
| noonedeadpunk | or well, a lot of | 09:41 |
| noonedeadpunk | but yes. overriding it is another way around | 09:44 |
| noonedeadpunk | We will fail at least here I guess https://review.opendev.org/c/openstack/ansible-collections-openstack/+/845304 | 09:44 |
| noonedeadpunk | *here - on image upload in tempest role | 09:45 |
| noonedeadpunk | oh my, for god's sake https://pypi.org/project/ansible-config-template/ | 09:47 |
| noonedeadpunk | how we should deprecate that//// | 09:47 |
| jrosser_ | i know | 09:51 |
| jrosser_ | did we do that my mistake or has it come from the tripleo side? | 09:52 |
| jrosser_ | *by mistake | 09:52 |
| noonedeadpunk | I think we did it intentionally before galaxy times | 09:54 |
| noonedeadpunk | as that was the neatest way to install plugin I guess | 09:54 |
| jrosser_ | i do remember a time when there was lots of discussion if to publish ansible things to pypi | 09:55 |
| noonedeadpunk | but then galaxy came in I believe | 09:56 |
| noonedeadpunk | ok. then let's update readme, publish on galaxy and drop from pypi as a next step... | 10:00 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-config_template master: Update documentation about installation methods https://review.opendev.org/c/openstack/ansible-config_template/+/847302 | 10:16 |
| noonedeadpunk | Regarding config_template. My suggestion is to merge deperecation notice of pypi and publish to galaxy. Release 2.0.0 version with that. Afterwards drop publishing to pypi and setup.py and do another release with that | 10:23 |
| noonedeadpunk | regarding master branch - maybe we should land several bugfixes like https://review.opendev.org/c/openstack/openstack-ansible-os_manila/+/847276 https://review.opendev.org/c/openstack/openstack-ansible/+/846799 and https://review.opendev.org/c/openstack/openstack-ansible/+/846580 and backport them? | 10:25 |
| noonedeadpunk | there's also bug with etcd/zun which I will work on today | 10:25 |
| noonedeadpunk | regarding sdk/collections let's see how bad master is regarding collections. Maybe it's close to working state today and jsut needs small push | 10:26 |
| noonedeadpunk | So for config_template basically we need 2 things now I guess? https://review.opendev.org/c/openstack/ansible-config_template/+/847302 and https://review.opendev.org/c/openstack/ansible-config_template/+/846391 | 10:27 |
| noonedeadpunk | damiandabrowski[m]: having some reviews would be helpful | 10:28 |
| damiandabrowski[m] | okok | 11:04 |
| *** dviroel|afk is now known as dviroel | 11:20 | |
| noonedeadpunk | jrosser_: yes, we need to install sdk<0.99 in ansible venv... | 11:21 |
| noonedeadpunk | https://hackmd.io/7NtovjRkRn-tKraBXfz9jw?view that's todo list | 11:21 |
| jrosser_ | we need a way to override parts of u-c for the ansible venv then | 11:22 |
| jrosser_ | i think we should have a list of things which should be excluded from u-c | 11:25 |
| jrosser_ | and then another list of things that we add in | 11:25 |
| jrosser_ | for the ansible venv | 11:25 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Continue Zed development https://review.opendev.org/c/openstack/openstack-ansible/+/847272 | 11:31 |
| jrosser_ | the messaging around sdk and collection versions is not totally clear | 11:42 |
| opendevreview | Merged openstack/openstack-ansible-os_manila stable/yoga: Updated from OpenStack Ansible Tests https://review.opendev.org/c/openstack/openstack-ansible-os_manila/+/846778 | 12:16 |
| opendevreview | Merged openstack/openstack-ansible-os_masakari stable/yoga: Updated from OpenStack Ansible Tests https://review.opendev.org/c/openstack/openstack-ansible-os_masakari/+/846779 | 12:30 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Continue Zed development https://review.opendev.org/c/openstack/openstack-ansible/+/847272 | 12:41 |
| spatel | jrosser_ noonedeadpunk didn't we fix this issue in wallaby ? - TASK [python_venv_build : Fail if wheels are not built for multiple hosts] | 12:41 |
| spatel | This is related rally | 12:41 |
| noonedeadpunk | spatel: nope https://review.opendev.org/c/openstack/openstack-ansible/+/845685 | 12:42 |
| spatel | I meant disable rally | 12:42 |
| spatel | oh!! so its in WIP | 12:42 |
| noonedeadpunk | seems that rabbitmq install is broken on bionic for Wallaby | 12:43 |
| noonedeadpunk | so CI's not passing and didn't have time to follow that up | 12:43 |
| spatel | good to know, its kind of annoying to see that error every time build new deployment :) | 12:43 |
| spatel | I will see if i can fix it in my free time.. | 12:45 |
| jrosser_ | noonedeadpunk: didnt we have a patch to switch wallaby to cloudsmith, maybe i think of something else | 12:53 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Continue Zed development https://review.opendev.org/c/openstack/openstack-ansible/+/847272 | 12:54 |
| noonedeadpunk | I made a backport but it failed close to same way | 12:54 |
| jrosser_ | aha https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/846122 | 12:54 |
| noonedeadpunk | ok it failed way worse :) | 12:55 |
| *** ysandeep is now known as ysandeep|afk | 13:39 | |
| *** ysandeep|afk is now known as ysandeep | 13:47 | |
| *** ThiagoCMC is now known as UltraForgot | 14:15 | |
| UltraForgot | Hi folks! I'm trying to deploy OSA AIO, but the `scripts/bootstrap-ansible.sh` is failing to clone `python_venv_build`, saying "fatal: reference is not a tree". Any idea? I'm trying to deploy OSA 23.0.0 | 14:17 |
| UltraForgot | I also tried 22.0.0, same error. | 14:17 |
| noonedeadpunk | UltraForgot: if I was you, I'd try latest tag in each series | 14:19 |
| noonedeadpunk | So 22.4.4 or 23.3.1 | 14:19 |
| noonedeadpunk | I believe this is actually depth issue (ie, how deep git clones repo). By default we set it to 20. But since 22.0.0 for some repos there could be more then 20 patches landed | 14:20 |
| noonedeadpunk | Now we hopefully fixed that with replacing depth with shallow_since | 14:21 |
| noonedeadpunk | but anyway I'd say it's worth using latest version in series | 14:21 |
| noonedeadpunk | Whole 22.*.* is still Victoria, whole 23.*.* is still Wallaby | 14:22 |
| UltraForgot | Hmm... I see, makes sense, thanks! I'll try. | 14:24 |
| UltraForgot | I was planning to experiment with upgrades from 22.0.0, some minor versions pump first, then 23.0.0, and etc... For a PoC | 14:25 |
| noonedeadpunk | UltraForgot: oh, well. Then try setting `role_clone_default_depth: 50` https://opendev.org/openstack/openstack-ansible/src/branch/master/scripts/get-ansible-role-requirements.yml#L176 | 14:26 |
| noonedeadpunk | likely, you can set that also in user_variables.yml but not 100% sure | 14:27 |
| noonedeadpunk | but if you want to jump between releases it's usually not required to do minor upgrade first. So you for example you can jump from 22.0.0 to 23.3.1 directly | 14:28 |
| jrosser_ | 22.0.0 will be the very first release of that branch that we made | 14:29 |
| noonedeadpunk | There was 1 issue though with rabbitmq version, that if you upgrade from bionic to focal, you must upgrade to some minor release first. | 14:29 |
| jrosser_ | i can bet almost 100% there will be something broken | 14:29 |
| noonedeadpunk | oh, yes, it will :D | 14:29 |
| jrosser_ | don't we generally make 22.1.0 be the first stable release we have increased confidence in | 14:30 |
| UltraForgot | Cool, 23.2.0 worked lol | 14:36 |
| *** ysandeep is now known as ysandeep|out | 15:21 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Don not try to run etcd in serial mode https://review.opendev.org/c/openstack/openstack-ansible/+/847362 | 15:32 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Do not try to run etcd in serial mode https://review.opendev.org/c/openstack/openstack-ansible/+/847362 | 15:32 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_zun master: Move etcd installation to playbook https://review.opendev.org/c/openstack/openstack-ansible-os_zun/+/847363 | 15:34 |
| UltraForgot | Something else is happening now... lol | 16:15 |
| UltraForgot | tested multiple versions | 16:15 |
| UltraForgot | The TASK systemd_mount Download EPEL gpg keys is failing, but I'm on Ubuntu 20.04. | 16:15 |
| jrosser_ | having some context around that would help | 16:21 |
| jrosser_ | which playbook and pasting some helpful output to paste.opendev.org | 16:21 |
| *** frenzy_friday is now known as frenzyfriday|PTO | 16:21 | |
| UltraForgot | I'm trying to run `scripts/bootstrap-aio.sh` and that TASK fails. | 16:22 |
| UltraForgot | OSA version 23.3.0 | 16:23 |
| UltraForgot | It says: "The conditional check ''s3fs' in systemd_mount_types' failed. | 16:24 |
| spatel | Quick question do you guys prefer hw_disk_bus=scsi property for glance image? | 16:28 |
| spatel | or hw_disk_bus=virtio ? | 16:28 |
| UltraForgot | VirtIO SCSI FTW | 16:28 |
| spatel | ? | 16:29 |
| jrosser_ | noonedeadpunk: from a focal job: `2021-12-14T09:41:52,047 Looking in indexes: https://mirror-int.dfw.rax.opendev.org/pypi/simple, https://mirror-int.dfw.rax.opendev.org/wheel/centos-8-x86_64` | 16:31 |
| jrosser_ | UltraForgot: thats really not enough information to be able to help unfortunately | 16:35 |
| UltraForgot | ok, well, I'm following the bootstrap AIO and it's stopping on that task | 16:43 |
| UltraForgot | This one: https://docs.openstack.org/openstack-ansible/wallaby/user/aio/quickstart.html | 16:44 |
| UltraForgot | I'll try again (it's a test in a VM, no big deal)... =) | 16:44 |
| jrosser_ | we have some requirements trouble with horizon https://github.com/fusionbox/django-pyscss/blob/master/setup.py#L14 | 16:57 |
| jrosser_ | vs. https://github.com/openstack/requirements/blob/master/upper-constraints.txt#L75 | 16:58 |
| jrosser_ | which results in this https://paste.opendev.org/show/bMIVHl4BBfo6TUo05bqm/ | 16:59 |
| UltraForgot | Ok, tried again bootstrap-aio.sh from scratch and it passed that part! Now it failed in TASK [systemd_networkd : Create systemd-networkd network network(s)], error: no filter named 'ipaddr' - Any idea? | 17:18 |
| UltraForgot | Seems fixed in 22.4.2, gonna try again! | 17:21 |
| jrosser_ | UltraForgot: this is the price to pay for using older versions | 17:24 |
| jrosser_ | in this case the early 22.x releases were ok at the time that they were made | 17:24 |
| jrosser_ | but a ansibke unfortunately made breaking changes particularly to the ipaddr module so that’s only fixed in newer tags | 17:25 |
| noonedeadpunk | huh, metal jobs passed https://review.opendev.org/c/openstack/openstack-ansible/+/847272 with master collections.... | 17:28 |
| UltraForgot | jrosser_, okay, thanks! | 17:29 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/yoga: Return facts gathering to pre_tasks for masakari https://review.opendev.org/c/openstack/openstack-ansible/+/847260 | 17:31 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/yoga: Fix facts gathering for zun https://review.opendev.org/c/openstack/openstack-ansible/+/847261 | 17:33 |
| damiandabrowski[m] | noonedeadpunk: please let me know if I missed reviews for some urgent changes | 17:35 |
| jrosser_ | noonedeadpunk: yes I saw the metal all passed - does that mean we still don’t run horizon for metal jobs….. | 17:53 |
| jrosser_ | horizon fails certainly for the focal lxc one | 17:53 |
| UltraForgot | Oops... Just realized I'm running on M1 Parallels, it's arch64... Nobody building an amr64-based cloud around, right? lol | 17:55 |
| UltraForgot | No 22.4.2/ubuntu-20.04-aarch64 ... :-D | 17:56 |
| jrosser_ | I did work with OSA for aarch64 a while ago but could never get a CI job to complete inside 3 hours | 18:13 |
| jrosser_ | that’s probably better now that there should be aarch64 python wheels already built in the CI environment | 18:14 |
| UltraForgot | Any tips to point it to a repository that contains the aarch64 container images? | 18:17 |
| noonedeadpunk | oh, well, that's yet another thing that likely not needed on newer releases :D | 18:17 |
| UltraForgot | Oh no... lol | 18:17 |
| UltraForgot | Which one was the first? | 18:17 |
| noonedeadpunk | but eventually, for lxc images (if you mean them) arch should be respected from what I see | 18:18 |
| UltraForgot | Yep, the lxc images | 18:18 |
| UltraForgot | I didn't work for 22.4.2 | 18:18 |
| noonedeadpunk | at least it's stated there https://opendev.org/openstack/openstack-ansible-lxc_hosts/src/branch/stable/victoria/vars/ubuntu-20.04.yml#L18 | 18:18 |
| noonedeadpunk | but we switched to building images with debootstrap starting Wallaby only | 18:19 |
| UltraForgot | Cool! It looks like it should work in Victoria - http://cdimage.ubuntu.com/ubuntu-base/releases/20.04/release have arm64 base image | 18:21 |
| jrosser_ | why not use newer code :) | 18:21 |
| noonedeadpunk | I think issue there is that ubuntu names it arm64 and ansible detects as aarch64 | 18:21 |
| UltraForgot | True | 18:21 |
| jrosser_ | I had this working and committed a bunch of stuff for arm | 18:21 |
| noonedeadpunk | But you can override that easily in usera_variables | 18:21 |
| jrosser_ | so the problems will be relatively small | 18:22 |
| UltraForgot | I'm building a PoC to prove how awesome is to upgrade OpenStack with OSA | 18:22 |
| UltraForgot | So, start with old and upgrade to new | 18:22 |
| UltraForgot | ;-) | 18:22 |
| noonedeadpunk | Just put that content under variable `lxc_cache_map` https://opendev.org/openstack/openstack-ansible-lxc_hosts/src/branch/stable/victoria/vars/ubuntu-20.04.yml#L21-L32 | 18:22 |
| noonedeadpunk | except adjust arch value to arm64 | 18:22 |
| UltraForgot | Nice! | 18:22 |
| jrosser_ | here is my effort for arm https://review.opendev.org/c/openstack/openstack-ansible/+/618305 | 18:25 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/yoga: Update doc index page https://review.opendev.org/c/openstack/openstack-ansible/+/847397 | 18:27 |
| jrosser_ | https://review.opendev.org/q/topic:arm64+owner:jrosser | 18:27 |
| UltraForgot | Cool, I'll give it a try. | 18:49 |
| *** dviroel is now known as dviroel|out | 21:14 | |
| opendevreview | Merged openstack/ansible-config_template master: Update documentation about installation methods https://review.opendev.org/c/openstack/ansible-config_template/+/847302 | 22:15 |
| opendevreview | Merged openstack/openstack-ansible-os_zun stable/yoga: Updated from OpenStack Ansible Tests https://review.opendev.org/c/openstack/openstack-ansible-os_zun/+/846780 | 23:56 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!