| noonedeadpunk | mornings | 09:23 |
|---|---|---|
| hamidlotfi_ | Hi there, | 09:25 |
| hamidlotfi_ | I am looking for a comparison between OSA and other deployment methods for deploying OpenStack, can anyone help me? | 09:25 |
| jrosser | good morning | 09:27 |
| noonedeadpunk | nah, I don't think that exists. and that might be slightly intentionally as well | 09:27 |
| noonedeadpunk | as it's more or less matter of personal preferences | 09:28 |
| noonedeadpunk | that will be different depending whom you ask and will be subjective as well | 09:29 |
| hamidlotfi_ | noonedeadpunk: oh, Does this mean that technical issues cannot be involved? | 09:31 |
| noonedeadpunk | um, like what? | 09:32 |
| hamidlotfi_ | noonedeadpunk like performan | 09:32 |
| hamidlotfi_ | sorry , | 09:33 |
| hamidlotfi_ | like performance, security, flexibility, scalibility | 09:33 |
| hamidlotfi_ | and somethings like this | 09:34 |
| jrosser | who is going to make that judgement because they are not absolute | 09:34 |
| jrosser | your view of what is important for security in your context is probably different from mine | 09:34 |
| hamidlotfi_ | It was quite convincing | 09:35 |
| hamidlotfi_ | The only option can be ease of work and use from the user side | 09:35 |
| hamidlotfi_ | I mean less technical involvement of the end user to use this | 09:36 |
| hamidlotfi_ | Can this be a criterion? | 09:37 |
| noonedeadpunk | I'd say all deployment tools are more or less toolboxes that does implement user design. If user not aware of what they want to reach it would be tough with any tool, IMO | 09:43 |
| hamidlotfi_ | Yes you are right. | 09:53 |
| noonedeadpunk | as still you need to know how openstack works overall to operate it further | 09:54 |
| noonedeadpunk | and then it's more about what you like most - docker, k8s, bare metal deployments, puppet, etc | 09:55 |
| noonedeadpunk | execution speed or flexability | 09:55 |
| noonedeadpunk | read docker files or ansible code, etc | 09:55 |
| hamidlotfi_ | Got it. thank you | 09:58 |
| noonedeadpunk | so it's really hard to compare such things to come up with smth viable. And then comparing features is useless, since this changes very quickly | 10:01 |
| hamidlotfi_ | Yes absolutely right. | 10:02 |
| noonedeadpunk | and toolings more or less are in mode - come up with a usecase and it will be merged in many cases | 10:04 |
| jrosser | thats probably most true for the community tools rather than the vendor ones | 10:07 |
| noonedeadpunk | yeah, true | 10:08 |
| hamidlotfi_ | So considering that these are all tools, why does one tool have more users and one less? | 10:09 |
| noonedeadpunk | Or well. Depending on how big customer you are :D | 10:09 |
| hamidlotfi_ | It can't be just personal taste. | 10:10 |
| noonedeadpunk | some are backed buy bigger orgs focusing on private deployments with plenty of their own customers | 10:13 |
| noonedeadpunk | and working with some vendors more closely | 10:13 |
| noonedeadpunk | And yes - plenty of PPL just love Docker for $reason | 10:15 |
| hamidlotfi_ | I had not looked at the matter from this point of view. | 10:17 |
| noonedeadpunk | So they kinda see "Docker" on the main page and just go with that | 10:18 |
| jrosser | hamidlotfi_: is this all still to do with needing the storage network on your controllers? | 10:20 |
| hamidlotfi_ | 😅 | 10:22 |
| jrosser | or was that someone else? i am forgetting | 10:22 |
| hamidlotfi_ | No, the problem was with the network provider. | 10:24 |
| opendevreview | Merged openstack/openstack-ansible-os_ironic stable/2023.2: Stop generating ssh keypair for ironic user https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/903543 | 11:32 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/2023.2: Bump SHAs for 2023.2 (Bobcat) https://review.opendev.org/c/openstack/openstack-ansible/+/905584 | 11:59 |
| noonedeadpunk | Would be great to review bumps: https://review.opendev.org/q/topic:%22bump_osa%22+status:open | 12:01 |
| jrosser | i need to decide how to proceed with https://review.opendev.org/c/openstack/openstack-ansible-lxc_container_create/+/893229 | 12:05 |
| jrosser | and a question is, do we expect there to be one, and only one entry for `lxc.mount.auto` in the config file | 12:05 |
| noonedeadpunk | that is actually good question | 12:15 |
| noonedeadpunk | jrosser: it kinda feels that we're doing things terribly wrong atm | 12:18 |
| jrosser | yes, as you can absolutely make a huge mess here | 12:19 |
| jrosser | like create container, update this var, re-run playbook -> fail | 12:19 |
| noonedeadpunk | Like - we are placing LXC config template in lxc_hosts role: https://opendev.org/openstack/openstack-ansible-lxc_hosts/src/branch/master/templates/lxc-openstack.conf.j2 | 12:19 |
| noonedeadpunk | then we're trying to do lineinfile instead of blockinfile or just a proper template | 12:20 |
| jrosser | oh well that is different | 12:20 |
| noonedeadpunk | is it? because we pass it as "lxc config"? https://opendev.org/openstack/openstack-ansible-lxc_container_create/src/branch/master/tasks/lxc_container_create_dir.yml#L21 | 12:20 |
| jrosser | this is in /var/lib/lxc/<container-name>/config | 12:20 |
| jrosser | right but then in the lxc_container_create role there is a whole ton of config specific to that containers host_vars/group_vars | 12:21 |
| jrosser | like the bind mounts or whatever | 12:21 |
| noonedeadpunk | I guess I'm trying to understand now if(and what) community.general.lxc_container adds to the config | 12:23 |
| noonedeadpunk | And why we're not using `container_config` instead, for example... | 12:25 |
| noonedeadpunk | Or not template the config file to the destination directly and feed it to lxc_container | 12:25 |
| noonedeadpunk | Looking at code it feels that some refactoring there is needed to allow altering anything in there | 12:29 |
| noonedeadpunk | (and will reduce amount of tasks a lot) | 12:31 |
| jrosser | i expect that internals of lxc create /var/lib/lxc/<container-name>/* | 12:31 |
| noonedeadpunk | We run lineinfile towards "/var/lib/lxc/{{ inventory_hostname }}/config" in 6 tasks there | 12:32 |
| jrosser | and updating anything at all in the config cannot use community.general.lxc_container | 12:32 |
| noonedeadpunk | So you mean this can't be used? https://docs.ansible.com/ansible/latest/collections/community/general/lxc_container_module.html#parameter-container_config | 12:32 |
| noonedeadpunk | `A list of key=value options to use when configuring a container.` | 12:33 |
| jrosser | i expect we can use that at container creation | 12:33 |
| jrosser | but currently we support adding a bind mount or something after creation | 12:33 |
| jrosser | update config / restart container -> new mount | 12:33 |
| jrosser | so i suspect this is the reason that the code is the way it is | 12:34 |
| noonedeadpunk | ah, ok, gotcha what you mean. that lxc_container will not alter config afterwards. true | 12:34 |
| jrosser | totally true that its a bit gross though | 12:34 |
| jrosser | for this patch if mine, the lxc docs are not really saying if lxc.mount.auto can appear more than once | 12:35 |
| noonedeadpunk | iirc in the default config they're present only once | 12:36 |
| * jrosser asks in #lxc | 12:36 | |
| jrosser | that would make the patch simpler | 12:37 |
| jrosser | just replace any line starting with `lxc.mount.auto` | 12:37 |
| noonedeadpunk | ok, so that's the default config: https://paste.openstack.org/show/blQdgWOQXWD7bpQs0nly/ | 13:30 |
| noonedeadpunk | I actually think that we should combine all settings we need together and place them as "blockinfile"... | 13:31 |
| noonedeadpunk | I just clean forgot how we handled ordering issue for shadowing mounts... | 13:34 |
| noonedeadpunk | oh crap: https://opendev.org/openstack/openstack-ansible/src/branch/master/playbooks/common-tasks/os-lxc-container-setup.yml#L51-L86 | 13:39 |
| noonedeadpunk | so in fact no blockinfile possible then. as we insert in random places later.... | 13:41 |
| jrosser | yes, theres messing config for creation, post-creation and service configuration | 13:55 |
| noonedeadpunk | I"d say - let's assume it can be used only once | 14:11 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-lxc_container_create master: Allow LXC container auto mounts to be customised https://review.opendev.org/c/openstack/openstack-ansible-lxc_container_create/+/893229 | 15:04 |
| opendevreview | James Denton proposed openstack/openstack-ansible master: [WIP] Add support for Octavia testing with OVS/OVN https://review.opendev.org/c/openstack/openstack-ansible/+/894811 | 15:27 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_magnum master: Add job to test Vexxhost cluster API driver https://review.opendev.org/c/openstack/openstack-ansible-os_magnum/+/905199 | 15:35 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-ops master: WIP - Bootstrapping playbook https://review.opendev.org/c/openstack/openstack-ansible-ops/+/902178 | 15:45 |
| opendevreview | Merged openstack/openstack-ansible stable/2023.1: Bump SHAs for 2023.1 https://review.opendev.org/c/openstack/openstack-ansible/+/905346 | 16:46 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Bump ansible version to 2.15.8 https://review.opendev.org/c/openstack/openstack-ansible/+/905619 | 17:43 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Update ansible collections. https://review.opendev.org/c/openstack/openstack-ansible/+/905620 | 17:47 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Update global-requirements-pins https://review.opendev.org/c/openstack/openstack-ansible/+/905621 | 17:51 |
| noonedeadpunk | jrosser: fwiw, bump tool is capable of bumping collections | 17:53 |
| jrosser | oooh | 17:53 |
| jrosser | i was just looking also at galera | 17:53 |
| jrosser | we could do 11.2.2 i think | 17:53 |
| noonedeadpunk | but it's making slightly extra changes, which are weird... but still better then doing it manually | 17:53 |
| noonedeadpunk | that was diff of running it: https://paste.openstack.org/ | 17:54 |
| noonedeadpunk | ugh, paste broke :( | 17:55 |
| noonedeadpunk | https://paste.ubuntu.com/p/78RJxdZ6mv/ | 17:55 |
| noonedeadpunk | so nasty v6.0.0 > 6.0.0 | 17:55 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-galera_server master: Bump version to latest stable release of MariaDB 10.11 https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/905622 | 17:56 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-galera_server master: Bump version to 11.2 stable release https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/905623 | 17:56 |
| jrosser | i also could not remember why we had held back the pip version recently | 17:56 |
| noonedeadpunk | jrosser: regarding 11.2 - meh, I'd use just LTS there | 17:56 |
| jrosser | i wonder if the was centos/rocky8 | 17:56 |
| noonedeadpunk | 11.2 supported until Nov 2024 | 17:57 |
| jrosser | oh i completely missed the LTS thing | 17:57 |
| noonedeadpunk | rabbitmq 3.13 released iirc | 17:57 |
| noonedeadpunk | they removed? cqv1 (or just switched default there) | 17:58 |
| jrosser | speaking of which i guess oslo still not merging the other quorum queues stuff | 17:58 |
| noonedeadpunk | yeah, something was merged, but far from everything | 17:59 |
| jrosser | not sure i see packages for rabbitmq 3.13 | 18:01 |
| noonedeadpunk | dunno, was just reading their blogpost yesterday https://blog.rabbitmq.com/posts/2024/01/3.13-release/ | 18:11 |
| noonedeadpunk | from interesting, is that mirrored queues v1 are performming better then when switched to v2 | 18:12 |
| noonedeadpunk | so in fact, we shouldn't rush there until switch to quorum queues | 18:12 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!