| spotz | prometheanfire or admin0 either of you know when we released OSA on Ussuri? | 00:00 |
|---|---|---|
| prometheanfire | spotz: 11 days ago for https://github.com/openstack/openstack-ansible/releases/tag/21.0.0 | 00:02 |
| *** npalladium has quit IRC | 00:02 | |
| spotz | prometheanfire: I saw https://opendev.org/openstack/openstack-ansible/commit/b6b6f6eb751a7828192ff109e38bbd7c83556a62 but could Ussuri only have been released 3 months ago?!!! I have no sense of time:( | 00:03 |
| prometheanfire | the branch date was around 3 months ago | 00:04 |
| prometheanfire | iirc | 00:04 |
| spotz | Ok I'll go with the date on this and will patch if wrong | 00:04 |
| prometheanfire | depends on what you call a 'release' | 00:05 |
| prometheanfire | I've been running it for a while, lol | 00:05 |
| openstackgerrit | Amy Marrich (spotz) proposed openstack/openstack-ansible master: Update docs for Ussuri release https://review.opendev.org/748053 | 00:06 |
| spotz | wrong date is better then wrong release:) | 00:07 |
| spotz | I feel all upstreamy todau:) | 00:07 |
| prometheanfire | I should probably update when the nova hash is updated (for the cve) | 00:09 |
| spotz | yeah | 00:10 |
| *** gyee has quit IRC | 00:26 | |
| *** spatel has quit IRC | 00:51 | |
| *** cshen has joined #openstack-ansible | 01:18 | |
| *** cshen has quit IRC | 01:22 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/ussuri: Do not override tempest_test_blacklist in CI https://review.opendev.org/747982 | 02:01 |
| ThiagoCMC | BTW, I'm seeing that Ubuntu 20.04 isn't listed for Ussuri at the documentation... Is it also accurate? | 02:07 |
| *** spatel has joined #openstack-ansible | 02:45 | |
| spatel | jrosser: noonedeadpunk when do you think this patch can get merge? - https://review.opendev.org/#/c/745696/ | 02:45 |
| spatel | everytime i have to patch manually when i re-build my lab :( | 02:47 |
| *** chenhaw has quit IRC | 02:51 | |
| *** chenhaw has joined #openstack-ansible | 02:52 | |
| *** rh-jelabarre has quit IRC | 03:08 | |
| *** chenhaw has quit IRC | 04:00 | |
| *** chenhaw has joined #openstack-ansible | 04:00 | |
| *** dave-mccowan has quit IRC | 04:07 | |
| *** spatel has quit IRC | 04:25 | |
| *** evrardjp has quit IRC | 04:33 | |
| *** evrardjp has joined #openstack-ansible | 04:33 | |
| *** shyamb has joined #openstack-ansible | 04:58 | |
| *** raukadah is now known as chandankumar | 05:44 | |
| *** redrobot has quit IRC | 05:53 | |
| *** miloa has joined #openstack-ansible | 06:04 | |
| jrosser | ThiagoCMC: spotz there are patches for updating the docs here https://review.opendev.org/746741 and https://review.opendev.org/746729 which need reviews | 06:38 |
| *** shyamb has quit IRC | 06:48 | |
| *** rohit02 has joined #openstack-ansible | 06:55 | |
| jrosser | noonedeadpunk: this fails consistently and i'm not really seeing why https://zuul.opendev.org/t/openstack/build/935188a821544edb9fc927702eefdf47/log/job-output.txt#13341 | 06:58 |
| jrosser | like the python interpreter is not pointing to the utility venv or something..... | 06:58 |
| jrosser | its from this https://review.opendev.org/#/c/747183/ | 06:58 |
| noonedeadpunk | I think I know why... | 07:00 |
| rohit02 | hey there! openstack-ansible ussuri installation failed on ubuntu 20 [focal] << http://paste.openstack.org/show/797155/>> | 07:01 |
| noonedeadpunk | oh, wait, it's on master:( | 07:01 |
| *** d34dh0r53 has quit IRC | 07:01 | |
| noonedeadpunk | and it fails on U from what I see (during the first run) | 07:03 |
| noonedeadpunk | I can recall some fix there but I believe that it has been backported... | 07:04 |
| noonedeadpunk | rohit02: good catch. we don't test it with ceph yet, so yeah, missed that | 07:05 |
| rohit02 | @noonedeadpunk: when we will expect a fix for ceph | 07:07 |
| *** cshen has joined #openstack-ansible | 07:10 | |
| noonedeadpunk | give me several mins | 07:13 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-ceph_client master: Fix ceph-client deployment for Ubuntu Focal https://review.opendev.org/748128 | 07:15 |
| noonedeadpunk | rohit02: ^ you can try cherry-picking that patch and re-running the role | 07:15 |
| noonedeadpunk | I hope we will get this fixed in 21.0.1 | 07:16 |
| *** d34dh0r53 has joined #openstack-ansible | 07:17 | |
| noonedeadpunk | jrosser: I will try to take a deeper look at this designate thing today | 07:18 |
| jrosser | oh ok, cool. just looked weird and like something we already fixed | 07:18 |
| jrosser | but i just don't see it :/ | 07:19 |
| noonedeadpunk | Me neither but I believe we did as well | 07:22 |
| *** miloa has quit IRC | 07:23 | |
| *** andrewbonney has joined #openstack-ansible | 07:26 | |
| *** tosky has joined #openstack-ansible | 07:40 | |
| openstackgerrit | Andrew Bonney proposed openstack/openstack-ansible-openstack_hosts master: Allow 'rp_filter' to be configured by individual deployments https://review.opendev.org/748132 | 07:58 |
| *** jbadiapa has joined #openstack-ansible | 08:11 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_congress stable/ussuri: Replace py2 packages with py3 https://review.opendev.org/746659 | 08:11 |
| *** shyamb has joined #openstack-ansible | 08:16 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_congress stable/ussuri: Updated from OpenStack Ansible Tests https://review.opendev.org/736346 | 08:24 |
| admin0 | been waiting since yesterday :) https://review.opendev.org/#/c/747586/ | 08:26 |
| openstackgerrit | Andrew Bonney proposed openstack/openstack-ansible stable/ussuri: Bump SHAs for stable/ussuri https://review.opendev.org/747586 | 08:30 |
| admin0 | is there a quick way to get the host where the l3-agent is just from an ip/uuid ? | 08:30 |
| *** tosky has quit IRC | 08:32 | |
| *** andrewbonney has quit IRC | 08:33 | |
| *** tosky has joined #openstack-ansible | 08:33 | |
| openstackgerrit | Merged openstack/openstack-ansible-haproxy_server stable/ussuri: Add haproxy_ssl_all_vips variable https://review.opendev.org/747891 | 08:38 |
| *** andrewbonney has joined #openstack-ansible | 08:38 | |
| jrosser | admin0: you can always deploy stable/ussuri if you are desperate, the release is just a tag on the branch | 08:40 |
| admin0 | yeah .. since focal is supported, started to upgrade some and reinstall some | 08:40 |
| *** npalladium has joined #openstack-ansible | 08:42 | |
| npalladium | Managed to get a working production setup of OpenStack with Ironic setup. Just wanted to thank you all for answering my queries and being supportive. Especially jrosser without whom I'd have been stuck. Cheers and thanks again! | 08:44 |
| jrosser | npalladium: thats great! particularly for ironic is there anything you think we need to add to the documentation, i suspect we could be better there | 08:45 |
| *** shyamb has quit IRC | 08:48 | |
| npalladium | A few things that come to mind (not just Ironic; and take them with a pinch of salt as a I'm a complete newcomer to OpenStack/similar offerings): 1. I had trouble understanding what the lb IPs were supposed to do and that they should be added to the used IPs list. 2. It wasn't immediately clear that I could turn on the Ironic UI by adding a user variable; I only found it because I was looking for it. These are the only things I can think | 08:59 |
| npalladium | off of the top of my head (I'm on cloud 9 :P). Other than that it's been a fun and illuminating experience. I will try and go through my steps to see if I can find something else; if I do I will post here or try and send a PR. | 08:59 |
| npalladium | It goes without saying that pulling off the setup would've been near impossible without OSA. I tried my hand at writing a playbook for Swift in the spirit of curiosity. Learnt a lot but couldn't pull it off. :D | 09:07 |
| npalladium | Talking of Swift, it's playbook seemned to expect Ceilometer installed but ceilometer playbook is called after Swift playbook. It worked after switching the two but the reason could be something else as I didn't look too deeply. The playbook wasn't able to get the gnocchi project id either but I just disabled it as I didn't need it. (Sorry for the spam; signing off for now) | 09:11 |
| *** spatel has joined #openstack-ansible | 09:26 | |
| *** spatel has quit IRC | 09:30 | |
| *** SecOpsNinja has joined #openstack-ansible | 09:33 | |
| SecOpsNinja | hello. so regarding the error of TASK [os_glance : Perform a Glance DB sync] "Can't locate revision identified by 'ussuri_contract01'" when upgrading to the latest version of train release, is there any sugestion how to resolv this or its better that i try to force the upgrade do ussuri release 21.0? | 09:40 |
| openstackgerrit | Merged openstack/openstack-ansible-os_masakari stable/ussuri: Add masakari-monitors to constraints https://review.opendev.org/747886 | 10:00 |
| noonedeadpunk | npalladium: there was some weirdness about swift and gnocchi relation tbh, but would be great if you could give some references where things went bad so I can check this out | 10:01 |
| noonedeadpunk | As eventually telemetry is not a requirement for swift | 10:01 |
| npalladium | noonedeadpunk: `FAILED - RETRYING: Get Gnocchi project id` is what I get and the play fails towards the end. Worked when I disabled Gnocchi so I'm assuming it's because of that. I can shared verbose logs if necessary as I've been religiously keeping track of it. | 10:06 |
| noonedeadpunk | I mean can you paste task name for which this failer message was shown? | 10:07 |
| npalladium | Get Gnocchi project id in swift_proxy_hosts.yml | 10:09 |
| noonedeadpunk | SecOpsNinja: hm, on train I don't really see how you could get required to use ussuri migrations... I think this patch has added them, so until it there should be no references.. Unless alembic itself is somehow broken https://opendev.org/openstack/glance/commit/d41362517238b9ea3741c4a565e0b73a12c55b92 | 10:09 |
| openstackgerrit | Merged openstack/openstack-ansible-os_gnocchi stable/ussuri: Fix run task conditionals https://review.opendev.org/747884 | 10:12 |
| npalladium | As for ceilometer, it failed at `TASK [os_swift : Add service users to the role]` with `FAILED - RETRYING: Add service users to the role` and "msg": "User ceilometer is not valid". | 10:18 |
| SecOpsNinja | noonedeadpunk, yep didn't find any errors when i did the checkout to latest tag of stable/train before doing ./scripts/bootstrap-ansible.sh but after resolving another error in openstack i will try to upgrade to ussuri. | 10:23 |
| openstackgerrit | Merged openstack/openstack-ansible-os_gnocchi stable/train: Fix run task conditionals https://review.opendev.org/747885 | 10:30 |
| SecOpsNinja | now im finding another strange behaviour: i had the wrong "host_bind_override" in my vlan networks, but i have updated today the new one and deployed openstack ansible again (it failed in previous glance taks in openstack-ansible setup-openstack.yml) but in my compute and network hosts there is still the incorrect mapping in /openstack/venvs/neutron-20.1.0.dev235/etc/neutron/plugins/ml2/lin | 10:30 |
| SecOpsNinja | uxbridge_agent.ini. when is this change applied in openstack-ansible? | 10:30 |
| openstackgerrit | Merged openstack/openstack-ansible master: Remove tempest_test_distro_blacklist https://review.opendev.org/747981 | 10:33 |
| openstackgerrit | amolkahat proposed openstack/openstack-ansible-os_tempest master: Added tempest ironic resources setup. https://review.opendev.org/720705 | 10:40 |
| noonedeadpunk | npalladium: but this task is run only when you have defined gnocchi hosts in openstack_user_config or set swift_gnocchi_enabled variable to true... https://opendev.org/openstack/openstack-ansible-os_swift/src/branch/master/tasks/swift_proxy_hosts.yml#L26 | 11:22 |
| noonedeadpunk | so um... are you sure you don't have gnocchi depoyed?:) | 11:22 |
| noonedeadpunk | jrosser: have you ever reserched on the way how to prohibit ppl to use ephemeral storage? | 11:33 |
| *** rh-jelabarre has joined #openstack-ansible | 11:50 | |
| *** npalladium has quit IRC | 11:51 | |
| *** Mensis89 has joined #openstack-ansible | 11:52 | |
| *** sshnaidm|afk is now known as sshnaidm|mtg | 11:56 | |
| *** rohit02 has quit IRC | 12:00 | |
| *** rohit02 has joined #openstack-ansible | 12:00 | |
| *** rohit02 has quit IRC | 12:04 | |
| noonedeadpunk | answering my questions, it's with https://docs.openstack.org/nova/latest/configuration/config.html#DEFAULT.max_local_block_devices | 12:12 |
| *** rohit02 has joined #openstack-ansible | 12:16 | |
| spotz | noonedeadpunk jrosser - We need to get https://review.opendev.org/#/c/748053/ merged as we were way out of date. I used the date we cut Ussuri but if anyone knows the real date let me know. We never announced on the lists | 12:26 |
| jrosser | there are already two unmerged patches to fix the docs? | 12:26 |
| spotz | Ok let's abandon mine and get the other merged | 12:27 |
| noonedeadpunk | yeah. I kind of wrote in ML but it was squashed with Andrew welcoming email | 12:27 |
| noonedeadpunk | my bad I guess | 12:27 |
| *** aedc has joined #openstack-ansible | 12:27 | |
| spotz | noonedeadpunk: You mind if I clarify one thing? | 12:29 |
| noonedeadpunk | sure! | 12:29 |
| noonedeadpunk | I mean sure go on :) | 12:29 |
| *** npalladium has joined #openstack-ansible | 12:29 | |
| spotz | And add the victoria stuff from mine? | 12:29 |
| noonedeadpunk | I think I have victoria part here https://review.opendev.org/#/c/746741/3 | 12:30 |
| noonedeadpunk | I just want this one to be backported to the stable/ussuri | 12:30 |
| noonedeadpunk | where Victoria bit does not have much sense | 12:30 |
| spotz | looking | 12:31 |
| spotz | Ok let's leave this one as is then | 12:32 |
| spotz | and get it through quick and then the backport so it'll make sense and be up to date | 12:33 |
| npalladium | noonedeadpunk: To clarify I do have gnocchi deployed. I was facing the issue I talked about. I later set the swift_gnocchi_enabled: False . | 12:35 |
| noonedeadpunk | Ok, but at the time you execute swift role, gnocchi should be already deployed | 12:35 |
| spotz | Ok reviewed they look good we just need to get them in | 12:36 |
| noonedeadpunk | npalladium: https://opendev.org/openstack/openstack-ansible/src/branch/master/playbooks/setup-openstack.yml#L26-L27 | 12:37 |
| noonedeadpunk | so that's the order.... | 12:37 |
| noonedeadpunk | or you had gnocchi deployed, but still was facing issue? | 12:37 |
| * noonedeadpunk needs to reproduce that as not totally clear.... | 12:39 | |
| npalladium | Gnocchi was deployed. os-gnocchi-install.yml was run before setting swift up. | 12:40 |
| noonedeadpunk | oh. I think the problem might be in missing openstacksdk or smth... can you remove no_log: true and give output? https://opendev.org/openstack/openstack-ansible-os_swift/src/branch/master/tasks/swift_proxy_hosts.yml#L27 | 12:41 |
| jrosser | looks like that task is missing delegation to the setup host | 12:42 |
| noonedeadpunk | yeah | 12:42 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_swift master: Delegate gnocchi retrievement task to setup host https://review.opendev.org/748217 | 12:50 |
| noonedeadpunk | npalladium: can you check if that solves the issue? ^ | 12:50 |
| npalladium | I already have it lying somewhere. I'll get back to you as soon as I can login to the server. Could you elaborate on the delegation part? | 12:51 |
| *** spatel has joined #openstack-ansible | 12:51 | |
| *** Adri2000 has joined #openstack-ansible | 13:05 | |
| *** cshen has quit IRC | 13:12 | |
| *** dave-mccowan has joined #openstack-ansible | 13:17 | |
| noonedeadpunk | yeah I eventually just proposed patch for that so would be awesome if you could try it one when comfortable | 13:33 |
| *** cshen has joined #openstack-ansible | 13:49 | |
| *** sshnaidm|mtg is now known as sshnaidm | 13:52 | |
| Adri2000 | hello | 14:21 |
| spatel | hello | 14:22 |
| admin0 | jrosser, https://review.opendev.org/#/c/747586/ -- is this good to go now ? | 14:22 |
| Adri2000 | on my OSA deployment what would be a proper way to override the version of a python dependency that is installed? namely I'd like to bump the version of the amqp module, which is a dependency of oslo.messaging... (because of known issue at https://docs.openstack.org/releasenotes/oslo.messaging/rocky.html#relnotes-8-1-3-stable-rocky) | 14:22 |
| spatel | noonedeadpunk: do you think we can cherry pick and add into current stable/ussuri? https://review.opendev.org/#/c/745696/ | 14:23 |
| openstackgerrit | amolkahat proposed openstack/openstack-ansible-os_tempest master: Added tempest ironic resources setup. https://review.opendev.org/720705 | 14:24 |
| noonedeadpunk | spatel: we already did? https://review.opendev.org/#/c/746006/5 | 14:24 |
| noonedeadpunk | we just need another U release to gat this change tagged | 14:24 |
| spatel | last Sunday i deployed environment with stable/ussuri branch and hit this bug.. hmm | 14:25 |
| spatel | i will give it a try again.. | 14:26 |
| noonedeadpunk | yeah... let us know if it's still an issue | 14:26 |
| spatel | noonedeadpunk: yesterday i was looking for senlin and look like OSA doesn't have that project, am i right? | 14:26 |
| noonedeadpunk | yep. you are right, we don't have that role yet | 14:27 |
| spatel | I really need that project, so what is the process to onboard one? I can totally work on that project and you guys can just provide guidance | 14:29 |
| ThiagoCMC | jrosser, awesome, thank you! :-D | 14:38 |
| *** cshen has quit IRC | 14:46 | |
| *** cshen has joined #openstack-ansible | 14:46 | |
| noonedeadpunk | spatel: sounds good:) can this wait till next tuesday so we discuss that during our regular meeting? | 14:51 |
| noonedeadpunk | I can create a repo and stuff for you once we take a decision :) | 14:51 |
| noonedeadpunk | as I'm not sure about the way it works tbh | 14:51 |
| admin0 | guys .. /etc/default/lxc-net , line#11 .. https://gist.github.com/a1git/940245645dad2a3fa5e76c292f25e4ea -- what causes ansible to put 8.8.8.8 in that line in place of config-file ? | 14:52 |
| *** alvinstarr has joined #openstack-ansible | 14:52 | |
| admin0 | in other clusters, its null .. here, osa always puts 8.8.8.8 and then lxc-dnsmasq fails saying cofig file 8.8.8.8 does not exist | 14:52 |
| spatel | noonedeadpunk: sure, No rush, right now i am just playing with to see how integration goes so far. (its awesome project and provide EC2 level auto scaling for cloud) | 14:56 |
| spatel | noonedeadpunk: This is cool video of power of senlin - https://www.youtube.com/watch?v=e9Mj7k-BoGU | 14:59 |
| jrosser | that is a really good talk in that video, i went to that in denver | 15:16 |
| admin0 | spatel, get on it .. i can help with the tests .. | 15:17 |
| jrosser | admin0: here is where the value you ask about comes from http://codesearch.openstack.org/?q=LXC_DHCP_CONFILE | 15:18 |
| jrosser | i don't know why it's 8.8.8.8 though | 15:18 |
| jrosser | becasue the default seems to be an empty string http://codesearch.openstack.org/?q=lxc_net_dhcp_config | 15:19 |
| spatel | admin0: +1 (Totally, we are doing lots of autoscaling with AWS and missing that piece in our private cloud so this is right time to play) | 15:19 |
| admin0 | i am waiting for ovn to land as well | 15:19 |
| admin0 | also what happened to the octavia documentation | 15:21 |
| ThiagoCMC | Senlin is AWESOME! | 15:21 |
| *** gyee has joined #openstack-ansible | 15:26 | |
| * noonedeadpunk wathing video | 15:28 | |
| noonedeadpunk | I'm all for adding more projects support:) Esp if we have more maintainers as well :p | 15:28 |
| *** dirk has quit IRC | 15:29 | |
| noonedeadpunk | as tbh right now I'm a bit sad about sunsetting some of them (like watcher)... | 15:31 |
| noonedeadpunk | but I think the driver to sunsetting was lack of maintainers... | 15:31 |
| *** cshen has quit IRC | 15:32 | |
| *** andrea15 has quit IRC | 15:34 | |
| spatel | senlin support webhooks so you can ship direct metrics from Prometheus to trigger scale up or down. | 15:36 |
| *** andrea15 has joined #openstack-ansible | 15:38 | |
| admin0 | spatel, are you using openstack-ansible-ops to setup monitoring ? | 15:39 |
| spatel | no, i am using influx+telegraf | 15:40 |
| spatel | Grafana | 15:40 |
| admin0 | an integrated monitoring osa will make it perfect | 15:41 |
| admin0 | something that can detect a l3-agent being down and can alert/show in dashboard | 15:41 |
| spatel | That would be good. I am using influx+telegraf for mysql/rabbitmq/memcache etc + each compute nodes cpu/memory/disk/network etc.. | 15:43 |
| spatel | I would be great if openstack have some kind of inventory module, currently openstack doesn't have that feature (Ex: you can see number of compute nodes but it don't tell you which kind of hardware its, like server model, CPU arch, memory type etc..) | 15:45 |
| spatel | we have 600 compute nodes and they are all kind of mix and match so sometime its hard to manage inventory outside openstack eco system | 15:46 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_horizon master: Add option to override horizon policies https://review.opendev.org/748264 | 15:47 |
| *** cshen has joined #openstack-ansible | 15:48 | |
| *** rohit02 has quit IRC | 15:50 | |
| noonedeadpunk | spatel: I think there should be ansible netbox module | 15:51 |
| noonedeadpunk | and you can fill in all inventory with ansible there iirc | 15:52 |
| noonedeadpunk | spatel: btw https://github.com/vexxhost/ansible-role-netbox | 15:52 |
| ThiagoCMC | Netbox can also be used as a Dynamic Inventory for Ansible, if I remember well. | 15:53 |
| noonedeadpunk | yep. you just need to decide what was first - chicken or egg I guess :) | 15:54 |
| spatel | noonedeadpunk: hmm.. let me try out, we are using zabbix at present. | 15:54 |
| noonedeadpunk | zabbix is awesome imo | 15:55 |
| noonedeadpunk | but not for handling inventory :) | 15:55 |
| admin0 | what is the best way to monitor l3-agent up/down status ? are you using zabbix for it ? | 15:55 |
| noonedeadpunk | I yes | 15:55 |
| noonedeadpunk | I have kind of re-written zabbix-agent role so that it install required packages and distribute custom scripts, and it works really nicely with osa | 15:56 |
| noonedeadpunk | admin0: but I think you can use prometheus for that as well | 15:57 |
| noonedeadpunk | https://github.com/openstack-exporter/openstack-exporter | 15:57 |
| spatel | noonedeadpunk: that designate patch still not working for me, check this out - http://paste.openstack.org/show/797180/ | 15:58 |
| *** dirk has joined #openstack-ansible | 15:58 | |
| noonedeadpunk | ah.... | 15:59 |
| noonedeadpunk | I guess I know why.... | 15:59 |
| noonedeadpunk | that's why we have bumps https://review.opendev.org/#/c/747586/ | 15:59 |
| spatel | does that means its not merge yet right? | 16:01 |
| openstackgerrit | Merged openstack/openstack-ansible master: Adjust docs for Ussuri https://review.opendev.org/746729 | 16:02 |
| openstackgerrit | Merged openstack/openstack-ansible master: [doc] Adjustments for Victoria release cycle https://review.opendev.org/746741 | 16:04 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible stable/ussuri: Adjust docs for Ussuri https://review.opendev.org/748268 | 16:08 |
| noonedeadpunk | spatel: change is merged, but bootstrap-ansible relies on the SHA's provided in the ansible-role-requirements.yml | 16:09 |
| spatel | oh! so we need to bump SHA in order to fix it | 16:10 |
| noonedeadpunk | and that patch is exactly the bump) | 16:18 |
| noonedeadpunk | and we do releases on top of the bumps just in case | 16:18 |
| spatel | Thank you, good to know | 16:22 |
| npalladium | noonedeadpunk: (and spatel) if you go for a more recent version of netbox and need redis, you might want to check out geerlingguy's redis role; the one which is there in lae's netbox role doc installs redis from source. | 16:46 |
| spatel | npalladium: does openstack-ansible-ops has playbook to install netbox or i have to source is from outside? | 16:50 |
| *** jbadiapa has quit IRC | 16:50 | |
| npalladium | spatel: I don't think it does but I'm not at all sure of this as I never gave it more than a cursory glance, I ended up using a separate role. | 17:01 |
| spatel | how does netbox collect hardware inventory? does it has agent reporting to server or use snamp? | 17:01 |
| noonedeadpunk | spatel: nah, kind of manually, but, you have really a lot of thing as ansible facts | 17:08 |
| npalladium | At our datacentre, a staff member ended up manually entering the data from our old DCIM. If there are integrations that automate that I would be interested too. | 17:08 |
| noonedeadpunk | and since there's an ansible module (and netbox has an api) you can populate it from them | 17:08 |
| noonedeadpunk | eventually you can populate zabbix from ansible facts as well, just need to use curl instead of fancy module | 17:09 |
| npalladium | noonedeadpunk: Could you please elaborate on what exactly you wanted me to test wrt Gnocchi? | 17:11 |
| *** SecOpsNinja has left #openstack-ansible | 17:12 | |
| noonedeadpunk | npalladium: can you cd to /etc/ansible/roles/os_swift, and run there cherry-pick command from https://review.opendev.org/#/c/748217/ | 17:18 |
| noonedeadpunk | and then try running swift role again with swift_gnocchi_enabled: true | 17:19 |
| *** andrewbonney has quit IRC | 17:28 | |
| npalladium | noonedeadpunk: Apparently someone's using the service and I've been asked to not run any scripts against it as of now. I will try to do it later, but not sure if I'll be able to anytime soon. | 17:39 |
| noonedeadpunk | ok then) | 17:39 |
| noonedeadpunk | will hope it fix the issue | 17:39 |
| npalladium | Sorry about that. | 17:40 |
| *** aedc_ has joined #openstack-ansible | 17:42 | |
| *** cshen has quit IRC | 17:45 | |
| *** aedc has quit IRC | 17:45 | |
| *** cshen has joined #openstack-ansible | 18:01 | |
| noonedeadpunk | np at all | 18:02 |
| *** cshen has quit IRC | 18:05 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_magnum master: Use the utility host for db setup tasks https://review.opendev.org/747211 | 18:08 |
| *** sshnaidm is now known as sshnaidm|afk | 18:11 | |
| npalladium | I'll have to deploy another setup of OpenStack which will have different config; both of them will be used in prod simultaneously. Any suggestions on how I should organise the config files for the two? Or should I just use a different deployment host altogether? | 18:14 |
| *** aedc_ has quit IRC | 18:15 | |
| noonedeadpunk | You can use the same deployment host, but put OSA config into another dir. ie /etc/openstack/cloud1 and /etc/openstack/cloud2 | 18:19 |
| noonedeadpunk | But before running openstack-ansilbe you should set OSA_CONFIG_DIR env var, ie `export OSA_CONFIG_DIR=/etc/openstack/cloud2` | 18:20 |
| npalladium | Sounds good. I was hoping and something like this would be there. :) Thanks! (Sidenote I'll be able to test Gnocchi in that install). | 18:28 |
| noonedeadpunk | jrosser: I have a crazy idea that we need https://review.opendev.org/#/c/747586/ to get https://review.opendev.org/#/c/747183/ upgrade job passing :) | 18:31 |
| noonedeadpunk | we can't unfrotunatelly set depends-on since I didn't implement it working on lower branches | 18:31 |
| *** Mensis89 has quit IRC | 18:47 | |
| jrosser | noonedeadpunk: done! | 18:50 |
| *** npalladium has quit IRC | 19:16 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-galera_server master: Update galera to 10.5.5 https://review.opendev.org/742105 | 19:23 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: [WIP] Bind services to mgmt network addresses https://review.opendev.org/670051 | 19:26 |
| *** NewJorg has quit IRC | 19:30 | |
| *** NewJorg has joined #openstack-ansible | 19:32 | |
| *** tosky has quit IRC | 19:38 | |
| *** cshen has joined #openstack-ansible | 19:40 | |
| *** cshen has quit IRC | 19:44 | |
| *** spatel has quit IRC | 20:47 | |
| *** ChiTo has joined #openstack-ansible | 21:05 | |
| ChiTo | Hi openstack-ansible team, is there any possibility to tell the infrastucture playbooks to not install galera-cluster/mysql on LXC but on a metal/vm? | 21:06 |
| ChiTo | or even other openstack control plane service or memcached or haproxy | 21:14 |
| jrosser | ChiTo: you can deploy components to metal rather than LXC | 21:26 |
| jrosser | this can be per component, or globally like here in the AIO configuration https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/openstack_user_config.yml.aio.j2#L38 | 21:27 |
| jrosser | I would suggest that starting by building an AIO with the environment variable SCENARIO=aio_metal is a good starting point | 21:28 |
| jrosser | currently not all of the services can co-exist on the same metal host but that is something which is planned to be fixed in the future | 21:29 |
| jrosser | for your specific question about making galera on metal you can do it per service with the env.d configuratiom, | 21:31 |
| jrosser | https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/env.d/aio_metal.yml.example#L36-L38 | 21:31 |
| ChiTo | jrosser: thank you very much for your prompt response, I was plpaying with the is_metal in true for specific properties | 21:35 |
| ChiTo | thanks for the links you provided me | 21:35 |
| *** cshen has joined #openstack-ansible | 21:41 | |
| *** cshen has quit IRC | 21:45 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/ussuri: Adjust docs for Ussuri https://review.opendev.org/748268 | 22:03 |
| admin0 | ChiTo, is there a reason why you want them on metal ? | 22:10 |
| admin0 | having them on containers provide you more flexibility .. plus allows you to install packages on the host OS without worring about breaking up the services | 22:11 |
| ChiTo | admin0: Totally agreed, I just have some dependencies on a different operational teams for databases, and they want to control those assets on an isolated galera cluster | 22:29 |
| admin0 | in that case, you can easily move the db off the controllers to its own dedicated servers | 22:29 |
| ChiTo | the rest of the services I feel comfortable on LXC | 22:29 |
| ChiTo | exactly but I wanted to manage the installation leveraging the openstack-ansible playbooks | 22:30 |
| admin0 | ansible will still manage and operate it .. it will be in the servers provided by the db team | 22:31 |
| ChiTo | I mean at the end my architecture for DBs is not "converged" with the rest of the controllers | 22:31 |
| ChiTo | agreed, now with the metal feature I will still install them through ansible but put them isolated in dedicated servers | 22:31 |
| admin0 | 4 years ago, i had the same stuff .. where the db team wanted operational control over mysql .. so had to install db in thier provided server .. which ansible will install, but was not a part of the openstack controllers | 22:32 |
| ChiTo | understood, thanks for the feedback and clarification | 22:33 |
| admin0 | i documented that process here: https://www.openstackfaq.com/openstack-liberty-private-cloud-howto/ and the configs are here: https://github.com/a1git/openstack-cloud/tree/master/openstack_deploy | 22:33 |
| admin0 | in case you want to move db and rabbit or what you want to servers managed by diff teams | 22:34 |
| ChiTo | got it, thanks! | 22:39 |
| *** ChiTo has quit IRC | 22:42 | |
| *** hamzaachi has joined #openstack-ansible | 22:55 | |
| *** hamzaachi has quit IRC | 22:59 | |
| *** jamesdenton has quit IRC | 23:36 | |
| *** cshen has joined #openstack-ansible | 23:41 | |
| *** cshen has quit IRC | 23:45 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!