| *** chkumar|rover is now known as chandankumar | 03:26 | |
| *** ysandeep|out is now known as ysandeep | 05:38 | |
| noonedeadpunk | mornings | 06:12 |
|---|---|---|
| noonedeadpunk | I think it's time to fix stable branches:) | 06:12 |
| noonedeadpunk | I assume it would be easier if we merge rally backport to Wallaby first | 06:13 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_rally stable/wallaby: Control rally-openstack installed version https://review.opendev.org/c/openstack/openstack-ansible-os_rally/+/848721 | 06:16 |
| *** ysandeep is now known as ysandeep|holiday | 06:18 | |
| jrosser | morning | 06:44 |
| jrosser | "extra keys not allowed" what? | 06:54 |
| noonedeadpunk | Well, I did more then one mistake there... | 07:11 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_rally stable/wallaby: Control rally-openstack installed version https://review.opendev.org/c/openstack/openstack-ansible-os_rally/+/848721 | 07:12 |
| jrosser | oh doh :) i see it now | 07:19 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Bump ansible-core to 2.13.0 https://review.opendev.org/c/openstack/openstack-ansible/+/842387 | 07:20 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Bump ansible-core to 2.13.1 https://review.opendev.org/c/openstack/openstack-ansible/+/842387 | 07:21 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Bump ansible-core to 2.13.1 https://review.opendev.org/c/openstack/openstack-ansible/+/842387 | 07:22 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Update pip and wheel to latest versions https://review.opendev.org/c/openstack/openstack-ansible/+/849279 | 07:24 |
| noonedeadpunk | jrosser: regarding https://lists.openstack.org/pipermail/openstack-discuss/2022-July/029492.html but you don't really need to use br-vlan? Nobody stopps you from defining flat network as bond0.100 or br-vxlan the same and don't have br-vlan at all | 08:36 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-os_tempest master: Fix the behavior of defining fixed_network_name https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/846954 | 09:00 |
| jrosser | noonedeadpunk: i did say "but alternatively can be any interface of your choice" | 09:15 |
| noonedeadpunk | ah yes | 09:15 |
| noonedeadpunk | but it won't even by default be br-vlan? | 09:16 |
| jrosser | well no, until you do some calls to neutron API there is no provider network at all | 09:16 |
| jrosser | but we do leave people to figure this all out for themselves | 09:17 |
| noonedeadpunk | but with https://opendev.org/openstack/openstack-ansible/src/branch/master/etc/openstack_deploy/openstack_user_config.yml.example#L290-L298 there will be just `flat` network? | 09:17 |
| noonedeadpunk | it's confusing about br-vlan since only `host_bind_override` is smth that makes sense here | 09:17 |
| noonedeadpunk | until you place neutron agents in containers... | 09:18 |
| jrosser | well yes, though the whole eth12 business feels somehow like a hangover from things that make the AIO work, getting into the example prod config | 09:18 |
| noonedeadpunk | Doh, we have really confusing setup... | 09:18 |
| jrosser | flat network is pretty unhelpful also imho | 09:19 |
| jrosser | it all sounds so "easy" for first time deployments then brings nothing but misery after that | 09:19 |
| noonedeadpunk | But yeah. I just kind of never had br-vlan, as you actually don't need it to be bridge, or well, it doesn't make sense to have it as bridge. As what will neutron do with it is add bridge to another bridge... | 09:19 |
| jrosser | maybe we need some docs about external networks | 09:20 |
| noonedeadpunk | yeah, I think we should add more details to the docs | 09:23 |
| damiandabrowski[m] | I'm using br-vlan and I don't have any extra bridge, I define `network_interface` to specify the interface that OVS should link to | 09:33 |
| damiandabrowski[m] | maybe I missed something but i don't see anything confusing here :D | 09:33 |
| damiandabrowski[m] | my config: https://paste.openstack.org/show/bjmUJornWynpyXLJCEMF/ | 09:34 |
| noonedeadpunk | damiandabrowski[m]: confusing things are: container_interface that is never used, that you must set just bare interface to br-vlan (bond0, not bond0.100), and then bond0.100 as flat network if you have only one network card. And that br-vlan not really required and optional one, but then for flat network you still say it's br-vlan... | 10:39 |
| jrosser | or flat network is br-vlan untagged | 10:41 |
| jrosser | at the same time as vlans being tagged things on br-vlan | 10:41 |
| damiandabrowski[m] | ahh relation between br-vlan and flat networks may be weird, I've never used flat though | 10:41 |
| *** dviroel|out is now known as dviroel|rover | 11:43 | |
| mgariepy | morning everyone | 12:00 |
| damiandabrowski[m] | hi! | 12:00 |
| noonedeadpunk | o/ | 12:08 |
| jrosser | hello | 12:09 |
| mgariepy | how are you guys doing? | 12:14 |
| damiandabrowski[m] | not so good considering the temperature in PL dropped to 14 degrees :D | 12:22 |
| mgariepy | 14 is better than 42 :) | 12:25 |
| damiandabrowski[m] | ok, that's a fair point :D | 12:27 |
| mgariepy | here currently we do have quite nice weather. aroung 15 at night,. and 22-25 during the day. | 12:28 |
| mgariepy | my garden starts to look like it's finally growing :D | 12:28 |
| damiandabrowski[m] | good to heard that, You live in Canada, right? | 12:31 |
| mgariepy | yep. | 12:32 |
| *** damiandabrowski[m] is now known as damiandabrowski | 12:45 | |
| noonedeadpunk | and indeed I'd prefer 14 over 34.... | 13:10 |
| noonedeadpunk | I dream about rainy weather these days | 13:10 |
| mgariepy | jrosser, https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/848657/ | 13:42 |
| nixbuilder | I am having a problem with, I think, a missing repo an infra nodes. It seems like the setup-infrastructure script is attempting to mount glusterfs, yet when I go in to the infra nodes there is an error stating that 'gfs-repo' does not exist... I am still looking but could use a little help. | 14:00 |
| nixbuilder | Specifically this error is within the 'infra1-repo-container' | 14:01 |
| noonedeadpunk | I think we install gluster inside repo-containers, not on the infra host themselves | 14:14 |
| noonedeadpunk | can you also paste error you get while running playbook? | 14:15 |
| noonedeadpunk | It's a new thing we introduced quite recently, so there might be some bugs | 14:16 |
| noonedeadpunk | oh no, I need to run now, so will be back in several hours. I believe that folks that are around will help you ;) | 14:17 |
| nixbuilder | OK... I'll post the errors in a minute. | 14:17 |
| nixbuilder | TASK [systemd_mount : Set the state of the mount] ****************************** fatal: [infra1_repo_container-73f82fbb]: FAILED! => {"changed": false, "cmd": "systemctl reload-or-restart $(systemd-escape -p --suffix=\"mount\" \"/var/www/repo\")", "delta": "0:00:00.118438", "end": "2022-07-11 06:56:02.590359", "msg": "non-zero return code", "rc": 1, "start": "2022-07-11 06:56:02.471921", "stderr": "Job failed. See \"journalctl -xe\" for detai | 14:21 |
| nixbuilder | root@infra1-repo-container-73f82fbb:~# tail /var/log/glusterfs/glusterd.log [2022-07-11 11:55:41.741549] I [MSGID: 106477] [glusterd.c:181:glusterd_uuid_generate_save] 0-management: generated UUID: 35e48598-de55-4b0c-8bfc-f72e03261e99 [2022-07-11 11:55:50.313556] I [MSGID: 106499] [glusterd-handler.c:4263:__glusterd_handle_status_volume] 0-management: Received status volume req for volume gfs-repo [2022-07-11 11:55:50.314274] E [MSGID: 106 | 14:22 |
| mgariepy | you can use paste.openstack.org | 14:22 |
| nixbuilder | Oops... ok. | 14:22 |
| nixbuilder | https://paste.opendev.org/show/b7iRC8xQZm715okFdTdD/ | 14:25 |
| nixbuilder | I grabbed version '25.0.0.0rc1' of the openstack ansible script from 'https://opendev.org/openstack/openstack-ansible' | 14:27 |
| nixbuilder | Also I am using Ubuntu 20.04.4 LTS (GNU/Linux 5.4.0-121-generic x86_64) | 14:30 |
| jrosser | nixbuilder: can you try `gluster volume info gfs-repo` | 14:45 |
| nixbuilder | Volume gfs-repo does not exist | 14:48 |
| nixbuilder | So where do I get this repo from? | 14:49 |
| jrosser | it's not a repo - its the name of a glusterfs volume | 14:52 |
| jrosser | did anything else go wrong earlier in the playbook run | 14:52 |
| jrosser | and is this a fresh installation, or an upgrade? | 14:53 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server master: Allow customisation of stick-tables for each service. https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/848657 | 14:54 |
| nixbuilder | Fresh installation... here is the the output of the playbook run... https://paste.opendev.org/show/bWNwwdLrYJht5VIQBXz3/ | 14:55 |
| nixbuilder | It looks like several errors of "Probe returned with Transport endpoint is not connected" | 14:56 |
| jrosser | yeah | 14:58 |
| jrosser | one moment | 14:58 |
| jrosser | nixbuilder: what does `journalctl -u glusterd` say? | 15:02 |
| nixbuilder | -- Logs begin at Fri 2022-07-08 10:12:35 CDT, end at Mon 2022-07-11 09:28:51 CDT. --Jul 11 06:55:40 infra1-repo-container-73f82fbb systemd[1]: Starting GlusterFS, a clustered file-system server... Jul 11 06:55:41 infra1-repo-container-73f82fbb glusterfs-wait.sh[960]: No volumes present Jul 11 06:55:41 infra1-repo-container-73f82fbb systemd[1]: Started GlusterFS, a clustered file-system server | 15:03 |
| jrosser | then i have checked the connectivity between one of my infra hosts and the repo container (i only have one in an AIO here) https://paste.opendev.org/show/bU8j6kRB9i56hwOv7bv9/ | 15:04 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-os_tempest master: Use network UUID for floating_network_name https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/849420 | 15:05 |
| jrosser | the repo containers need to be able to communicate with each other over the mgmt network for this to work | 15:05 |
| jrosser | deploying the gluster cluster is the first point that things would go bad if there was some networking issue with br-mgmt | 15:06 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-os_tempest master: Use network UUID for floating_network_name https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/849420 | 15:06 |
| *** dviroel|rover is now known as dviroel|rover|lunch | 15:07 | |
| nixbuilder | jrosser - Thank you! That is the problem. Thanks so much for your help!!! https://paste.opendev.org/show/bi3bGTkcmiMIgVOsEUht/ | 15:09 |
| jrosser | ah! that won't help | 15:10 |
| jrosser | if you get the network running i would be interested to know if re-running the playbook for the repo server is sufficient to get things all in order | 15:10 |
| jrosser | corner cases like this can throw up some surprises | 15:10 |
| nixbuilder | I will make sure to report back on what happens when I re-run the playbook after fixing the networking. | 15:11 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-os_tempest master: Make router creation independent from tempest_public_net_create https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/849421 | 15:13 |
| jrosser | just to be completely clear you don't necessarily need that connectivity between the deploy host and the infra hosts (but you might actually have the deploy host on the same network as br-mgmt for simplicity), gluster requires infra host repo container to infra host repo container in this case and I think that is whats causing the error | 15:14 |
| nixbuilder | OK... thanks for the info... working on it now. | 15:15 |
| jrosser | br-mgmt ends up on eth1 inside the containers | 15:16 |
| *** dviroel|rover|lunch is now known as dviroel|rover | 16:11 | |
| mnaser | i'm doing some unrelated ansibel stuff but since my ansible friends are here :) | 18:58 |
| mnaser | is there any other way of accessing a register'd variable for a task from another host | 18:58 |
| mnaser | without set_fact'ing beforehand? | 18:58 |
| mnaser | (so to avoid that task) | 18:59 |
| noonedeadpunk | mnaser: I think you can always access other host vars through hostvars[inventory_hostname], but not sure you asked that... | 19:10 |
| mnaser | noonedeadpunk: i mean if you do like shell: lscpu \n register: _lscpu | 19:10 |
| mnaser | then you wanted to do access the value of `lscpu` from another host that ran it | 19:10 |
| mnaser | usually i do set_fact, then access it from hostvars.. but im not sure if there is a way to avoid doing set_fact | 19:11 |
| * noonedeadpunk thinking | 19:15 | |
| noonedeadpunk | I think no. Ansible doc says, that `unlike registered variables, facts can be gathered independently and cached for repeated use. With cached facts, you can refer to facts from one system when configuring a second system` which is exactly what you wanna do here but with registered variable... | 19:17 |
| noonedeadpunk | I also tried several ideas but none worked | 19:17 |
| noonedeadpunk | If it's something you always do, you likely can gather as facts by having own module for gather_facts that would inherit it, but provide more option what to collect as fact. | 19:19 |
| noonedeadpunk | might be quite an overkill though | 19:22 |
| mnaser | yeah i think it'll just have to be an extra task noonedeadpunk | 19:52 |
| mnaser | thanks for giving it some thought too :) | 19:52 |
| *** dviroel|rover is now known as dviroel|rover|biab | 20:36 | |
| *** dviroel|rover|biab is now known as dviroel|rover | 21:15 | |
| *** dviroel|rover is now known as dviroel|rover|biab | 21:41 | |
| *** dviroel|rover|biab is now known as dviroel|rover| | 21:59 | |
| *** dviroel|rover| is now known as dviroel|out | 22:01 | |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-os_tempest master: Make _tempest_plugins a list of dicts. https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/848642 | 23:20 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-os_tempest master: Allow to define extra tempest plugins https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/848643 | 23:21 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-os_tempest master: Fix the behavior of defining fixed_network_name https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/846954 | 23:21 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-os_tempest master: Use network UUID for floating_network_name https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/849420 | 23:21 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!