| dmsimard | tested ara 1.5.7rc4 successfully, I'll release it monday | 00:55 |
|---|---|---|
| dmsimard | if you want to look at all the playbooks run for a single patchset to see what it looks like: https://demo.recordsansible.org/?path=openstack-ansible&started_after=2021-07-28T00%3A54%3A38.758605 | 00:56 |
| dmsimard | just two failures out of 62 playbooks, not bad :p | 00:56 |
| dmsimard | one is a systemd module failure with no output, those make me sad: https://demo.recordsansible.org/playbooks/3076.html?status=failed&status=unreachable#results | 00:57 |
| dmsimard | the other looks like openvswitch packaging shenanigans https://demo.recordsansible.org/playbooks/3028.html?status=failed&status=unreachable#results | 00:58 |
| *** rpittau|afk is now known as rpittau | 07:21 | |
| depasquale | ciao guys. trying to install wallaby during the bootstrap of openstack-ansible there are issues related to ceph | 07:36 |
| depasquale | https://paste.opendev.org/show/807816/ | 07:36 |
| depasquale | any idea? | 07:36 |
| depasquale | it seems it could be an issue related to some action in ansible... not so much related to openstack-ansible | 07:36 |
| depasquale | I have to admit that setup-hosts.yml works without any problem (because no ceph tasks are included)... but during setup-infrastructure the procedure stops with an error related to a missing role for ansible-ceph | 07:37 |
| depasquale | just for sharing further information, also changing the commit hashcode of the version to download, or changing from opendev to github does not recover the issue | 07:39 |
| depasquale | another useful info -> ansible version 2.10.12 | 07:54 |
| *** kleini_ is now known as kleini | 08:49 | |
| kleini | depasquale: check the /etc/ansible/roles/ceph-ansible directory for that failing ceph-ansible role. you can clone it manually and the referenced commit ID exists for me. so you should be able to manually check it out. if that fails manually, you will get more information about the failure from git. | 08:51 |
| jrosser | the ceph-ansible issue is this https://review.opendev.org/c/openstack/openstack-ansible/+/802143 | 09:31 |
| jrosser | they commit loads and loads of patches to their stable branches which upsets the shallow cloning | 09:35 |
| kleini | oh, shallow cloning is used. yes, then commit IDs may be missing | 09:47 |
| jrosser | backporting that to stable/wallaby would be super helpful (i'm not here today!) | 10:10 |
| depasquale | ciao jrosser | 10:24 |
| depasquale | I had a problem with internet connection and i can't see messages above. | 10:25 |
| jrosser | depasquale: see https://review.opendev.org/c/openstack/openstack-ansible/+/802143 | 10:25 |
| depasquale | you are the best | 10:26 |
| depasquale | let me check | 10:26 |
| jrosser | this fix is already on the stable/wallaby branch | 10:32 |
| jrosser | but the 23.0.0 tag won't pick it up (see git log in your openstack-ansible directory to see why) | 10:33 |
| jrosser | so checkout stable/wallaby instead of a tag | 10:33 |
| depasquale | ok I will investigate | 10:39 |
| depasquale | I confirm the fix solves the problem | 10:39 |
| depasquale | I have to switch my mind in using openstack-ansible... thanks for your help | 10:40 |
| jrosser | there will be a tag 23.0.1 in maybe 1-2 weeks time | 10:51 |
| spatel | jrosser can we set Calico for non-voting to unblock ourself until we have good solution to handle metadata or change SSL vip to non-SSL ? | 14:05 |
| jrosser | why not just switch the metadata haproxy endpoint back to http like we talked about yesterday? | 14:07 |
| *** rpittau is now known as rpittau|afk | 14:07 | |
| jrosser | i think there is confusion on the github felix bug becasue i'm not sure the neutron agent is even running at all in a calico deployment, which is why there's no haproxy | 14:07 |
| jrosser | but for the time being we can't do anything about that anyway | 14:07 |
| spatel | jrosser do you mean set this in user_variables.yml haproxy_ssl_all_vips: false ? | 14:08 |
| jrosser | hold on | 14:09 |
| spatel | k | 14:09 |
| -opendevstatus- NOTICE: There will be a brief outage of the Gerrit service on review.opendev.org starting at 15:00 UTC today as part of a routine project rename maintenance: http://lists.opendev.org/pipermail/service-announce/2021-July/000023.html | 14:10 | |
| jrosser | you have a calico AIO? | 14:12 |
| spatel | yes i do | 14:12 |
| jrosser | does it run the neutron agents at all? | 14:12 |
| spatel | let me check | 14:12 |
| spatel | i can see only neutron-server | 14:13 |
| jrosser | right | 14:13 |
| spatel | hmm? | 14:17 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Do not use https for metadata service with calico networking https://review.opendev.org/c/openstack/openstack-ansible/+/803041 | 14:19 |
| jrosser | spatel: ^ that should be fairly quick to try | 14:19 |
| spatel | oh! i can see what did you insert | 14:19 |
| spatel | let me run in my AIO and see | 14:19 |
| spatel | jrosser that works! that was neat!! | 14:26 |
| jrosser | awesome - thanks for testing it | 14:27 |
| spatel | in this patch why i can't see centos-8-steam jobs running https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/802701 ? | 14:29 |
| spatel | ovs job specially | 14:30 |
| jrosser | is there a project.yml? I think you have to use the jobs after defining them | 14:42 |
| -opendevstatus- NOTICE: There will be a brief outage of the Gerrit service on review.opendev.org in the next few minutes as part of a routine project rename maintenance: http://lists.opendev.org/pipermail/service-announce/2021-July/000023.html | 14:59 | |
| spatel | jrosser no i didn't add that in project.yml file i just added in jobs.yml | 16:03 |
| spatel | jrosser here right? https://review.opendev.org/plugins/gitiles/openstack/openstack-ansible-os_neutron/+/refs/changes/01/802701/2/zuul.d/project.yaml | 16:05 |
| spatel | can i remove centos-8 jobs if we don't care about that | 16:06 |
| jrosser | yes so update the job definitions in jobs.yml, update project.yml to reference the new ones | 16:18 |
| spatel | ok let me do that | 16:41 |
| b1tsh1ft3r | Is there any straight forward way to rebuild the repo containers? Going through a point release upgrade from 20.2.1 to 20.2.6 and finding that the /var/www/repo/os-releases/20.2.6/ directory is almost empty entirely with no constraints files... | 16:41 |
| opendevreview | Satish Patel proposed openstack/openstack-ansible-os_neutron master: Adding centos-8-stream job for ovs https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/802701 | 16:44 |
| spatel | done https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/802701/3/zuul.d/project.yaml | 16:44 |
| opendevreview | Satish Patel proposed openstack/openstack-ansible-os_neutron master: Use list of cluster member ipaddr for ovn ml2 agent to connect https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/803060 | 17:34 |
| jrosser | b1tsh1ft3r: the repo server contents are built as you install other things, not up front | 17:52 |
| jrosser | for example, deploying glance should build the necessary wheels and create the constraints files for glance | 17:53 |
| jrosser | is that what you see? | 17:53 |
| *** melwitt is now known as jgwentworth | 17:55 | |
| b1tsh1ft3r | jrosser: for example its going through i believe the install-utility.yml play inside setup-infrastructure.yml. Im seeing the "slurp up the constraints file for later re-deployment" and it fails to find the utility-20.2.6-constraints.txt file in the repo. Ive destroyed and re-created all utility and repo containers using the lxc-container-destroy/lxc-container-create plays | 18:06 |
| b1tsh1ft3r | when looking in the repo containers in the /var/www/repo/os-release/20.2.6/ directory there are no constraints files at all. What ever happened to rebuil-build.yml? when was that deprocated? | 18:07 |
| jrosser | this is train i think? | 18:08 |
| b1tsh1ft3r | yes | 18:08 |
| b1tsh1ft3r | on 20.2.1 and trying to upgrade to 20.2.6 | 18:08 |
| jrosser | if i remember right it was maybe Q->R when the repo server was completely overhauled | 18:08 |
| jrosser | it used to build all the wheels in advance with repo-build | 18:08 |
| jrosser | but now it does them on demand | 18:08 |
| b1tsh1ft3r | oddly it seems to be skipping the "Build wheels for the packages to be install into the venv" and "Index built wheels" parts of the playbook just before the failure. | 18:10 |
| jrosser | ok well thats certainly not right | 18:10 |
| jrosser | that points to this? https://github.com/openstack/ansible-role-python_venv_build/blob/stable/train/tasks/python_venv_wheel_build.yml#L127 | 18:13 |
| jrosser | b1tsh1ft3r: my suggestion would be to work back through the python_venv_build role to find why those conditionals are not met | 18:29 |
| spatel | jrosser what is the deal with tox doc here, its failing - https://review.opendev.org/c/openstack/openstack-ansible/+/803041 | 19:00 |
| *** odyssey4me is now known as Guest2849 | 19:12 | |
| opendevreview | Satish Patel proposed openstack/openstack-ansible-os_neutron master: Set ovn hostname using nodename facts https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/802134 | 19:14 |
| jrosser | spatel: it's this for docs errors http://lists.openstack.org/pipermail/openstack-discuss/2021-July/023886.html | 19:27 |
| jrosser | i don't know what the fix is though | 19:27 |
| spatel | curious why its just pop up.. look like something new | 19:28 |
| spatel | that means bunch of builds going to fail because of it | 19:29 |
| jrosser | becasue the version of sphinx has been moved ahead for all openstack projects in the requrements repo i think | 19:30 |
| jrosser | or something like that | 19:30 |
| spatel | yes something just happened today :) | 19:31 |
| spatel | Finally i pushed this patch so hope all OVN issue has been resolved now - https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/803060 | 19:31 |
| spatel | Now all ovn agent will directly talk to cluster member and not using haproxy | 19:32 |
| spatel | Just like rabbitMQ | 19:32 |
| jrosser | thats great - nice work | 19:32 |
| jrosser | we should try to get all this merged and backport anything needed | 19:33 |
| jrosser | time to make a release for W at least to roll up all these fixes | 19:33 |
| spatel | Yes we are ready to merge them and i think OVN work has been completed pretty much! last thing i am testing which is SSL support for ovn (hope by next week it will be done) | 19:34 |
| spatel | I really want to finish all remaining OVN stuff so we can make it stable and deployable in production | 19:34 |
| spatel | I gotta go! i will catch you next week.. Have a good weekend | 19:36 |
| spatel | Hope noonedeadpunk will be back from vacation :) | 19:36 |
| *** kleini_ is now known as kleini | 21:16 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!