| jrosser | noonedeadpunk: i am super confused about this loop error https://paste.opendev.org/show/bLIL6EZRZYxoBb7p6qdo/ | 08:42 |
|---|---|---|
| jrosser | i'm not sure how this has ever worked properly for adding keystone roles | 08:52 |
| jrosser | https://paste.opendev.org/show/b2BGsEPIsnN1WVe44Cy9/ | 08:52 |
| frickler | jrosser: https://github.com/ansible/ansible/issues/78759 suggests using thing["foo"] instead of thing.foo for the inner loop | 09:06 |
| frickler | or maybe that's a special thing only with .values | 09:07 |
| frickler | but I think you wat "when: thing.foo is string" in your "STRING" block | 09:08 |
| jrosser | trouble is that it seems to try to evaluate the loop: even though the task conditionals say it should be skipped | 09:08 |
| frickler | *want even | 09:08 |
| jrosser | ah yes that is wrong in the STRING block | 09:08 |
| jrosser | my confusion is why specifically now do i see CI failures like this, rather than at any time in the past] | 09:10 |
| frickler | jrosser: do you have a link handy to a build where this is happening? | 09:11 |
| jrosser | this is a paste of the relevant log https://paste.opendev.org/show/bLIL6EZRZYxoBb7p6qdo/ | 09:12 |
| jrosser | the first iteration works where role: is a list | 09:12 |
| jrosser | then the subsequent ones where it is a string it fails | 09:13 |
| jrosser | this last change in the code was a long time ago https://opendev.org/openstack/openstack-ansible-plugins/commit/15e596cb8eda024c4079191634edbc26d44f3249 | 09:14 |
| jrosser | i think this is the build i noticed it on https://zuul.opendev.org/t/openstack/build/aa070b2621a44957a884d94dad06e5b6 | 09:15 |
| jrosser | but then it's also breaking * in here too the same way https://review.opendev.org/c/openstack/openstack-ansible-os_ceilometer/+/895696?tab=change-view-tab-header-zuul-results-summary | 09:15 |
| beny | Hi all | 09:21 |
| beny | I can't find informations so I come here to look for it... | 09:21 |
| beny | My client is considering moving from a production openstack from OSAnsible to Kolla. Do you know if the database schema is identical if we stay in the same version of openstack (v train in this case)? | 09:21 |
| beny | We reinstall the OS (ubuntu), we'll deploy kolla in train version, then once operational, a large restore of a database dump from OSA in the gallery. | 09:21 |
| beny | Do you think this is possible or do we have to use a tool like https://github.com/os-migrate/os-migrate? | 09:21 |
| beny | We can stop the service, downtime of 2 days possible, so no need to do this hot. | 09:21 |
| beny | Thank you for your feedback | 09:21 |
| frickler | jrosser: I have no idea what changed, but using the "wantlist=True" mentioned in the hint does sound reasonable to me | 09:32 |
| jrosser | frickler: i'm not sure where to put that | 09:33 |
| frickler | beny: neither osa nor kolla do any changes to the database schemes. but train is also very old, so you'll likely be on your own if you run into any issue | 09:33 |
| frickler | jrosser: oh, that's a parameter for lookup only. hmm, gotta run your test locally | 09:35 |
| jrosser | frickler: so this works, by trickery to avoid needing the task conditional https://paste.opendev.org/show/b9ViLen2dAgPMkp4Kv12/ | 09:36 |
| jrosser | oh maybe not..... | 09:37 |
| beny | frickler: Thanks, i know the rules, out of support, i fixed issues without commit them/PR because EOL... | 09:38 |
| jrosser | huh interesting, ultimately the whole thing can be totally simplified https://paste.opendev.org/show/btx20Su8sbHrRRzohCju/ | 09:43 |
| frickler | jrosser: that similar to just doing loop: "{{ things | flatten(1) }}" on the outside I guess. but it doesn't explain why it worked earlier. I get the same error with some old ansible 2.10 for your test | 09:48 |
| admin1 | beny, best to install both dbs in a single controller in lab and test it out | 10:05 |
| admin1 | maybe do a db dump without data of both and then diff them | 10:05 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-plugins master: Simplfy addition of keystone users to roles https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/896017 | 10:20 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: WIP - test Vexxhost CAPI driver for magnum https://review.opendev.org/c/openstack/openstack-ansible/+/893240 | 10:21 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-plugins master: Simplfy addition of keystone users to roles https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/896017 | 11:24 |
| deflateder | Hi all, managed to get a few hours of prep work on this (the joys of waiting for other teams) so from the docs i think i have this right and almost ready for a test run, could someone look over my settings? https://paste.opendev.org/show/b7p5z8FvoPyp2XrpQF78/ Would really like to get this right the first time. Cheers people | 12:55 |
| admin1 | deflateder, i have not seen a netplan config like this .. i also run ovs/ovn .. but interesting :) | 13:04 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-plugins master: Simplfy addition of keystone users to roles https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/896017 | 13:05 |
| admin1 | is ceph using cephadm ? | 13:06 |
| admin1 | you don't need ceph_mons block | 13:07 |
| admin1 | when gnocchi is via ceph, i have gnocchi_ceph_username and gnocchi_ceph_pool and i try to use gnocchi_incoming_driver to redis to prevent the osd being written to death | 13:08 |
| jrosser | deflateder: i would make your openstack_user_config as minimal as you can | 13:14 |
| jrosser | it looks like there is a lot in there which you are not actually using | 13:14 |
| jrosser | deflateder: `when this goes to production it'll be about 20 servers so at least one can be used for ironic.` <- you know that ironic is about deploying baremetal hosts, so you need a pool of hardware to dedicate to that | 13:18 |
| deflateder | its not netplan its plain old network/interfaces | 13:22 |
| deflateder | ceph is using cephadm yes | 13:23 |
| deflateder | i will be looking at setting up redis but from what i can see that is a step for after initial deployment | 13:23 |
| deflateder | and i meant as the node that runs it, from my understanding it cant be the same as compute so need at least one system dedicated to it | 13:24 |
| jrosser | the ironic service runs on the control plane | 13:25 |
| jrosser | but the point of it is to PXEboot and provision other hardware | 13:25 |
| deflateder | that was the plan for when we expand, ironic seems to make expansion easier in the future | 13:25 |
| deflateder | i've sidelined it for now anyway i just forgot to remove it from my paste | 13:26 |
| deflateder | so forgetting ironic, any obvious mistakes? networking seems to be making me question myself more than anything here as i'm still not sure i fully understand it but i am still going through and reading the docs over and over again | 13:29 |
| admin1 | well, we can't comment on the specific tags etc.. but go ahead and set it up .. if the controllers and bridges can talk to each other and with 9000 mtu, should be good | 13:41 |
| admin1 | if they can't , setup-infra step will show some issues somwhere, or even before | 13:42 |
| deflateder | thanks, i won't get to try the proper deployment until i'm back off holidays but i'll keep reading the docs in the mean time and then try and enjoy my holiday | 13:50 |
| deflateder | i'll likely be back if i run into problems, thanks again, bye for now | 13:51 |
| opendevreview | Merged openstack/openstack-ansible master: Bump ansible collection versions https://review.opendev.org/c/openstack/openstack-ansible/+/892373 | 14:13 |
| -opendevstatus- NOTICE: The lists.openinfra.dev and lists.starlingx.io sites will be offline briefly for migration to a new server | 15:30 | |
| opendevreview | Merged openstack/openstack-ansible-galera_server master: Added vars to override systemd for mariabackup https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/894530 | 19:12 |
| admin1 | trying to add gnocchi .. fatal: [r1c1_gnocchi_container-7b774098 -> r1c1_repo_container-31ce4d5e(172.29.239.63)]: FAILED! => {"changed": false, "msg": "file not found: /var/www/repo/os-releases/26.1.0/ubuntu-22.04-x86_64/requirements/gnocchi-26.1.0-constraints.txt"} -- i need to somehow rebuild the repo ? | 19:20 |
| admin1 | gnocchi only has: gnocchi-26.1.0-global-constraints.txt gnocchi-26.1.0-requirements.txt gnocchi-26.1.0-source-constraints.txt but not gnocchi-26.1.0-constraints.txt | 19:42 |
| admin1 | i copied the file from an identical other cluster | 19:54 |
| admin1 | redis still needs a manual install using pip inside the venv | 20:04 |
| jrosser | didnt we give you a hook to do that? | 20:04 |
| admin1 | i think that was a few releases ago .. this one is a new install .. i am sure i did the config correctly | 20:05 |
| admin1 | let me share | 20:05 |
| admin1 | https://gist.githubusercontent.com/a1git/639e6bb0495769fe4a154bc4ed79bb24/raw/ebcaa7be623f6851e7438bc89a74e27aae2b3fe4/gistfile1.txt | 20:07 |
| admin1 | oh wait .. it was somewhere else .. | 20:07 |
| jrosser | so you've not actually manually installed it? | 20:09 |
| admin1 | :D | 20:09 |
| admin1 | i went inside each venv and did pip install redis | 20:09 |
| admin1 | on all 3 gnochhi venvs | 20:10 |
| admin1 | i mean on all 3x controllers | 20:10 |
| admin1 | i am syncing the irc logs .. i will grep redis and figure it out in which file it was | 20:10 |
| jrosser | i think it would be good to figure out if this is wrong https://opendev.org/openstack/openstack-ansible-os_gnocchi/src/branch/master/vars/main.yml#L42-L52 | 20:12 |
| jrosser | specifically this https://opendev.org/openstack/openstack-ansible-os_gnocchi/src/branch/master/vars/main.yml#L50 | 20:13 |
| jrosser | but i think you are one of very few people using this | 20:13 |
| admin1 | or large clouds ( public), i have seen the osds getting hammered by every second metrics .. so gnocchi_incoming_driver: redis solves that part | 20:15 |
| admin1 | set extra_package = gnocchi_storage_driver -- shouldn't it be set extra_package = gnocchi_incoming_driver ? | 20:17 |
| admin1 | my use case is gnocchi_storage_driver: ceph gnocchi_incoming_driver: redis | 20:22 |
| admin1 | explained here -> https://docs.openstack.org/openstack-ansible/latest/user/prod/gnocchi_redis.html | 20:49 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!