| *** mriosfer has quit IRC | 00:01 | |
| *** spatel_ has joined #openstack-ansible | 00:48 | |
| *** spatel_ is now known as spatel | 00:48 | |
| *** gyee has quit IRC | 01:01 | |
| *** jamesdenton has quit IRC | 01:04 | |
| *** jamesden_ has joined #openstack-ansible | 01:04 | |
| *** spatel has quit IRC | 01:08 | |
| *** spatel_ has joined #openstack-ansible | 01:15 | |
| *** spatel_ is now known as spatel | 01:15 | |
| *** gshippey has quit IRC | 01:35 | |
| *** evrardjp has quit IRC | 02:33 | |
| *** evrardjp has joined #openstack-ansible | 02:33 | |
| *** akahat has quit IRC | 03:08 | |
| *** akahat has joined #openstack-ansible | 03:22 | |
| *** macz_ has joined #openstack-ansible | 03:58 | |
| *** macz_ has quit IRC | 04:03 | |
| *** rohit02 has joined #openstack-ansible | 04:21 | |
| *** rohit02 has quit IRC | 04:53 | |
| *** rohit02 has joined #openstack-ansible | 04:54 | |
| *** zbr|rover4 has joined #openstack-ansible | 05:04 | |
| *** zbr|rover has quit IRC | 05:06 | |
| *** zbr|rover4 is now known as zbr|rover | 05:06 | |
| *** yasemind has joined #openstack-ansible | 05:15 | |
| *** spatel has quit IRC | 05:44 | |
| *** openstackgerrit has joined #openstack-ansible | 06:17 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_neutron master: [goal] Deprecate the JSON formatted policy file https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/782278 | 06:17 |
|---|---|---|
| jrosser | centos-8 distro jobs look in a bad way | 06:23 |
| *** miloa has joined #openstack-ansible | 06:24 | |
| jrosser | oh well thats game over for centos-8 classic distro jobs | 06:37 |
| jrosser | the rdo packages now have dependancies on versions only available in centos8-stream | 06:38 |
| jrosser | "nothing provides python3-rtslib >= 2.1.74 needed by python3-cinder-common-1:18.0.0-0.20210325222515.7c4b626.el8.noarch" | 06:38 |
| jrosser | https://pkgs.org/search/?q=python3-rtslib | 06:38 |
| *** kleini_ has joined #openstack-ansible | 06:39 | |
| *** luksky has joined #openstack-ansible | 06:42 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Switch centos-8 distro jobs to centos-8-stream nodes https://review.opendev.org/c/openstack/openstack-ansible/+/784009 | 06:47 |
| *** lkoranda has joined #openstack-ansible | 06:51 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Switch centos-8 distro jobs to centos-8-stream nodes https://review.opendev.org/c/openstack/openstack-ansible/+/784009 | 06:52 |
| *** jamesden_ has quit IRC | 07:08 | |
| *** jamesdenton has joined #openstack-ansible | 07:09 | |
| noonedeadpunk | mornings | 07:15 |
| noonedeadpunk | I'm wondering if that's master or also for W... | 07:16 |
| noonedeadpunk | As eventually it's kind of branching time | 07:17 |
| noonedeadpunk | but I guess it's for W already :( | 07:17 |
| noonedeadpunk | nice part is that for distro stream we don't need lxc :) | 07:18 |
| noonedeadpunk | as we test distro only on metal | 07:19 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_tempest master: Switch tempestconf centos-8 job to centos-8-stream nodess https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/784019 | 07:30 |
| *** tosky has joined #openstack-ansible | 07:33 | |
| noonedeadpunk | btw, there's tempestconf failure https://zuul.opendev.org/t/openstack/build/a404af3a06664e16a440a079732f9133/log/logs/openstack/aio1-utility/tempestconf.log.txt#48 | 07:37 |
| jrosser | oh dear, i do remember spending some time in an AIO trying to understand that | 07:43 |
| jrosser | iirc tempestconf made bad choices from the service catalog | 07:43 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_tempest master: Use internal endpoint for tempestconf and respect tempest_keystone_interface_insecure https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/769966 | 07:43 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_tempest master: Use internal endpoint for tempestconf and respect tempest_keystone_interface_insecure https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/769966 | 07:44 |
| jrosser | it was kind of two layer problem "read the service catalog from internal endpoint" was first issue | 07:44 |
| jrosser | but then "use the external endpoints listed in the service catalog" was the second thing it did | 07:45 |
| noonedeadpunk | Yeah but then it picked up wrong endpoint from it... | 07:45 |
| noonedeadpunk | I just saw that happening with trove client yestarday :) | 07:45 |
| jrosser | senlin does the same i thnk | 07:46 |
| noonedeadpunk | so from some side it's good we don't have trusted CA since we can catch services that don't respect endpoints.... | 07:46 |
| noonedeadpunk | doesn't make our life easier though | 07:46 |
| noonedeadpunk | Also I realised that there are services that don't have config reference, which is super annoying... | 07:49 |
| * noonedeadpunk talking about trove | 07:49 | |
| *** miloa has quit IRC | 07:54 | |
| jrosser | oh thats nice, openstack-ansible-deploy-aio_distro_metal-centos-8-stream has passed | 08:03 |
| noonedeadpunk | nice ideed | 08:04 |
| *** dpawlik0 is now known as dpawlik | 08:08 | |
| *** fanfi has quit IRC | 08:17 | |
| *** fanfi has joined #openstack-ansible | 08:22 | |
| *** fanfi has quit IRC | 08:26 | |
| CeeMac | jonher: is this tempest issue the same issue you were seeing with cloudkitty tests where its calling the external endpoint with ssl instead of the internal endpoint? | 09:37 |
| jonher | yeah so https://github.com/openstack/sahara-tests/blob/master/sahara_tempest_plugin/config.py#L31 is using the publicURL and that endpoint has https in the catalog, but not actually on the metal deployment, i've edited to use the internal URL but my keystone replies back that the service is unavailable | 09:41 |
| jonher | that's where i'm at currently in regards to the tempest tests | 09:42 |
| CeeMac | https://opendev.org/openstack/cloudkitty-tempest-plugin/src/branch/master/cloudkitty_tempest_plugin/config.py#L29 | 09:44 |
| CeeMac | using publicURL here too | 09:44 |
| CeeMac | what services are showing in your endpoint list for cloudkitty? | 09:45 |
| jonher | wops wrong link yeah :) | 09:45 |
| jonher | that's the one i meant to link :) | 09:45 |
| CeeMac | no worries :) | 09:45 |
| jonher | http://paste.openstack.org/show/804070/ this is basically what i'm facing now | 09:48 |
| jonher | cloudkitty is configured to contact http://172.29.236.101:5000 and that can be reached with curl | 09:49 |
| CeeMac | what do you see in ' openstack endpoint list' for cloudkitty | 09:49 |
| jonher | http://paste.openstack.org/show/804071/ | 09:49 |
| CeeMac | is "http://172.29.236.101:8089/v1/rating/module_config/pyscripts/scripts/" responding to curl? | 09:49 |
| jonher | yep | 09:50 |
| jonher | "The request you have made requires authentication." | 09:50 |
| CeeMac | hmm | 09:51 |
| CeeMac | so its sending the request with a token, but can't validate the auth request as its getting 503 from keystone service | 09:52 |
| CeeMac | but the keystone service is reachable | 09:52 |
| CeeMac | its not using the external keystone url though is it? | 09:52 |
| jonher | what i'm thinking too, looking at hte keystone logs currently | 09:52 |
| jonher | it is not | 09:53 |
| CeeMac | did you check haproxy logs? | 09:53 |
| jonher | i might have to put it against the haproxy published keystone, but so far i didn't | 09:55 |
| jonher | "CRITICAL keystonemiddleware.auth_token [-] Unable to validate token: Unable to establish connection to https://127.0.0.1:35357: HTTPSConnectionPool(host='127.0.0.1', port=35357)" that sure does not look like what i'd expect from the cloudkitty-api logs :) | 09:59 |
| *** jamesdenton has quit IRC | 10:20 | |
| *** jamesden_ has joined #openstack-ansible | 10:20 | |
| jrosser | thats the admin endpoint, and i think we quite deliberatly don't bind to 127.0.0.1 | 10:28 |
| *** yasemind has quit IRC | 10:34 | |
| openstackgerrit | Sebastian Gumprich proposed openstack/openstack-ansible-galera_server master: add support for encryption https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/784069 | 10:43 |
| *** rndmh3ro has joined #openstack-ansible | 10:45 | |
| rndmh3ro | hi! I recently opened https://bugs.launchpad.net/openstack-ansible/+bug/1921861 and noonedeadpunk helpfully answered! So I created a patchset for the new galera encryption feature here: https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/784069 | 10:50 |
| openstack | Launchpad bug 1921861 in openstack-ansible "Add table encryption support?" [Undecided,Triaged] | 10:50 |
| noonedeadpunk | rndmh3ro: hey! will review now | 10:55 |
| noonedeadpunk | rndmh3ro: btw, do we need these key files to be present during service restart? | 10:58 |
| noonedeadpunk | or they're store somewhere internally? | 10:58 |
| jonher | now i just need tempest to ignore ssl validation or make it trust the cert, but first meetings | 10:59 |
| noonedeadpunk | jonher: just make tmepest plugin to use internal endpoint instead of public which is default | 11:00 |
| jonher | pretty sure that cuased the attempts to 127.0.0.1 but i'll verify once more after meetings etc. | 11:00 |
| *** mugsie__ is now known as mugsie | 11:01 | |
| rndmh3ro | noonedeadpunk: they're needed on service-restart. mysql needs to read the (optionally encrypted) keyfile from a file or an environment variable to start. storing them somewhere would probably defeat the whole purpose of this encryption at rest | 11:01 |
| noonedeadpunk | then we can;t probably jsut drop key with handler? | 11:01 |
| noonedeadpunk | and we need the same key to be present on all cluster members, not only for single one? | 11:02 |
| noonedeadpunk | aha, it's delegated, sorry | 11:03 |
| rndmh3ro | .noonedeadpunk: it has to be the same key on all nodes, yes. every node decrypts its own files. what do you mean with a handler? | 11:03 |
| noonedeadpunk | I just expected block to be delegated :) | 11:03 |
| rndmh3ro | let me check | 11:03 |
| rndmh3ro | the first task needs to be run on all machines. then the key gets generated and copied to all hosts. | 11:04 |
| noonedeadpunk | yeah, got it | 11:04 |
| rndmh3ro | and I just noticed that the last task (Set permissions of keyfile on servers) is redundant since the mode is set correctly in the tasks before | 11:05 |
| openstackgerrit | Sebastian Gumprich proposed openstack/openstack-ansible-galera_server master: add support for encryption https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/784074 | 11:08 |
| noonedeadpunk | rndmh3ro: I posted a comment there | 11:11 |
| noonedeadpunk | rndmh3ro: um, and I guess you've created new patch, instead of editing old one:) | 11:12 |
| rndmh3ro | noonedeadpunk: damn. I'm new to gerrit. I did a rebase, then a git-review. should I just have pushed? | 11:13 |
| noonedeadpunk | you should git amend and git review again | 11:13 |
| noonedeadpunk | or, at least, leave `Change-Id:` the same in the commit msg | 11:13 |
| noonedeadpunk | as that's an identifier for gerrit | 11:14 |
| rndmh3ro | I thought I did, here are two ids: https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/784074 | 11:14 |
| noonedeadpunk | eventually what I do when I need to change smth - do changes; git add .; git commit --amend; git review | 11:15 |
| noonedeadpunk | yeah. there're 2 Change-Ids which confused gerrit I guess | 11:16 |
| rndmh3ro | okay, how can I use the old changeset? or should I just continue with the new changeset and amend the commit? | 11:16 |
| rndmh3ro | or I'll delete the second change-id | 11:17 |
| noonedeadpunk | in your current HEAD, you can delete second change-id and make sure that first one is the last line is msg | 11:17 |
| rndmh3ro | thanks, I'll try! | 11:17 |
| noonedeadpunk | and when you run git review, new patchset should be placed to the original patch | 11:18 |
| noonedeadpunk | and yeah, you can abandon change in UI | 11:19 |
| noonedeadpunk | In case I completely confused you, there's another doc regarding gerrit usage https://docs.openstack.org/contributors/code-and-documentation/using-gerrit.html | 11:21 |
| rndmh3ro | You did not confuse me, so far. :) | 11:21 |
| noonedeadpunk | also would be great if you could reference bug in commit msg, with "Closes-Bug: #1921861" | 11:29 |
| openstack | bug 1921861 in openstack-ansible "Add table encryption support?" [Undecided,Triaged] https://launchpad.net/bugs/1921861 | 11:29 |
| rndmh3ro | noonedeadpunk: will do | 11:29 |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder stable/victoria: Fix condition when to create backends https://review.opendev.org/c/openstack/openstack-ansible-os_cinder/+/782963 | 11:41 |
| *** sshnaidm|off is now known as sshnaidm | 11:57 | |
| openstackgerrit | Sebastian Gumprich proposed openstack/openstack-ansible-galera_server master: add support for encryption https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/784084 | 12:04 |
| rndmh3ro | great. a third one... | 12:05 |
| noonedeadpunk | lol | 12:05 |
| noonedeadpunk | so git-review adds git hook, that generates `Change-Id` when it can;t find one.. Maybe it's somehow misbehaving for you | 12:06 |
| openstackgerrit | Sebastian Gumprich proposed openstack/openstack-ansible-galera_server master: add support for encryption https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/784085 | 12:06 |
| noonedeadpunk | 4th | 12:06 |
| rndmh3ro | I ammended one and set the change-id explicitly. | 12:06 |
| noonedeadpunk | It wasn\'t last line, was it? | 12:07 |
| rndmh3ro | no, that's the problem? | 12:07 |
| rndmh3ro | should i retry? could be the 5th | 12:07 |
| noonedeadpunk | it's better to understand once:) | 12:08 |
| noonedeadpunk | we can abandon others at the end of the day | 12:08 |
| * rndmh3ro < https://matrix.org/_matrix/media/r0/download/matrix.org/MtdBxXtuMYHAOnsPqQtxnUgJ/message.txt > | 12:08 | |
| noonedeadpunk | lgtm at this point... | 12:09 |
| * rndmh3ro < https://matrix.org/_matrix/media/r0/download/matrix.org/SbjoKSPJrPuHEoyGUWNHxPOl/message.txt > | 12:09 | |
| rndmh3ro | I'll probabyly do this though: `git push gerrit HEAD:refs/for/master%topic=galera_encryption` | 12:10 |
| noonedeadpunk | -n is dry-run iirc? | 12:10 |
| rndmh3ro | yes | 12:10 |
| rndmh3ro | better yet: `git review -t galera_encryption -n` | 12:11 |
| rndmh3ro | without the -n | 12:12 |
| openstackgerrit | Sebastian Gumprich proposed openstack/openstack-ansible-galera_server master: add support for encryption https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/784069 | 12:15 |
| rndmh3ro | this worked | 12:15 |
| noonedeadpunk | I will abandon other patches if you don't mind:) | 12:16 |
| rndmh3ro | thank you! | 12:16 |
| noonedeadpunk | and I'm not sure if you've covered my comments for it | 12:17 |
| rndmh3ro | I answered them now. Had to find the save-button | 12:19 |
| noonedeadpunk | rndmh3ro: wait. is keyfile jsut a rand string? | 12:23 |
| * rndmh3ro < https://matrix.org/_matrix/media/r0/download/matrix.org/yZEpZKllCszfaKSirTWKrbEg/message.txt > | 12:25 | |
| rndmh3ro | this gets encrypted with a password. if mysql starts, this file is decrypted and used. | 12:25 |
| noonedeadpunk | and then we remove it after server start from galera server? | 12:26 |
| noonedeadpunk | sorry for asking stupid questions, just trying to fuly understand intention :) | 12:27 |
| rndmh3ro | we remove the password-file that decrypts the encrypted file (with the hex-strings). | 12:27 |
| rndmh3ro | good questions! :) | 12:27 |
| rndmh3ro | also the docs are pretty good on this topic: https://mariadb.com/kb/en/file-key-management-encryption-plugin/ | 12:30 |
| noonedeadpunk | yeah, I'm just not super sure how "safe" it is to drop key from galera containers | 12:35 |
| noonedeadpunk | I mean we should at least say, that you can't manually restart instance when this is used at least | 12:35 |
| rndmh3ro | definetly! | 12:36 |
| noonedeadpunk | but with all bits on the server it doesn't make much sense | 12:36 |
| rndmh3ro | that's the problem. :) | 12:36 |
| rndmh3ro | I'd add it to the docs | 12:36 |
| noonedeadpunk | yeah | 12:37 |
| *** spatel_ has joined #openstack-ansible | 13:01 | |
| *** spatel_ is now known as spatel | 13:01 | |
| *** macz_ has joined #openstack-ansible | 13:08 | |
| *** macz_ has quit IRC | 13:12 | |
| *** cp- has quit IRC | 13:31 | |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/victoria: Bump SHAs for stable/victoria https://review.opendev.org/c/openstack/openstack-ansible/+/783580 | 13:33 |
| spatel | noonedeadpunk morning! question, i am still running RC1 victoria on production, do you think i should upgrade victoria with whatever latest stable we have or i can directly go to wallaby ? | 14:09 |
| noonedeadpunk | I won't expect wallaby in the nearest month for sure | 14:14 |
| noonedeadpunk | I'd say with things that are now, we will release pretty close to the deadline... | 14:15 |
| *** mmercer has quit IRC | 14:17 | |
| *** macz_ has joined #openstack-ansible | 14:17 | |
| openstackgerrit | Sebastian Gumprich proposed openstack/openstack-ansible-galera_server master: add support for encryption https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/784069 | 14:17 |
| openstackgerrit | Sebastian Gumprich proposed openstack/openstack-ansible-galera_server master: add support for encryption https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/784069 | 14:20 |
| rndmh3ro | that took longer than expected, but the change works now for: completetly new cluster-installations, for single node reinstallations and it can be run multiple times without triggering a restart each time. | 14:21 |
| *** macz_ has quit IRC | 14:21 | |
| spatel | noonedeadpunk should i wait for just do minor upgrade for safety :) | 14:23 |
| spatel | s/for/or | 14:23 |
| *** jamesden_ has quit IRC | 14:24 | |
| noonedeadpunk | let me just release another one as it fixes nasty cinder bug | 14:24 |
| *** jamesdenton has joined #openstack-ansible | 14:24 | |
| noonedeadpunk | it's hitting new deployments only though | 14:25 |
| spatel | perfect i can wait | 14:25 |
| noonedeadpunk | https://review.opendev.org/c/openstack/openstack-ansible/+/783580 <- this is the patch I'm aiming to tag with 22.1.1 | 14:25 |
| spatel | nice then i will go with 22.1.1 before upgrading to wallaby | 14:27 |
| noonedeadpunk | rndmh3ro: if you don't mind I will take another round tomorrow on fresh head | 14:29 |
| rndmh3ro | noonedeadpunk: sure thing! | 14:29 |
| *** fresta has joined #openstack-ansible | 14:30 | |
| arxcruz | noonedeadpunk: hey, can I add a topic for ptg ? if yes, do you have the etherpad ? | 14:31 |
| noonedeadpunk | sure thing | 14:31 |
| noonedeadpunk | https://etherpad.opendev.org/p/osa-xena-ptg | 14:31 |
| noonedeadpunk | wasn't ML it because want to fill it in a bit | 14:32 |
| noonedeadpunk | jrosser: hm, we do smth wrong with gather facts in `containers-lxc-create.yml` http://paste.openstack.org/show/804075/ | 14:35 |
| noonedeadpunk | it's without your patch for min facts by the way | 14:35 |
| jrosser | thats a vars include i think | 14:36 |
| noonedeadpunk | no, not really | 14:36 |
| noonedeadpunk | we run lxc_container_create role twice https://opendev.org/openstack/openstack-ansible/src/branch/master/playbooks/containers-lxc-create.yml#L42 | 14:36 |
| noonedeadpunk | and eventually, I think we should run against containers during first set of tasks | 14:37 |
| noonedeadpunk | *we should not run | 14:37 |
| jrosser | where does it run twice? | 14:38 |
| noonedeadpunk | oh, sorry, indeed, looked at tags instead of roles lol | 14:38 |
| jrosser | it would help if the vars include tasks was templated for the os/distro/version | 14:39 |
| jrosser | then it would be super clear if there was missing/bogus data it was looking for | 14:39 |
| jrosser | the task name, i mean | 14:39 |
| noonedeadpunk | yeah, agree | 14:41 |
| noonedeadpunk | eventually it would be also great to add prefix to python_venv_build tasks and systemd_service and etc | 14:41 |
| *** macz_ has joined #openstack-ansible | 14:49 | |
| noonedeadpunk | hm, eventually facts should be there... | 14:50 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Fix facts collection for LXC hosts https://review.opendev.org/c/openstack/openstack-ansible/+/784112 | 14:59 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add trove to br-dbaas https://review.opendev.org/c/openstack/openstack-ansible/+/784113 | 15:02 |
| jrosser | noonedeadpunk: do you think that instead we should remove the TTL on the facts cache? | 15:20 |
| jrosser | it's somehow unhelpful that they are all present but ignored after 24hours | 15:20 |
| *** johanssone has quit IRC | 15:26 | |
| *** noonedeadpunk has quit IRC | 15:27 | |
| *** noonedeadpunk has joined #openstack-ansible | 15:28 | |
| *** johanssone has joined #openstack-ansible | 15:29 | |
| noonedeadpunk | jrosser: dunno, that's kind of default ansible behaviour that can be easily changed? | 15:33 |
| jrosser | we set it currently in the openstack-ansible.rc | 15:33 |
| noonedeadpunk | and eventually in case of os upgrade you will need to manually drop facts (or in cases where you add extra interface or etc) | 15:33 |
| jrosser | seems like two different things | 15:34 |
| jrosser | ensure facts are gathered properly when needed, like you say for OS upgrade, by fixes you already did with tags: always | 15:34 |
| jrosser | but stuff goes wrong when you want to use a random playbook which relies on facts > 24hours later | 15:34 |
| jrosser | thats really kind of unexpected | 15:35 |
| jrosser | i think in particular when we need facts for a host that is delegated to, like the repo server | 15:35 |
| jrosser | even if we added always facts gathering everywhere that doesnt cover things that are not targets of the play, but are delegated to | 15:36 |
| noonedeadpunk | So you mean to set facts cache time to 0 but not remove gather_facts tasks? | 15:38 |
| noonedeadpunk | *setup | 15:39 |
| jrosser | we should have setup / gather_facts where we need them | 15:39 |
| jrosser | but it does seem we have a couple of times now where folk have had unexpected results by facts being present, but too old | 15:39 |
| jrosser | i was wondering if that was the root cause of the osa DOS on the infra git servers recently | 15:40 |
| noonedeadpunk | but wouldn't timeout to facts make things even worse in this terms? | 15:40 |
| noonedeadpunk | *absent timeout | 15:40 |
| jrosser | becasue they are no longer valid? | 15:41 |
| noonedeadpunk | well, yes | 15:41 |
| noonedeadpunk | I think it's kind of tradeoff anyway? | 15:41 |
| jrosser | it is | 15:42 |
| jrosser | i was thinking about python_venv_build regarding this | 15:42 |
| jrosser | and adding some kind of assert that if groups['repo_all']|length>0 and venv_build_host==localhost then something has gone bad with facts | 15:43 |
| jrosser | thats maybe my one specific case where the current fact timeout is a bad thing | 15:43 |
| jrosser | but as you say for other situations it's not good to have old data | 15:44 |
| noonedeadpunk | but not sure we can add that in python_venv_build role itself as we should kind of leave it more agnostic? | 15:44 |
| jrosser | yeah, it's tricky | 15:45 |
| noonedeadpunk | I also have a question, about what would you feel to do here. We have dbaas network set like this in sample https://opendev.org/openstack/openstack-ansible/src/branch/master/etc/openstack_deploy/openstack_user_config.yml.aio#L3 But in role defaults we use this https://opendev.org/openstack/openstack-ansible-os_trove/src/branch/master/defaults/main.yml#L166 | 15:48 |
| noonedeadpunk | not sure where to change default.... | 15:48 |
| noonedeadpunk | I think in trove role, but a bit afraid to broke envs during upgrade | 15:48 |
| *** prometheanfire has joined #openstack-ansible | 15:49 | |
| prometheanfire | noonedeadpunk: https://review.opendev.org/783715 still needed? | 15:49 |
| noonedeadpunk | yes?:) | 15:50 |
| noonedeadpunk | except I did the same with importlib-resources | 15:51 |
| noonedeadpunk | *I would do the same | 15:51 |
| noonedeadpunk | for gates things resolved as for now, but eventually we'd love these to get fixed in u-c | 15:52 |
| *** cp- has joined #openstack-ansible | 15:53 | |
| *** spotz has joined #openstack-ansible | 15:53 | |
| prometheanfire | ok, this one is fine, but I need something more than 'I think' for other changes :P | 15:53 |
| prometheanfire | /bu/buffer 20 | 15:54 |
| noonedeadpunk | prometheanfire: our main concern was that things are fixed for py3.6 but they are not for py3.7. dataclasses are already integrated in 3.7, so it;s only backport for 3.6. While importlib-resources and importlib-metadata are not till 3.8 | 15:56 |
| noonedeadpunk | which may sooner or later break envs | 15:56 |
| prometheanfire | ok, added it | 15:57 |
| noonedeadpunk | we don't use these packages directly, but services that use them do break then (which affects our gates in it's turn) | 15:58 |
| noonedeadpunk | thanks | 15:58 |
| jonher | noonedeadpunk "jonher: just make tmepest plugin to use internal endpoint instead of public which is default" http://paste.openstack.org/show/804080/ i'm not sure where it's getting that 127.0.0.1 from when i set it to admin or internal but not on "public" | 15:58 |
| jrosser | noonedeadpunk: well i'm not sure about the dbaas network, this makes things even less clear i guess https://github.com/openstack/openstack-ansible-os_octavia/blob/master/defaults/main.yml#L328 | 15:59 |
| noonedeadpunk | doh | 15:59 |
| jrosser | yeah, thats unfortunate | 16:00 |
| noonedeadpunk | (╯°□°)╯︵ ┻━┻ | 16:00 |
| jrosser | i guess octavia is more established in OSA stuff, so maybe that is the defaults/main.yml we do not change | 16:00 |
| noonedeadpunk | yeah... | 16:01 |
| noonedeadpunk | I guess we don't test octavia in lxc? | 16:01 |
| noonedeadpunk | we don't yes | 16:02 |
| jrosser | the thing that we risk breaking is a deployment where the network from defaults/main.yml was used without an override | 16:02 |
| noonedeadpunk | and I can bet most of them are like that | 16:03 |
| noonedeadpunk | (at least I never changed that) | 16:03 |
| *** rohit02 has quit IRC | 16:03 | |
| jrosser | it depends how many people copy/paste the example config | 16:03 |
| jrosser | btw i had a patch to remove nspawn which clears a bunch of this up, as that used totally different networks again :( | 16:04 |
| noonedeadpunk | like https://opendev.org/openstack/openstack-ansible/src/branch/master/tests/roles/bootstrap-host/templates/user_variables_translations.yml.j2 ? | 16:05 |
| noonedeadpunk | this all makes me feel pretty sick... | 16:09 |
| noonedeadpunk | Can I at least get votes for https://review.opendev.org/q/topic:%22bump_osa%22+(status:open) | 16:14 |
| *** rohit02 has joined #openstack-ansible | 16:18 | |
| noonedeadpunk | doh https://opendev.org/openstack/ansible-role-systemd_networkd/src/branch/master/examples/playbook.yml#L148-L156 | 16:24 |
| *** jamesdenton has quit IRC | 16:24 | |
| *** jamesdenton has joined #openstack-ansible | 16:25 | |
| noonedeadpunk | what a mess.... | 16:25 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Map dbaas and lbaas with role defaults https://review.opendev.org/c/openstack/openstack-ansible/+/784113 | 17:06 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_trove master: Change default pool subnet https://review.opendev.org/c/openstack/openstack-ansible-os_trove/+/784145 | 17:31 |
| *** rohit02 has quit IRC | 17:44 | |
| noonedeadpunk | btw, also would be super great to merge https://review.opendev.org/c/openstack/openstack-ansible/+/784009 | 17:45 |
| *** jamesdenton has quit IRC | 18:56 | |
| *** jamesden_ has joined #openstack-ansible | 18:57 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/victoria: Bump SHAs for stable/victoria https://review.opendev.org/c/openstack/openstack-ansible/+/783580 | 19:27 |
| openstackgerrit | Merged openstack/openstack-ansible stable/ussuri: Bump SHAs for stable/ussuri https://review.opendev.org/c/openstack/openstack-ansible/+/783578 | 19:27 |
| openstackgerrit | Merged openstack/openstack-ansible stable/train: Bump SHAs for stable/train https://review.opendev.org/c/openstack/openstack-ansible/+/783579 | 19:33 |
| *** fanfi has joined #openstack-ansible | 20:11 | |
| *** spatel has quit IRC | 20:18 | |
| *** fanfi has quit IRC | 20:28 | |
| *** mriosfer has joined #openstack-ansible | 20:29 | |
| mriosfer | Hi guys, i continue testing openstack-ansible lab envirioment for victoria release, im getting this error in setup-infraestructure.yml https://pastebin.com/Hr15r6bE | 20:30 |
| *** d34dh0r53 has quit IRC | 20:35 | |
| jrosser | mriosfer: which operating system? | 20:35 |
| mriosfer | Ubuntu 20.0.4 lts | 20:38 |
| *** jamesden_ has quit IRC | 20:38 | |
| *** jamesdenton has joined #openstack-ansible | 20:39 | |
| jrosser | i think i would wonder if something else has failed earlier in the playbook | 20:40 |
| jrosser | handlers run and the end | 20:40 |
| *** d34dh0r53 has joined #openstack-ansible | 20:40 | |
| jrosser | *at teh | 20:40 |
| mriosfer | im re-running again setup-infraestructure after a infras servers reboot | 20:41 |
| jrosser | best to look back through the log and see what failed | 20:46 |
| *** mriosfer has quit IRC | 20:49 | |
| *** mriosfer has joined #openstack-ansible | 20:51 | |
| *** lkoranda has quit IRC | 21:18 | |
| *** mriosfer has quit IRC | 21:34 | |
| *** mriosfer has joined #openstack-ansible | 21:34 | |
| *** luksky has quit IRC | 22:30 | |
| *** rh-jlabarre has quit IRC | 22:48 | |
| *** jamesdenton has quit IRC | 22:57 | |
| *** jamesden_ has joined #openstack-ansible | 22:58 | |
| *** NewJorg has quit IRC | 23:08 | |
| *** NewJorg has joined #openstack-ansible | 23:10 | |
| *** tosky has quit IRC | 23:30 | |
| *** macz_ has quit IRC | 23:31 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!