opendevreview | Doug Goldstein proposed openstack/ironic master: inspection: define inventory shape with TypedDict https://review.opendev.org/c/openstack/ironic/+/944868 | 00:18 |
---|---|---|
opendevreview | Iury Gregory Melo Ferreira proposed openstack/sushy master: Don't fail to eject vmedia in case there is no vmedia https://review.opendev.org/c/openstack/sushy/+/952730 | 02:14 |
opendevreview | Doug Goldstein proposed openstack/ironic master: inspection: define inventory shape with dataclasses https://review.opendev.org/c/openstack/ironic/+/944868 | 04:26 |
opendevreview | Doug Goldstein proposed openstack/ironic master: fix port vif attach with networks with dynamic segments https://review.opendev.org/c/openstack/ironic/+/952168 | 05:01 |
rpittau | good morning ironic! o/ | 06:28 |
rpittau | cardoe: re py3.9 and up: we're still testing with python 3.9, so we should be good for now | 06:30 |
opendevreview | Abhishek Bongale proposed openstack/ironic-tempest-plugin master: WIP Add Tempest tests for inspection rules in Ironic https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/951761 | 08:21 |
abongale | Good morning Ironic! | 08:23 |
opendevreview | cid proposed openstack/sushy master: Add retry logic to connection failures https://review.opendev.org/c/openstack/sushy/+/953248 | 08:26 |
opendevreview | Abhishek Bongale proposed openstack/ironic-tempest-plugin master: WIP Add Tempest tests for inspection rules in Ironic https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/951761 | 09:50 |
dtantsur | guilhermesp: it does not look like my patch is applied | 09:55 |
dtantsur | > raid_utils.create_raid_device(index, logical_disk) | 09:55 |
dtantsur | my patch adds the 3rd argument there, it's not in your traceback | 09:55 |
dtantsur | cardoe: I'm totally fine with 3.9+ at this point | 09:57 |
stephenfin | cid: fyi https://bugs.launchpad.net/ironic/+bug/2115332 | 10:47 |
stephenfin | cid: spotted while working on https://review.opendev.org/c/openstack/openstacksdk/+/942705/ The ironic job wasn't voting so we missed the fact that quite a few of the tests in https://review.opendev.org/c/openstack/openstacksdk/+/945700 weren't working. We live and learn | 10:49 |
cid | 👀 | 10:49 |
cid | We do live and learn. I will check that out, hopefully before the end of the week. | 10:51 |
cid | Thanks, stephenfin. | 10:51 |
stephenfin | nw | 10:51 |
iurygregory | good morning ironic | 11:03 |
opendevreview | Abhishek Bongale proposed openstack/ironic-tempest-plugin master: WIP Add Tempest tests for inspection rules in Ironic https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/951761 | 11:07 |
Continuity | Hey Ironic | 11:48 |
Continuity | Any idea what ironic-neutron-agent-member-manager-pool queue is used for | 11:48 |
Continuity | It seems to have a lot of messages which are not being consumed | 11:48 |
Continuity | We have multiple of them, all with lots of messages in them | 11:51 |
opendevreview | Riccardo Pittau proposed openstack/ironic-python-agent-builder master: [WIP] Build ironic-python-agent DIB image with CentOS Stream 10 https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/950965 | 11:52 |
iurygregory | Continuity, I've found https://storyboard.openstack.org/#!/story/2004938 (seems similar to what you are describing), but it was fixed already.. | 12:00 |
opendevreview | cid proposed openstack/ironic master: WIP: Migrate conductor threads to a thread pool https://review.opendev.org/c/openstack/ironic/+/952939 | 12:18 |
guilhermesp | dtantsur: yep my bad, i did a fetch instead of a pull in master branch. I just ran a clean here and looks like it completed. I will inspect logs soon and deploy a system and provide you some feedback :) | 13:15 |
dtantsur | ++ | 13:15 |
TheJulia | Continuity: Interesting.... Once I've caffinated my brain I'll take a look | 13:22 |
* TheJulia caffinates | 13:22 | |
Continuity | Iurygregory: thanks, yeah I was reading that as well | 13:23 |
TheJulia | oh wow, yeah | 13:23 |
Continuity | TheJulia: sounds good. We came across it when investigating high disk usage. We are on 2023.2 | 13:23 |
TheJulia | Continuity: Are they continuing to grow? | 13:48 |
guilhermesp | https://www.irccloud.com/pastebin/8k8TQf9x/ | 13:52 |
guilhermesp | dtantsur: bingo? :P | 13:53 |
dtantsur | \o/ | 13:55 |
guilhermesp | good stuff! thank you for that | 13:56 |
opendevreview | Verification of a change to openstack/networking-generic-switch master failed: Replace LOG.warn https://review.opendev.org/c/openstack/networking-generic-switch/+/953056 | 14:02 |
dtantsur | np, I'll finish the patch when I have a minute | 14:03 |
abongale | I'm writing a Tempest test for the *Inspection Rule API* to verify updates to `actions`, `conditions`, and `priority`. | 14:19 |
abongale | The test is currently failing due to a mismatch in the `actions` field during assertion. Specifically, the returned patched inspection rule `actions` include an unexpected extra variable: `loop`. | 14:19 |
abongale | I guess it's wrong? . AFAIK, `loop` should be a valid variable since we're following the Ansible pattern for both `actions` and `conditions` | 14:19 |
abongale | Ref: Test https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/951761/30/ironic_tempest_plugin/tests/api/admin/test_inspection_rules.py#127 | 14:19 |
abongale | https://paste.openstack.org/show/b55InrObK3MgxI2iJmm2/ | 14:20 |
abongale | traceback call ^^ | 14:20 |
abongale | I believe we might have to update the `convert_actions` method and make it similar to `convert_conditions` https://paste.openstack.org/show/bXGGzI9wRSl9J2GdjB5S/ | 14:21 |
TheJulia | Continuity: I suspect that they may be from a prior upgrade or state, but I'm also looking at the code wondering if it is just still happening based upon the state of the code, so some orientation would be helpful. | 14:23 |
TheJulia | abongale: so to reframe it, your not expecting loop to be added to the field by the API, but it ?is? getting added in it seems | 14:25 |
TheJulia | abongale: offhand, it seems like you might have found the issue in the ironic code, cid any thoughts? | 14:26 |
abongale | TheJulia : I am not expecting to update the actions with loop field but returned updated response have a loop variable in the actions. | 14:30 |
TheJulia | Are there any internal tests which place the expectation? I guess I'm wondering where the expectation came from which is why I'm hoping cid can chime in | 14:31 |
opendevreview | Merged openstack/ironic-specs master: Trait based port selection and dynamic portgroups https://review.opendev.org/c/openstack/ironic-specs/+/945642 | 14:31 |
cid | abongale, TheJulia, I remember there was some special handling of a field. It's likely the 'loop' field that's added during rule creation. | 14:33 |
cid | I will need to look at the code to refresh my memory | 14:33 |
* cid Goes looking... | 14:35 | |
cid | abongale, I think it makes sense to update convert_actions, clearly the spec does not indicate it as a required field. | 14:45 |
cid | I could push that update now, unless you want to take care of it. | 14:45 |
abongale | +++, I will create PR | 14:46 |
abongale | I can make changes. Thank you @cid | 14:47 |
cid | Alright. More thanks to you, abongale. | 14:49 |
cid | By the way, stephenfin found something that I'm picking up right about now. Maybe I will just push fix your finding alongside. | 14:49 |
Continuity | Thejulia: I shall investigate and come back to you... | 14:52 |
TheJulia | Continuity: Thanks! | 14:52 |
Continuity | TheJulia: yes they do seem to be growing | 14:55 |
TheJulia | okay, yeah, that kind of lines up with what I'd expect based upon the code | 14:55 |
Continuity | ironic-neutron-agent-member-manager-pool-c9d73d2d-3445-454d-a633-106e114c1a92 4792513 | 14:55 |
Continuity | 1 min | 14:55 |
Continuity | ironic-neutron-agent-member-manager-pool-c9d73d2d-3445-454d-a633-106e114c1a92 4792519 | 14:55 |
Continuity | Do you want me to throw a bug into launchpad | 14:56 |
TheJulia | oh, so c9d73d2d-3445-454d-a633-106e114c1a92 is the agent uuid? | 14:56 |
Continuity | Yeah | 14:56 |
Continuity | Thats the output from rabbitmqctl | 14:56 |
TheJulia | Yeah, go ahead and create a bug in launchpad then | 14:56 |
Continuity | Wilco | 14:57 |
Continuity | Do you need any other detail | 14:57 |
TheJulia | Just sort of the facts we've identified here | 14:58 |
TheJulia | the fact it is the agent uuid based entry which is getting recorded but not read/removed. I might need to get hjensas to chime in but it is clearly a bug we then have downstream as well which means we're also going to need to track it | 14:58 |
opendevreview | Abhishek Bongale proposed openstack/ironic master: fix: handle unexpected 'loop' in actions field https://review.opendev.org/c/openstack/ironic/+/953323 | 15:04 |
opendevreview | Abhishek Bongale proposed openstack/ironic master: fix: handle unexpected 'loop' in actions field https://review.opendev.org/c/openstack/ironic/+/953323 | 15:34 |
mnaser | dtantsur: thanks for helping guilhermesp with the patch, can i help in finishing up that patch if you've got a bunch of stuff on your table, maybe if you put some guidance on whats needed for https://review.opendev.org/c/openstack/ironic-python-agent/+/953122 to land? :) | 15:47 |
opendevreview | Merged openstack/networking-generic-switch stable/2025.1: Cast numeric Netmiko kwargs to native types. https://review.opendev.org/c/openstack/networking-generic-switch/+/952600 | 15:56 |
opendevreview | Merged openstack/networking-generic-switch stable/2024.1: Cast numeric Netmiko kwargs to native types. https://review.opendev.org/c/openstack/networking-generic-switch/+/952602 | 15:56 |
opendevreview | Merged openstack/networking-generic-switch stable/2024.2: Cast numeric Netmiko kwargs to native types. https://review.opendev.org/c/openstack/networking-generic-switch/+/952601 | 15:56 |
dtantsur | mnaser: oh, that would be nice! I think I've handled the unit tests, so it only needs a release note and check if the docs need any update (either in IPA itself or a follow-up in Ironic). | 16:08 |
mnaser | dtantsur: ok cool, let me try and check that out | 16:09 |
dtantsur | thanks! | 16:09 |
opendevreview | Mohammed Naser proposed openstack/ironic-python-agent master: Fix software RAID creation on different physical devices https://review.opendev.org/c/openstack/ironic-python-agent/+/953122 | 16:23 |
mnaser | dtantsur: ^ done =) | 16:23 |
dtantsur | \o/ | 16:24 |
guilhermesp | ++ | 16:27 |
dtantsur | rpittau: 2025 OpenInfra Summit Europe - Session Accepted! \o/ | 16:36 |
opendevreview | Harald Jensås proposed openstack/ironic master: Stop using oslo.service graceful_shutdown_timeout https://review.opendev.org/c/openstack/ironic/+/953335 | 16:44 |
opendevreview | Maksim Malchuk proposed openstack/bifrost master: Do not pass empty values to instance_info https://review.opendev.org/c/openstack/bifrost/+/953336 | 16:48 |
Continuity | TheJulia: https://bugs.launchpad.net/networking-baremetal/+bug/2086303 looks like a copy of what we were seeing | 17:19 |
iurygregory | dtantsur, rpittau congrats! | 17:24 |
TheJulia | Continuity: are you using kolla? | 17:25 |
iurygregory | the one I have with janders was also accepted :D | 17:25 |
TheJulia | Congrats guys | 17:25 |
iurygregory | now into the complicated part, budget... :D | 17:25 |
dtantsur | congrats too! | 17:38 |
iurygregory | ty | 17:42 |
opendevreview | Doug Goldstein proposed openstack/ironic master: fix port vif attach with networks with dynamic segments https://review.opendev.org/c/openstack/ironic/+/952168 | 17:53 |
FreemanBoss[m] | Congratulations rpittau | 18:15 |
opendevreview | Doug Goldstein proposed openstack/ironic master: inspection: define inventory shape with dataclasses https://review.opendev.org/c/openstack/ironic/+/944868 | 18:32 |
abongale | congrats rpittau!! | 18:33 |
opendevreview | Merged openstack/networking-generic-switch master: Clean up add_subports_on_trunk, del_subports_on_trunk https://review.opendev.org/c/openstack/networking-generic-switch/+/949833 | 18:36 |
cardoe | So TheJulia you might not like my inventory shape patch. I know dtantsur won't like it. | 18:39 |
cardoe | Hoping someone might have some suggestions. | 18:39 |
cardoe | I've found a bunch of code paths that the fields in the inventory matter cause we read them with code like inventory["some_key_here"] | 18:40 |
cardoe | Now I'm enforcing a shape... whether its the right shape I dunno | 18:41 |
cardoe | But if it doesn't match it treats it like the box has no inventory. | 18:41 |
cardoe | Which is certainly better than conductor crashing I think. | 18:41 |
TheJulia | cardoe: cool cool, I'll try to take a look tomorrow at this point, I'm sort of heads down on some eventlet stuff at the moment | 18:42 |
cardoe | No rush cause it's not 100% ready. | 18:42 |
cardoe | There's at least 5 shapes I've found. ironic-inspector migrated data. IPA web documented fields. IPA code itself. IPA's example for implementing your own hardware manager. And lastly redfish inspection. | 18:44 |
cardoe | Haseeb on my team is fixing up redfish fields because we want to have inspector hooks and rules run on redfish still. | 18:45 |
cardoe | I figure once we can agree on the shape, I can also add JSON schema validation into the API endpoint that IPA hits to submit the data. | 18:47 |
opendevreview | Harald Jensås proposed openstack/ironic master: Stop using oslo.service graceful_shutdown_timeout https://review.opendev.org/c/openstack/ironic/+/953335 | 18:50 |
opendevreview | Doug Goldstein proposed openstack/ironic master: fix port vif attach with networks with dynamic segments https://review.opendev.org/c/openstack/ironic/+/952168 | 18:58 |
TheJulia | cid: regarding eventlet removal, any thoughts on starting to peel it out from services in general. I ask because the ironic-novncproxy is a bit of an all or nothing removal, seems to work fine if we remove the patching, but there is no clean intermediate runtime state short of adding logic to handle the state where it is still being patched to use eventlet explicitly. | 19:04 |
TheJulia | stevebaker[m]: ^^^ | 19:04 |
TheJulia | I guess I'm comfortable doing so with ironic-novncproxy in general since it has minimal (none, yet?) usage given its such a new feature | 19:06 |
TheJulia | but that means changing the other pattern and does sort of also line us up for a per-service removal, once we're happy with removing it from the API (and it does seem to work, FWIW as well... my dev machine only monkey_patches the conductor process.) | 19:06 |
stevebaker[m] | Is there a way of detecting if eventlet is being used without importing eventlet? | 19:07 |
TheJulia | the pattern in the libraries is I think imports, but not monkey patches | 19:08 |
TheJulia | and then there is a method to determine if it is patched, or not | 19:08 |
TheJulia | I'm strongly thinking just removing __init__.py and moving the contents into each service for now | 19:08 |
opendevreview | Doug Goldstein proposed openstack/ironic master: fix port vif attach with networks with dynamic segments https://review.opendev.org/c/openstack/ironic/+/952168 | 19:11 |
TheJulia | steve has reminded me downstream of the single process launcher | 19:14 |
* TheJulia hunts for it | 19:18 | |
* cid sees that single process launcher might be benefitting from the module-level eventlet patching, but feels it might not require any major updates. | 19:29 | |
cid | TheJulia, I think we literally just rip out the module-level eventlet patching, like we're doing with the threadpool change. Looking through the codebase, novncproxy might need some independent work, but I honestly don't know how to test it. Aside from that, are there even any other services? | 19:29 |
TheJulia | I'm working on it right now, singleprocess is the only hurdle so I think it will have some conditional code until we can remove the generalized monkey patch invokation | 19:30 |
TheJulia | invocation | 19:30 |
cid | Cool, tks | 19:34 |
TheJulia | my goal was to rip it out, but... looks like we can't just go directly to remove so... one step at a time :) | 19:34 |
TheJulia | okay, that seems to work, now to just re-test by turning eventlet back on :) | 19:44 |
opendevreview | Doug Goldstein proposed openstack/ironic master: fix port vif attach with networks with dynamic segments https://review.opendev.org/c/openstack/ironic/+/952168 | 19:51 |
opendevreview | Julia Kreger proposed openstack/ironic master: Change the novncproxy service code to handle eventlet-lessness https://review.opendev.org/c/openstack/ironic/+/953351 | 20:21 |
TheJulia | stevebaker[m]: cid: ^^^ | 20:36 |
cid | \o/ | 20:38 |
opendevreview | Adam McArthur proposed openstack/ironic master: Update firmware and shard schemas to require 'created_at' and 'updated_at' fields https://review.opendev.org/c/openstack/ironic/+/953352 | 20:49 |
adamcarthur5 | TheJulia appreciate your early review. Jay was pretty keen for me to get this out in any form before he left. Its very pre-baked so give me some time before your next review? I.e. I want to make a tonne of changes | 21:06 |
TheJulia | Cool cool | 21:07 |
adamcarthur5 | I'll get in touch via IRC when its ready. Your comments will help me greatly in shaping it so appreciate it a lot | 21:09 |
opendevreview | Adam McArthur proposed openstack/ironic-specs master: Add emergency bulk operations https://review.opendev.org/c/openstack/ironic-specs/+/952533 | 21:11 |
iurygregory | if anyone has some time https://review.opendev.org/c/openstack/sushy/+/952730 o/ | 21:33 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!