Tuesday, 2025-11-04

opendevreviewMerged openstack/ironic master: Enable fake-graphical console for all hardware types  https://review.opendev.org/c/openstack/ironic/+/96315100:31
cardoeWell I totally expected to see some failures but I didn’t.01:51
opendevreviewVerification of a change to openstack/ironic stable/2025.1 failed: Memoize calls to bcrypt.checkpw  https://review.opendev.org/c/openstack/ironic/+/95831603:39
Abhi123Good morning ironic !04:54
rpittaugood morning ironic! o/07:05
himanigood afternoon, ironic!!07:11
Ayo[m]Good morning ironic :)08:32
jandersgood morning Ironic o/09:48
opendevreviewRiccardo Pittau proposed openstack/ironic master: Handle HTTP 400 and 409 race condition in Redfish power operations  https://review.opendev.org/c/openstack/ironic/+/96524809:51
rpittauhey when someone has a moment can I get a quick review on https://review.opendev.org/c/openstack/bifrost/+/964752 ? we need that to migrate out of tinyipa09:52
opendevreviewRiccardo Pittau proposed openstack/ironic-python-agent master: Test advertised ip reachability before assigning it  https://review.opendev.org/c/openstack/ironic-python-agent/+/96367009:59
opendevreviewMerged openstack/ironic-specs master: Update standalone networking initial conditions  https://review.opendev.org/c/openstack/ironic-specs/+/96055010:25
*** himani is now known as himani-sahu10:49
opendevreviewVerification of a change to openstack/ironic master failed: api: Add schema for inspection rules API (requests)  https://review.opendev.org/c/openstack/ironic/+/95403910:51
opendevreviewMerged openstack/ironic master: Add generic redfish compatible vendor name  https://review.opendev.org/c/openstack/ironic/+/96487711:16
opendevreviewMerged openstack/ironic master: Increase default sync_power_state_interval  https://review.opendev.org/c/openstack/ironic/+/96155411:28
opendevreviewMerged openstack/ironic master: Remove bespoke logic for handling redirects while validating URLs  https://review.opendev.org/c/openstack/ironic/+/96348511:28
opendevreviewMerged openstack/ironic master: Fixes for console containers functional testing  https://review.opendev.org/c/openstack/ironic/+/96332311:28
opendevreviewDmitry Tantsur proposed openstack/ironic stable/2025.2: Remove bespoke logic for handling redirects while validating URLs  https://review.opendev.org/c/openstack/ironic/+/96602711:45
iurygregorygood morning ironic11:55
opendevreviewMerged openstack/networking-generic-switch master: fix: segment binding level for VLAN support  https://review.opendev.org/c/openstack/networking-generic-switch/+/96594811:57
opendevreviewMerged openstack/ironic-python-agent stable/2025.2: Fix skip block devices for RAID arrays  https://review.opendev.org/c/openstack/ironic-python-agent/+/96259112:13
opendevreviewMerged openstack/ironic-python-agent master: Fix for matching hints with lists of strings  https://review.opendev.org/c/openstack/ironic-python-agent/+/96579712:13
opendevreviewMorten Stephansen proposed openstack/ironic-python-agent stable/2025.2: Fix for matching hints with lists of strings  https://review.opendev.org/c/openstack/ironic-python-agent/+/96603112:26
opendevreviewMorten Stephansen proposed openstack/ironic-python-agent stable/2025.1: Fix for matching hints with lists of strings  https://review.opendev.org/c/openstack/ironic-python-agent/+/96603212:30
opendevreviewMorten Stephansen proposed openstack/ironic-python-agent master: Implement functionality for the is_root_volume RAID config  https://review.opendev.org/c/openstack/ironic-python-agent/+/96320013:06
opendevreviewMorten Stephansen proposed openstack/ironic-python-agent master: Implement functionality for the is_root_volume RAID config  https://review.opendev.org/c/openstack/ironic-python-agent/+/96320013:11
opendevreviewRiccardo Pittau proposed openstack/bifrost master: Use cirros 0.6.1 as default cirros version  https://review.opendev.org/c/openstack/bifrost/+/93217513:35
alegacyJayF: clif: TheJulia: so far most serious comments on my reviews are 1) size, and 2) naming of the network interface.  14:01
alegacyFor 1), if I split it into 3 chunks (rpc service, network driver, switch driver) would that alleviate some concern?  14:01
alegacyFor 2) the rational behind using the word "experimental" in the driver name was to allow for dropping it without support if/when we formalize the API away from port.extra14:01
alegacyI'm not tied to the name (standalone-experimental)... maybe just "ironic", or "ironic-rpc"... i'm open to suggestions.  #namingisfun14:02
dtantsuralegacy: for (1), that's what we usually do to make it manageable, unless there is way to split it into a more simple feature followed by additions of extras.14:02
opendevreviewMerged openstack/ironic master: api: Add schema for inspection rules API (requests)  https://review.opendev.org/c/openstack/ironic/+/95403914:03
alegacydtantsur: thanks, ok... i'll split into those 3 unless there is a better suggestion.  The service will still be large, i'll look for a way to split it one more time.14:06
opendevreviewMilan Fencik proposed openstack/sushy master: Handle missing storage controller mode attribute  https://review.opendev.org/c/openstack/sushy/+/96607214:10
TheJuliagood morning14:12
dtantsurmorning TheJulia14:12
cliffrom a reviewer's persepective: 10k lines of change represents a huge time commitment14:28
clifto properly review, and honestly I'm going to get tired halfway through and start to rush and skip over things14:29
clifthe more chunk-able the better14:29
TheJulia++14:32
opendevreviewcid proposed openstack/ironic-prometheus-exporter master: Restore `entity_id` label in Redfish/iDRAC metrics  https://review.opendev.org/c/openstack/ironic-prometheus-exporter/+/96608215:02
cardoeDid my email reply to the sushy thread ever hit the list?15:41
cardoeclif: so the weird thing is I expect https://review.opendev.org/c/openstack/ironic/+/965985 to fail cause now there's extra data there.15:42
cardoeah no there's not. we never refresh the portgroup from the DB in the tests so null == null still15:43
JayFsounds like a good reason to update those teses15:43
JayF**tests15:43
* JayF pokes john garbutt downstream again to test https://review.opendev.org/c/openstack/ironic/+/964502 (glance checksum bugfix)16:35
opendevreviewJay Faulkner proposed openstack/ironic-specs master: 2026.1 Ironic work items  https://review.opendev.org/c/openstack/ironic-specs/+/96586516:40
JayFplease also review https://review.opendev.org/c/openstack/ironic-specs/+/96524916:40
TheJuliaJayF: did you see my comments by chance?17:21
TheJuliacardoe: I don't think I've seen a post from you on the ML, at least one hasn't made it past gmail. Did you check the archive?17:22
TheJuliaJayF: w/r/t the checksum17:22
cardoeYeah I didn't see it there either17:22
JayFit matched my findings17:27
JayFand seemed non-actionable? 17:27
JayFunless I misunderstood17:28
TheJuliaI guess its a regression of the sdk compared with the client library that it doesn't support checksumming downloads with more advanced algos. 17:31
JayFI agree17:41
JayFbut I don't think that changes the shape of the ironic fix in the interim17:41
TheJuliano, definitely not17:42
JayFif you want me to go back up to openstacksdk, I can look into that17:42
TheJulianah, not asking you to, just wondering if we have enough consensus to raise a bug there17:42
JayFprobably17:43
JayFI'll own that17:43
TheJuliak, thanks17:46
TheJuliacardoe: I feel like https://bugs.launchpad.net/ironic/+bug/2130667 begins to set some of the groundwork for internalizing the support into ironic so you can drop your debug log to message bus to automation logic, that being said I could have sworn I've looked at a machine and saw nvme specific settings exposed. It feels like you might have stumbled upon such, but if we can get more clarity out there then it might be 18:56
TheJuliaeasier to collaborate on that.18:56
cardoeAbsolutely.18:57
cardoeI'll be honest bandwidth wise right now I wanna try and steer this VXLAN stuff and then James has thrown my network stuff for a loop cause he's wanting me to just let him use cinder inside of his hypervisor.18:58
cardoeI was actually wondering if maybe I should instead aim to add my docs and ML2 mechanism driver to networking-baremetal ?18:58
cardoeInstead of neutron proper18:58
cardoeI'd really like it in neutron proper because it makes them aware of the use case.18:59
TheJuliacardoe: so, on the nova side, it looks like we quite literally just need to have nqn as an option off the field? am I groking that right?19:14
cardoeYep.19:14
cardoeI'm actually working on a set of patches to nova and cinder for that.19:14
TheJuliafair, I just want to get enough notes in place so we can go back to storage stuff and not be starting from ground level again19:15
cardoeI discussed it with the cinder folks that there's a lot of hardcoded magical strings that are dict keys which makes all this hard to nail down.19:15
TheJuliaon the nova side, it looks like we could only load an initial nqn... but ironic could always sync the rest later19:15
TheJuliawhich might be totally okay19:15
cardoeThey no longer have cinder-lib so I'm defining them in os-brick and changing os-brick to use those. Then we'll use the defines in nova.19:15
TheJuliaso just loop through a list of contstants instead from the library and let that sync, perhaps?19:16
cardoeFrom a fast approach, absolutely.... we could just send along nqn or host_nqn?19:16
cardoeWell some backends need host_nqn and some need nqn19:16
TheJulialist provided by os-brick :)19:16
TheJuliaexactly19:16
TheJuliawe just look for both values?19:16
TheJuliaor so, the code pattern does it19:17
cardoeWell we have to set both.19:17
cardoeMy small nova patch locally just sets initiator, nqn, and host_nqn to the same value19:17
TheJuliaso two lists, translation mappings and actual mappings19:17
cardoeI've made a bug for cinder that they need to use constants and have their backends use the constants instead of free form.19:18
TheJuliaand then the nova code could just loop/translate as necessary from lists and get the specific labels gone from the code19:18
TheJuliaat least, nova.virt.ironic code19:18
TheJuliaanyway, sounds like your chasing on that path, and I can see if James is deploying whole hypervisors it makes sense to let the software do the needful, vxlan networking definitely seems higher up on my priority list19:19
cardoeYou want me to dump my notes in that bug?19:19
cardoeSo we can collaborate.19:20
TheJuliaregarding your mech driver and all, I looked this morning and didn't see anything. I can get they might be... hesitant to engage on the physical front but progress is better than no progress and if we need to drop it into the netowrking-baremetal plugin, I could be good with that too19:20
TheJuliaby all means19:20
*** mnaser[m] is now known as mnaser20:46
cardoeMan. These tests don’t want to fail. I don’t like it.20:54
cardoeclif: I dunno. It’s awkward that the only enforcement point is in the RPC and nothing at the lower object level.20:59
* TheJulia feels like a packet got lost in the woods21:13
clifthat means my changes are perfect ;)21:30
clifcardoe: so you're worried about a portgroup and its constiuent ports not having the same physical_network because there's not lower-level enforcement beyond rpc?21:31
clifI'm happy to put up a patch with something along those lines if you let me know at what level the enforcement should take place21:32
cardoeeh just chasing my tail with the tests there's only one place where things can get broken right now and that's in ironic/drivers/modules/inspector/hooks/physical_network.py21:39
cardoeOtherwise its just all in the tests making all kinds of mock/fake data. Like not real portgroup objects or things like that.21:40
cardoeSo at this point nothing is really broken. When I added the physical_network is sent in the binding:profile to Neutron, I read the physical_network from the portgroup. As you had commented on my patch. But no test ever sets the physical_network on a portgroup. They always use the value of the first port on the portgroup. So my test failed. And I keep poking places thinking aha this will fail. And it doesn't fail.21:43
TheJuliacardoe: wouldn't it just make sense to update the base test mock data used for tests to represent it as such. That is the only way to really get it across a swath of tests aside from explicitly creating tests or a class of tests as that being the starting state.22:01
opendevreviewSteve Baker proposed openstack/ironic master: Improve logging when console unit doesn't start  https://review.opendev.org/c/openstack/ironic/+/96332422:07
opendevreviewSteve Baker proposed openstack/ironic master: Replace Chrome/Selenium console with Firefox extension  https://review.opendev.org/c/openstack/ironic/+/96143422:07
opendevreviewSteve Baker proposed openstack/ironic master: Switch graphical console image to ubuntu by default  https://review.opendev.org/c/openstack/ironic/+/96420122:07
opendevreviewSteve Baker proposed openstack/ironic master: Add a kubernetes provider for console container  https://review.opendev.org/c/openstack/ironic/+/96224222:07
opendevreviewMerged openstack/ironic master: docs: cover what happens if someone changes iscsi passwords  https://review.opendev.org/c/openstack/ironic/+/96344222:33
cardoeTheJulia: there's not really a common spot like this function https://opendev.org/openstack/ironic/src/commit/9305e273a886a793a397d7244daa09ff0174dda4/ironic/tests/unit/drivers/modules/network/test_common.py#L30422:37
cardoeIt's fine. Like I said. It's not really impacting anything.22:37
cardoeThe only place it can break today is in the inspection hook.22:37
opendevreviewDoug Goldstein proposed openstack/ironic master: fix: ensure that portgroup physical_network is updated for tests  https://review.opendev.org/c/openstack/ironic/+/96598522:39
opendevreviewDoug Goldstein proposed openstack/ironic master: pass along physical_network to neutron from the baremetal port  https://review.opendev.org/c/openstack/ironic/+/96457022:39
TheJuliacardoe: well, it really depends on the test, but I'd change obj_utils.create_test_portgroup23:42
cardoeI did in the first patch23:42
cardoeThe issue is port groups are created without a physical_network all over the place and then ports are created referencing the portgroup with a physical_network23:42
cardoeThere we go. Now it failed how I expected it to.23:53
TheJuliaokay, that is making more sense then23:58

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!