| opendevreview | Kaifeng Wang proposed openstack/ironic master: Follow up to fix anaconda deployment on multitenant network https://review.opendev.org/c/openstack/ironic/+/988371 | 01:21 |
|---|---|---|
| opendevreview | Verification of a change to openstack/networking-generic-switch master failed: GenericSwitchNetmikoConfigError error details https://review.opendev.org/c/openstack/networking-generic-switch/+/985029 | 02:43 |
| opendevreview | Merged openstack/python-ironicclient master: trivial: fix author in latex doc output https://review.opendev.org/c/openstack/python-ironicclient/+/988367 | 02:45 |
| opendevreview | Merged openstack/ironic master: pecan: Drop redundant force_canonical https://review.opendev.org/c/openstack/ironic/+/987096 | 03:53 |
| opendevreview | Merged openstack/networking-baremetal master: Disable ovn-metadata-agent https://review.opendev.org/c/openstack/networking-baremetal/+/987041 | 04:02 |
| *** ykarel_ is now known as ykarel | 05:04 | |
| opendevreview | Merged openstack/ironic stable/2026.1: security: move file url validation up into deploy_utils main path https://review.opendev.org/c/openstack/ironic/+/988355 | 05:36 |
| nuhakala | Hi, I have deployed Ironic with bifrost using the default image "deployment_image.qcow2" but I cannot SSH into the image. What user/password should I use, or does the image even have those configured? | 06:44 |
| opendevreview | Takashi Kajinami proposed openstack/ironic-python-agent-builder master: Replace tabs with spaces https://review.opendev.org/c/openstack/ironic-python-agent-builder/+/988397 | 08:30 |
| opendevreview | Merged openstack/ironic stable/2026.1: Fix Redfish BIOS settings on systems without BootProgress https://review.opendev.org/c/openstack/ironic/+/988363 | 09:16 |
| opendevreview | Bertrand Lanson proposed openstack/ironic master: fix: skip storage controllers without Volumes link https://review.opendev.org/c/openstack/ironic/+/987853 | 09:24 |
| blanson[m] | cid: I tried something in ^ , not sure if this is the right way to do it ? | 09:26 |
| blanson[m] | (hello !) | 09:26 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/34.0: Configure CI for bugfix/34.0 https://review.opendev.org/c/openstack/ironic/+/988293 | 09:41 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/33.0: Configure CI for bugfix/33.0 https://review.opendev.org/c/openstack/ironic/+/988416 | 09:46 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/33.0: Configure CI for bugfix/33.0 https://review.opendev.org/c/openstack/ironic/+/988416 | 10:37 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/33.0: Configure CI for bugfix/33.0 https://review.opendev.org/c/openstack/ironic/+/988416 | 10:38 |
| dtantsur | cardoe, janders, hey, is this actually a reasonable thing to assume? https://opendev.org/openstack/ironic/src/branch/master/ironic/drivers/modules/redfish/bios.py#L432-L433 | 10:49 |
| dtantsur | i.e. is anything Redfish says for sure that Sessings are wiped after application? it's not the case for sushy-tools, so BIOS settings don't work on it still, and I wonder what needs to be fixed | 10:49 |
| dtantsur | At least https://www.supermicro.com/manuals/other/redfish-ref-guide-html/Content/general-content/bios-configuration.htm sounds like sushy-tools need to clear its pending settings | 10:58 |
| dtantsur | rpittau: hi! do I get it right that bugfix 31.0 and older should be EOL at this point? | 11:08 |
| dtantsur | (in fact, it's only 31.0, there are no older branches) | 11:09 |
| dtantsur | TheJulia, JayF, https://review.opendev.org/c/openstack/ironic/+/988293 fixes bugfix/34.0 and https://review.opendev.org/c/openstack/ironic/+/988416 should hopefully fix 33.0 too. I believe that 31.0 should not be updated any more (pending confirmation from rpittau) | 11:10 |
| rpittau | dtantsur: yes 31 is EOL, same for 11.1, if you give me 5 minutes I will remove them | 11:22 |
| rpittau | dtantsur: done, it will propagate to opendev in the next hour(s) | 11:26 |
| iurygregory | good morning ironic | 11:27 |
| opendevreview | Riccardo Pittau proposed openstack/networking-generic-switch master: Exclude .cache from pep8 tests https://review.opendev.org/c/openstack/networking-generic-switch/+/988431 | 11:30 |
| cardoe | dtantsur: yes that’s how implementations are supposed to work per the spec. | 11:36 |
| dtantsur | cardoe: ah, do you have a link to that? I could only find general references (Redfish for you..) | 11:37 |
| cardoe | Not at the moment. But janders and I looked it up and validated it against hardware | 11:49 |
| dtantsur | Okie, tasking Claude with fixing sushy-tools then. | 11:54 |
| opendevreview | Dmitry Tantsur proposed openstack/sushy-tools master: Handle pending BIOS settings the same way as real hardware https://review.opendev.org/c/openstack/sushy-tools/+/988434 | 11:59 |
| dtantsur | rpittau: thanks! Could you send a brief EOL announcement to metal3-dev (and generally do it each time with EOL Ironic branches)? | 12:13 |
| rpittau | dtruong: sure, I will include that in the release announcement, we usually update the version matrix, not sure if this time was done | 12:27 |
| rpittau | whoops | 12:27 |
| rpittau | dtantsur: & | 12:27 |
| dtantsur | rpittau: I think it's separate from IrSO and other version matrices since people may follow old versions of Ironic despite our recommendations | 12:28 |
| rpittau | I'm talking about this https://book.metal3.io/version_support.html#ironic-image | 12:28 |
| rpittau | it was not updated | 12:28 |
| dtantsur | rpittau: right, it's a part of https://github.com/metal3-io/metal3-docs/pull/688 which is blocked by CAPM3 | 12:29 |
| rpittau | oh! ok | 12:29 |
| opendevreview | Riccardo Pittau proposed openstack/ironic-prometheus-exporter master: Fix DevStack plugin timing issue with sensor data collection https://review.opendev.org/c/openstack/ironic-prometheus-exporter/+/988442 | 12:50 |
| TheJulia | good morning | 13:03 |
| dtantsur | Looking for quick opinions (cc iurygregory): if bios.apply_configuration receives something like "42" (a string with an integer) for a field that has type "Integer", should it convert or reject the value? | 13:14 |
| dtantsur | Currently we do the worst of both: convert while validating but send the original value to the BMC. That does not work well. | 13:14 |
| iurygregory | OH WOW | 13:15 |
| dtantsur | Context (should be public): https://redhat.atlassian.net/browse/OCPBUGS-44902?focusedCommentId=16968474 | 13:15 |
| iurygregory | I thought we would validate and send the converted value lol | 13:15 |
| dtantsur | Claude suggests we do the same. But I'm undecided on whether we want to be strict with JSON types or more flexible. | 13:16 |
| iurygregory | dtantsur, I would say Ironic should accept, so we would send the converted value, but worth having LOG saying "sending converted value" ... | 13:21 |
| iurygregory | or BMO could send the converted value to ironic already?! | 13:21 |
| iurygregory | in case it has the information that the field is from type int for example, and the user had "42" .. bmo would send to ironic 42 | 13:22 |
| dtantsur | BMO can. The question is whether we consider it a flaw in Metal3 (then we need to harden Ironic to reject wrong values) or a flaw in Ironic (and we need to convert the values properly) | 13:22 |
| dtantsur | We cannot leave Ironic as it is because it's currently inconsistent | 13:22 |
| iurygregory | I would say we fix in ironic... | 13:22 |
| iurygregory | like other people would get the benefit in case they put "42" ... | 13:23 |
| iurygregory | via some other automation tools etc | 13:23 |
| dtantsur | Yeah. I dislike YAML as much as the next person, but having INI files in the picture makes nothing better... | 13:23 |
| iurygregory | yup | 13:23 |
| dtantsur | Okay, I'm telling Claude to run with an Ironic fix | 13:24 |
| dtantsur | (it also discovered that we don't do anything meaningful with booleans) | 13:24 |
| iurygregory | ops =X | 13:25 |
| dtantsur | ikr? bios and firmware stuff is one big oops | 13:25 |
| iurygregory | yup | 13:25 |
| TheJulia | JayF: dtantsur: regarding https://review.opendev.org/c/openstack/ironic/+/986597 I have a growing concern that the approach is wrong, or that its a flow control issue as well. I'm really not sure. The one thing I'm sure of is we've previously had a similar bug with lenovo hardware we managed the last step in the flow (link included in comments on the review), the actual invocation path is not entirely clear, I'm sure | 14:12 |
| TheJulia | dtantsur can bring clarity to that, but ultimately with that change we're also making a super deep default behavior change which I worry will break flows. The question is if this is coming in from "restore_boot_device" or "set_boot_device" and ultimately the higher level caller is only aware of where in the flow we are at. My inherent worry is the patch as is breaks deployments in general. | 14:12 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Convert string values in BIOS settings before sending them to the BMC https://review.opendev.org/c/openstack/ironic/+/988467 | 14:14 |
| dtantsur | TheJulia: I don't have this bug in my context, I can dive into it whenever i have time next (sigh) if you need me to | 14:16 |
| TheJulia | dtantsur: i know that feeling | 14:16 |
| TheJulia | iurygregory: you have more context ^ | 14:16 |
| TheJulia | dtantsur: ignore my ping then, iurygregory can chime in and look | 14:16 |
| JayF | TheJulia: I mean, you basically echoed and amplified by -1 so I am not in disagreement with you | 14:16 |
| JayF | TheJulia: in general this felt like a "I did this and it fixed this one machine, therefore all lenovos" which is why I was nervous | 14:17 |
| TheJulia | well, so I went hunting and I was like "oh, we did something similar in the agent code path already..." | 14:17 |
| TheJulia | and then I was like, wait a second, this is super deeeeep in the code | 14:17 |
| TheJulia | oh, its going to break everything | 14:17 |
| TheJulia | I think the clean path is install_openshift ; set_boot_to_disk in the higher level flow, but someone needs to wrap their head into the union of the flow and calls | 14:18 |
| opendevreview | Doug Goldstein proposed openstack/ironic master: Update packaging for package discovery https://review.opendev.org/c/openstack/ironic/+/988468 | 14:18 |
| cardoe | stephenfin: just cause I'm pinging ya everywhere.. ^ that's the update I'll do for the places I know need it. | 14:19 |
| cid | blanson[m], o/. that matches what I had in mind. Thanks! | 14:21 |
| blanson[m] | cid: zuul unhappy tho, I will have a look | 14:22 |
| stephenfin | cardoe: Tested +1d | 14:23 |
| stephenfin | *and | 14:23 |
| cid | Yep. It's a real failure. Other than that, looks good to me. | 14:24 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: Fix periodic tasks not running after pickle/unpickle https://review.opendev.org/c/openstack/ironic/+/988469 | 14:27 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/2025.2: security: move file url validation up into deploy_utils main path https://review.opendev.org/c/openstack/ironic/+/988356 | 14:53 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/2025.1: security: move file url validation up into deploy_utils main path https://review.opendev.org/c/openstack/ironic/+/988357 | 14:54 |
| opendevreview | Julia Kreger proposed openstack/ironic unmaintained/2024.1: security: move file url validation up into deploy_utils main path https://review.opendev.org/c/openstack/ironic/+/988359 | 15:16 |
| opendevreview | Julia Kreger proposed openstack/ironic unmaintained/2023.1: security: move file url validation up into deploy_utils main path https://review.opendev.org/c/openstack/ironic/+/988480 | 15:19 |
| opendevreview | Merged openstack/ironic bugfix/34.0: Configure CI for bugfix/34.0 https://review.opendev.org/c/openstack/ironic/+/988293 | 15:38 |
| opendevreview | Verification of a change to openstack/bifrost master failed: Drop support for ironic-staging-drivers https://review.opendev.org/c/openstack/bifrost/+/986131 | 15:39 |
| dtantsur | I'd really appreciate some eyes on https://review.opendev.org/c/openstack/sushy-tools/+/988434 since it's blocking my effort to add BIOS settings CI coverage | 15:42 |
| TheJulia | Welcome Jad_Haj ! | 15:49 |
| TheJulia | Since you were not here when we were discussing the issue with your change, even though at a super high level but maybe with an idea forward: see the channel log: https://meetings.opendev.org/irclogs/%23openstack-ironic/%23openstack-ironic.2026-05-13.log.html#openstack-ironic.2026-05-13.log.html#t2026-05-13T14:12:02 | 15:51 |
| opendevreview | Merged openstack/ironic bugfix/33.0: Configure CI for bugfix/33.0 https://review.opendev.org/c/openstack/ironic/+/988416 | 15:57 |
| dtantsur | TheJulia: rebase all the things \o/ (bugfix 33.0 and 34.0 I mean) | 15:58 |
| TheJulia | \o/ | 15:58 |
| TheJulia | speaking of bugfix, I still need to do the backports for the DoS ossa | 15:58 |
| Jad_Haj | Thanks TheJulia | 15:59 |
| dtantsur | TheJulia: in case you missed above: rpittau intends to EOL 31.0 immediately, so you don't need to bother with it | 16:00 |
| TheJulia | already landed | 16:00 |
| Jad_Haj | i wnet throgh the logs and tend to aggree the approach is broken to begin with. looks like it need more planning | 16:00 |
| Jad_Haj | TheJulia: TheJulia: should abandon and start over taking review into account | 16:03 |
| Jad_Haj | TheJulia: meant: should I abandon and start over taking review notes into account ? | 16:06 |
| TheJulia | you can revise the current change if it makes sense to do so | 16:06 |
| Jad_Haj | TheJulia: Ok, Will try that first then | 16:08 |
| TheJulia | Jad_Haj: I think the question boils down to how are we getting into this state, which is sort of the same issue, but I think its is rooted in the steps being invoked in the flow to deploy openshift | 17:03 |
| cardoe | someone broke the docs | 17:04 |
| cardoe | they're failing on all branches | 17:04 |
| opendevreview | cid proposed openstack/ironic master: Follow-Up: Keep anaconda on provisioning network https://review.opendev.org/c/openstack/ironic/+/988505 | 17:31 |
| JayF | dtantsur: need your +1 here https://review.opendev.org/c/openstack/project-config/+/984159 | 17:38 |
| Jad_Haj_ | TheJulia: Will push a fix that adds the same Lenovo UEFI logic to management.py with the safety | 17:42 |
| Jad_Haj_ | checks you mentioned (persistent only, UEFI only, single NVRAM change). | 17:42 |
| TheJulia | Jad_Haj_: I'm not sure your getting what I'm saying then | 17:42 |
| TheJulia | Jad_Haj_: specifically, in management.py, your changing the low level setting which impacts the state of even getting a node ready to be deployed | 17:42 |
| TheJulia | so you'll effectively break that | 17:42 |
| TheJulia | In ironic's proper flow, i.e. not with whatever is happening in openshift with the custom-agent plugin and steps resulting, the flow should already be setting the stage for that in ironic proper, functionally your undoing the setting and creating the conflict upfront when we likely need to step away from management.py and understand if the problem is actually the flow with the custom-agent and should another step be | 17:44 |
| TheJulia | invoked in that case or not, and if we can model that such that touching the lower level of management.py is really not necessary | 17:44 |
| Jad_Haj_ | TheJulia: so my fix is in the wrong place entirely | 17:47 |
| TheJulia | yes, I suspect so | 17:47 |
| TheJulia | I think your hyper-focusing which is fine, but we need to ask why are you in this weird state | 17:47 |
| TheJulia | and then determine exactly how your sort of getting into it | 17:48 |
| TheJulia | In essence, it might be similar to the right move, but we ideally don't want _set_boot_something to have to know the state to properly navigate the issue | 17:49 |
| TheJulia | I think your hitting https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent.py#L392 | 17:50 |
| Jad_Haj_ | TheJulia: so i more to investigate with reporter of issue. for now, will Step back from trying to add vendor logic there | 17:50 |
| TheJulia | https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent.py#L239-L250 is the flow we normally take and expect, via AgentDeploy, not CustomAgentDeploy | 17:52 |
| TheJulia | so if your in CustomAgentDeploy, calling prepare_instance_boot, you end up calling task.driver.boot.prepare_instance(task), so if that is for example vmedia, that would be into this call and method https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/redfish/boot.py#L1131C22-L1131C38 which just maps to the wider scaled wrapper | 17:55 |
| TheJulia | https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/redfish/boot.py#L1245 in manager utils, at that point your already too far zoomed in to really know to short circuit | 17:55 |
| TheJulia | The right place is likely to add similar "if not lenovo logic" guard rails into redfish/boot.py and if PXE derived similarly in the pxe prepare_instance code path, but since your saying secure boot I'm leaning towards thinking your doing vmedia only. | 17:59 |
| Jad_Haj_ | TheJulia: thanks for your insights. will consider it and resume the work tomorrow | 18:01 |
| TheJulia | okay, have a good evening | 18:01 |
| opendevreview | cid proposed openstack/ironic master: Follow-Up: Keep anaconda on provisioning network https://review.opendev.org/c/openstack/ironic/+/988505 | 19:42 |
| cardoe | TheJulia: look at you all optimistic that master passes | 19:54 |
| TheJulia | ... how bad is it? | 19:58 |
| cardoe | the docs build is broken I'm just on back to back to back to back calls | 20:34 |
| JayF | <unknown>:1: WARNING: more than one target found for cross-reference 'type': ironic.objects.volume_connector.VolumeConnector.type, ironic.objects.volume_connector.VolumeConnectorCRUDPayload.type, ironic.db.sqlalchemy.models.VolumeConnector.type [ref.python] | 20:40 |
| JayF | that's the warning | 20:40 |
| JayF | I'm going to point claude at it | 20:40 |
| JayF | I suspect this is related to a recent oslo.versionedobjects change, but not 100% sure | 20:48 |
| TheJulia | ugh, I bet its the objects expanding out for the notification messages | 20:53 |
| JayF | yeah | 20:53 |
| JayF | this is ideal claude case though | 20:53 |
| JayF | easy to reproduce, I can point it at a root cause | 20:53 |
| TheJulia | ++ | 20:53 |
| JayF | I will say this: if the plan is like "lol remove -W" I may chuck electronics out the window lol | 20:53 |
| opendevreview | Jay Faulkner proposed openstack/ironic master: Fix docs build breakage caused by new o.vo https://review.opendev.org/c/openstack/ironic/+/988538 | 20:57 |
| JayF | cardoe: TheJulia ^ | 20:57 |
| cardoe | +2/+w... thanks for fixing that JayF | 21:31 |
| opendevreview | Merged openstack/ironic master: Fix docs build breakage caused by new o.vo https://review.opendev.org/c/openstack/ironic/+/988538 | 21:54 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Add TLS hardening options for the API service https://review.opendev.org/c/openstack/ironic/+/985878 | 23:13 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Add TLS hardening options for JSON-RPC service https://review.opendev.org/c/openstack/ironic/+/985879 | 23:13 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Add TLS hardening for agent client connections https://review.opendev.org/c/openstack/ironic/+/985880 | 23:13 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Add docs for TLS hardening options https://review.opendev.org/c/openstack/ironic/+/985881 | 23:13 |
| opendevreview | Julia Kreger proposed openstack/ironic master: novncproxy - Standardize ssl->tls for parameters https://review.opendev.org/c/openstack/ironic/+/988546 | 23:13 |
| opendevreview | Merged openstack/sushy-tools master: Handle pending BIOS settings the same way as real hardware https://review.opendev.org/c/openstack/sushy-tools/+/988434 | 23:54 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!