| openstackgerrit | Julia Kreger proposed openstack/ironic-tempest-plugin master: Actually log state transition timeouts in debug log https://review.opendev.org/720620 | 00:16 |
|---|---|---|
| *** cdearborn has quit IRC | 00:28 | |
| *** derekh has quit IRC | 00:42 | |
| *** ociuhandu has joined #openstack-ironic | 01:01 | |
| *** ociuhandu has quit IRC | 01:05 | |
| *** ricolin has joined #openstack-ironic | 01:14 | |
| *** tzumainn has quit IRC | 01:15 | |
| *** yaawang has quit IRC | 01:46 | |
| *** yaawang has joined #openstack-ironic | 01:47 | |
| *** yaawang has quit IRC | 01:52 | |
| *** yaawang has joined #openstack-ironic | 01:53 | |
| *** Qianbiao has joined #openstack-ironic | 01:54 | |
| *** Goneri has quit IRC | 01:59 | |
| openstackgerrit | Merged openstack/ironic master: Update jobs description https://review.opendev.org/718722 | 02:32 |
| openstackgerrit | Merged openstack/ironic master: Update iDRAC doc about vendor passthru timeout https://review.opendev.org/720173 | 02:32 |
| openstackgerrit | Verification of a change to openstack/ironic failed: Generalize ISO building for virtual media driver https://review.opendev.org/677996 | 02:37 |
| *** igordc has joined #openstack-ironic | 02:46 | |
| *** mkrai has joined #openstack-ironic | 02:50 | |
| *** igordc has quit IRC | 02:52 | |
| *** igordc has joined #openstack-ironic | 02:52 | |
| *** igordc has quit IRC | 02:57 | |
| *** rcernin has quit IRC | 03:10 | |
| *** rcernin has joined #openstack-ironic | 03:24 | |
| *** yaawang has quit IRC | 03:40 | |
| *** yaawang has joined #openstack-ironic | 03:41 | |
| *** ociuhandu has joined #openstack-ironic | 03:44 | |
| *** rh-jelabarre has quit IRC | 03:48 | |
| *** ociuhandu has quit IRC | 03:49 | |
| openstackgerrit | Merged openstack/ironic stable/train: Do not autoescape all Jinja2 templates https://review.opendev.org/716963 | 04:01 |
| openstackgerrit | Verification of a change to openstack/ironic failed: Update iDRAC doc about soft power off timeout https://review.opendev.org/718761 | 04:01 |
| *** mkrai has quit IRC | 04:53 | |
| *** mkrai has joined #openstack-ironic | 04:54 | |
| *** mkrai has quit IRC | 04:56 | |
| *** mkrai has joined #openstack-ironic | 04:57 | |
| *** mnasiadka has quit IRC | 05:10 | |
| *** Anticimex has quit IRC | 05:10 | |
| *** sziviani has quit IRC | 05:10 | |
| *** shadower has quit IRC | 05:10 | |
| *** jtomasek has quit IRC | 05:14 | |
| *** mnasiadka has joined #openstack-ironic | 05:15 | |
| *** Anticimex has joined #openstack-ironic | 05:15 | |
| *** sziviani has joined #openstack-ironic | 05:15 | |
| *** shadower has joined #openstack-ironic | 05:15 | |
| *** jtomasek has joined #openstack-ironic | 05:16 | |
| *** k_mouza has joined #openstack-ironic | 05:19 | |
| *** k_mouza has quit IRC | 05:23 | |
| *** akahat|ruck is now known as akahat|rover | 05:31 | |
| *** mkrai_ has joined #openstack-ironic | 05:33 | |
| *** mkrai has quit IRC | 05:33 | |
| *** ociuhandu has joined #openstack-ironic | 05:37 | |
| *** mkrai_ has quit IRC | 05:37 | |
| *** ociuhandu has quit IRC | 05:47 | |
| *** gyee has quit IRC | 05:53 | |
| *** mkrai has joined #openstack-ironic | 06:02 | |
| *** Qianbiao has quit IRC | 06:03 | |
| *** Qianbiao has joined #openstack-ironic | 06:04 | |
| *** mkrai has quit IRC | 06:10 | |
| *** mkrai_ has joined #openstack-ironic | 06:10 | |
| iurygregory | good morning Ironic o/ Happy Friday! | 06:23 |
| kaifeng | hey iurygregory, happy friday \o/ | 06:24 |
| iurygregory | hey kaifeng o/ | 06:25 |
| openstackgerrit | Merged openstack/ironic master: Fix issue where server fails to reboot https://review.opendev.org/714491 | 06:31 |
| *** ociuhandu has joined #openstack-ironic | 06:37 | |
| *** ociuhandu has quit IRC | 06:43 | |
| arne_wiebalck | Good morning iurygregory kaifeng and ironic! | 06:44 |
| *** dougsz has joined #openstack-ironic | 06:48 | |
| iurygregory | morning arne_wiebalck o/ | 06:49 |
| kaifeng | hey arne_wiebalck :) | 06:50 |
| kaifeng | arne_wiebalck: i am just about to ask how do you change process/thread for ironic conductor | 06:50 |
| kaifeng | it seems multiprocess can only be achieved by hacking code, we pass worker=N to service launcher | 06:51 |
| arne_wiebalck | kaifeng: Yeah, I was wondering this myself when I mentioned this the other day :) | 06:51 |
| arne_wiebalck | kaifeng: I don't remember, I need to check. | 06:51 |
| kaifeng | while for threads, i am not sure which one you have changed, our conductor uses green threads only. | 06:51 |
| arne_wiebalck | I don't think I changed the code. | 06:52 |
| kaifeng | okay, so the workers are just coroutines as i can see it, raising workers doesn't have obvious impact on the memory usage as expected | 06:54 |
| arne_wiebalck | I think I changed the httpd config. | 06:55 |
| kaifeng | arne_wiebalck: hmm, so it's api service | 06:57 |
| arne_wiebalck | kaifeng: my notes are not very detailed, but this was around the time we changed from eventlet to wsgi | 06:58 |
| arne_wiebalck | kaifeng: yes, it is the API workers | 06:58 |
| *** yolanda has quit IRC | 06:59 | |
| *** yolanda has joined #openstack-ironic | 07:00 | |
| kaifeng | arne_wiebalck: ack, for large scale nodes, i think we didn't utilize pagination, this could be an improvment | 07:00 |
| *** jhesketh has quit IRC | 07:04 | |
| *** ociuhandu has joined #openstack-ironic | 07:09 | |
| *** rpittau|afk is now known as rpittau | 07:19 | |
| rpittau | good morning ironic! o/ | 07:19 |
| iurygregory | morning rpittau o/ | 07:23 |
| rpittau | hey iurygregory :) | 07:24 |
| openstackgerrit | Aija Jaunteva proposed openstack/ironic master: Update iDRAC doc about soft power off timeout https://review.opendev.org/718761 | 07:26 |
| *** akahat|rover is now known as akahat|rover|lun | 07:29 | |
| *** akahat|rover|lun is now known as akahat|rover|brb | 07:29 | |
| *** ociuhandu has quit IRC | 07:30 | |
| kaifeng | rpittau o/ | 07:35 |
| rpittau | hey kaifeng :) | 07:35 |
| *** akahat|rover|brb is now known as akahat|rover | 07:53 | |
| *** threestrands has quit IRC | 07:56 | |
| *** derekh has joined #openstack-ironic | 08:02 | |
| *** ociuhandu has joined #openstack-ironic | 08:04 | |
| Qianbiao | hello, i find raid delete_configuration support both clean_step and deploy_step. | 08:06 |
| Qianbiao | is there any document about when it will be used in deploy step | 08:06 |
| *** dtantsur|afk is now known as dtantsur | 08:13 | |
| dtantsur | morning ironic | 08:13 |
| dtantsur | TGIF! | 08:13 |
| *** alexmcleod has joined #openstack-ironic | 08:17 | |
| iurygregory | morning dtantsur o/ | 08:18 |
| *** ociuhandu has quit IRC | 08:23 | |
| *** ociuhandu has joined #openstack-ironic | 08:23 | |
| *** lucasagomes has joined #openstack-ironic | 08:24 | |
| openstackgerrit | Ilya Etingof proposed openstack/ironic master: Do boot validation on cleaning https://review.opendev.org/720661 | 08:24 |
| openstackgerrit | Ilya Etingof proposed openstack/ironic master: Add `network_data` field to ironic node object https://review.opendev.org/687910 | 08:26 |
| kaifeng | hey dtantsur o/ | 08:26 |
| openstackgerrit | Ilya Etingof proposed openstack/ironic master: Add `get_node_network_data` method to NetworkInterface https://review.opendev.org/688184 | 08:26 |
| openstackgerrit | Ilya Etingof proposed openstack/ironic master: Add `get_node_network_data` to non-Neutron NetworkInterface https://review.opendev.org/688266 | 08:27 |
| openstackgerrit | Ilya Etingof proposed openstack/ironic master: Make redfish boot interface to use `network_data` https://review.opendev.org/688272 | 08:27 |
| kaifeng | dtantsur: i was wrong on the not-working-gmr, it seems unrelated with signal, we just didn't pass configuration to it, it doesn't have default log_dir | 08:28 |
| openstackgerrit | Ilya Etingof proposed openstack/ironic master: Add `get_node_network_data` to Neutron NetworkInterface https://review.opendev.org/688896 | 08:28 |
| openstackgerrit | Ilya Etingof proposed openstack/ironic master: Add L3 boot section to the docs https://review.opendev.org/689844 | 08:28 |
| dtantsur | ah | 08:28 |
| kaifeng | dtantsur: since we make gmr as optional, i want to get your opinion on whether we should add oslo.reports to our config generator? | 08:29 |
| dtantsur | kaifeng: we probably should if it doesn't require adding it to requirements | 08:34 |
| kaifeng | dtantsur: ack, thanks, i will propose a patch after local test | 08:40 |
| *** derekh has quit IRC | 08:43 | |
| *** ociuhandu has quit IRC | 08:45 | |
| *** tkajinam has quit IRC | 08:51 | |
| *** derekh has joined #openstack-ironic | 08:56 | |
| *** mkrai_ has quit IRC | 09:01 | |
| *** k_mouza has joined #openstack-ironic | 09:02 | |
| openstackgerrit | Aija Jaunteva proposed openstack/ironic master: Update iDRAC doc about soft power off timeout https://review.opendev.org/718761 | 09:08 |
| *** yolanda has quit IRC | 09:23 | |
| *** ociuhandu has joined #openstack-ironic | 09:25 | |
| *** yolanda has joined #openstack-ironic | 09:29 | |
| *** mkrai has joined #openstack-ironic | 09:33 | |
| openstackgerrit | Merged openstack/ironic master: Release reservation when stoping the ironic-conductor service https://review.opendev.org/718342 | 09:34 |
| openstackgerrit | Merged openstack/ironic-tempest-plugin master: Split Tempest tests between ipmi and redfish https://review.opendev.org/712041 | 09:34 |
| openstackgerrit | Merged openstack/ironic-python-agent stable/train: A boot partition on a GPT disk should be considered an EFI partition https://review.opendev.org/720549 | 09:34 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic master: Add ironic-standalone-redfish https://review.opendev.org/712042 | 09:40 |
| *** jtomasek has quit IRC | 09:41 | |
| *** jtomasek has joined #openstack-ironic | 09:41 | |
| *** ociuhandu has quit IRC | 09:43 | |
| *** k_mouza has quit IRC | 09:47 | |
| *** ociuhandu has joined #openstack-ironic | 09:49 | |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic-tempest-plugin master: Add standalone redfish jobs https://review.opendev.org/720675 | 09:49 |
| *** k_mouza has joined #openstack-ironic | 09:58 | |
| *** yolanda has quit IRC | 10:10 | |
| *** uzumaki has joined #openstack-ironic | 10:17 | |
| uzumaki | o/ ironic | 10:17 |
| openstackgerrit | Merged openstack/ironic-python-agent-builder master: Cleanup py27 support https://review.opendev.org/720554 | 10:23 |
| *** rpittau is now known as rpittau|bbl | 10:30 | |
| iurygregory | hey uzumaki | 10:42 |
| uzumaki | heyy! how you doing iurygregory ? | 10:42 |
| uzumaki | It's such a fine day here, rain, a bit chilly, perfect spring whether! | 10:42 |
| iurygregory | uzumaki, doing good, and you? | 10:42 |
| *** k_mouza has quit IRC | 10:43 | |
| uzumaki | I'm great! | 10:43 |
| openstackgerrit | Noor Muhammad Malik proposed openstack/ironic master: Implement the bios-interface for idrac-wsman driver https://review.opendev.org/713450 | 10:44 |
| dtantsur | uzumaki: ideally, after a patch is approved, minor issues should be addresses in a new one | 10:46 |
| dtantsur | to avoid us re-reviewing the whole change from scratch | 10:46 |
| *** ociuhandu has quit IRC | 10:48 | |
| *** ociuhandu has joined #openstack-ironic | 10:48 | |
| dtantsur | unless you deem an issue critical enough, that is | 10:50 |
| uzumaki | ah! I see, I'll keep that in mind, sorry about that. | 10:52 |
| *** ociuhandu has quit IRC | 10:52 | |
| dtantsur | no worries :) | 11:02 |
| *** k_mouza has joined #openstack-ironic | 11:04 | |
| uzumaki | I really loved working in this community, I'll be contributing more surely :D (starting with the todo's ;)) | 11:08 |
| *** mkrai has quit IRC | 11:09 | |
| *** mkrai has joined #openstack-ironic | 11:09 | |
| dtantsur | great :) | 11:15 |
| *** ociuhandu has joined #openstack-ironic | 11:23 | |
| *** Qianbiao has quit IRC | 11:23 | |
| *** Qianbiao has joined #openstack-ironic | 11:23 | |
| openstackgerrit | Merged openstack/ironic-inspector master: Add variable to set discovery timeout in tempest tests https://review.opendev.org/720394 | 11:33 |
| iurygregory | nice \o/ | 11:33 |
| uzumaki | \o/ | 11:34 |
| *** dougsz has quit IRC | 11:35 | |
| *** uzumaki has quit IRC | 11:38 | |
| iurygregory | If any core can approve https://review.opendev.org/#/c/717684/ https://review.opendev.org/#/c/717689/ before we do the release for the exporter =) | 11:46 |
| patchbot | patch 717684 - ironic-prometheus-exporter - Add lower constraints jobs - 2 patch sets | 11:46 |
| patchbot | patch 717689 - ironic-prometheus-exporter - Use hacking for pep8 tests - 5 patch sets | 11:46 |
| * dtantsur has +2 already | 11:47 | |
| iurygregory | yeah | 11:47 |
| * iurygregory is wondering if we have etherpads for the Victoria PTG | 11:48 | |
| etingof | why don't we have `validate_cleaning` method? | 11:50 |
| dtantsur | etingof: why should we? we add special methods for processes that are not on the default path (like inspection) | 11:51 |
| *** hjensas has quit IRC | 11:52 | |
| etingof | that's my question - if we got `validate_inspection`, should we also have `validate_cleaning`? | 11:53 |
| dtantsur | no, because cleaning is on the default path (and identical to deploying) | 11:53 |
| *** jcoufal has joined #openstack-ironic | 11:54 | |
| etingof | hmm, deployment might require image among other params, no? while cleaning would not | 11:54 |
| *** rh-jelabarre has joined #openstack-ironic | 11:54 | |
| dtantsur | maybe that's why we don't call driver.boot.validate :) | 11:55 |
| dtantsur | yeah, it's the same thing I hit with inspection, I think: the image is required | 11:55 |
| etingof | ++ | 11:55 |
| *** tkajinam has joined #openstack-ironic | 11:56 | |
| etingof | so boot validation for cleaning - does it make sense at all? | 11:57 |
| dtantsur | well, it does. I wonder if we can implement it without big changes to the boot interface itself | 11:58 |
| dtantsur | maybe we need a sort of validate_ramdisk (with validate_inspection using it internally) | 11:58 |
| etingof | yes! | 11:58 |
| *** jcoufal has quit IRC | 11:59 | |
| etingof | that's what I came up with in my dreams | 11:59 |
| dtantsur | man, you have scary dreams :D | 12:00 |
| dtantsur | it would be awesome if you could untangle this mess | 12:00 |
| etingof | perhaps I have to, otherwise network_data validation is stuck | 12:00 |
| dtantsur | etingof: cannot you proceed without boot interface validation? | 12:02 |
| etingof | well, by calling validation from prepare_ramdisk...? | 12:03 |
| etingof | that works, but it's you know... | 12:03 |
| dtantsur | etingof: I mean, why do you even need to call validation so badly? | 12:03 |
| etingof | to make sure that ramdisk will be able to consume network configuration | 12:04 |
| *** tkajinam has quit IRC | 12:04 | |
| etingof | the alternative is to pull that far away from the drivers, to the API level... | 12:04 |
| etingof | logically, it feels like network_data is part of boot info which makes sense to validate closer to boot interface... | 12:05 |
| dtantsur | etingof: so, yes. the latter. we cannot make API structure validation depend on the current driver, so just move it to the API. | 12:05 |
| dtantsur | as a bonus, you won't be blocked any more :) | 12:06 |
| etingof | that's that old argument... | 12:06 |
| dtantsur | (not that I mind you figuring out validation of the boot interface..) | 12:06 |
| etingof | I see it as more logical to stick to boot interface, tbh | 12:06 |
| etingof | and not making network_data JSON part of the API | 12:07 |
| dtantsur | It's part of an API, so it has to be validated in the API. | 12:07 |
| dtantsur | you don't have an option of not making it | 12:07 |
| dtantsur | because that's the task in question | 12:07 |
| dtantsur | as a thought exercise, you can try implementing this feature without anyhow exposing it in the API :) | 12:08 |
| *** Qianbiao has quit IRC | 12:09 | |
| *** Qianbiao has joined #openstack-ironic | 12:09 | |
| *** rcernin has quit IRC | 12:11 | |
| etingof | my desire is not to make it operator-ramdisk kind of interaction, without making ironic understanding the details of the format | 12:16 |
| etingof | the rationale is to account for custom-built ramdisks | 12:17 |
| dtantsur | etingof: let's for a thing assume it's a real problem to account for (it's not). then moving the validation around won't help. | 12:18 |
| *** rpittau|bbl is now known as rpittau | 12:19 | |
| etingof | considering network_data format as part of the API - this is what I think is not good enough | 12:19 |
| etingof | not moving code around indeed | 12:20 |
| dtantsur | same with boot interface validation | 12:20 |
| dtantsur | anyway, that's not a real problem to solve. people don't rebuild ramdisks just to mess with us. | 12:20 |
| *** hjensas has joined #openstack-ironic | 12:21 | |
| dtantsur | let's please concentrate on delivering the MVP. we've already missed Ussuri while discussing future enhancements nobody's asked for. | 12:21 |
| dtantsur | blocking on reworking the boot interface validation is hardly justified at this point | 12:21 |
| dtantsur | also keep in mind that it's easier to relax restrictions than to tighten them | 12:21 |
| etingof | I see it more as a design consideration than as a feature... | 12:23 |
| etingof | anyway, I will move it somewhere else for now | 12:24 |
| openstackgerrit | Aija Jaunteva proposed openstack/ironic master: Fix RAID configuration with idrac-wsman interface https://review.opendev.org/720702 | 12:26 |
| *** mkrai has quit IRC | 12:28 | |
| openstackgerrit | Merged openstack/ironic master: Generalize ISO building for virtual media driver https://review.opendev.org/677996 | 12:32 |
| * dtantsur -> lunch | 12:38 | |
| *** Goneri has joined #openstack-ironic | 12:41 | |
| *** ociuhandu has quit IRC | 12:48 | |
| *** ociuhandu has joined #openstack-ironic | 12:49 | |
| *** rloo has joined #openstack-ironic | 13:02 | |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic master: Add ironic-standalone-redfish https://review.opendev.org/712042 | 13:07 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic master: Add ironic-standalone-redfish https://review.opendev.org/712042 | 13:11 |
| dtantsur | etingof: if you're seriously interested in incompatible 3rd party ramdisks and drivers, let's do this. Let's implement the feature ASAP in the most strict way (so that we have an option of relaxing it). Let's create an RFE in storyboard to track supporting non-nova formats and evaluate interest (after letting the MVP out of the door). | 13:16 |
| dtantsur | WDYT? | 13:16 |
| *** gokhani has joined #openstack-ironic | 13:17 | |
| etingof | I can do that... but i am not optimistic tbh | 13:18 |
| dtantsur | about what exactly? | 13:19 |
| etingof | well, one thing that bothers me is having Nova network_data monster being tied up settled as part of ironic API (long, old dispute) | 13:20 |
| etingof | the other thing is that I am not sure we will ever find time to rework this | 13:20 |
| etingof | for example, with API validation approach we ensure the validity of Node.network_data | 13:21 |
| dtantsur | etingof: for one, we need to settle on *something*. I think it was use who proposed network_data initially, but I may be mistaken. | 13:21 |
| dtantsur | second, if we never get priority to rework this, it means that nobody cares. if nobody cares, why do the work? | 13:22 |
| etingof | ...if we eventually relax this API validation that might mean that Node.network_data might go bad | 13:22 |
| etingof | but other drivers might not be prepared for bad Node.network_data | 13:23 |
| dtantsur | see, you're already arguing for having a strict API validation :) | 13:23 |
| etingof | I am just thinking over the amount of work that will be involved to change the design | 13:23 |
| dtantsur | Note that it was not me who suggested the design will ever change\ | 13:24 |
| dtantsur | I personally believe it wont' | 13:24 |
| dtantsur | but "the amount of work" doesn't depend on whether we have API validations or ANY validations | 13:24 |
| dtantsur | we'll still have to make sure that the already stored network_data doesn't become invalid | 13:25 |
| dtantsur | we'll still have to make sure that old applications setting node.network_data keep working | 13:25 |
| dtantsur | changing the format and/or the backend will be a lot of work, no matter what we decide now | 13:26 |
| etingof | I am OK with (1) using nova metadata, (2) having it validated strictly before it hits ramdisk. I am not happy about having this whole chain hardcoded into ironic, I do not see that as a benefit to the project going forward. | 13:26 |
| etingof | I think that's the only contention that we have. | 13:26 |
| dtantsur | I cannot understand how you implement an input format without hardcoding it... | 13:27 |
| etingof | well, we discussed a few ways... the one I favor is to have validation performed by the boot interface against user-configurable schema | 13:29 |
| *** mkrai has joined #openstack-ironic | 13:32 | |
| etingof | anyway, I move the validation to node API to get it out of the door | 13:33 |
| hjensas | ironic-cores, can you please add these inspector patches to your queues - https://review.opendev.org/#/q/topic:port_physnet_hook+(status:open+OR+status:merged) | 13:34 |
| dtantsur | etingof: that doesn't result in great user experience (no discoverability), but worry not: if we deem it needed, we can always relax the requirement later. | 13:34 |
| dtantsur | going from a hardcoded schema to per-driver one is a user-invisible change (initially) | 13:34 |
| dtantsur | we just need to make sure we have demand for that | 13:34 |
| dtantsur | hjensas: I was about to ask, but never got time to review properly: why do you need a base hook there? | 13:35 |
| dtantsur | do you envision future extensions? | 13:35 |
| hjensas | dtantsur: stackhpc have two more hooks, one using a switchname map, and another for infiniband that they may want to upstream. | 13:36 |
| dtantsur | okay, I'll look into your patches | 13:36 |
| hjensas | dtantsur: thanks. | 13:36 |
| *** cdearborn has joined #openstack-ironic | 13:39 | |
| erbarr | hello ironic | 13:45 |
| *** tzumainn has joined #openstack-ironic | 13:47 | |
| etingof | I am definitely worrying about this prescriptive approach. In the context of ironic being a general-purpose provisioning tool that becomes part of other tool sets, the assumption that nobody packs their own ramdisk does not look too convincing to me | 14:00 |
| etingof | It does not even require to mess with IPA per-se, it can be only about repacking ramdisk OS | 14:01 |
| etingof | not to mention that we have official conductor/IPA API which in theory promotes alternative IPA implementation | 14:02 |
| etingof | I am not arguing about this (I move things to API for now). Just to explain why I think this is a mistake. | 14:03 |
| *** ociuhandu has quit IRC | 14:07 | |
| dtantsur | etingof: "the assumption that nobody packs their own ramdisk" is incorrect. the assumption that nobody will overwrite our features with a new implementation just because is quite reasonable. | 14:07 |
| dtantsur | and even so, people hate repacking ramdisks. most of the attempts fail. they only do it when left with no other choice. | 14:07 |
| dtantsur | the cases include nearly exclusively troubleshooting and missing kernel modules | 14:07 |
| dtantsur | ah, and custom hardware managers | 14:09 |
| *** dtantsur is now known as dtantsur|brb | 14:09 | |
| *** ociuhandu has joined #openstack-ironic | 14:11 | |
| TheJulia | dtantsur|brb: looks like https://review.opendev.org/#/c/713450/ got updated while in workflow | 14:11 |
| patchbot | patch 713450 - ironic - Implement the bios-interface for idrac-wsman driver - 24 patch sets | 14:11 |
| etingof | quite reasonable sounds somewhat subjective to me | 14:11 |
| *** gokhani has quit IRC | 14:12 | |
| etingof | "640K is more memory than anyone will ever need on a computer" | 14:12 |
| TheJulia | heh | 14:21 |
| openstackgerrit | Harald Jensås proposed openstack/ironic-inspector master: Port physical network CIDR map hook https://review.opendev.org/718594 | 14:32 |
| *** dtantsur|brb is now known as dtantsur | 15:03 | |
| dtantsur | etingof: "I don't feel it's correct" is also pretty subjective; doesn't prevent you from using it as your main argument ;) | 15:04 |
| openstackgerrit | Aija Jaunteva proposed openstack/ironic master: Update iDRAC doc about soft power off timeout https://review.opendev.org/718761 | 15:05 |
| etingof | indeed, we are still humans despite the circumstances | 15:05 |
| * dtantsur is an owl, but never mind | 15:06 | |
| etingof | though "my gut feeling" is not my argument | 15:06 |
| dtantsur | my proposal is simply: let's solve requests in the order we receive them | 15:07 |
| dtantsur | we cannot implement a perfect thing in informational vacuum | 15:07 |
| dtantsur | let's ship a small thing and see how people react | 15:07 |
| * dtantsur has had too much experience with implementing features that nobody ended up caring about | 15:08 | |
| *** ociuhandu has quit IRC | 15:10 | |
| *** ociuhandu has joined #openstack-ironic | 15:13 | |
| *** mkrai has quit IRC | 15:18 | |
| *** ricolin has quit IRC | 15:19 | |
| etingof | as for listening for feedback, I also have some reservations... may be we do not actually get much of the feedback... may be they either manage to use the tool or just walk away to try their luck with other tools? | 15:26 |
| erbarr | i started seeing this error in the ironic conductor log | 15:26 |
| erbarr | ERROR ironic.drivers.modules.agent_base [None req-163d125e-c81f-4b7a-a466-4676afbfc415 None None] Failed to install a bootloader when deploying node 48463eec-31f9-474f-92f8-a3ed6bdc9e37. Error: {'message': "install_bootloader() got an unexpected keyword argument 'target_boot_mode'", 'code': 500, 'type': 'TypeError', 'details': ''} | 15:26 |
| dtantsur | erbarr: older IPA? does it result in an actual failure or only in this message? | 15:34 |
| dtantsur | arne_wiebalck: do we have any compatibility there ^^? | 15:34 |
| *** Qianbiao has quit IRC | 15:34 | |
| erbarr | failure, and yea I think old | 15:35 |
| arne_wiebalck | dtantsur: yes, this is old IPA and new conductor | 15:35 |
| arne_wiebalck | dtantsur: I'd say | 15:35 |
| arne_wiebalck | dtantsur: erm ... | 15:36 |
| arne_wiebalck | dtantsur: the code on the conductor side should handle this | 15:36 |
| dtantsur | erbarr: do you have a traceback? | 15:36 |
| arne_wiebalck | erbarr: what does the conductor say | 15:37 |
| erbarr | https://www.irccloud.com/pastebin/xbJBzWI4/ | 15:38 |
| *** ociuhandu has quit IRC | 15:40 | |
| erbarr | full log is here https://stash.dellemc-community.org/logs/dellemc-poweredge-R730-R630-tempest-dsvm-ironic-idrac/2020-04-17:03:26/687910/28//logs/screen-ir-cond.txt | 15:40 |
| *** ociuhandu has joined #openstack-ironic | 15:42 | |
| dtantsur | arne_wiebalck: I don't see any handling of target_boot_mode being absent | 15:42 |
| dtantsur | s/absent/not supported/ | 15:42 |
| dtantsur | we may need to update agent_base | 15:43 |
| arne_wiebalck | it got merged 2 days ago or so ... | 15:43 |
| arne_wiebalck | well, the attempt was merged :) | 15:44 |
| arne_wiebalck | https://github.com/openstack/ironic/commit/65b5ac6a7a985dd430614be03d168d5a7921ab2c | 15:44 |
| dtantsur | arne_wiebalck: `if target_boot_mode == 'uefi'` is problematic | 15:45 |
| dtantsur | we cannot break compatibility for everyone who is using UEFI.. | 15:45 |
| dtantsur | is the breakage justified? cannot new ironic work at all with old IPA in UEFI mode? | 15:46 |
| *** ociuhandu has quit IRC | 15:47 | |
| arne_wiebalck | I think the assumption was this path is only exercised for UEFI and s/w RAID | 15:48 |
| arne_wiebalck | which may be wrong | 15:49 |
| openstackgerrit | Aija Jaunteva proposed openstack/ironic master: Fix RAID configuration with idrac-wsman interface https://review.opendev.org/720702 | 15:49 |
| dtantsur | arne_wiebalck: you only check for UEFI, maybe the RAID check is missing? | 15:50 |
| arne_wiebalck | dtantsur: is install_bootloader called in the non-RAID case? | 15:50 |
| arne_wiebalck | erbarr: you do not use software RAID, do you? | 15:50 |
| * dtantsur is not sure | 15:51 | |
| erbarr | i don't believe so | 15:51 |
| erbarr | let me check | 15:51 |
| dtantsur | arne_wiebalck: it's called for whole disk images + UEFI | 15:51 |
| dtantsur | OR software RAID | 15:51 |
| dtantsur | arne_wiebalck: I'd probably move the logic from agent_client to agent_base to make it more explicit | 15:52 |
| arne_wiebalck | dtantsur: right, wdi and uefi | 15:54 |
| erbarr | i'm seeing no-raid | 15:54 |
| arne_wiebalck | erbarr: thanks | 15:55 |
| arne_wiebalck | dtantsur: I think the IPA does not do anything except for s/w RAID | 15:56 |
| dtantsur | arne_wiebalck: yeah, but we're passing a non-supported argument | 15:57 |
| dtantsur | nothing even gets to actual code in IPA | 15:57 |
| arne_wiebalck | dtantsur: sure, just explaining where the mistake comes from | 15:57 |
| dtantsur | ah | 15:57 |
| arne_wiebalck | not sure if we can check for s/w RAID on the conductor easily ... | 15:58 |
| *** dustinc has joined #openstack-ironic | 16:00 | |
| *** ociuhandu has joined #openstack-ironic | 16:05 | |
| *** lucasagomes has quit IRC | 16:05 | |
| arne_wiebalck | we should rather break s/w RAID than all UEFI | 16:07 |
| arne_wiebalck | we could pass the software RAID flag also to agent_client | 16:10 |
| dtantsur | arne_wiebalck: it's not impossible to check on the conductor side | 16:10 |
| dtantsur | or is it?.. | 16:10 |
| dtantsur | mmmm | 16:10 |
| arne_wiebalck | well, we just did in agent_base | 16:10 |
| arne_wiebalck | see, the 'software_raid' flag | 16:11 |
| dtantsur | mmm, we assume that target_raid_config does not change between cleaning and deployment | 16:11 |
| arne_wiebalck | in agent_base | 16:11 |
| dtantsur | fair enough, although not 100% robust | 16:11 |
| dtantsur | then yes, we can only break in case of UEFI+sw RAID | 16:11 |
| arne_wiebalck | right | 16:12 |
| arne_wiebalck | we pass the flag with the boot mode to agent_client or do the same dance in agent_client again? | 16:12 |
| dtantsur | arne_wiebalck: I'd rather use a flag or move the retry logic to agent_base | 16:13 |
| arne_wiebalck | dtantsur: ok | 16:13 |
| *** ociuhandu has quit IRC | 16:15 | |
| *** ociuhandu has joined #openstack-ironic | 16:16 | |
| *** dtantsur is now known as dtantsur|afk | 16:16 | |
| dtantsur|afk | have a great weekend | 16:16 |
| rpittau | bye everyone, have a nice weekend! o/ | 16:17 |
| *** rpittau is now known as rpittau|afk | 16:17 | |
| *** ociuhandu has quit IRC | 16:21 | |
| *** ianychoi_ has joined #openstack-ironic | 16:29 | |
| *** hamzy__ has joined #openstack-ironic | 16:31 | |
| *** gyee has joined #openstack-ironic | 16:31 | |
| *** ianychoi has quit IRC | 16:31 | |
| *** hamzy_ has quit IRC | 16:31 | |
| *** yaawang_ has joined #openstack-ironic | 16:33 | |
| *** dtruong has quit IRC | 16:33 | |
| *** yaawang has quit IRC | 16:34 | |
| *** dtruong has joined #openstack-ironic | 16:34 | |
| *** ociuhandu has joined #openstack-ironic | 16:40 | |
| arne_wiebalck | erbarr: I'll submit a patch trying to fix this. | 16:47 |
| erbarr | thank you, much appreciated | 16:48 |
| openstackgerrit | Arne Wiebalck proposed openstack/ironic master: Don't break UEFI install with older IPAs https://review.opendev.org/720781 | 16:55 |
| arne_wiebalck | erbarr: dtantsur|afk ^^ ... hot-fix :) | 16:56 |
| arne_wiebalck | bye everyone, have a good weekend o/ | 16:56 |
| erbarr | bye, enjoy your weekeknd | 16:57 |
| *** bdodd has quit IRC | 16:58 | |
| *** bdodd has joined #openstack-ironic | 16:59 | |
| *** ociuhandu has quit IRC | 17:03 | |
| *** igordc has joined #openstack-ironic | 17:04 | |
| *** igordc has quit IRC | 17:09 | |
| *** derekh has quit IRC | 17:12 | |
| *** k_mouza has quit IRC | 17:16 | |
| *** ociuhandu has joined #openstack-ironic | 17:30 | |
| *** ociuhandu has quit IRC | 17:43 | |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic master: Stop configuring install_command in tox. https://review.opendev.org/720790 | 17:55 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic-inspector master: Stop configuring install_command in tox. https://review.opendev.org/720791 | 17:58 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic-prometheus-exporter master: Stop configuring install_command in tox. https://review.opendev.org/720793 | 18:08 |
| *** ociuhandu has joined #openstack-ironic | 18:09 | |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/sushy master: Stop configuring install_command in tox. https://review.opendev.org/720795 | 18:14 |
| *** ociuhandu has quit IRC | 18:22 | |
| *** ociuhandu has joined #openstack-ironic | 18:23 | |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic-prometheus-exporter master: Stop configuring install_command in tox. https://review.opendev.org/720793 | 18:23 |
| *** ociuhandu has quit IRC | 18:28 | |
| *** Lucas_Gray has joined #openstack-ironic | 18:38 | |
| *** alexmcleod has quit IRC | 18:43 | |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic-inspector master: Stop configuring install_command in tox. https://review.opendev.org/720791 | 18:43 |
| iurygregory | yay lower-constraints have some problems http://lists.openstack.org/pipermail/openstack-discuss/2020-April/014237.html | 18:44 |
| * iurygregory will push the changes to tox.ini to see *boom* and fix it =) | 18:44 | |
| TheJulia | :( | 18:45 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/networking-baremetal master: Stop configuring install_command in tox. https://review.opendev.org/720806 | 18:46 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/sushy-tools master: Stop configuring install_command in tox. https://review.opendev.org/720807 | 18:46 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/networking-generic-switch master: Stop configuring install_command in tox. https://review.opendev.org/720808 | 18:47 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/bifrost master: Stop configuring install_command in tox. https://review.opendev.org/720809 | 18:48 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/sushy-cli master: Stop configuring install_command in tox. https://review.opendev.org/720810 | 18:49 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic-python-agent master: Stop configuring install_command in tox. https://review.opendev.org/720811 | 18:50 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic-ui master: Stop configuring install_command in tox. https://review.opendev.org/720813 | 18:53 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/virtualbmc master: Stop configuring install_command in tox. https://review.opendev.org/720816 | 18:55 |
| TheJulia | iurygregory: thanks | 19:09 |
| *** dougsz has joined #openstack-ironic | 19:15 | |
| iurygregory | TheJulia, yw, the sad part is that locally nothing blows up, but in the CI it complain about some constraints so I'm waitting for results to update o/ | 19:18 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/bifrost master: Stop configuring install_command in tox. https://review.opendev.org/720809 | 19:20 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic-prometheus-exporter master: Stop configuring install_command in tox. https://review.opendev.org/720793 | 19:24 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/networking-generic-switch master: Stop configuring install_command in tox. https://review.opendev.org/720808 | 19:26 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/networking-baremetal master: Stop configuring install_command in tox. https://review.opendev.org/720806 | 19:28 |
| *** ociuhandu has joined #openstack-ironic | 19:35 | |
| *** ociuhandu has quit IRC | 19:42 | |
| *** jtomasek has quit IRC | 20:03 | |
| TheJulia | fun fun :( | 20:13 |
| *** dustinc has quit IRC | 20:29 | |
| *** dougsz has quit IRC | 20:33 | |
| *** dougsz has joined #openstack-ironic | 20:34 | |
| openstackgerrit | Verification of a change to openstack/python-ironic-inspector-client failed: Convert job to dib https://review.opendev.org/718698 | 20:35 |
| *** dougsz has quit IRC | 20:38 | |
| rloo | hi TheJulia, wrt https://review.opendev.org/#/c/665835/. Wondering if you were good with my comment about us assuming that the port's internal_info stuff is not specific to a particular network driver, but it must have those entries (the *_vif_port_id) for any portlikeobj, regardless of the network driver. | 20:44 |
| patchbot | patch 665835 - ironic - Block port deletions where vif is present - 12 patch sets | 20:44 |
| *** ociuhandu has joined #openstack-ironic | 20:45 | |
| *** ociuhandu has quit IRC | 21:06 | |
| *** ociuhandu has joined #openstack-ironic | 21:08 | |
| TheJulia | well, the key is for them not to, but I'm not sure i saw/grok a comment to that effect | 21:12 |
| *** ociuhandu has quit IRC | 21:13 | |
| rloo | TheJulia: I commented in patch set 11. In that case, we need to go back to some earlier version. | 21:14 |
| TheJulia | reading | 21:15 |
| TheJulia | I see it was on the body of the reply | 21:15 |
| TheJulia | So vifs are not specific to the network driver interface, but the reason I did it there was because it seemed like we should represent underlying behavior, so if one changes to noop networking not care | 21:16 |
| TheJulia | but... I think the proper thing is to just overall maintain the behavior of "there be data here, for some reason, something is wrong because it shouldn't be here" | 21:17 |
| *** rh-jelabarre has quit IRC | 21:31 | |
| rloo | TheJulia: huh? Is that hedging our bets? :) So you are good with that PR then? Am taking off, please comment in the pr & I'll look on Monday. Have a great weekend! | 21:55 |
| *** rloo has quit IRC | 21:55 | |
| *** ociuhandu has joined #openstack-ironic | 22:32 | |
| *** Lucas_Gray has quit IRC | 22:32 | |
| *** ociuhandu has quit IRC | 22:38 | |
| *** igordc has joined #openstack-ironic | 23:08 | |
| *** gyee has quit IRC | 23:10 | |
| *** stevebaker has quit IRC | 23:21 | |
| *** igordc has quit IRC | 23:27 | |
| *** tzumainn has quit IRC | 23:46 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!