| opendevreview | Iury Gregory Melo Ferreira proposed openstack/ironic master: Add DB API for Firmware https://review.opendev.org/c/openstack/ironic/+/883062 | 01:06 |
|---|---|---|
| opendevreview | Iury Gregory Melo Ferreira proposed openstack/ironic master: FW Object structure https://review.opendev.org/c/openstack/ironic/+/883093 | 01:07 |
| opendevreview | Merged openstack/ironic master: follow-up on DPU change api-ref https://review.opendev.org/c/openstack/ironic/+/884301 | 01:58 |
| *** jroll3 is now known as jroll | 04:20 | |
| opendevreview | Merged openstack/ironic master: Remove model_query use from general dbapi calls https://review.opendev.org/c/openstack/ironic/+/883968 | 04:43 |
| opendevreview | Merged openstack/ironic master: Remove unused get_not_versions from dbapi https://review.opendev.org/c/openstack/ironic/+/883969 | 04:43 |
| rpittau | good morning ironic! o/ | 07:33 |
| Sandzwerg[m] | hmm I tried to change a MAC after a NIC replacement on a active node and it was possible for the ports itself after I set the node in maintenance but I got an error when I did the same for the port group. Is that expected or did I hold it wrong? This is on xena | 09:10 |
| Sandzwerg[m] | okay a colleague found it. Neutron is to blame. You can't update the mac of a bound port (sounds resonable). | 09:36 |
| opendevreview | Graeme Moss proposed openstack/networking-generic-switch master: Fix fallback to plug_port_to_network on 802_3ad bonds https://review.opendev.org/c/openstack/networking-generic-switch/+/881959 | 10:01 |
| opendevreview | Merged openstack/ironic master: Add ironic-grenade-skip-level Job https://review.opendev.org/c/openstack/ironic/+/836966 | 10:45 |
| opendevreview | waleed mousa proposed openstack/ironic-python-agent master: Follow-up Add documentation for MellanoxDeviceHardwareManager https://review.opendev.org/c/openstack/ironic-python-agent/+/884357 | 10:55 |
| iurygregory | good morning Ironic | 11:16 |
| iurygregory | if anyone has time for reviews today https://review.opendev.org/c/openstack/ironic/+/883031/ https://review.opendev.org/c/openstack/ironic/+/883062/ https://review.opendev.org/c/openstack/ironic/+/883093/ | 12:34 |
| iurygregory | tks! | 12:35 |
| iurygregory | dtantsur, regarding the delete call, don't we need to delete the reference in the DB when necessary? (i.e the node is removed, for example) | 13:03 |
| dtantsur | iurygregory: you already have code to remove the references when the node is removed, check dbapi | 13:07 |
| dtantsur | it's done as part of the node deletion code, not separately | 13:07 |
| iurygregory | gotcha, I was following a bit what we have for bios seetings | 13:09 |
| dtantsur | I'm not sure why we have these in BIOS either :) | 13:09 |
| iurygregory | oh wow | 13:09 |
| TheJulia | should easy enough fix :) | 13:36 |
| * TheJulia tries to wake up | 13:37 | |
| TheJulia | Sandzwerg[m]: using vif detach/attaach? | 13:37 |
| *** dking is now known as Guest1321 | 13:45 | |
| Sandzwerg[m] | TheJulia: yeah probably. I haven't tried it but unbinding the port, doing the change and rebinding the port should work. | 13:48 |
| Sandzwerg[m] | Ideally we want to use such a change to reimport a node to make sure all is in order but we might have customers were that might not be possible so we thought if it would be possible to just update the mac | 13:49 |
| TheJulia | okay, yeah | 13:57 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/wallaby: DNM Fix Cinder Integration fallout from CVE-2023-2088 https://review.opendev.org/c/openstack/ironic/+/883581 | 14:01 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/2023.1: Fix self_owned_node policy check https://review.opendev.org/c/openstack/ironic/+/884371 | 14:16 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/zed: Fix self_owned_node policy check https://review.opendev.org/c/openstack/ironic/+/884402 | 14:18 |
| TheJulia | getting those backports merged should allow us to completely fix the tempest plugin | 14:19 |
| TheJulia | hjensas: if you want to disable zed for now on the tempest plugin to fix the policy, I'm good with that personally | 14:19 |
| TheJulia | as long as we get the fixes merged | 14:19 |
| opendevreview | Merged openstack/ironic master: Add DB model for Firmware https://review.opendev.org/c/openstack/ironic/+/883031 | 14:23 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Add hold steps https://review.opendev.org/c/openstack/ironic/+/879060 | 14:27 |
| TheJulia | has anyone looked at the metal3 job recently? no exceptions, but looks like the job is just not really getting started?! | 14:30 |
| rpittau | TheJulia: w00t | 14:31 |
| iurygregory | I was checking but couldn't really figure out the problem .-. | 14:32 |
| rpittau | I see it's failing | 14:33 |
| rpittau | sinec a while | 14:34 |
| rpittau | I don't know why I thought it was moved to voting | 14:34 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Enable vendor interfaces to be called as steps https://review.opendev.org/c/openstack/ironic/+/879089 | 14:35 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Add wait step https://review.opendev.org/c/openstack/ironic/+/879061 | 14:35 |
| TheJulia | rebasing | 14:35 |
| TheJulia | iurygregory: okay, that doesn't make me feel that bad that you kind of hit the same point I did as well | 14:36 |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: Make metal3 job voting https://review.opendev.org/c/openstack/ironic/+/882073 | 14:36 |
| rpittau | there ^ rebased | 14:36 |
| rpittau | let's see | 14:36 |
| TheJulia | it looks like a node is never enrolled | 14:36 |
| rpittau | I think I see what the problem is | 14:36 |
| TheJulia | which makes me think something with the operator might be unhappy itself | 14:36 |
| rpittau | yeah | 14:36 |
| * TheJulia goes back to her rebase party | 14:37 | |
| rpittau | " Error from server: etcdserver: request timed out" well ok | 14:38 |
| iurygregory | <insert it's fine gif> | 14:38 |
| rpittau | last success was on 14/05 | 14:38 |
| *** Guest1321 is now known as dking | 14:39 | |
| iurygregory | TheJulia, did I? I thought your other patches wouldn't have merge conflict at least with the DB model for Firmware | 14:40 |
| dking | Does anybody know the channel for diskimage-builder? | 14:41 |
| TheJulia | etcd change? | 14:41 |
| TheJulia | dking: #openstack-dib | 14:41 |
| TheJulia | hjensas: so the rbac job change is my own job... doh! | 14:41 |
| dking | TheJulia: Oh, right! Thank you very much. | 14:42 |
| opendevreview | Julia Kreger proposed openstack/ironic-tempest-plugin master: CI: Add RBAC specific tempest jobs for main branch exec https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/882312 | 14:45 |
| TheJulia | iurygregory: rebasing largely because I had missed incrementing a field (although, I'm thinking I might need to change the patches again | 14:46 |
| TheJulia | oh, wait, it was the primary API version, and I don't think yours has merged. yet, since I did mine as a all-in-one | 14:47 |
| iurygregory | oh wow | 14:47 |
| iurygregory | the API version I'm still working on it | 14:47 |
| hjensas | TheJulia: :) ack, I voted on it again. | 14:48 |
| TheJulia | any objection if we backport https://review.opendev.org/c/openstack/sushy/+/867675 on a private method change ? | 14:50 |
| TheJulia | note it changes _op | 14:50 |
| iurygregory | I'm ok with it | 14:56 |
| dtantsur | private method should be fine | 14:58 |
| TheJulia | okay, that makes me feel more comfortable with backporting it and the ilo fix | 14:58 |
| opendevreview | Julia Kreger proposed openstack/ironic-python-agent master: Allow md5 to be disabled from the conductor https://review.opendev.org/c/openstack/ironic-python-agent/+/882367 | 14:59 |
| opendevreview | Merged openstack/ironic-python-agent master: Follow-up Add documentation for MellanoxDeviceHardwareManager https://review.opendev.org/c/openstack/ironic-python-agent/+/884357 | 15:03 |
| TheJulia | dtantsur: at some point, we likely just need to get on a call and discuss https://review.opendev.org/c/openstack/ironic-python-agent/+/879984 | 15:07 |
| dtantsur | TheJulia: possibly? Cannot we just use !address.is_global or something like that? | 15:09 |
| TheJulia | temporary addresses are global | 15:10 |
| dtantsur | then what's the problem with having them? | 15:10 |
| TheJulia | They will only be valid for so many seconds | 15:10 |
| TheJulia | I want to think the default is 300 seconds, trying to hunt down what the knob is | 15:11 |
| opendevreview | Mahnoor Asghar proposed openstack/ironic master: Add to Redfish hardware inventory collection https://review.opendev.org/c/openstack/ironic/+/881783 | 15:11 |
| dtantsur | TheJulia: I'm a bit on the fence whether we should go so far to guess the intention | 15:12 |
| TheJulia | well, if a temporary address gets picked | 15:13 |
| dtantsur | maybe then people should NOT enable this? :-/ | 15:13 |
| TheJulia | because the order list can basically be random based upon time, then we could actually have agent communications get broken due to the address drop | 15:13 |
| dtantsur | I mean, if you enable feature designed for mobile phones on a server... | 15:13 |
| TheJulia | so kernel default is off, distros might be turning it on | 15:14 |
| TheJulia | I have debian at home and it is on by default | 15:14 |
| TheJulia | I can look later today or this weekend when I swing by my house | 15:14 |
| TheJulia | I've got a few v6 enabled vms there too I can check, if it is just generally disabled, then meh and move forward | 15:16 |
| TheJulia | I didn't think the extensions were for phones, but more so to make it harder to pin down end users to specific hardware with slaac usage.= | 15:16 |
| dtantsur | yeah, which is 99% a case for phones IMO | 15:17 |
| dtantsur | do we even have a reliable way to distinguish SLAAC from SLAAC-with-privacy? | 15:17 |
| TheJulia | well, not just for phones, but any network hopping device really | 15:17 |
| TheJulia | only by getting info from the kernel for each entry | 15:18 |
| TheJulia | since it all gets kernel managed | 15:18 |
| TheJulia | "magic" | 15:18 |
| dtantsur | TheJulia: how do I check my system? | 15:18 |
| TheJulia | if it is enabled, or if you have any addresses present? | 15:18 |
| dtantsur | the latter | 15:18 |
| TheJulia | ip -6 |grep temp | 15:18 |
| TheJulia | err | 15:18 |
| TheJulia | ip -6 addr |grep temp | 15:18 |
| TheJulia | that should work, on a csb machine, it is disabled | 15:19 |
| dtantsur | nothing on my fedora | 15:20 |
| TheJulia | ack, I'll check my machines at home, I have yet to setup a vpn | 15:20 |
| TheJulia | ... for better or worse. | 15:21 |
| TheJulia | (also, sleep is enabled because electricity is expensive | 15:21 |
| TheJulia | ) | 15:21 |
| TheJulia | and no cats are there to walk across keybaords turning the machines back on | 15:22 |
| dtantsur | heh | 15:23 |
| * TheJulia now has a cat inspector named Peaches who is instead looking for birds out the window | 15:24 | |
| opendevreview | Riccardo Pittau proposed openstack/ironic master: [DNM] troubleshoot metal3-integration job https://review.opendev.org/c/openstack/ironic/+/884413 | 15:32 |
| * TheJulia ponders taking tomorrow off | 15:35 | |
| opendevreview | Merged openstack/ironic master: Don't return the in-flight SQL handler https://review.opendev.org/c/openstack/ironic/+/882732 | 15:40 |
| JayF | Heads up: I'm out sick today. Caught whatever cold my wife had earlier this week. | 15:45 |
| iurygregory | dtantsur, we have the same problem with create_bios_settings_list in case of the DBDuplicateEntry to raise BIOSSettingAlreadyExists <O> | 15:47 |
| iurygregory | <insert it's fine gif> | 15:47 |
| rpittau | JayF: take care! | 15:47 |
| iurygregory | JayF, hope you feel better soon, take care! | 15:48 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Permit Ironic to notify IPA it can support MD5 https://review.opendev.org/c/openstack/ironic/+/882168 | 15:48 |
| rpittau | bye everyone and don't panic! o/ | 15:54 |
| iurygregory | 42! | 15:59 |
| iurygregory | happy geek pride day \o/ | 16:00 |
| opendevreview | Merged openstack/ironic stable/2023.1: [ci] [stable-only] Cinder fixed; make BFV job vote https://review.opendev.org/c/openstack/ironic/+/884280 | 16:05 |
| opendevreview | Merged openstack/ironic stable/zed: [ci] [stable-only] Cinder fixed; make BFV job vote https://review.opendev.org/c/openstack/ironic/+/884220 | 16:06 |
| opendevreview | Merged openstack/ironic stable/2023.1: CI: Disable mysql counters for grenade https://review.opendev.org/c/openstack/ironic/+/883699 | 16:06 |
| opendevreview | Merged openstack/ironic-tempest-plugin master: Clarify checksum for wholedisk images. https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/882096 | 16:06 |
| *** tosky_ is now known as tosky | 16:20 | |
| TheJulia | NobodyCam: in regard to your need, would a "execute service step with an "expect_heartbeat=false" option work? | 18:17 |
| NobodyCam | it might. but form what I looked at yesterday I was changing this line: https://github.com/openstack/ironic/blob/32532eeda56a1d9cfbe30bf8492cceccede82b0c/ironic/drivers/modules/agent.py#L836 | 18:20 |
| NobodyCam | to something like `return states.RESCUE` | 18:20 |
| TheJulia | so maybe a patch might be viable, since it could be viable, and reasonable to sort of do the same in service steps | 18:21 |
| NobodyCam | it worked for me in a hacked together test | 18:21 |
| TheJulia | yeah | 18:21 |
| NobodyCam | yea | 18:21 |
| TheJulia | I think the definitive way would be to have service steps | 18:21 |
| TheJulia | but your case is just a little bit different (yet, super valid actually) | 18:21 |
| TheJulia | just didn't think of the heartbeat part | 18:21 |
| NobodyCam | :) | 18:21 |
| TheJulia | oooh a formatting error of sort on service steps spec | 18:23 |
| TheJulia | 44 failing tests, fix one, 104 failing tests | 19:14 |
| TheJulia | fix another, and now 5 :) | 19:25 |
| NobodyCam | crazy question... does long lived IPA still exist? | 21:15 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/wallaby: DNM Fix Cinder Integration fallout from CVE-2023-2088 https://review.opendev.org/c/openstack/ironic/+/883581 | 21:25 |
| *** janders1 is now known as janders | 21:34 | |
| TheJulia | NobodyCam: sort of depends on where, there is fast track where the agent can be heartbeating, and we turn it over into deploy the node without a reboot | 21:42 |
| TheJulia | NobodyCam: time line wise, there is no enforce limit that I can recall, as long as the agent can still heartbeat for it to sit in manageable/available states https://docs.openstack.org/ironic/latest/_images/states.svg Conundrum is this was largely modeled/intended for standalone users, so a move to a cleaning network and back might not be desirable | 21:49 |
| opendevreview | Mahnoor Asghar proposed openstack/python-ironicclient master: Add support for node inventory https://review.opendev.org/c/openstack/python-ironicclient/+/883833 | 21:51 |
| *** sfinucan is now known as stephenfin | 21:52 | |
| opendevreview | Julia Kreger proposed openstack/ironic stable/wallaby: DNM Fix Cinder Integration fallout from CVE-2023-2088 https://review.opendev.org/c/openstack/ironic/+/883581 | 21:57 |
| TheJulia | NobodyCam: hope that helps | 21:58 |
| NobodyCam | we are looking to eliminate one of the post cycles in the deployment process. some of the new gen hardware has some very long post cycles | 22:03 |
| NobodyCam | TheJulia: are there doc I can search for? I assume "fast track" is what I should search for? | 22:05 |
| TheJulia | https://docs.openstack.org/ironic/latest/admin/fast-track.html | 22:07 |
| iurygregory | thoughts on how we should handle the creation of a list of firmware components? I was following the idea we have in create_bios_setting_list but Dmitry pointed something important in the review https://review.opendev.org/c/openstack/ironic/+/883062/4/ironic/db/sqlalchemy/api.py#2687 we don't handle properly if there is a DuplicatedEntry because it's always the last component, I'm thinking in have a private method | 23:32 |
| iurygregory | that would handle creation one by one and would be called in create_firmware_component_list, thoughts? | 23:32 |
| iurygregory | so yeah, currently we have a bug in create_bios_setting_list | 23:33 |
| TheJulia | Move flush in closet on the loop so each time it gets applied, that should then trigger as expected | 23:34 |
| TheJulia | Err, closer | 23:34 |
| TheJulia | At least, that is my immediate thought | 23:35 |
| TheJulia | Honestly, I’d abstract it one step away from the db | 23:37 |
| iurygregory | like we calling multiple times a single create? | 23:38 |
| TheJulia | Pull current, reconcile as finite transactions based upon just the diff | 23:38 |
| TheJulia | Yeah, the premise I guess here is that once the node is already known at all, it is just updates | 23:38 |
| TheJulia | Then again, is components everything, or just the delta | 23:39 |
| TheJulia | Are you working tomorrow? | 23:39 |
| iurygregory | yes | 23:39 |
| iurygregory | I still have 3hrs to work today | 23:39 |
| iurygregory | so I'm trying to update this patch and get to the slides :D | 23:40 |
| TheJulia | Heh | 23:40 |
| TheJulia | https://meet.google.com/hhu-eopz-cya | 23:41 |
| TheJulia | lets talk it through | 23:41 |
| continuity | Hey, does anyone have any clues on how to build a rocky linux image for use with ironic, and software raid. I have managed to get something that will deploy, but it hangs part way through the boot cycle on rocky 8? | 23:41 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
