| opendevreview | Merged openstack/ironic-python-agent master: Allow md5 to be disabled from the conductor https://review.opendev.org/c/openstack/ironic-python-agent/+/882367 | 03:53 |
|---|---|---|
| opendevreview | yatin proposed openstack/ironic master: [DNM] Attempt source install dnsmasq https://review.opendev.org/c/openstack/ironic/+/888121 | 06:09 |
| opendevreview | yatin proposed openstack/ironic master: [DNM] Attempt source install dnsmasq https://review.opendev.org/c/openstack/ironic/+/888121 | 08:07 |
| opendevreview | Merged openstack/bifrost master: Fix schema[meta] linter warnings https://review.opendev.org/c/openstack/bifrost/+/874855 | 08:24 |
| opendevreview | Huy Mai proposed openstack/sushy-tools master: Add fake_ipa inspection, lookup and heartbeater to fake system https://review.opendev.org/c/openstack/sushy-tools/+/875366 | 12:22 |
| opendevreview | Huy Mai proposed openstack/sushy-tools master: Add fake_ipa inspection, lookup and heartbeater to fake system https://review.opendev.org/c/openstack/sushy-tools/+/875366 | 12:41 |
| TheJulia | good morning | 13:19 |
| TheJulia | is it some sort of holiday?! | 13:20 |
| JayF | it's jay in the uk day | 13:23 |
| opendevreview | Julia Kreger proposed openstack/ironic master: DB: Fix result set locking with periodics https://review.opendev.org/c/openstack/ironic/+/888188 | 13:25 |
| * TheJulia shrugs | 13:26 | |
| * TheJulia attempts to wake up | 13:26 | |
| dtantsur | good morning TheJulia | 13:38 |
| opendevreview | Jay Faulkner proposed openstack/ironic master: DNM: Testing sharding against Ironic https://review.opendev.org/c/openstack/ironic/+/888277 | 13:41 |
| JayF | TheJulia: you have a sec? | 13:57 |
| TheJulia | sort of, whats up? | 13:57 |
| JayF | TheJulia: I'm here with johnthetubaguy IRL and we have a question about automatic_lessee | 13:57 |
| JayF | mainly he's confused about how we're getting the user information and is concerned it'd be a service user | 13:57 |
| JayF | and I don't understand how it works well enough to explain it and we've been looking for quite a bit lol | 13:57 |
| JayF | mainly he's concerned it's getting the project owned by the service user | 13:58 |
| JayF | but I thought that's not how it worked? | 13:58 |
| TheJulia | hmmmm | 13:59 |
| TheJulia | rewinding brain, hold please | 13:59 |
| JayF | we can do a gmeet if you want? | 13:59 |
| JayF | The concern is this: automatic_lessee may use the service project instead of the one that called nova | 14:03 |
| JayF | and tbh it made me wonder if that might have been by design | 14:03 |
| TheJulia | sigh, sure | 14:03 |
| TheJulia | sorry, trying to get a proliant but fix posted and a gerrit is fighting with me | 14:04 |
| JayF | yeah it's okay, this is no rush I can async it to john g when/if you have time | 14:05 |
| JayF | we are just having a brain itch for the answer | 14:05 |
| TheJulia | like 5 minutes or so | 14:05 |
| JayF | ** it only works for standalone, doesn't it? | 14:07 |
| TheJulia | it should work fully integrated | 14:07 |
| JayF | okay so then the brain itch persists lol | 14:07 |
| TheJulia | because when you authenticate with a service token, your supposed to transmit "I'm doing this on behalf of" data | 14:07 |
| TheJulia | we do that with neutron, and the pattern is repeated all over the place in various projects | 14:07 |
| JayF | John is under the impression that nova is doing that ~wrong | 14:08 |
| JayF | for the ironic api client bits | 14:08 |
| TheJulia | ... that wouldn't surprise me | 14:08 |
| JayF | and looking at the code, I think he might be right | 14:08 |
| TheJulia | wow that was absurdly painful, thanks gerrit! | 14:09 |
| JayF | we are on the make it work/fix it party | 14:10 |
| JayF | so don't sweat it :D | 14:10 |
| TheJulia | turns out caching was broken on train for ilo | 14:12 |
| TheJulia | https://review.opendev.org/c/x/proliantutils/+/694448 | 14:12 |
| TheJulia | :( | 14:12 |
| JayF | I added a followup item to explicitly test the behavior in devstack and either 1) document it if good or 2) get johnthetubaguy to fix it on nova side and then document it | 14:13 |
| JayF | automatic_lessee is wonderful for one of my use cases so it'll be nice to see it working :D | 14:13 |
| JayF | (also John didn't know about it, and it's not everyday you get to show john something he didn't know lol) | 14:14 |
| TheJulia | okay, I'm available now | 14:14 |
| JayF | we have 15 minutes | 14:14 |
| JayF | https://us06web.zoom.us/j/85169384929?pwd=N1FGdm5qUjhpbTdYdEMzUnBTUGRNQT09 | 14:15 |
| opendevreview | Huy Mai proposed openstack/sushy-tools master: Add fake_ipa inspection, lookup and heartbeater to fake system https://review.opendev.org/c/openstack/sushy-tools/+/875366 | 14:16 |
| opendevreview | Verification of a change to openstack/bifrost stable/xena failed: Skip unnecessary SDK get_machine calls https://review.opendev.org/c/openstack/bifrost/+/883118 | 14:57 |
| * TheJulia wonders if tuple(row) just works | 15:02 | |
| TheJulia | I guess calling tuple(Row) (as in Row the class) does just work | 15:10 |
| opendevreview | Julia Kreger proposed openstack/ironic master: DB: Fix result set locking with periodics https://review.opendev.org/c/openstack/ironic/+/888188 | 15:13 |
| JayF | we should steal the autopep8 with a diff feature from nova's linting setup | 15:24 |
| JayF | extremely nice | 15:24 |
| TheJulia | huh? | 15:24 |
| JayF | if you do `tox -epep8` in nova | 15:26 |
| JayF | and it fails | 15:26 |
| JayF | it returns a diff of what autopep8 would've done to fix it | 15:26 |
| JayF | so basically not just a pointer to where you failed lint, but also explicit example of fixing it | 15:26 |
| JayF | which I'm sure is even nicer in those places where the indentation is fiddly | 15:26 |
| JayF | TheJulia: I'm writing that patch for lessee we just talked about right meow | 15:26 |
| TheJulia | right meow() | 15:27 |
| TheJulia | is there an eventlet meow() method?! | 15:27 |
| JayF | yeah I just found it and when I called it | 15:31 |
| JayF | https://review.opendev.org/c/openstack/nova/+/888290 WIP: [ironic] Optionally set lessee on deployment [NEW] | 15:31 |
| JayF | magically appeared ;) | 15:31 |
| JayF | needs testing in devstack and unit tests, but I was hoping that John could ensure it's the Right Thing(tm) before I go too far | 15:32 |
| TheJulia | did you check the version of the api nova was using? | 15:38 |
| TheJulia | just to make sure the lessee field was visible | 15:38 |
| JayF | I'll check that when I load it into devstack to test | 15:39 |
| JayF | I will probably enable this on one of the ironic CI jobs once it lands | 15:39 |
| TheJulia | yeah | 15:39 |
| JayF | but for now, will setup a temp devstack to test and all | 15:39 |
| TheJulia | an ironic tempest test could do it | 15:39 |
| JayF | but will wait until I'm back in the us for it for sure | 15:39 |
| JayF | I'm going to doc this for Ironic before I get off, explictily saying "use automated_lessee for standalone; use nova's [ironic]set_lessee_on_deploy for integrated" | 15:41 |
| JayF | if someone wants this feature + mixed standalone and nova-driven is the only place this falls over | 15:41 |
| JayF | which bluntly, if you're doing that you're probably sophisticated enough to figure out a way around it | 15:41 |
| JayF | e.g. all nodes in nova have an owner and nova only sees those | 15:41 |
| JayF | yeah actually that isn't a workaround, that's just doing it right in this case, I think lol | 15:42 |
| TheJulia | I think there is a nuance there | 15:42 |
| TheJulia | "standalone applies direct ironic usage", but there may be different configurations operators have where lessee usage is valid as-is | 15:43 |
| TheJulia | just not directly mapped from nova, that is | 15:43 |
| JayF | oh, 100% absolutely | 15:43 |
| TheJulia | at least, in a prime nova configuration with only one nova | 15:43 |
| JayF | oh 10000% | 15:43 |
| JayF | I'm just talking about the automagic | 15:43 |
| TheJulia | multiple novas, with distinct openstacks, yeah | 15:43 |
| TheJulia | k | 15:43 |
| JayF | I am so stoked about this | 15:43 |
| JayF | I always talk about building blocks | 15:43 |
| TheJulia | did we create an ilo6 bug anywhere? | 15:43 |
| JayF | and this is great | 15:44 |
| JayF | TheJulia: not afaik; I have zero access or knowledge of that hardware beyond the press release about ilo v obmc | 15:44 |
| JayF | TheJulia: and Ididn't wanna file a bug that was nothing but FUD | 15:44 |
| TheJulia | yeah, I'd almost prefer actual hardware to play with | 15:45 |
| TheJulia | I'm going to create a patch and see if nisha can provide guidance, I found the chat where it came up | 15:45 |
| opendevreview | Huy Mai proposed openstack/sushy-tools master: Add fake_ipa inspection, lookup and heartbeater to fake system https://review.opendev.org/c/openstack/sushy-tools/+/875366 | 15:50 |
| opendevreview | Julia Kreger proposed openstack/ironic master: DNM/WIP: Detect ilo6 and redirect? https://review.opendev.org/c/openstack/ironic/+/888297 | 16:14 |
| opendevreview | Merged openstack/bifrost stable/2023.1: Skip unnecessary SDK get_machine calls https://review.opendev.org/c/openstack/bifrost/+/883115 | 16:16 |
| TheJulia | JayF: did your eventlet experiment yield results? | 16:26 |
| opendevreview | Merged openstack/bifrost stable/zed: Skip unnecessary SDK get_machine calls https://review.opendev.org/c/openstack/bifrost/+/883116 | 16:32 |
| opendevreview | Merged openstack/bifrost stable/yoga: Skip unnecessary SDK get_machine calls https://review.opendev.org/c/openstack/bifrost/+/883117 | 16:32 |
| opendevreview | Mahnoor Asghar proposed openstack/ironic master: Add the initial skeleton of the agent inspect interface https://review.opendev.org/c/openstack/ironic/+/877814 | 16:33 |
| opendevreview | Mahnoor Asghar proposed openstack/ironic master: [WIP] Very basic in-band inspection with the "agent" interface https://review.opendev.org/c/openstack/ironic/+/885450 | 16:33 |
| opendevreview | Mahnoor Asghar proposed openstack/ironic master: WIP: Add inspection (processing) hooks https://review.opendev.org/c/openstack/ironic/+/887554 | 16:33 |
| opendevreview | Merged openstack/ironic-specs master: Follow-up on DPU Management Change https://review.opendev.org/c/openstack/ironic-specs/+/882760 | 16:43 |
| TheJulia | so we have a weird... frustrating... weirdness in the vendor passthru stuffs | 17:55 |
| TheJulia | tl;dr we expect an http_method argument on the style of some, albeit *not* required except through the existence of the variable | 17:57 |
| TheJulia | some vendor drivers which have implemented passthrough steps just do **kwargs and move on in life | 17:57 |
| TheJulia | .... I think it is just a result of confusion... | 18:00 |
| TheJulia | yeah, lots of confusion, I can't find a case where it is used nor stated as expected the a parameter | 18:07 |
| opendevreview | Merged openstack/ironic stable/ussuri: Fix Cinder Integration fallout from CVE-2023-2088 https://review.opendev.org/c/openstack/ironic/+/885063 | 18:17 |
| opendevreview | Merged openstack/ironic stable/train: Fix Cinder Integration fallout from CVE-2023-2088 https://review.opendev.org/c/openstack/ironic/+/885065 | 18:17 |
| TheJulia | okay, it is explicitly optional but always included in practicality | 18:21 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Enable vendor interfaces to be called as steps https://review.opendev.org/c/openstack/ironic/+/879089 | 19:01 |
| opendevreview | Julia Kreger proposed openstack/ironic master: DB: Fix result set locking with periodics https://review.opendev.org/c/openstack/ironic/+/888188 | 19:07 |
| opendevreview | Merged openstack/ironic master: Disable spanning tree https://review.opendev.org/c/openstack/ironic/+/886863 | 19:55 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Add a little variability for heartbeating https://review.opendev.org/c/openstack/ironic/+/888359 | 22:31 |
| TheJulia | ^ might be a little over the top, given the other change, but trying to start a bunch of stuff all at once and then having this also trigger... dunno | 22:33 |
| TheJulia | I think we should add sort of similar logic around saving node updates, although it is not a deadlock as captured by oslodb | 22:36 |
| TheJulia | so we would need to carry it ourselves, which is not horrible really, as long as we keep it to the "most needful" places | 22:36 |
| TheJulia | really, with all locking centered around node update db calls, it really only makes complete sense there | 22:52 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!