| arne_wiebalck | Good morning, Ironic! | 07:25 |
|---|---|---|
| stendulker | arne_wiebalck: Good morning ! | 07:27 |
| arne_wiebalck | hey stendulker, good morning! | 07:27 |
| *** bodgix1 is now known as bodgix | 10:07 | |
| janders | good morning arne_wiebalck stendulker and Ironic o/ | 10:21 |
| arne_wiebalck | hey janders o/ | 10:21 |
| stendulker | good morning janders ! | 10:25 |
| iurygregory | good morning Ironic o/ | 11:24 |
| dtantsur | okay, fun, fixing the tempest plugin exposes our inability to test partition images with local boot.. | 11:30 |
| opendevreview | Mark Goddard proposed openstack/tenks master: Support UEFI boot mode https://review.opendev.org/c/openstack/tenks/+/827479 | 11:34 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: CI: expose default_boot_option to the tempest plugin https://review.opendev.org/c/openstack/ironic/+/827481 | 11:38 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: CI: set the tempest plugin's partition_netboot option https://review.opendev.org/c/openstack/ironic/+/827481 | 11:42 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic-tempest-plugin master: Apply the correct image reference when booting an instance https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/827340 | 11:50 |
| ajya | Hi Ironic, a quick Q - after manual cleaning a neutron port is deleted. What's the purpose of that? And what's intended workflow if want to do another cleaning or do a deployment? Recreate neutron port? | 12:18 |
| dtantsur | ajya: ironic creates it and ironic deletes it. for a deployment either a user or nova creates a new port, they can be even on a different network. | 12:19 |
| ajya | dtantsur: will it also be ok if tempest recreates the port as part of the test? There is one test in mind to revert BIOS settings on hardware after a testing BIOS apply_configuration. | 12:22 |
| ajya | dtantsur: and when you say Ironic creates it - do you mean it happens as part of the manual cleaning process? So far I create it explicitly with `openstack port create` and then attach baremetal port. | 12:28 |
| ajya | not expecting that neutron port will be deleted. Maybe something wrong with my workflow? | 12:29 |
| dtantsur | ajya: you're not expected to create ports for manual cleaning | 12:58 |
| dtantsur | the fact that it works is probably an accident | 12:58 |
| ajya | dtantsur: ok, thanks, will reassess the scripts | 12:59 |
| dtantsur | metalsmith stable/wallaby is quite broken... | 13:34 |
| opendevreview | Mark Goddard proposed openstack/tenks master: Support UEFI boot mode https://review.opendev.org/c/openstack/tenks/+/827479 | 13:41 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: [DNM] Testing the CI https://review.opendev.org/c/openstack/ironic/+/827500 | 13:53 |
| opendevreview | Mark Goddard proposed openstack/tenks master: Support UEFI boot mode https://review.opendev.org/c/openstack/tenks/+/827479 | 14:07 |
| opendevreview | Mark Goddard proposed openstack/tenks master: Support UEFI boot mode https://review.opendev.org/c/openstack/tenks/+/827479 | 14:08 |
| iurygregory | dtantsur, maybe is related to the pip issue? (not sure if is affecting wallaby) | 14:39 |
| opendevreview | Mark Goddard proposed openstack/tenks master: Support UEFI boot mode https://review.opendev.org/c/openstack/tenks/+/827479 | 14:42 |
| dtantsur | iurygregory: maybe? which pip issue? | 14:52 |
| iurygregory | [infra] Issue with the new pip | 14:52 |
| iurygregory | pip 22.0.0 | 14:52 |
| iurygregory | I only saw the email today ( /me needs to create better filters...) | 14:53 |
| iurygregory | seems like its not compatible with py3.6 ofc :D | 14:54 |
| dtantsur | oh no | 14:54 |
| iurygregory | oh yes... | 14:54 |
| dtantsur | no, the metalsmith wallaby bug is with pxe booting | 14:54 |
| iurygregory | ogw | 14:55 |
| iurygregory | D: | 14:55 |
| iurygregory | I need to grab some lunch now, I have a meeting about tempest in 1hr, I can try to look at things later | 14:55 |
| dtantsur | cool, thanks! (I'll need to leave early today, as always on Wednesdays) | 14:56 |
| iurygregory | https://review.opendev.org/c/openstack/metalsmith/+/827234 this patch right? | 14:56 |
| dtantsur | exactly | 14:58 |
| mnasiadka | dtantsur: any outlook on getting https://review.opendev.org/c/openstack/bifrost/+/825053 and https://review.opendev.org/c/openstack/bifrost/+/825054 merged? | 15:14 |
| mnasiadka | Kolla-Ansible's patch to drop centos8 jobs in stable/train depends on those ;-) | 15:14 |
| dtantsur | mnasiadka: done | 15:21 |
| opendevreview | Merged openstack/bifrost bugfix/10.0: [CI] Drop all config https://review.opendev.org/c/openstack/bifrost/+/825053 | 15:23 |
| opendevreview | Merged openstack/bifrost bugfix/10.1: [CI] Drop all config https://review.opendev.org/c/openstack/bifrost/+/825054 | 15:24 |
| * TheJulia kind of wakes up | 16:33 | |
| opendevreview | Mark Goddard proposed openstack/tenks master: Support UEFI boot mode https://review.opendev.org/c/openstack/tenks/+/827479 | 16:38 |
| opendevreview | Mark Goddard proposed openstack/tenks master: Support UEFI boot mode https://review.opendev.org/c/openstack/tenks/+/827479 | 16:41 |
| * iurygregory thinks TheJulia should go back to bed | 16:49 | |
| TheJulia | I'm feeling a lot better | 16:51 |
| TheJulia | Minus early migraine symptoms, but I have medication for that | 16:52 |
| iurygregory | nice \o/ | 16:52 |
| opendevreview | Mark Goddard proposed openstack/tenks master: Support UEFI boot mode https://review.opendev.org/c/openstack/tenks/+/827479 | 17:50 |
| arne_wiebalck | TheJulia: just revived introspection on active nodes via a container to update missing inventory data, seems to work ok :) | 18:05 |
| arne_wiebalck | TheJulia: we will run this on some 200 nodes to update the missing RAM info (lshw bug from some months ago), then potentially on a few thousand nodes to move from basic to extended inventory data ... | 18:06 |
| TheJulia | arne_wiebalck: awesome | 18:07 |
| * arne_wiebalck just wanted to report sth positive with all the breakages :) | 18:08 | |
| arne_wiebalck | bye everyone, see you tomorrow o/ | 18:08 |
| iurygregory | bye arne_wiebalck o/ | 18:20 |
| stevebaker[m] | good morning | 19:09 |
| iurygregory | good morning stevebaker[m] | 19:20 |
| TheJulia | good morning stevebaker[m] | 20:14 |
| TheJulia | stevebaker[m]: so I've been thinking about rbac and I'm thinking we might be able to use the pre-existing methods and just make list/arguments... so I'm wondering if something like this could be sane enough to articulate it | 20:16 |
| TheJulia | https://etherpad.opendev.org/p/ironic-rbac-testing | 20:16 |
| TheJulia | hmm, I'm not sure we get result codes | 20:18 |
| TheJulia | no, not all the way back | 20:18 |
| stevebaker[m] | TheJulia: that would make the existing tests more clear also. I assume 'valid' and 'invalid' can just be arbitrary labels because the actual assertion is the code | 20:19 |
| stevebaker[m] | ... or the response content | 20:19 |
| TheJulia | well, since each one would loop through all of the roles in each, it would be how to convey a positive scenario expectation, and a negative exception I guess | 20:20 |
| * TheJulia wonders how to deal with when we need a piece of data | 20:20 | |
| opendevreview | Mark Goddard proposed openstack/tenks master: Change default boot mode to UEFI https://review.opendev.org/c/openstack/tenks/+/827553 | 20:20 |
| stevebaker[m] | TheJulia: well, I'm in favor of one test entry testing more than one scenario, in general. Just need to settle on the format | 20:23 |
| TheJulia | yeah | 20:24 |
| TheJulia | which has to be extensible enough to make it work basically | 20:24 |
| TheJulia | updated etherpad, thoughts? | 20:26 |
| TheJulia | I believe clients get setup per class, so the entire ddt class run should have consistent project idents and all | 20:28 |
| TheJulia | and a third... | 20:32 |
| rpioso | Good morning, ironic :) | 20:39 |
| rpioso | TheJulia: Is there a hook that supports including a custom data blob in the ISO created for DHCP-less ramdisk booting? | 20:40 |
| TheJulia | a custom data blob, outside of network configuration I don't think so | 20:41 |
| TheJulia | why? | 20:41 |
| rpioso | TheJulia: For services which need configuration data and/or licenses. Similar to networking :) | 20:44 |
| TheJulia | per node dynamically, I don't really see a way. | 20:44 |
| TheJulia | If you can embed something into the ramdisk, I think that would make the most sense | 20:45 |
| TheJulia | Additional services implies additional things occuring outside of the typical deployment case, so there may be need to do something additional. | 20:45 |
| rpioso | TheJulia: Yes, per node, and allocated during the deployment workflow. | 20:47 |
| TheJulia | stevebaker[m]: I guess it would make sense to do things like list nodes as well as just get the singular node | 20:47 |
| TheJulia | rpioso: so in order to perform deployment, before the final os boots? | 20:48 |
| rpioso | TheJulia: Embedding it into the ramdisk would mean building a new ISO per node, which strikes me as costly. | 20:50 |
| rpioso | TheJulia: The data is needed by the service hosted by the running, final OS to do its job. | 20:51 |
| rpioso | TheJulia: It's analogous to networking configuration. | 20:52 |
| TheJulia | rpioso: we already have to do that anyway for config injection with networking | 20:57 |
| TheJulia | final os configuration can just go into instance config drive user-data | 20:58 |
| TheJulia | that is the design-intended use, fwiw | 20:58 |
| stevebaker[m] | TheJulia: I've proposed a syntax in the ether pad, see #83 vs #102 | 20:59 |
| TheJulia | stevebaker[m]: i kind of like the per-case idea | 21:00 |
| TheJulia | the one thing though, we won't have any headers, we'll have to drive the tempest client | 21:00 |
| TheJulia | so it will automatically unwind operations as well | 21:00 |
| stevebaker[m] | ok, so it will be some higher-level concept than actual headers? | 21:01 |
| TheJulia | yeah | 21:02 |
| TheJulia | so in tempest, we can ask the class to initialize clients attached to the base class | 21:03 |
| TheJulia | and each has access/roles and any operation through them uses that role/rights/access | 21:03 |
| TheJulia | so we end up with a thing like... self.os_system_member.baremetal.BaremetalClient() which will become a fully fledged baremetal client which we can assign to self.client and do things like self.client.create_node() | 21:05 |
| TheJulia | self.client is the style in the plugin | 21:05 |
| TheJulia | so thinking we could loop through $things, and use the expected client | 21:07 |
| TheJulia | stevebaker[m]: you know... a setup case would likely be good separately | 21:07 |
| TheJulia | since we have so many things only a system_admin or sysetm_member or owner_admin can really do | 21:08 |
| * TheJulia wonders if the touchpad ont he work laptop has gone haywire | 21:09 | |
| rpioso | TheJulia: Ahhhh ... I see, similar to how first-boot services, such as Ignition, can be provided its configuration data. | 21:11 |
| TheJulia | yup | 21:11 |
| rpioso | TheJulia: Thank you! | 21:11 |
| TheJulia | cloud-init as well | 21:11 |
| TheJulia | oh, so ignition does *prefer* a URL to download from instead of read from the user-data | 21:12 |
| TheJulia | thre is a case where it is limited, but I don't remember | 21:12 |
| TheJulia | dtantsur might remember | 21:12 |
| * TheJulia lays down for a little while | 21:13 | |
| rpioso | TheJulia: I suppose it could vary from service to service. I expect fetching it from a URL would be common. Thanks, again, and take care of yourself. | 21:15 |
| TheJulia | rpioso: yeah. Ignition really expects to be handed a bootstrap or config load url to download configuration from | 21:16 |
| TheJulia | and I think the size limit is so you can include a certificate iwth it | 21:16 |
| TheJulia | but I don't remember, and that might have changed as of recent | 21:16 |
| TheJulia | anyway, this wind storm is kind of crazy, hopefully it is just the pressure changes | 21:17 |
| JayF | TheJulia: congrats, chairwoman | 21:28 |
| janders | good morning Ironic o/ | 21:44 |
| rpioso | janders: \o | 21:45 |
| janders | hey rpioso o/ | 21:47 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!