| *** osmanlicilegi is now known as Guest0 | 01:48 | |
| opendevreview | Merged openstack/ironic master: Change molds option to appropriate class https://review.opendev.org/c/openstack/ironic/+/852253 | 01:58 |
|---|---|---|
| opendevreview | Verification of a change to openstack/ironic-inspector master failed: Ignore connection errors when getting ironic ports https://review.opendev.org/c/openstack/ironic-inspector/+/836847 | 01:58 |
| opendevreview | Merged openstack/networking-baremetal master: Add OpenConfig classes for interface aggregate https://review.opendev.org/c/openstack/networking-baremetal/+/835160 | 02:56 |
| opendevreview | Merged openstack/networking-baremetal master: Add OpenConfig classes for LACP https://review.opendev.org/c/openstack/networking-baremetal/+/835161 | 03:05 |
| opendevreview | Merged openstack/networking-baremetal master: OpenConfig YANG Model, python-bindings releasenote https://review.opendev.org/c/openstack/networking-baremetal/+/852468 | 03:20 |
| opendevreview | Merged openstack/ironic-lib master: Replace the netboot job with a local-boot one https://review.opendev.org/c/openstack/ironic-lib/+/848966 | 03:45 |
| opendevreview | Merged openstack/ironic-python-agent bugfix/8.6: Gather details about bond interfaces if present https://review.opendev.org/c/openstack/ironic-python-agent/+/848866 | 03:45 |
| opendevreview | Merged openstack/ironic-python-agent stable/xena: Gather details about bond interfaces if present https://review.opendev.org/c/openstack/ironic-python-agent/+/848868 | 03:45 |
| opendevreview | Merged openstack/ironic-python-agent bugfix/8.3: Gather details about bond interfaces if present https://review.opendev.org/c/openstack/ironic-python-agent/+/848867 | 03:48 |
| opendevreview | Verification of a change to openstack/ironic-python-agent bugfix/8.1 failed: Gather details about bond interfaces if present https://review.opendev.org/c/openstack/ironic-python-agent/+/848869 | 04:09 |
| opendevreview | Merged openstack/ironic-python-agent stable/wallaby: Gather details about bond interfaces if present https://review.opendev.org/c/openstack/ironic-python-agent/+/848990 | 05:31 |
| opendevreview | Merged openstack/ironic-python-agent stable/wallaby: Ignore efi grub2-install failure https://review.opendev.org/c/openstack/ironic-python-agent/+/841275 | 05:31 |
| opendevreview | Merged openstack/ironic-python-agent stable/wallaby: Catch ismount not being handled https://review.opendev.org/c/openstack/ironic-python-agent/+/841333 | 05:35 |
| opendevreview | chenwei proposed openstack/ironic-python-agent master: Remove unicode literal from code https://review.opendev.org/c/openstack/ironic-python-agent/+/852512 | 06:08 |
| opendevreview | Vanou Ishii proposed openstack/ironic master: Modify test code to avoid CONF modification affection https://review.opendev.org/c/openstack/ironic/+/852250 | 06:36 |
| kubajj | good morning ironic! | 06:57 |
| rpittau | good morning ironic! o/ | 08:17 |
| *** mat_fechner is now known as matfechner | 09:11 | |
| opendevreview | Merged openstack/ironic-inspector master: remove unicode literal from code https://review.opendev.org/c/openstack/ironic-inspector/+/852343 | 09:25 |
| frickler | iurygregory: any progress regarding the jsonschema update? I didn't spot any patch mentioning it | 09:30 |
| opendevreview | chenwei proposed openstack/ironic-python-agent master: Remove unicode literal from code https://review.opendev.org/c/openstack/ironic-python-agent/+/852512 | 10:33 |
| *** tosky is now known as Guest71 | 10:46 | |
| *** tosky_ is now known as tosky | 10:46 | |
| iurygregory | frickler, not yet, going to continue today, sorry o/ | 10:58 |
| dtantsur | Morning folks! I've lost scrollback since yesterday, please repeat any pings. | 10:59 |
| iurygregory | dtantsur, rpittau not sure if you had time to check https://review.opendev.org/c/openstack/ironic/+/852234 to see if we can agree on a path forward =) | 11:48 |
| *** tosky is now known as Guest77 | 12:01 | |
| *** tosky__ is now known as tosky | 12:01 | |
| rpittau | iurygregory: I'm reading the comments, again | 12:02 |
| rpittau | maybe we're really looking at resolving a problem from the wrong side | 12:02 |
| iurygregory | yeah, I understand we have other options (all of them have the good/bad ofc) | 12:08 |
| iurygregory | just thinking on what would be better.. | 12:08 |
| rpittau | ajya: hi! do you expect to remove the upper constraints for dracclient requirement during this cycle? | 12:12 |
| iurygregory | https://paste.opendev.org/show/bs6D42mTAAjaDh7AIuDW/ this are some of the ideas I had after discussion, maybe I'm missing other possible options also | 12:14 |
| ajya | hi rpittau, what's the context? Do you mean changing <9.0.0 to <10.0.0? | 12:19 |
| rpittau | ajya: yes, increasing the version in general | 12:20 |
| ajya | rpittau, good question. Generally, in previous cycles new major version was released and mapped to OpenStack cycle. This cycle looks to be the 1st one when there are no changes to dracclient and I don't know if it makes sense to bump the version to keep the pattern. | 12:24 |
| ajya | Have to think if it can cause issues later, though in longer term dracclient will lose its usage as wsman itself is deprecated in idrac and in future likely to be removed altogether. | 12:25 |
| rpittau | ajya: if there are no changes at all I think it makes sense to keep the current version, less confusion | 12:25 |
| opendevreview | Merged openstack/ironic-python-agent bugfix/8.1: Gather details about bond interfaces if present https://review.opendev.org/c/openstack/ironic-python-agent/+/848869 | 12:26 |
| ajya | rpittau, yup, and don't plan to work on features in wsman unless something extraordinary happens. Most likely it will be bugs if anything. | 12:27 |
| rpittau | ajya: perfect, thanks | 12:27 |
| *** tosky is now known as Guest78 | 12:30 | |
| *** tosky_ is now known as tosky | 12:30 | |
| dtantsur | Could we get a 2nd +2 on https://review.opendev.org/c/openstack/ironic/+/852393 please? It fixes a recent regression | 12:52 |
| *** tosky is now known as Guest85 | 13:13 | |
| *** tosky__ is now known as tosky | 13:13 | |
| opendevreview | Harald Jensås proposed openstack/networking-baremetal master: Device management driver iface https://review.opendev.org/c/openstack/networking-baremetal/+/835170 | 14:23 |
| opendevreview | Harald Jensås proposed openstack/networking-baremetal master: Add netconf-openconfig device driver https://review.opendev.org/c/openstack/networking-baremetal/+/835324 | 14:23 |
| opendevreview | Harald Jensås proposed openstack/networking-baremetal master: Add LACP support to Netconf OpenConfig driver https://review.opendev.org/c/openstack/networking-baremetal/+/835378 | 14:23 |
| opendevreview | Harald Jensås proposed openstack/networking-baremetal master: Add support for pre-configured link aggregates https://review.opendev.org/c/openstack/networking-baremetal/+/840533 | 14:23 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/yoga: Change molds option to appropriate class https://review.opendev.org/c/openstack/ironic/+/852506 | 14:24 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/xena: Change molds option to appropriate class https://review.opendev.org/c/openstack/ironic/+/852507 | 14:24 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/wallaby: Change molds option to appropriate class https://review.opendev.org/c/openstack/ironic/+/852508 | 14:24 |
| opendevreview | Merged openstack/ironic master: Update raid_type handling for Redfish raid_config https://review.opendev.org/c/openstack/ironic/+/852393 | 14:32 |
| opendevreview | Merged openstack/ironic master: Modify test code to avoid CONF modification affection https://review.opendev.org/c/openstack/ironic/+/852250 | 14:32 |
| opendevreview | Merged openstack/ironic-tempest-plugin master: remove unicode from code https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/852313 | 14:32 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/yoga: Modify test code to avoid CONF modification affection https://review.opendev.org/c/openstack/ironic/+/852509 | 14:55 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/yoga: Modify test code to avoid CONF modification affection https://review.opendev.org/c/openstack/ironic/+/852509 | 14:56 |
| iurygregory | JayF, just wondering if my comment was matching the idea you had in the review it was something along those lines? | 14:56 |
| JayF | Gonna be honest, I couldn't tell. | 14:56 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/yoga: Modify test code to avoid CONF modification affection https://review.opendev.org/c/openstack/ironic/+/852509 | 14:57 |
| JayF | The big problem I have with that PR is the language around v4 and v6. IMO we shouldn't have any language or logic to that effect at all. | 14:57 |
| JayF | The change should be about serving up URLs that work to node BMCs, and the contents should be (mostly) opaque to Ironic. | 14:57 |
| JayF | The way it's structured now, I think an ill-informed operator could walk away with the idea they need to configure both v4 and v6 addresses and that dual stack doesn't work. | 14:57 |
| TheJulia | so there is a delineation here I didn't pickup on, only for the BMC for virtual media | 15:10 |
| JayF | I'm not sure why that matters for my concern. | 15:25 |
| JayF | I literally had *this exact class of problem* at a previous job, and it was not along v4/v6 lines | 15:26 |
| JayF | we fixed it in a strange way (think emulated conductor groups with a duplicated driver) | 15:26 |
| JayF | but like, this is not a v4/v6 specific issue, and couching it in those terms is going to be super misleading | 15:26 |
| dtantsur | so, sorta node[driver_info][external_http_url]? | 15:27 |
| JayF | for the immediate problem: yes. | 15:28 |
| JayF | I wish this existed: node[driver_info][bmc_network_scope] = datacenter_17 (then somewhere in config: network_settings { datacenter_17 { external_http_url = banana } } | 15:28 |
| JayF | but that's obviously not well hashed out yet | 15:28 |
| JayF | but that type of construct would've been crazy useful | 15:29 |
| dtantsur | hmm | 15:29 |
| JayF | doing a single override in driver_info per node becomes borderline-impossible to update (because it's a json dict in a db) at extreme scale | 15:29 |
| dtantsur | yeah, and I have worries about the rate we add options there | 15:29 |
| dtantsur | does it make sense to stop at bmc_network_scope? | 15:30 |
| JayF | I would assume we could also set a network_scope on ports or something? | 15:30 |
| JayF | like I said, it's not a fully formed idea yet | 15:30 |
| dtantsur | I mean.. we could generalize it further | 15:30 |
| dtantsur | to just configuration scopes | 15:30 |
| JayF | oh. OH. | 15:30 |
| JayF | I think I like this. | 15:31 |
| JayF | Because you might need to change, e.g. default cleaning image across network boundaries too | 15:31 |
| dtantsur | or, say, the peering settings for Arne's network burn-in | 15:31 |
| JayF | I think this is a great example of the sort of thing where 6 years ago we might have tried to farm some of this out to e.g. neutron | 15:32 |
| JayF | adding this kind of functionality for a standalone ironic would leap it forward a lot in being able to operate in more complex environments I think | 15:32 |
| iurygregory | back from lunch | 15:44 |
| iurygregory | reading | 15:44 |
| iurygregory | dtantsur, can you elaborate a bit more in "configuration scopes" ? | 15:46 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Log successful clean up in image cache https://review.opendev.org/c/openstack/ironic/+/852603 | 16:00 |
| dtantsur | iurygregory: Jay's idea generalized to any configuration, not just networking | 16:02 |
| iurygregory | you mean this one "node[driver_info][bmc_network_scope] = datacenter_17 (then somewhere in config: network_settings { datacenter_17 { external_http_url = banana } }" correct? | 16:04 |
| dtantsur | yeah, this idea | 16:05 |
| iurygregory | but more generic it would be.. not really using "bmc_network_scope" | 16:05 |
| iurygregory | the config we are talking would be in ironic.conf? | 16:06 |
| rpittau | see ya tomorrow! o/ | 16:13 |
| iurygregory | JayF, the configs would be stored in ironic.conf? trying to understand this to see if I can push an initial patch for it | 16:34 |
| JayF | iurygregory: yeah. Honestly syntax for the config is going to be what's going to be toughest. | 16:35 |
| JayF | because of how the settings are grouped in ironic.conf | 16:35 |
| JayF | I admit my idea is not completely fully formed | 16:35 |
| TheJulia | I do like the idea of the config file, primarily from the standpoint of one of the big things we get hit on is DB activity, so anything that is a separate data set/structure which needs to be accessed is going to come at a higher cost if it is not already on the node object or joined in on the base query and still populated on a full object | 16:39 |
| TheJulia | And I've recently had a couple operators ask for like global config file overrides for things unrelated to deploy/rescue ramdisks/kernels | 16:39 |
| iurygregory | I'm trying to imagine how the config should look like... | 16:48 |
| iurygregory | something like we have for our drivers? the value from the new field in driver_info would represent a section and we would add variables there that we want to override ... https://paste.opendev.org/show/bZKU0hw05Ux0Ncl1myJ4/ ? | 16:50 |
| iurygregory | what would happen if we don't find the section etc | 16:51 |
| TheJulia | likely log an error and ignore the failure and move on | 16:51 |
| TheJulia | well, we would log the error on the node events | 16:52 |
| JayF | iurygregory: it'd probably need to be `conductor/external_http_url=banana` if you used that format | 16:58 |
| JayF | iurygregory: or you could do [conductor:datacenter_17] as the header | 16:58 |
| JayF | in any event, you have to account for *the original group* that setting would be in, right? | 16:58 |
| TheJulia | I'm not entirely sure : will be grokked in oslo_cfg | 16:59 |
| JayF | s/:/[some delimiter here]// | 16:59 |
| TheJulia | but it would be more clear if it did | 17:00 |
| iurygregory | JayF, the config is in [deploy] not [conductor] | 17:00 |
| JayF | shows how long it's been since I hand-edited an ironic.conf :| | 17:00 |
| iurygregory | =) | 17:00 |
| JayF | the idea is still valid tho | 17:00 |
| * TheJulia makes bread because ... broken fridge | 17:01 | |
| JayF | I'll also say: idk how urgent your problem is iurygregory, but I'd be OK with the [external_http_url] in driver_info to get you or your customer unbroken, then we can spec out the more general solution | 17:01 |
| JayF | but I don't have any sense of if there's urgency to fix that or something already worked around | 17:01 |
| iurygregory | JayF, quite urgent when looking at the downstream deadlines... | 17:02 |
| iurygregory | JayF, gotcha! | 17:02 |
| JayF | I can't speak for other cores, but I can tell you my insistent -1 was simply around the v4/v6 nomenclature and documentation | 17:02 |
| iurygregory | yeah, adding the field in driver_info would workaround the issue for the customer | 17:03 |
| JayF | but that being said, config namespaces is a kick ass idea and we shouldn't let it drop | 17:03 |
| TheJulia | so.. harald has done some config namespace stuffs recently | 17:03 |
| iurygregory | JayF, totally agree ++ | 17:03 |
| TheJulia | in his networking-baremetal stuffs | 17:03 |
| TheJulia | in part, specifically because he needed to support specific switch level configs | 17:04 |
| JayF | the easiest way to know something is a good idea is for someone to have alraedy made sure it's a good idea by doing it first ;) | 17:04 |
| TheJulia | based upon the driver loaded with driver dynamic options... *blarg* | 17:04 |
| iurygregory | yeah, the enabled_devices and the section for each one available | 17:05 |
| iurygregory | I think this is what your are talking about https://review.opendev.org/c/openstack/networking-baremetal/+/835324/20/networking_baremetal/config.py | 17:05 |
| iurygregory | TheJulia, dtantsur thoughts on this workaround so we don't rush with an implementation for the namespaces ? | 17:13 |
| * dtantsur reads | 17:14 | |
| TheJulia | iurygregory: yeah | 17:14 |
| dtantsur | iurygregory: I guess it's fine | 17:15 |
| iurygregory | ok, I'll proceed with it | 17:16 |
| opendevreview | Julia Kreger proposed openstack/ironic master: Redfish: Consider password part of the session cache https://review.opendev.org/c/openstack/ironic/+/850259 | 17:32 |
| *** cstone3 is now known as cstone | 17:46 | |
| hjensas | hm, config_namespaces, make sure [conductor] is loaded before [conductor:namespace], register the same opts and use [conductor]/_opts_ for default values in [conductor:namespace]/_opts_? | 18:00 |
| JayF | That sounds roughly like what we're kicking around | 18:03 |
| JayF | but it's a very rough idea at this point | 18:03 |
| hjensas | It sure seems like it would add flexibility. | 18:06 |
| TheJulia | it would need to know namespaces... | 18:10 |
| TheJulia | somehow.... | 18:10 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/xena: Modify test code to avoid CONF modification affection https://review.opendev.org/c/openstack/ironic/+/852656 | 18:19 |
| *** lifeless_ is now known as lifeless | 18:24 | |
| hjensas | may use [conductor:default] as a special ns that always exist, CONF[task.config_ns].opt, if no namespace task.config_ns == 'default'? Or would it be possible to copy config values to the task once the ns is known, and then do 'task.conf_ns.group.opt' when getting the value? (IDK, just ideas...) | 18:29 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/wallaby: Modify test code to avoid CONF modification affection https://review.opendev.org/c/openstack/ironic/+/852657 | 18:39 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/wallaby: Modify test code to avoid CONF modification affection https://review.opendev.org/c/openstack/ironic/+/852657 | 18:40 |
| opendevreview | Harald Jensås proposed openstack/networking-baremetal master: Add netconf-openconfig device driver https://review.opendev.org/c/openstack/networking-baremetal/+/835324 | 18:47 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/victoria: Modify test code to avoid CONF modification affection https://review.opendev.org/c/openstack/ironic/+/852662 | 19:00 |
| JayF | hjensas: we'd have to fully support configs where [conductor] was set (to imply conductor:default). In fact, I'd strongly suggest we would avoid making people migrate that, ever -- we only want to expose this complexity to people who need it (if possible) | 19:58 |
| hjensas | JayF: I was thinking there would be an implicit copy of [conductor] to [conductor:default], i.e in general people would just use [conductor] and not be aware of the internal default thing. But yeah, may not be a good idea. | 20:18 |
| ashinclouds[m] | It might be a thing to do… at some point… if there was a helper… but my immediate thought was any migration, explicit or not, might not be a good idea | 20:22 |
| ashinclouds[m] | Just thinking out loud while waiting for noms | 20:22 |
| JayF | the difficulty is keeping ironic simple enough for the small case but complex enough, if you understand how, to support the scaled cases | 20:24 |
| JayF | if people start see conductor:default and conductor:lolcats in our sample config file, that's going to imply a complexity for the general case that just isn't there | 20:24 |
| JayF | but IMBW; maybe I'm the only op who reads through *.conf.sample :D | 20:24 |
| ashinclouds[m] | Some do… some don’t actually read the entries | 20:25 |
| ashinclouds[m] | It is an epic mixed bag | 20:25 |
| ashinclouds[m] | And a lot of operators are not fans of config files anymore | 20:26 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/ussuri: Modify test code to avoid CONF modification affection https://review.opendev.org/c/openstack/ironic/+/852676 | 21:38 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/train: Modify test code to avoid CONF modification affection https://review.opendev.org/c/openstack/ironic/+/852678 | 21:43 |
| TheJulia | iurygregory: ^ | 21:46 |
| iurygregory | TheJulia, going to check after dinner o/ | 22:02 |
| *** undefined is now known as Guest145 | 22:02 | |
| *** Guest145 is now known as rcastillo | 22:03 | |
| TheJulia | Ack, going to lay down for a little while, not feeling great | 22:03 |
| iurygregory | TheJulia, take care! | 22:04 |
| JayF | ashinclouds[m] / TheJulia - did I hear you just ask me to build an etcd backend for oslo.config? ;) | 23:49 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!