| *** trandles has quit IRC | 00:04 | |
| *** janders has quit IRC | 00:17 | |
| *** baha has quit IRC | 00:17 | |
| *** yoctozepto has quit IRC | 00:17 | |
| *** eandersson has quit IRC | 00:17 | |
| *** janders has joined #openstack-ironic | 00:17 | |
| *** eandersson has joined #openstack-ironic | 00:17 | |
| *** baha has joined #openstack-ironic | 00:17 | |
| *** yoctozepto has joined #openstack-ironic | 00:17 | |
| *** ociuhandu has joined #openstack-ironic | 01:11 | |
| *** ociuhandu has quit IRC | 01:15 | |
| *** rh-jlabarre has joined #openstack-ironic | 01:27 | |
| *** paras333 has joined #openstack-ironic | 01:58 | |
| *** paras333 has quit IRC | 02:02 | |
| *** livelace has joined #openstack-ironic | 02:26 | |
| *** rh-jlabarre has quit IRC | 02:53 | |
| *** ricolin has joined #openstack-ironic | 03:12 | |
| *** ricolin has quit IRC | 03:12 | |
| *** ricolin has joined #openstack-ironic | 03:14 | |
| *** livelace has quit IRC | 03:16 | |
| *** akahat is now known as akahat|ruck | 04:26 | |
| *** paras333_ has joined #openstack-ironic | 04:45 | |
| *** paras333_ has quit IRC | 04:51 | |
| *** early has quit IRC | 05:25 | |
| *** early has joined #openstack-ironic | 05:28 | |
| *** logan- has quit IRC | 06:12 | |
| *** logan- has joined #openstack-ironic | 06:15 | |
| timeu | window page_down | 06:23 |
|---|---|---|
| *** k_mouza has joined #openstack-ironic | 06:26 | |
| *** k_mouza has quit IRC | 06:30 | |
| iurygregory | good morning Ironic | 06:37 |
| *** mkowalski has quit IRC | 06:43 | |
| *** mkowalski has joined #openstack-ironic | 06:43 | |
| *** MahnoorAsghar_ has joined #openstack-ironic | 06:45 | |
| arne_wiebalck | Good morning, iurygregory and ironic! | 06:54 |
| iurygregory | arne_wiebalck, o/ | 06:56 |
| * iurygregory brb need to restart the notebook | 07:00 | |
| *** jawad_axd has joined #openstack-ironic | 07:05 | |
| *** iurygregory has quit IRC | 07:12 | |
| *** paras333 has joined #openstack-ironic | 07:20 | |
| *** iurygregory has joined #openstack-ironic | 07:21 | |
| *** paras333 has quit IRC | 07:25 | |
| *** tosky has joined #openstack-ironic | 07:32 | |
| *** lucasagomes has joined #openstack-ironic | 07:58 | |
| *** dtantsur|afk is now known as dtantsur | 08:09 | |
| dtantsur | morning ironic | 08:10 |
| *** ociuhandu has joined #openstack-ironic | 08:13 | |
| iurygregory | morning dtantsur | 08:16 |
| *** ociuhandu has quit IRC | 08:19 | |
| *** ociuhandu has joined #openstack-ironic | 08:28 | |
| janders | hey iurygregory arne_wiebalck dtantsur o/ | 08:40 |
| arne_wiebalck | good morning janders dtantsur o/ | 08:41 |
| iurygregory | hey janders o/ | 08:42 |
| *** MahnoorAsghar_ has quit IRC | 08:47 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: [WIP] Stop accepting duplicated configdrive https://review.opendev.org/c/openstack/ironic-python-agent/+/790471 | 08:57 |
| *** k_mouza has joined #openstack-ironic | 09:02 | |
| *** derekh has joined #openstack-ironic | 09:24 | |
| *** alexmcleod_ has quit IRC | 09:27 | |
| *** alexmcleod has joined #openstack-ironic | 09:27 | |
| *** dougsz has joined #openstack-ironic | 09:39 | |
| *** ociuhandu has quit IRC | 09:49 | |
| *** MahnoorAsghar__ has joined #openstack-ironic | 09:52 | |
| *** ociuhandu has joined #openstack-ironic | 10:15 | |
| *** ociuhandu has quit IRC | 10:15 | |
| *** ociuhandu has joined #openstack-ironic | 10:15 | |
| *** ociuhandu has quit IRC | 10:40 | |
| *** ociuhandu has joined #openstack-ironic | 10:43 | |
| *** ociuhandu has quit IRC | 10:44 | |
| *** ociuhandu has joined #openstack-ironic | 10:44 | |
| *** ociuhandu has quit IRC | 10:44 | |
| *** ociuhandu has joined #openstack-ironic | 10:46 | |
| *** ociuhandu has quit IRC | 10:50 | |
| *** ociuhandu has joined #openstack-ironic | 10:51 | |
| *** ociuhandu has quit IRC | 11:04 | |
| *** ociuhandu has joined #openstack-ironic | 11:14 | |
| openstackgerrit | Bob Fournier proposed openstack/sushy-tools master: Add support for returning message and attribute registries https://review.opendev.org/c/openstack/sushy-tools/+/790361 | 11:21 |
| *** ociuhandu has quit IRC | 11:23 | |
| *** paras333_ has joined #openstack-ironic | 11:24 | |
| *** MahnoorAsghar__ has quit IRC | 11:25 | |
| openstackgerrit | Merged openstack/ironic-lib master: Remove runtime dependency on pbr https://review.opendev.org/c/openstack/ironic-lib/+/787922 | 11:27 |
| *** paras333_ has quit IRC | 11:29 | |
| timeu | window page_down | 11:33 |
| *** timeu has quit IRC | 11:33 | |
| *** timeu has joined #openstack-ironic | 11:33 | |
| *** ociuhandu has joined #openstack-ironic | 11:35 | |
| openstackgerrit | Merged openstack/ironic master: Add bios_interface to api-ref spec https://review.opendev.org/c/openstack/ironic/+/791493 | 11:38 |
| openstackgerrit | Merged openstack/ironic-lib master: setup.cfg: Replace dashes with underscores https://review.opendev.org/c/openstack/ironic-lib/+/791619 | 11:39 |
| *** ociuhandu has quit IRC | 11:44 | |
| *** yoctozepto has quit IRC | 11:55 | |
| *** rh-jlabarre has joined #openstack-ironic | 12:14 | |
| iurygregory | I've updated the Agenda for the meeting today (removed some items from last meeting) | 12:16 |
| iurygregory | if anyone has topics feel free to add (the agenda is almost empty) | 12:16 |
| *** ociuhandu has joined #openstack-ironic | 12:18 | |
| *** tzumainn has joined #openstack-ironic | 12:19 | |
| janders | see you tomorrow Ironic o/ | 12:21 |
| *** ociuhandu has quit IRC | 12:23 | |
| ajya | For the meeting I added a bug for discussion | 12:24 |
| *** ociuhandu has joined #openstack-ironic | 12:27 | |
| iurygregory | ajya, ack =) | 12:29 |
| iurygregory | bye janders o/ | 12:29 |
| *** yoctozepto has joined #openstack-ironic | 12:31 | |
| *** ociuhandu has quit IRC | 12:31 | |
| *** ociuhandu has joined #openstack-ironic | 12:33 | |
| TheJulia | good morning | 12:54 |
| iurygregory | good morning TheJulia | 13:05 |
| dtantsur | morning TheJulia! how is it going? | 13:07 |
| TheJulia | It is going, just starting my workday | 13:07 |
| * dtantsur sighs and looks at https://review.opendev.org/c/openstack/ironic/+/791035 again | 13:08 | |
| *** ociuhandu has quit IRC | 13:09 | |
| * TheJulia tries to wake up the work laptop | 13:10 | |
| * dtantsur uses "horrible mess" in a bug description, which is not a good sign | 13:10 | |
| TheJulia | "organic horrible mess" | 13:10 |
| dtantsur | thank you, will fix | 13:10 |
| TheJulia | where all well everything was well intentioned | 13:11 |
| TheJulia | "Hi Laptop, Wake up!" | 13:11 |
| iurygregory | you can also use wormhole | 13:11 |
| dtantsur | yep, nothing of this has been created maliciously | 13:11 |
| iurygregory | TheJulia, maybe the corgi or the cat can help :D | 13:11 |
| dtantsur | just an incoherent growth of our JSON fields over the course of years | 13:12 |
| iurygregory | ouch | 13:12 |
| iurygregory | let's switch to yaml :D | 13:12 |
| TheJulia | oh jeeze | 13:12 |
| dtantsur | this is re kernel_append_params, I showed you on Wed | 13:12 |
| *** rloo has joined #openstack-ironic | 13:12 | |
| * dtantsur runs away screaming | 13:12 | |
| iurygregory | kidding :D | 13:12 |
| dtantsur | well, with the invention of ansible and k8s, yaml is officially a programming language, soo.. | 13:12 |
| *** ociuhandu has joined #openstack-ironic | 13:13 | |
| iurygregory | makes sense | 13:14 |
| openstackgerrit | Pierre Riteau proposed openstack/tenks master: Update pip and setuptools before installing requirements https://review.opendev.org/c/openstack/tenks/+/791751 | 13:15 |
| dtantsur | okay, collected my thoughts again https://storyboard.openstack.org/#!/story/2008902 | 13:16 |
| *** ociuhandu has quit IRC | 13:19 | |
| *** paras333_ has joined #openstack-ironic | 13:23 | |
| *** jamesdenton has joined #openstack-ironic | 13:27 | |
| *** jamesden_ has quit IRC | 13:30 | |
| *** ociuhandu has joined #openstack-ironic | 13:31 | |
| *** ociuhandu has quit IRC | 13:38 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Clean up kernel_append_params for PXE/iPXE https://review.opendev.org/c/openstack/ironic/+/791755 | 13:39 |
| *** ociuhandu has joined #openstack-ironic | 13:39 | |
| *** sdanni has joined #openstack-ironic | 13:43 | |
| TheJulia | we need to clean up credentials | 13:47 |
| dtantsur | yep, and credentials too | 13:49 |
| TheJulia | do we have things to review jam upon today? | 13:50 |
| dtantsur | maybe my user guide changes/additions.. but it may require a thorough reading more than a discussion, so not sure | 13:51 |
| TheJulia | I wouldn't even know what those changes are yet | 13:52 |
| TheJulia | Literally still getting my bearings after being out for a week | 13:52 |
| dtantsur | heh, then probably skipping it is fine | 13:52 |
| * dtantsur is very sleepy | 13:52 | |
| openstackgerrit | Pierre Riteau proposed openstack/tenks master: Update pip and setuptools before installing requirements https://review.opendev.org/c/openstack/tenks/+/791751 | 13:53 |
| *** lmcgann has joined #openstack-ironic | 14:00 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Clean up kernel_append_params for PXE/iPXE https://review.opendev.org/c/openstack/ironic/+/791755 | 14:04 |
| *** ociuhandu has quit IRC | 14:05 | |
| *** ociuhandu has joined #openstack-ironic | 14:07 | |
| *** jamesden_ has joined #openstack-ironic | 14:10 | |
| iurygregory | upstream meeting in 50min | 14:10 |
| dtantsur | iurygregory: who's chairing today in the end, you or me? | 14:11 |
| * dtantsur doesn't care much | 14:11 | |
| iurygregory | dtantsur, I assumed it was me =) | 14:11 |
| dtantsur | cool :) | 14:11 |
| *** ociuhandu has quit IRC | 14:12 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Clean up kernel_append_params for PXE/iPXE https://review.opendev.org/c/openstack/ironic/+/791755 | 14:13 |
| *** jamesdenton has quit IRC | 14:14 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Clean up kernel_append_params for Redfish and iLO https://review.opendev.org/c/openstack/ironic/+/791761 | 14:22 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Clean up kernel_append_params for Redfish and iLO https://review.opendev.org/c/openstack/ironic/+/791761 | 14:23 |
| *** ociuhandu has joined #openstack-ironic | 14:36 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Clean up kernel_append_params for iRMC https://review.opendev.org/c/openstack/ironic/+/791035 | 14:38 |
| *** ociuhandu has quit IRC | 14:41 | |
| TheJulia | I could also chair, I'm just wrapping up some of the emails I needed to reply to this morning | 14:42 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Clean up kernel_append_params for Redfish and iLO https://review.opendev.org/c/openstack/ironic/+/791761 | 14:43 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Clean up kernel_append_params for iRMC https://review.opendev.org/c/openstack/ironic/+/791035 | 14:44 |
| *** Qianbiao has joined #openstack-ironic | 14:45 | |
| openstackgerrit | Pierre Riteau proposed openstack/tenks master: Update pip and setuptools before installing requirements https://review.opendev.org/c/openstack/tenks/+/791751 | 14:58 |
| *** MahnoorAsghar has joined #openstack-ironic | 14:59 | |
| iurygregory | #startmeeting ironic | 15:00 |
| openstack | Meeting started Mon May 17 15:00:01 2021 UTC and is due to finish in 60 minutes. The chair is iurygregory. Information about MeetBot at http://wiki.debian.org/MeetBot. | 15:00 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 15:00 |
| *** openstack changes topic to " (Meeting topic: ironic)" | 15:00 | |
| openstack | The meeting name has been set to 'ironic' | 15:00 |
| iurygregory | o/ | 15:00 |
| dtantsur | o/ | 15:00 |
| iurygregory | Hello everyone, welcome to our weekly meeting, you can find our agenda in the wiki | 15:00 |
| iurygregory | #link https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_next_meeting | 15:00 |
| iurygregory | #topic Announcements/Reminder | 15:00 |
| *** openstack changes topic to "Announcements/Reminder (Meeting topic: ironic)" | 15:00 | |
| TheJulia | o/ | 15:01 |
| bfournie | o/ | 15:01 |
| rpioso | \o | 15:01 |
| iurygregory | Does anyone have anything to announce / reminder us? | 15:01 |
| rloo | o/ | 15:01 |
| dtantsur | I think we're slowly approaching the sprint 1 releases | 15:02 |
| dtantsur | in 2 weeks | 15:02 |
| TheJulia | we are I believe | 15:02 |
| *** stendulker has joined #openstack-ironic | 15:02 | |
| stendulker | o/ | 15:02 |
| iurygregory | #info sprint 1 release in 2 weeks | 15:02 |
| ajya | o/ | 15:02 |
| TheJulia | I need to try and crank out performance related patches since it will be a major version bump with the removal of iscsi | 15:02 |
| erbarr | o/ | 15:02 |
| iurygregory | yeah that makes sense | 15:03 |
| iurygregory | should we move for our next topic? | 15:03 |
| TheJulia | ++ | 15:04 |
| iurygregory | #topic Review action items from previous meeting | 15:04 |
| *** openstack changes topic to "Review action items from previous meeting (Meeting topic: ironic)" | 15:04 | |
| iurygregory | I was looking at the logs, we don't have any action item, moving on | 15:04 |
| iurygregory | #topic #Review subteam status reports | 15:04 |
| *** openstack changes topic to "#Review subteam status reports (Meeting topic: ironic)" | 15:04 | |
| iurygregory | Xena themes is not merged yet https://review.opendev.org/c/openstack/ironic-specs/+/784143 | 15:05 |
| dtantsur | very close actually, we may even apply lazy consensus | 15:05 |
| dtantsur | (I think we wanted it on the last meeting?) | 15:05 |
| iurygregory | according to last week meeting, the concensus was that we would merge by EOW | 15:05 |
| iurygregory | dtantsur, yeah | 15:05 |
| dtantsur | SHIP IT! | 15:05 |
| rloo | ++ | 15:06 |
| iurygregory | Done! | 15:06 |
| iurygregory | I will update the ironic board with the information, next week we can have the status, sounds good? | 15:06 |
| dtantsur | yep | 15:06 |
| rloo | thx iurygregory! | 15:06 |
| iurygregory | #topic Deciding on priorities for the coming week | 15:07 |
| *** openstack changes topic to "Deciding on priorities for the coming week (Meeting topic: ironic)" | 15:07 | |
| iurygregory | Anyone has patches that would like to be added to the priorities? | 15:07 |
| TheJulia | thanks iurygregory | 15:08 |
| dtantsur | I have a bunch of new things related to driver refactoring (kernel_append_params this time): https://review.opendev.org/c/openstack/ironic/+/791755 https://review.opendev.org/c/openstack/ironic/+/791761 https://review.opendev.org/c/openstack/ironic/+/791035 | 15:08 |
| dtantsur | any objections to adding them? they're not large, just.. annoying? | 15:08 |
| iurygregory | dtantsur, ++ to me it makes sense to have them | 15:08 |
| TheJulia | no objections from me | 15:09 |
| TheJulia | I'd like to get the db indexes added to the priority list for the week https://review.opendev.org/c/openstack/ironic/+/788625 | 15:09 |
| * dtantsur should probably stop generating more patches now that 1/3 of the weekly priority are his :D | 15:10 | |
| dtantsur | ++ to indexes | 15:10 |
| TheJulia | and my patch to cleanup the node santiziation for performance | 15:10 |
| TheJulia | https://review.opendev.org/c/openstack/ironic/+/790142 however, it will also need another patch to cleanup objects on selected columns | 15:10 |
| iurygregory | TheJulia, I think it's ok, the failures seems to be with tests and cover | 15:10 |
| TheJulia | so it is blocked until I get that patch created/uploaded | 15:11 |
| TheJulia | iurygregory: yeah, I need to cycle back to them later today, I was MIA all last week | 15:11 |
| iurygregory | TheJulia, no worries =) | 15:11 |
| bfournie | not sure if its a priority, but would like to get some feedback on new fields in bios api - https://review.opendev.org/c/openstack/ironic/+/786707 | 15:11 |
| iurygregory | bfournie, can be a priority without problems I would say | 15:12 |
| dtantsur | yep, the only problem is its size :) | 15:12 |
| bfournie | thanks iurygregory | 15:12 |
| iurygregory | dtantsur, yeah =) | 15:12 |
| bfournie | yeah, size | 15:12 |
| dtantsur | you could avoid a dependency on sushy (which blocks this patch until a release) by splitting it | 15:13 |
| dtantsur | into the generic parts and the redfish parts | 15:13 |
| iurygregory | humm I think we can release sushy | 15:13 |
| iurygregory | we already have the support merged | 15:13 |
| dtantsur | ah, have we? | 15:13 |
| iurygregory | I can push today a patch in releases | 15:13 |
| dtantsur | good, then let's release and replace depends-on with a bump of driver-requirements | 15:13 |
| bfournie | yeah the sushy patch merged | 15:13 |
| bfournie | k, cool | 15:14 |
| bfournie | thanks | 15:14 |
| iurygregory | np! | 15:14 |
| iurygregory | anyone have more patches? | 15:15 |
| iurygregory | or should we move on? | 15:15 |
| iurygregory | moving on =) | 15:16 |
| iurygregory | #topic Discussion | 15:17 |
| *** openstack changes topic to "Discussion (Meeting topic: ironic)" | 15:17 | |
| iurygregory | we have one topic added by ajya - Failed to inspect node created with inspect-interface as idrac-redfish | 15:17 |
| iurygregory | #link https://storyboard.openstack.org/#!/story/2008901 | 15:17 |
| ajya | mysql has changed how it reports duplicate key error by adding table name to key name. This does not work well with oslo.db. There is bug reported in oslo.db, but not fixed yet | 15:18 |
| ajya | I'm thinking it should be fixed there, but if not possible, add workaround in Ironic. Thoughts? Comments? | 15:18 |
| dtantsur | we probably rely on this property in a lot of places... | 15:18 |
| rpioso | dtantsur: yep | 15:18 |
| dtantsur | like, name duplicates, MAC duplicates everywhere.. these all probably raise HTTP 500 now? | 15:19 |
| dtantsur | do we have any engagement from oslo.db folks already? | 15:19 |
| ajya | not yet, they're having their weekly meeting now | 15:19 |
| ajya | I'll raise is there, no response earlier today when I asked | 15:19 |
| openstackgerrit | Merged openstack/ironic-specs master: Xena themes https://review.opendev.org/c/openstack/ironic-specs/+/784143 | 15:19 |
| iurygregory | doesn't seem according to the | 15:19 |
| iurygregory | https://bugs.launchpad.net/oslo.db/+bug/1896916 | 15:19 |
| openstack | Launchpad bug 1896916 in oslo.db "duplicate key error for mysql >=8.0.19 are not parsed correctly" [Undecided,New] | 15:19 |
| rloo | I'd be surprised if this is only an issue with ironic | 15:19 |
| TheJulia | Definitely needs to be fixed in oslo.db | 15:20 |
| dtantsur | that being said, I'm surprised we're trying to create ports twice in redfish-idrac inspection | 15:20 |
| ajya | there is another project that had failing CI and they applied workaround | 15:20 |
| dtantsur | sounds like we shouldn't do it on the iDRAC side? | 15:20 |
| ajya | there is some logic in iDRAC side why it's done | 15:20 |
| iurygregory | I don't see problem in our CI atm, Dell CI is reporting this? | 15:20 |
| rpioso | iurygregory: No coverage | 15:21 |
| ajya | no problems in CI, I meant this issue - https://bugs.launchpad.net/tacker/+bug/1896867 | 15:21 |
| openstack | Launchpad bug 1896867 in tacker "functional test failures due to unsuccessfully parsed duplicate key error" [Undecided,Fix released] - Assigned to Koichiro Den (koichiroden) | 15:21 |
| dtantsur | ajya: if you insist on creating all ports (which I think may be wrong, but okay), you should overwrite _create_ports instead | 15:21 |
| ajya | and yes, there is no coverage for our case | 15:21 |
| iurygregory | people can workaround using other release from mysql also, I'm wondering what is the mysql version we have in requirements | 15:22 |
| dtantsur | trying to add ports twice is at least sub-optimal | 15:22 |
| dtantsur | I think this particular issue can and should be fixed on our side | 15:22 |
| rpioso | dtantsur: This bug could affect other uses of inspect_utils.create_ports_if_not_exist | 15:22 |
| TheJulia | But it also does follow the "fail and fallback model" as opposed to check first reconcile, then create model | 15:22 |
| rpioso | TheJulia: +1 | 15:23 |
| dtantsur | rpioso: surely. but it doesn't mean that the iDRAC implementation shouldn't be fixed. | 15:23 |
| rpioso | dtantsur: Not the topic at hand. | 15:23 |
| dtantsur | mmm? | 15:23 |
| rpioso | As rloo pointed out, other projects have encountered this, including tacker. | 15:23 |
| dtantsur | the bug is triggered by calling create_ports_if_not_exist twice. It should not happen. | 15:23 |
| TheJulia | We should have consistency in duplicate key errors because the code path all the way to API consumers doing things is impacted | 15:23 |
| dtantsur | I agree that it has to be fixed in oslo.db, but idrac-redfish is also wrong IMO. | 15:24 |
| dtantsur | and the fix is very trivial btw, can be landed much faster than oslo.db | 15:24 |
| TheJulia | Lets delineate the issues, they are two separate things and we could fall into a disagreement of python programming models easily with the driver and the reliance upon the failure and then fallback. | 15:24 |
| TheJulia | in other words, we could bikeshed the driver easily | 15:24 |
| rpioso | TheJulia: ty | 15:24 |
| dtantsur | we're not touching these topics | 15:24 |
| rloo | is this for a new feature in idrac? | 15:24 |
| TheJulia | rloo: no | 15:25 |
| dtantsur | we're discussing a very trivial issue: calling the same function twice for no reason | 15:25 |
| rpioso | rloo: No, it's a regression since Focal was introduced. | 15:25 |
| dtantsur | which has been fine all the time, but started triggering the mysql regression now | 15:25 |
| TheJulia | indeed | 15:25 |
| rloo | if you want to backport it, i suspect it'll have to work with older oslo.db. unless oslo.db's fix is backported too. | 15:25 |
| dtantsur | it = ???? | 15:25 |
| rloo | oh, it is already failing. | 15:25 |
| rpioso | dtantsur: All calls to inspect_utils.create_port_if_not_exist could be affected. | 15:26 |
| rloo | if the code fix is easy as dtantsur sez, would be worth fixing in both places. | 15:26 |
| dtantsur | rpioso: I don't argue with that. Nor does it cancel anything I've said. | 15:26 |
| rloo | or wait for oslo.db to fix. | 15:26 |
| dtantsur | unless I'm missing something. | 15:26 |
| iurygregory | we should probably talk with oslo folks to see about it | 15:27 |
| rloo | dtantsur: that idrac code has merged; guessing you are suggesting an improvement here. | 15:27 |
| rpioso | dtantsur: https://github.com/openstack/ironic/blob/af94a3da1e3f66c70309bbb889c68dfc5bd67e9f/ironic/drivers/modules/inspect_utils.py#L49 | 15:27 |
| rpioso | dtantsur: The exception is not caught. | 15:27 |
| *** ociuhandu has joined #openstack-ironic | 15:28 | |
| rloo | So I think we're done? Someone needs to mention to oslo.db that this affects ironic, it'll be up to them to decide priority etc. and if folks choose to enhance idrac code, that is their choice? | 15:28 |
| ajya | Talking with oslo now about this issue | 15:28 |
| rpioso | rloo: Sounds workable to me :-) | 15:28 |
| dtantsur | I don't think that fixing bugs is a choice | 15:28 |
| * dtantsur just makes a patch | 15:28 | |
| rpioso | dtantsur: We'll handle it :-) | 15:29 |
| dtantsur | Great, thanks! | 15:29 |
| rpioso | dtantsur: yw | 15:29 |
| dtantsur | It boils down to implementing _create_ports instead of inspect_hardware | 15:29 |
| *** ociuhandu has quit IRC | 15:30 | |
| *** ociuhandu has joined #openstack-ironic | 15:30 | |
| iurygregory | should we move to our next topic? | 15:30 |
| dtantsur | likely yes | 15:31 |
| iurygregory | #topic Baremetal SIG | 15:31 |
| *** openstack changes topic to "Baremetal SIG (Meeting topic: ironic)" | 15:31 | |
| ajya | ok, np, so for oslo we will propose a patch as noone else plans to do it, but they seem ok fixing it | 15:31 |
| iurygregory | arne_wiebalck, do you have anything for the Baremetal SIG? | 15:31 |
| arne_wiebalck | Record attendance at last week's meeting due to rpioso making publicity, thanks again! | 15:31 |
| iurygregory | nice! | 15:31 |
| iurygregory | tks rpioso =) | 15:32 |
| arne_wiebalck | Video is already up, thanks stevebaker! | 15:32 |
| arne_wiebalck | That's it, I think. | 15:32 |
| iurygregory | \o/ | 15:32 |
| iurygregory | cool! | 15:32 |
| iurygregory | moving on | 15:32 |
| iurygregory | #topic RFE review | 15:32 |
| *** openstack changes topic to "RFE review (Meeting topic: ironic)" | 15:32 | |
| rpioso | arne_wiebalck: Thank you so much for sharing CERN's experience with ironic. Very informative and insightful! | 15:32 |
| arne_wiebalck | rpioso: ty :) | 15:32 |
| iurygregory | MahnoorAsghar, you added https://storyboard.openstack.org/#!/story/2008866 | 15:33 |
| MahnoorAsghar | iurygregory: Yes; looking for an approval for the RFE | 15:33 |
| MahnoorAsghar | Updated it now | 15:33 |
| dtantsur | I don't my concerns about it have been addressed | 15:34 |
| dtantsur | I've put a comment re what we previously discussed | 15:34 |
| MahnoorAsghar | Last we discussed this, a vendor passthru implementation was considered favorable | 15:34 |
| iurygregory | yes | 15:34 |
| dtantsur | yep, just keep in mind that I'll object to including it in metal3 | 15:35 |
| TheJulia | arne_wiebalck: out of curiosity, was the video linked to from the ironicbaremetal.org blog? | 15:35 |
| dtantsur | so if you're doing it because of metal3, you may end up in a weird position | 15:35 |
| arne_wiebalck | TheJulia: the blog post with the same topic was | 15:35 |
| MahnoorAsghar | dtantsur: Ive changed it to vendor passthru now...let me see your comment | 15:35 |
| rpioso | Does Redfish RAID work with vendors other than Dell Technologies? | 15:36 |
| arne_wiebalck | TheJulia: maybe could amend that link to mention the video as well? | 15:36 |
| *** stendulker has quit IRC | 15:36 | |
| dtantsur | rpioso: I'm not sure if anyone has tried | 15:36 |
| rpioso | Does the ironic RAID schema support controller and disk IDs? | 15:37 |
| TheJulia | arne_wiebalck: should enitrely be possible | 15:37 |
| MahnoorAsghar | rpioso: It does, as per the documentation | 15:37 |
| TheJulia | arne_wiebalck: it is just markdown in a git repo :) | 15:37 |
| rpioso | Couldn't a default implementation of the originally proposed API addition return empty or a status that indicates it's not supported? | 15:38 |
| *** MahnoorAsghar has quit IRC | 15:39 | |
| dtantsur | shouldn't maybe the driver implementation support simple indexes instead of moving to metal3? | 15:40 |
| TheJulia | rpioso: each raid interface has slightly different requirements and I don't believe the ilo interface has as strict requirements for names/data like the idrac driver does, but it has been a couple years since I last looked at that code/docs so it is a bit difficult to speak to at the moment. | 15:40 |
| dtantsur | if e.g. the iDRAC impelementation allows numbers in addition to string names, it will solve the metal3 problem without most of complications | 15:41 |
| rpioso | dtantsur: The driver IDs indicate where the devices are located in the system. It depends on the system design. | 15:41 |
| dtantsur | right. what I'm suggesting is to move the numbering logic from https://github.com/metal3-io/metal3-docs/pull/148 to the ironic driver | 15:42 |
| rpioso | dtantsur: Redfish schema include ID props. | 15:42 |
| dtantsur | and avoid a new API in ironic completely | 15:42 |
| dtantsur | then the feature proposed for metal3 will also work for other ironic consumers | 15:42 |
| rpioso | dtantsur: I don't see how the IDs could be made generic. | 15:42 |
| dtantsur | rpioso: exactly the way you're proposing it? | 15:43 |
| rpioso | dtantsur: Even Redfish did not attempt to normalize them. | 15:43 |
| dtantsur | I'm not saying cross-vendor | 15:43 |
| dtantsur | I'm saying: rather than inventing a new API to match numbers to names, accept numbers in your RAID implementation | 15:43 |
| dtantsur | in the physical_drives field (or how is it called?) | 15:43 |
| dtantsur | so pretty much what is proposed for metal3, but do it on ironic level, not in BMO | 15:44 |
| *** MahnoorAsghar has joined #openstack-ironic | 15:45 | |
| rpioso | dtantsur: What would the numbers mean? | 15:45 |
| dtantsur | rpioso: please check the metal3 proposal | 15:45 |
| dtantsur | they're proposing using numbers instead of names and using the new API to replace numbers with names | 15:46 |
| *** jawad_axd has quit IRC | 15:46 | |
| MahnoorAsghar | My laptop suddenly died without warning, sorry | 15:46 |
| rpioso | dtantsur: We could not assure that as new hardware is introduced. | 15:46 |
| dtantsur | okay, then the metal3 proposal should probably drop the conversion bit | 15:46 |
| *** jawad_axd has joined #openstack-ironic | 15:46 | |
| dtantsur | that being said, I'm okay with https://storyboard.openstack.org/#!/story/2008866 | 15:47 |
| dtantsur | just trying to figure out if it's going to be useful for metal3 (seems like no) | 15:47 |
| rpioso | dtantsur: The metal3 proposal to generically number controllers and disks seems misguided to me. | 15:47 |
| dtantsur | rpioso: could you leave a comment there please? you have more information than me. | 15:47 |
| rpioso | dtantsur: 'naive' might be a better choice. | 15:47 |
| MahnoorAsghar | I think its an Airship use case | 15:47 |
| rpioso | dtantsur: We'll see what we can do. | 15:48 |
| rpioso | Regardless, seems like ironic should provide an API which offers controller and disk IDs since the RAID schema supports that. | 15:49 |
| dtantsur | also for generic Redfish? | 15:50 |
| rpioso | dtantsur: yep | 15:50 |
| dtantsur | this is nice. yeah, I agree. | 15:50 |
| rpioso | dtantsur: At least for Dell. | 15:50 |
| MahnoorAsghar | rpioso: +1 | 15:51 |
| rpioso | Other vendors are welcome to join to the discussion :) | 15:51 |
| MahnoorAsghar | I was looking into the HP docs; I think stendulkar could shed some light on it? | 15:51 |
| Qianbiao | rpioso may u give a link to the spec you are disgussing about? | 15:51 |
| rpioso | MahnoorAsghar: Do you have the schema handy? | 15:52 |
| TheJulia | MahnoorAsghar: he likely could, but I suspect he is asleep at this time. :) | 15:52 |
| *** jawad_axd has quit IRC | 15:52 | |
| MahnoorAsghar | TheJulia: I see | 15:52 |
| Qianbiao | I know you are talking about raid config of metal3 but not clear about which part. | 15:52 |
| TheJulia | MahnoorAsghar: he is based in India | 15:52 |
| MahnoorAsghar | TheJulia: My neighbour | 15:53 |
| iurygregory | Qianbiao, https://github.com/metal3-io/metal3-docs/pull/148/files | 15:53 |
| MahnoorAsghar | rpioso: Let me try and find it | 15:53 |
| MahnoorAsghar | rpioso: You were talking about the Redfish spec? | 15:53 |
| rpioso | MahnoorAsghar: schemas | 15:54 |
| MahnoorAsghar | rpioso: okay | 15:54 |
| iurygregory | https://redfish.dmtf.org/redfish/schema_index | 15:54 |
| iurygregory | ^ you can probably find in this page | 15:55 |
| Qianbiao | rpioso dtantsur may you have a look at https://docs.openstack.org/ironic/latest/admin/drivers/ibmc.html#backing-physical-disks | 15:55 |
| Qianbiao | is this implementation matches rpioso proposed? | 15:56 |
| rpioso | The ID property in https://redfish.dmtf.org/schemas/v1/Storage.v1_10_0.json | 15:56 |
| Qianbiao | i think we should support drive-id too | 15:56 |
| MahnoorAsghar | Qianbiao: The implementation matches the proposal | 15:57 |
| rpioso | s/ID/Id/ | 15:57 |
| iurygregory | we are almost out of time =) 2min | 15:57 |
| rpioso | Same idea in https://redfish.dmtf.org/schemas/v1/Drive.v1_12_0.json | 15:58 |
| Qianbiao | Hmm, i think we should give "id" to vendor. | 15:58 |
| Qianbiao | iurygregory no worries, we may end the meeting first :) | 15:58 |
| rpioso | Qianbiao: Agree. Those would be returned by the new APIs which the spec originally proposed. | 15:58 |
| *** bburns has quit IRC | 15:58 | |
| iurygregory | yeah =) | 15:58 |
| iurygregory | #topic Who is going to run the next meeting? | 15:58 |
| *** openstack changes topic to "Who is going to run the next meeting? (Meeting topic: ironic)" | 15:58 | |
| iurygregory | Do we have any volunteers? | 15:59 |
| TheJulia | I can | 15:59 |
| iurygregory | tks TheJulia | 15:59 |
| iurygregory | #endmeeting | 15:59 |
| *** openstack changes topic to "Bare Metal Provisioning | Status: http://bit.ly/ironic-whiteboard | Docs: http://docs.openstack.org/ironic/ | Bugs: https://storyboard.openstack.org/#!/project_group/75 | Contributors are generally present between 6 AM and 12 AM UTC, If we do not answer, please feel free to pose questions to openstack-discuss mailing list." | 15:59 | |
| openstack | Meeting ended Mon May 17 15:59:19 2021 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 15:59 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/ironic/2021/ironic.2021-05-17-15.00.html | 15:59 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/ironic/2021/ironic.2021-05-17-15.00.txt | 15:59 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/ironic/2021/ironic.2021-05-17-15.00.log.html | 15:59 |
| arne_wiebalck | thanks iurygregory | 16:00 |
| *** bburns has joined #openstack-ironic | 16:00 | |
| Qianbiao | rpioso id here can stands any possible fields of a drive. For end user, everything they thinks can identify a drive could be "id". | 16:00 |
| Qianbiao | when i implement raid feature, it descripes "drive-id", drive-name, drive serial number, etc.. | 16:01 |
| lmcgann | I know the meeting is over but since there was no open discussion I'd like to ask now if anybody has the time to check out our security spec it'd be greatly appreciated: https://review.opendev.org/c/openstack/ironic-specs/+/576718 | 16:02 |
| MahnoorAsghar | Qianbiao: Which one of these is the disk identifier given by the vendor? I couldn't figure that part out from the DMTF pages | 16:03 |
| rpioso | Qianbiao: The idrac driver supports the name the iDRAC uses to uniquely identify the device, which is referred to as the FQDD (Fully Qualified Device Descriptor). | 16:03 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: Do not serialize command_params https://review.opendev.org/c/openstack/ironic-python-agent/+/791798 | 16:03 |
| Qianbiao | MahnoorAsghar every drive only has one id indeed. | 16:04 |
| TheJulia | I feel like I need a giant caffeine IV | 16:04 |
| MahnoorAsghar | TheJulia: Same xD | 16:04 |
| Qianbiao | but when we handle the logic, drive name (which is unique too) can be an "id" to end user. | 16:04 |
| MahnoorAsghar | Qianbiao: Okay... | 16:05 |
| Qianbiao | the End user does not care about real id, he just knows drive name can identify an unique drive, and which is most friendly to him, so he just use it. | 16:05 |
| rpioso | Qianbiao: The iDRAC Redfish service uses the FQDD as the value of the Redfish resource 'Id' property. FQDDs are unique identifiers in the context of the system being managed. | 16:06 |
| Qianbiao | ibmc use slot id if i am correct. | 16:06 |
| rpioso | Qianbiao: Please see the definition of 'Id' in https://redfish.dmtf.org/schemas/v1/Resource.json. | 16:07 |
| *** ociuhandu has quit IRC | 16:08 | |
| Qianbiao | rpioso the drive id in ibmc looks like "HDDPlaneDisk1" | 16:09 |
| Qianbiao | HDD means hdd slot, 1 means slot number. | 16:10 |
| Qianbiao | I am not professional with ibmc btw, i am just a freelancer ~~ | 16:10 |
| * arne_wiebalck tried fast-track for the first time, in the context of burn-in, *really* nice :) | 16:11 | |
| rpioso | Qianbiao: Here's an example iDRAC Id of a RAID controller: RAID.Embedded.1-1 | 16:11 |
| TheJulia | arne_wiebalck: \o/ | 16:11 |
| Qianbiao | Hmm, really quit different. | 16:11 |
| rpioso | Qianbiao: 'Embedded' and 1-1 have meaning. | 16:11 |
| rpioso | Qianbiao: They indicate where it's located in the system. | 16:12 |
| Qianbiao | Yes i can guess that. | 16:12 |
| *** lucasagomes has quit IRC | 16:12 | |
| arne_wiebalck | TheJulia: I use a "waiting room" conductor group to do inspection, burn-in and benchmark. This group has fast track enabled, as a requirement from the hardware colleagues. Once confirmed ok, the nodes move to their destination group. | 16:13 |
| openstackgerrit | Pierre Riteau proposed openstack/tenks master: Update pip and setuptools before installing requirements https://review.opendev.org/c/openstack/tenks/+/791751 | 16:13 |
| dtantsur | wow, cool | 16:13 |
| TheJulia | arne_wiebalck: please write about this :) | 16:13 |
| TheJulia | sounds super cool :) | 16:13 |
| Qianbiao | rpioso btw, everytime i join weekly meeting, which encourages me to buy some dell stock. :) | 16:13 |
| openstackgerrit | Pierre Riteau proposed openstack/tenks master: Update pip and setuptools before installing requirements https://review.opendev.org/c/openstack/tenks/+/791751 | 16:13 |
| rpioso | Qianbiao: And here's the ID of a drive controlled by that RAID controller: Disk.Bay.0:Enclosure.Internal.0-1:PCIeExtender.Slot.4 | 16:14 |
| Qianbiao | redhat too :), really nice team | 16:14 |
| TheJulia | :) | 16:14 |
| rpioso | Qianbiao: :-) Thank you so much! | 16:14 |
| arne_wiebalck | TheJulia: will do :) | 16:14 |
| TheJulia | I'd buy stock in CERN if it was a company | 16:15 |
| arne_wiebalck | :-D :-D | 16:15 |
| rpioso | I consider CERN one of the wonders of the world :-) | 16:15 |
| Qianbiao | :) we can not buy it, we can join it. | 16:15 |
| TheJulia | There is part of my brain that is just satisfied by helping improve humanity's knowledge of the universe. It sounds weird I guess. | 16:16 |
| Qianbiao | rpioso almost the same, id identifies the hardware position on mainboard | 16:16 |
| rpioso | Qianbiao: Could be a separate drive enclosure ... | 16:17 |
| *** jawad_axd has joined #openstack-ironic | 16:17 | |
| Qianbiao | right | 16:17 |
| rpioso | Qianbiao: separate, as in external | 16:17 |
| rpioso | It can get complicated. | 16:18 |
| mraineri | I don't think you can deterministcally use "Id" in a manner to describe location or other types of relationships | 16:18 |
| mraineri | While I know Dell does that, "Id" is defined as an opaque identifier as a unique value for a member of a collection | 16:18 |
| TheJulia | I believe the "it can get complicated" case is the same with hpe gear, although I think that is why there is also some of the "these are only hints, we may just focus on attempting to achieve x,y,z without a list of precise drives. | 16:19 |
| mraineri | Some systems might just use a number "Id" as a simple way to manage the keys in the collection | 16:19 |
| MahnoorAsghar | The purpose is to specify the disks for RAID configuration. As long as the BMCs or Redfish accepts it, its all good | 16:19 |
| mraineri | If you want to start applying relationships, you need to use properties in the resource designed for that purpose (like how in the Location property there is a ServiceLabel property that reflects some string a user can find on their enclosure) | 16:20 |
| JayF | MahnoorAsghar: I'm not familiar with this specific hardware, but I think the concern is that the "Id" field may not be deterministically numbered across runs | 16:20 |
| *** Qianbiao has quit IRC | 16:21 | |
| rpioso | mraineri: Do you feel a vendor-independent simple numbering scheme could be constructed which would make sense for all system designs from a user perspective? | 16:21 |
| MahnoorAsghar | JayF: Oh I see | 16:22 |
| rpioso | JayF: It is deterministic for Dell gear. | 16:22 |
| mraineri | rpioso: I think so; we really want to encourage users to rely more on the properties of the resource rather than inferring any real knowledge about the thing based on the value of "Id" | 16:23 |
| JayF | rpioso: nice. Too bad the spec doesn't require that :C | 16:23 |
| mraineri | All you can really know for sure from "Id" is that you have some unique value that identifies that resource in a collection | 16:23 |
| TheJulia | Then the "what may change that" may need to be documented as well, just from a standpoint if somebody plugs in a cable between deployments, does it potentially change things around unexpectedly. At least, that would be my operational worry and without documentation I'd likely struggle to understand | 16:24 |
| mraineri | That is not expected to change over the life of the system | 16:24 |
| mraineri | There is a "Resource tree stability" clause that recommends things remain static as appropriate across service restarts | 16:25 |
| mraineri | So, if you come across a URI like "/redfish/v1/Chassis/1/Drives/6", it's very reasonable that the same system will keep reporting that at that URI | 16:25 |
| TheJulia | as long as there is no physical change I guess | 16:26 |
| mraineri | Correct | 16:26 |
| *** alexmcleod has quit IRC | 16:26 | |
| rpioso | TheJulia: To be clear, a drive could be replaced in the system. The new drive's FQDD/Id would be the same as the drive it replaced. | 16:26 |
| TheJulia | k | 16:28 |
| mraineri | I would expect that too | 16:34 |
| mraineri | I would also expect that in the case where the drive was removed, "/redfish/v1/Chassis/1/Drives/6" would still return 200 OK, but the status would indicate the drive is not present | 16:35 |
| dtantsur | HTTP 244 OK NOT FOUND | 16:35 |
| dtantsur | ("Task failed successfully") | 16:36 |
| mraineri | :) | 16:36 |
| rpioso | LOL | 16:36 |
| dtantsur | on this positive note I'm wishing you a great evening | 16:36 |
| *** dtantsur is now known as dtantsur|afk | 16:36 | |
| mraineri | In all seriousness though, even if a device is not present, there's still good info a user might want to obtain | 16:37 |
| mraineri | Like, slot type, service label indicator, other capabilities of the slot... | 16:37 |
| openstackgerrit | Merged openstack/tenks master: Update pip and setuptools before installing requirements https://review.opendev.org/c/openstack/tenks/+/791751 | 16:37 |
| mraineri | We have a "Special resource situations" clause that helps describe this behavior | 16:39 |
| * TheJulia blinks at ERROR 244 | 16:41 | |
| mraineri | Don't worry, we don't count that high in Redfish :D | 16:42 |
| rpioso | TheJulia: How do our consumers figure out the storage controllers and disks a system has? | 16:43 |
| *** gyee has joined #openstack-ironic | 16:44 | |
| arne_wiebalck | bye everyone o/ | 16:48 |
| rpioso | arne_wiebalck: Good nite! :-) | 16:48 |
| MahnoorAsghar | arne_wiebalck: bye o/ | 16:49 |
| *** jawad_axd has quit IRC | 16:50 | |
| MahnoorAsghar | Thank you guys for the discussion, Im gonna call it a night now | 16:51 |
| MahnoorAsghar | Cheers! | 16:51 |
| rpioso | MahnoorAsghar: Hope you have a nice evening. | 16:53 |
| *** MahnoorAsghar has quit IRC | 16:53 | |
| *** MahnoorAsghar has joined #openstack-ironic | 16:54 | |
| MahnoorAsghar | thanks :D | 16:54 |
| *** MahnoorAsghar has quit IRC | 16:54 | |
| *** dougsz has quit IRC | 16:59 | |
| *** derekh has quit IRC | 17:00 | |
| iurygregory | enr I only saw this now https://review.opendev.org/c/openstack/releases/+/790748 | 17:03 |
| TheJulia | rpioso: inspection data or direct examination. Which kind of leads me mentally to, why do we keep inspection data as-is | 17:03 |
| iurygregory | we should probably get feedback to see if we are ok or not | 17:03 |
| TheJulia | Yeah, Steve Baker might have feels on this subject | 17:04 |
| iurygregory | from RH perspective I think we are -1 | 17:04 |
| iurygregory | I will add him as reviewer =) | 17:04 |
| TheJulia | I'm ambivalent as it is train, since we can still merge to it | 17:04 |
| iurygregory | yup | 17:05 |
| TheJulia | but the direct to EOL may be bad. | 17:05 |
| iurygregory | ++ | 17:05 |
| *** k_mouza has quit IRC | 17:12 | |
| *** k_mouza has joined #openstack-ironic | 17:13 | |
| TheJulia | I do think we need need to consider dropping pre-train branches | 17:17 |
| TheJulia | like, EOL them. | 17:18 |
| *** k_mouza has quit IRC | 17:18 | |
| iurygregory | sounds like a plan | 17:18 |
| iurygregory | this is new to me... "Zuul PTL-Approved +1" | 17:37 |
| iurygregory | https://review.opendev.org/c/openstack/releases/+/791804 | 17:38 |
| iurygregory | nvm, it shows the name of Zuul in other patches, I never noticed :D (I was thought it was the name of the PTL/liaison) | 17:42 |
| *** k_mouza has joined #openstack-ironic | 17:45 | |
| rpioso | TheJulia: Are you aware of any OOB inspection that offers that? | 17:49 |
| *** k_mouza has quit IRC | 17:50 | |
| *** k_mouza has joined #openstack-ironic | 17:51 | |
| *** k_mouza has quit IRC | 17:51 | |
| *** k_mouza has joined #openstack-ironic | 17:52 | |
| *** rloo has quit IRC | 17:56 | |
| *** rloo has joined #openstack-ironic | 17:57 | |
| *** jawad_axd has joined #openstack-ironic | 18:01 | |
| *** k_mouza has quit IRC | 18:03 | |
| *** jawad_axd has quit IRC | 18:06 | |
| TheJulia | rpioso: oob, no | 18:33 |
| rpioso | TheJulia: Is the in-band data accessible via this API: https://docs.openstack.org/api-ref/baremetal-introspection/#get-introspection-data? | 18:38 |
| TheJulia | Yes, but entirely optional component *and* restricted to system-admin level access | 18:38 |
| TheJulia | well, system-reader I guess would work | 18:39 |
| rpioso | TheJulia: And that's the only way we offer to our consumers to identify the storage controllers and disks in a system? | 18:39 |
| TheJulia | Yes. Granted, we never intended for anyone other than a system adminstrator to be setting up such fine grained details on a system for things such as raid where a high level of explicit detail may be required to request something architecturally specific if the driver doesn't support a common/general happy path of configuration | 18:42 |
| TheJulia | "make me a giant raid5 if I don't tell you anything specific limited to the first controller" and the driver just figures it out | 18:42 |
| TheJulia | Naturally, if someone is doing something highly detailed, they need that level of information to articulate exactly what and how they need it | 18:43 |
| rpioso | TheJulia: Hrm ... A not uncommon PowerEdge config includes a BOSS controller for the operating system and a PERC controller for all other storage. They're both RAID controllers. How would configuring them be expressed without such explicit detail? | 19:09 |
| rpioso | TheJulia: Also, btw, BOSS + software RAID controller is a thing. | 19:10 |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: Add additional node indexes https://review.opendev.org/c/openstack/ironic/+/788625 | 19:58 |
| JayF | I have seen RAID 10 implemented as the first layer done in PERC, and the second layer done in software (MD) | 20:04 |
| JayF | This was years and years and years ago and I cannot remember why they did it that way (PERC/5 to really date it) | 20:04 |
| TheJulia | rpioso: Well, I *suspect* boss would be implied as the boot device under most circumstances. Why would anyone do a boss card with a mirror and then software raid though? That just adds to the overall io queue of the OS slowing it down :\ | 20:49 |
| TheJulia | Granted, operational tools | 20:49 |
| TheJulia | err | 20:49 |
| TheJulia | not tools, but requirements can be crazy and varied | 20:49 |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: Add additional node indexes https://review.opendev.org/c/openstack/ironic/+/788625 | 21:28 |
| -openstackstatus- NOTICE: The Zuul service at zuul.opendev.org will be offline for a few minutes (starting now) in order for us to make some needed filesystem changes; if the outage lasts longer than anticipated we'll issue further notices | 21:32 | |
| * TheJulia sighs | 21:32 | |
| * iurygregory time to go sleep, tks Zuul! | 21:34 | |
| *** paras333_ has quit IRC | 21:37 | |
| *** paras333 has joined #openstack-ironic | 21:38 | |
| *** paras333 has quit IRC | 21:42 | |
| *** lmcgann has quit IRC | 22:10 | |
| *** iurygregory has quit IRC | 22:10 | |
| *** iurygregory has joined #openstack-ironic | 22:19 | |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: Add additional node indexes https://review.opendev.org/c/openstack/ironic/+/788625 | 22:38 |
| *** uzumaki has joined #openstack-ironic | 22:45 | |
| *** uzumaki has quit IRC | 22:46 | |
| *** uzumaki has joined #openstack-ironic | 22:47 | |
| rpioso | TheJulia: In case you missed it ... A not uncommon PowerEdge config includes a BOSS controller for the operating system and a PERC controller for all other storage. They're both RAID controllers. How would configuring their logical disks be expressed without such explicit detail? | 23:28 |
| *** k_mouza has joined #openstack-ironic | 23:34 | |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: Add additional node indexes https://review.opendev.org/c/openstack/ironic/+/788625 | 23:35 |
| TheJulia | rpioso: It is somewhat logical to decide that the drive to be deployed to is the boss card resulting volume, but separate devices would need to be delienated so you'd basically need to know some additional information regardless | 23:38 |
| rpioso | TheJulia: I'm referring to the RAID logical disk configuration. | 23:39 |
| TheJulia | I know | 23:39 |
| TheJulia | controllers would need to be able to be delineated, then again if that can be done "under the hood" of the driver, maybe that is sufficent | 23:40 |
| * TheJulia shrugs | 23:40 | |
| rpioso | TheJulia: So the user just knows? | 23:40 |
| TheJulia | The API consumer would just have to know today, disks and controllers. if the driver just can't figure it out "under the hood" | 23:41 |
| rpioso | TheJulia: Perhaps an example target RAID config could help me understand what the driver would be expected to figure out. | 23:42 |
| *** tosky has quit IRC | 23:43 | |
| *** tzumainn has quit IRC | 23:44 | |
| *** uzumaki has quit IRC | 23:44 | |
| rpioso | dtantsur, MahnoorAsghar: I posted a comment to the Metal3 "raid-extension-for-physical-disks-and-controller" spec: https://github.com/metal3-io/metal3-docs/pull/148#pullrequestreview-661495303. Thank you, mraineri, for sharing your experience! I still feel that providing a means of obtaining the IDs from ironic is needed. | 23:45 |
| TheJulia | rpioso: examples 2 and 3 I think: https://docs.openstack.org/ironic/latest/admin/raid.html#examples-for-target-raid-config | 23:45 |
| rpioso | TheJulia: ty | 23:47 |
| NobodyCam | Good Afternoon Ironic Folks | 23:49 |
| NobodyCam | Any reason we don't emit a node moved to Available event.. | 23:52 |
| *** k_mouza has quit IRC | 23:57 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!