| *** strigazi has quit IRC | 00:01 | |
| *** vabada has quit IRC | 00:01 | |
| *** strigazi has joined #openstack-ironic | 00:02 | |
| *** vabada has joined #openstack-ironic | 00:02 | |
| *** gabys has joined #openstack-ironic | 00:17 | |
| *** sthussey has quit IRC | 00:18 | |
| *** gabys has quit IRC | 00:22 | |
| *** hamzy has joined #openstack-ironic | 00:25 | |
| *** gabys has joined #openstack-ironic | 00:44 | |
| *** gabys has quit IRC | 00:49 | |
| *** gyee has quit IRC | 00:51 | |
| *** gabys has joined #openstack-ironic | 01:01 | |
| *** phuongnh has joined #openstack-ironic | 01:04 | |
| *** gabys has quit IRC | 01:05 | |
| *** gabys has joined #openstack-ironic | 01:10 | |
| *** gabys has quit IRC | 01:15 | |
| *** JudeCross has quit IRC | 01:18 | |
| *** phuongnh has quit IRC | 01:30 | |
| *** gabys has joined #openstack-ironic | 01:33 | |
| *** gabys has quit IRC | 01:37 | |
| *** zzzeek has quit IRC | 01:48 | |
| *** zzzeek has joined #openstack-ironic | 01:49 | |
| *** gabys has joined #openstack-ironic | 01:49 | |
| *** rloo has quit IRC | 01:51 | |
| *** gabys has quit IRC | 01:54 | |
| *** JudeCross has joined #openstack-ironic | 02:05 | |
| *** hshiina has joined #openstack-ironic | 02:17 | |
| *** gabys has joined #openstack-ironic | 02:19 | |
| *** gabys has quit IRC | 02:24 | |
| *** gabys has joined #openstack-ironic | 02:39 | |
| *** gabys has quit IRC | 02:43 | |
| *** phuongnh has joined #openstack-ironic | 02:53 | |
| *** gabys has joined #openstack-ironic | 02:57 | |
| *** gabys has quit IRC | 03:01 | |
| madhuri[m] | Hi, can anyone help me understanding the adopt state in Ironic? Does a node in adopt state being provisioned on a boot request from Nova? | 03:14 |
|---|---|---|
| madhuri[m] | I mean re-provisioned | 03:15 |
| *** gabys has joined #openstack-ironic | 03:24 | |
| *** gabys has quit IRC | 03:29 | |
| *** dnuka has joined #openstack-ironic | 03:46 | |
| *** gabys has joined #openstack-ironic | 03:52 | |
| *** gabys has quit IRC | 03:57 | |
| *** gabys has joined #openstack-ironic | 04:21 | |
| *** gabys has quit IRC | 04:25 | |
| *** jaypipes has quit IRC | 04:27 | |
| *** gabys has joined #openstack-ironic | 04:34 | |
| *** gabys has quit IRC | 04:38 | |
| *** toabctl has quit IRC | 04:48 | |
| *** JudeCross has quit IRC | 04:53 | |
| *** gabys has joined #openstack-ironic | 05:00 | |
| *** cxc99 has quit IRC | 05:02 | |
| *** gabys has quit IRC | 05:04 | |
| *** cxc99 has joined #openstack-ironic | 05:09 | |
| *** gabys has joined #openstack-ironic | 05:10 | |
| *** gabys has quit IRC | 05:15 | |
| *** JudeCross has joined #openstack-ironic | 05:15 | |
| openstackgerrit | Merged openstack/ironic master: Minor fixes for docs on changing hardware types https://review.openstack.org/605656 | 05:21 |
| *** gabys has joined #openstack-ironic | 05:40 | |
| *** gabys has quit IRC | 05:44 | |
| openstackgerrit | Nguyen Van Trung proposed openstack/ironic master: Fix not detach boot image for virtual CDROM https://review.openstack.org/604704 | 05:51 |
| *** e0ne has joined #openstack-ironic | 05:52 | |
| *** JudeCross has quit IRC | 06:04 | |
| *** gabys has joined #openstack-ironic | 06:09 | |
| *** pcaruana has joined #openstack-ironic | 06:11 | |
| *** gabys has quit IRC | 06:14 | |
| *** skazi has quit IRC | 06:21 | |
| *** dsneddon has quit IRC | 06:23 | |
| *** verdurin has quit IRC | 06:23 | |
| *** e0ne has quit IRC | 06:24 | |
| *** gabys has joined #openstack-ironic | 06:28 | |
| *** JudeCross has joined #openstack-ironic | 06:29 | |
| *** gabys has quit IRC | 06:33 | |
| *** jtomasek has joined #openstack-ironic | 06:44 | |
| *** e0ne has joined #openstack-ironic | 06:51 | |
| *** gabys has joined #openstack-ironic | 06:52 | |
| openstackgerrit | Mallikarjuna Reddy Kolagatla proposed openstack/ironic master: Add admin documentation for rescue mode in iLO driver https://review.openstack.org/605647 | 06:56 |
| *** gabys has quit IRC | 06:57 | |
| *** hshiina has quit IRC | 06:59 | |
| *** gabys has joined #openstack-ironic | 07:04 | |
| openstackgerrit | Derek Higgins proposed openstack/ironic master: [WIP] Switch ironic-tempest-dsvm-ipa-wholedisk-agent_ipmitool-tinyipa-multinode https://review.openstack.org/587365 | 07:08 |
| *** gabys has quit IRC | 07:09 | |
| *** rcernin has quit IRC | 07:12 | |
| *** MattMan_1 has quit IRC | 07:16 | |
| *** MattMan_1 has joined #openstack-ironic | 07:16 | |
| *** gkadam has joined #openstack-ironic | 07:23 | |
| *** gabys has joined #openstack-ironic | 07:26 | |
| *** gabys has quit IRC | 07:31 | |
| *** S4ren has joined #openstack-ironic | 07:31 | |
| *** skazi has joined #openstack-ironic | 07:31 | |
| *** JudeCross has quit IRC | 07:42 | |
| *** hoangcx has quit IRC | 07:43 | |
| *** phuongnh has quit IRC | 07:44 | |
| *** gabys has joined #openstack-ironic | 07:46 | |
| *** gabys has quit IRC | 07:50 | |
| *** dnuka has quit IRC | 07:55 | |
| *** hoangcx has joined #openstack-ironic | 07:57 | |
| *** dougsz has joined #openstack-ironic | 08:05 | |
| *** gabys has joined #openstack-ironic | 08:07 | |
| *** lenka has joined #openstack-ironic | 08:09 | |
| *** gabys has quit IRC | 08:12 | |
| *** mgoddard has joined #openstack-ironic | 08:16 | |
| *** serlex has joined #openstack-ironic | 08:16 | |
| *** sai_p has quit IRC | 08:26 | |
| *** gabys has joined #openstack-ironic | 08:33 | |
| *** gabys has quit IRC | 08:37 | |
| *** derekh has joined #openstack-ironic | 08:37 | |
| *** olivierb has joined #openstack-ironic | 08:38 | |
| *** AJaeger has joined #openstack-ironic | 08:59 | |
| AJaeger | ironic team, all your python3-first changes that were proposed are passing, see https://review.openstack.org/#/q/is:open+topic:python3-first+(project:%255E.*ironic.*+OR+project:openstack/molteniron+OR+project:openstack/networking-generic-switch+) (note that list contains an ansible one and some cleanups from your cores) | 09:00 |
| AJaeger | Could you get all these merged, please? | 09:00 |
| AJaeger | TheJulia: who's can help reviewing these? ^ | 09:00 |
| *** gabys has joined #openstack-ironic | 09:01 | |
| olivierb | Good morning everyone, sorry but https://review.openstack.org/#/c/599544/ keeps failing and I really do not get why the VM does not start properly furthermore as the CI does not seem that crowded anymore :-( | 09:03 |
| patchbot | patch 599544 - bifrost - Add voting and gating test with Keystone enabled - 10 patch sets | 09:03 |
| *** gabys has quit IRC | 09:06 | |
| openstackgerrit | Derek Higgins proposed openstack/ironic master: [WIP] Switch ironic-tempest-dsvm-ipa-wholedisk-agent_ipmitool-tinyipa-multinode https://review.openstack.org/587365 | 09:20 |
| *** gabys has joined #openstack-ironic | 09:22 | |
| *** gabys has quit IRC | 09:26 | |
| *** gabys has joined #openstack-ironic | 09:42 | |
| *** tssurya has joined #openstack-ironic | 09:46 | |
| *** gabys has quit IRC | 09:47 | |
| *** jtomasek has quit IRC | 09:56 | |
| *** S4ren1 has joined #openstack-ironic | 10:16 | |
| *** e0ne has quit IRC | 10:16 | |
| *** S4ren has quit IRC | 10:17 | |
| *** S4ren1 has quit IRC | 10:17 | |
| *** hoangcx has quit IRC | 10:23 | |
| *** e0ne has joined #openstack-ironic | 10:25 | |
| *** gabys has joined #openstack-ironic | 10:39 | |
| *** gabys has quit IRC | 10:43 | |
| *** olivierb has quit IRC | 10:47 | |
| *** AJaeger has quit IRC | 10:56 | |
| *** dtantsur|afk is now known as dtantsur | 10:57 | |
| dtantsur | morning ironic | 10:57 |
| *** hjensas has joined #openstack-ironic | 11:05 | |
| *** AJaeger has joined #openstack-ironic | 11:08 | |
| openstackgerrit | Derek Higgins proposed openstack/ironic master: [WIP] Switch ironic-tempest-dsvm-ipa-wholedisk-agent_ipmitool-tinyipa-multinode https://review.openstack.org/587365 | 11:33 |
| *** yolanda has joined #openstack-ironic | 11:44 | |
| *** gabys has joined #openstack-ironic | 11:52 | |
| *** mgoddard has quit IRC | 11:54 | |
| *** gabys has quit IRC | 11:56 | |
| *** Bhujay has joined #openstack-ironic | 11:58 | |
| jroll | AJaeger: we'll get to them, thanks for the ping | 11:59 |
| *** mgoddard has joined #openstack-ironic | 12:00 | |
| *** mmethot has joined #openstack-ironic | 12:00 | |
| openstackgerrit | Mark Goddard proposed openstack/ironic-specs master: Deploy Templates https://review.openstack.org/504952 | 12:03 |
| *** mvpnitesh has joined #openstack-ironic | 12:07 | |
| *** vmud213 has joined #openstack-ironic | 12:08 | |
| dtantsur | morning jroll. both of these patches are already +2 for me, so you can get some stats if you need them :) | 12:08 |
| dtantsur | s/both/most/ | 12:08 |
| *** skazi has quit IRC | 12:10 | |
| *** serlex has quit IRC | 12:11 | |
| *** vmud213 has quit IRC | 12:12 | |
| *** rh-jelabarre has joined #openstack-ironic | 12:15 | |
| AJaeger | thanks, jroll | 12:17 |
| jroll | dtantsur: yeah, will get them shortly | 12:17 |
| openstackgerrit | Merged openstack/ironic-python-agent-builder master: import zuul job settings from project-config https://review.openstack.org/592396 | 12:17 |
| *** gabys has joined #openstack-ironic | 12:19 | |
| openstackgerrit | Merged openstack/ironic-python-agent-builder master: switch documentation job to new PTI https://review.openstack.org/592397 | 12:20 |
| *** gabys has quit IRC | 12:23 | |
| *** jcoufal has joined #openstack-ironic | 12:34 | |
| *** gabys has joined #openstack-ironic | 12:35 | |
| * TheJulia tries to wake up | 12:38 | |
| * AJaeger thanks jroll ! Looks like you got them all ;) | 12:39 | |
| jroll | AJaeger: you're welcome :) | 12:39 |
| *** gabys has quit IRC | 12:39 | |
| *** trown|outtypewww is now known as trown | 12:43 | |
| TheJulia | jroll: thanks! | 12:53 |
| *** dnuka has joined #openstack-ironic | 12:53 | |
| *** serlex has joined #openstack-ironic | 12:56 | |
| *** rloo has joined #openstack-ironic | 13:00 | |
| *** gabys has joined #openstack-ironic | 13:03 | |
| *** serlex has quit IRC | 13:03 | |
| TheJulia | hmm oliverb's issue is bizzar | 13:04 |
| TheJulia | madhuri[m]: I wrote the adopt feature, so I'm happy to answer questions about it. Adopt is completely outside the realm of nova interactions. It allows for an operator to migrate in machines into their ironic managed fleet | 13:05 |
| TheJulia | madhuri[m]: Adopt should go directly to adopting? which either goes to active or adopt failed | 13:06 |
| *** mvpnitesh has quit IRC | 13:06 | |
| *** gabys has quit IRC | 13:08 | |
| madhuri[m] | TheJulia: Thank you for the clarification. So when we change the state to adopt, it will automatically move to active on successful adoption? | 13:09 |
| TheJulia | madhuri[m]: yes | 13:12 |
| madhuri[m] | So the "rebuild" state makes it available for provisioning by Nova? | 13:13 |
| TheJulia | no | 13:13 |
| TheJulia | adopted nodes are invisible by nova until they are undeployed | 13:14 |
| TheJulia | s/by/from/ | 13:14 |
| *** gabys has joined #openstack-ironic | 13:19 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/sushy master: Update to use mapped field for Settings Message Severity field https://review.openstack.org/591612 | 13:21 |
| *** dnuka has quit IRC | 13:21 | |
| *** gabys has quit IRC | 13:24 | |
| *** derekh has quit IRC | 13:26 | |
| *** derekh has joined #openstack-ironic | 13:26 | |
| dtantsur | morning TheJulia | 13:26 |
| openstackgerrit | Mark Goddard proposed openstack/python-ironicclient master: Modify useful links to project resources in README https://review.openstack.org/606072 | 13:26 |
| openstackgerrit | Merged openstack/ironic-ui stable/rocky: import zuul job settings from project-config https://review.openstack.org/592465 | 13:30 |
| openstackgerrit | Merged openstack/ironic-ui stable/queens: import zuul job settings from project-config https://review.openstack.org/592451 | 13:30 |
| AJaeger | yeah! ^ | 13:30 |
| openstackgerrit | Merged openstack/ironic-ui stable/pike: import zuul job settings from project-config https://review.openstack.org/592440 | 13:34 |
| openstackgerrit | Merged openstack/ironic-ui stable/ocata: import zuul job settings from project-config https://review.openstack.org/592431 | 13:34 |
| dtantsur | :) | 13:36 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost master: Add voting and gating test with Keystone enabled https://review.openstack.org/599544 | 13:36 |
| *** lenka has quit IRC | 13:37 | |
| dtantsur | dhellmann: hi! we have a fix victim of the new release jobs: http://logs.openstack.org/24/591624/2/gate/test-release-openstack-python3/3cf3d56/job-output.txt.gz#_2018-09-28_13_27_15_328927 | 13:38 |
| dtantsur | a fix? I mean the first >_< | 13:38 |
| * dtantsur needs to see a doctor if this continues.... | 13:38 | |
| *** lenka has joined #openstack-ironic | 13:40 | |
| TheJulia | my brain sometimes does that. It is horrible when it is actual spoken words though :( | 13:40 |
| dtantsur | it happens to me all the time, even spoken | 13:41 |
| *** zzzeek has quit IRC | 13:43 | |
| openstackgerrit | yolanda.robla proposed openstack/ironic master: Add functionality for individual cleaning on nodes https://review.openstack.org/586277 | 13:44 |
| *** zzzeek has joined #openstack-ironic | 13:45 | |
| * TheJulia needs to get kaifeng into irc more often | 13:46 | |
| *** lenka has quit IRC | 13:48 | |
| *** zzzeek has quit IRC | 13:48 | |
| *** gabys has joined #openstack-ironic | 13:48 | |
| *** zzzeek has joined #openstack-ironic | 13:49 | |
| *** radez has joined #openstack-ironic | 13:50 | |
| *** lenka has joined #openstack-ironic | 13:52 | |
| radez | mgoddard: ping, hey Mark I spoke about networking-ansible in Denver, I wanted to follow up with you and see if you've poked at it since then or see if I could get you to spend a little time looking at it? | 13:52 |
| *** gabys has quit IRC | 13:53 | |
| openstackgerrit | Merged openstack/python-ironic-inspector-client stable/rocky: import zuul job settings from project-config https://review.openstack.org/592469 | 13:53 |
| openstackgerrit | Merged openstack/python-ironic-inspector-client stable/ocata: import zuul job settings from project-config https://review.openstack.org/592434 | 13:53 |
| openstackgerrit | Merged openstack/python-ironic-inspector-client stable/pike: import zuul job settings from project-config https://review.openstack.org/592444 | 13:53 |
| openstackgerrit | Merged openstack/python-ironic-inspector-client stable/queens: import zuul job settings from project-config https://review.openstack.org/592455 | 13:53 |
| radez | we're trying to collect requirements that folks have that would make it a viable alternative to what they're currenly using, so your feed back would be valuable to our development | 13:53 |
| *** odyssey4me has joined #openstack-ironic | 13:57 | |
| *** baha has joined #openstack-ironic | 14:00 | |
| TheJulia | Hi Brain, we were doing something... and then got derailed. What was it? | 14:01 |
| openstackgerrit | Merged openstack/ironic-ui master: Drop nose dependencies https://review.openstack.org/592558 | 14:02 |
| *** gabys has joined #openstack-ironic | 14:10 | |
| openstackgerrit | Derek Higgins proposed openstack/ironic master: [WIP] Switch ironic-tempest-dsvm-ipa-wholedisk-agent_ipmitool-tinyipa-multinode https://review.openstack.org/587365 | 14:12 |
| openstackgerrit | Merged openstack/python-ironic-inspector-client master: Change basepython to python3 https://review.openstack.org/591838 | 14:13 |
| *** gabys has quit IRC | 14:15 | |
| openstackgerrit | Merged openstack/networking-generic-switch stable/pike: import zuul job settings from project-config https://review.openstack.org/592442 | 14:19 |
| openstackgerrit | Merged openstack/networking-generic-switch stable/ocata: import zuul job settings from project-config https://review.openstack.org/592432 | 14:19 |
| *** bnemec is now known as beekneemech | 14:20 | |
| *** jaypipes has joined #openstack-ironic | 14:21 | |
| *** coreywright has joined #openstack-ironic | 14:21 | |
| jaypipes | mgoddard: hey, I'd like to say that you hit the nail on the head with your response. To quote you "Would you prefer a design in which these two things were separate? In that case, this spec would not need to mention traits at all. Deploy templates could be requested by (non-trait) name or UUID as a flavor extra spec, pushed to a 'deploy_templates' field in ironic instance_info by the virt driver, then acted upon by ironic. If scheduling | 14:22 |
| jaypipes | influence is required, this could be provided by traits, but this would be a separate concern." <-- that is PRECISELY what I would like to see and what I thought I had been talking about at the past two PTGs. | 14:22 |
| dtantsur | jaypipes: actually, I remember everyone in the room saying "Just use traits" at least once... | 14:24 |
| jaypipes | mgoddard: but pretty much the last thing I want to see is traits like CUSTOM_HAVE_MORE_THAN_2_GPUS | 14:24 |
| jaypipes | dtantsur: definitely wasn't me. | 14:25 |
| dtantsur | jaypipes: also I'm even more sure the idea of passing anything from flavor to ironic was shot down several times | 14:25 |
| openstackgerrit | yolanda.robla proposed openstack/ironic master: Add functionality for individual cleaning on nodes https://review.openstack.org/586277 | 14:25 |
| dtantsur | jaypipes: then my memory serves me wrong, because it includes you :) | 14:25 |
| jaypipes | dtantsur: you are confusing configuration key/value pairs with provider capabilities. | 14:25 |
| dtantsur | jaypipes: it's not me, it's reality of hardware | 14:25 |
| dtantsur | the reality is not binary, it's ternary | 14:26 |
| dtantsur | e.g. it's not that I have VT-X on my CPU or I do not | 14:26 |
| jaypipes | dtantsur: for things like whether a node supports booting via UEFI, that's a capability/trait. passing through disk partitioning information or RAID setup configuration is another thing entirely. | 14:26 |
| dtantsur | I have it, or I can have it or I cannot have it | 14:26 |
| dtantsur | disk partitioning - yes, RAID - not exactluy | 14:26 |
| dtantsur | again, ternary: I already have RAID-5, I can have RAID-5 on demand, I cannot have RAID-5 | 14:27 |
| jaypipes | dtantsur: I would just like to see a clean separation between things that are capabilities/traits and things that are configuration data. | 14:27 |
| dtantsur | I hear you in this regard. I'm just noting that it's not always trivial | 14:27 |
| jaypipes | dtantsur: no, you either support RAID5 or you don't. | 14:27 |
| dtantsur | jaypipes: support = ? | 14:28 |
| dtantsur | I can have it preconfigured or I can have it on demand | 14:28 |
| jaypipes | dtantsur: there is no such thing as "RAID5 on demand" | 14:28 |
| dtantsur | we're working on it | 14:28 |
| dtantsur | so yes, there is not, but there will be soon | 14:28 |
| jaypipes | dtantsur: if it's possible to do RAID5 (regardless of whether that functionality is "baked in" or whether it needs to be "enabled" or "activated" by some instruction, the capability is RAID5. | 14:29 |
| *** edleafe has quit IRC | 14:29 | |
| dtantsur | exactly. but then ironic needs a way to enforce it | 14:29 |
| jaypipes | dtantsur: there is no grey area there. | 14:29 |
| jaypipes | dtantsur: enforce it? you mean activate the functionality? | 14:29 |
| dtantsur | because "can do RAID-5" does not mean "there is RAID-5 there" | 14:29 |
| dtantsur | yep | 14:29 |
| jaypipes | dtantsur: right, sure, but that has nothing to do with placement or scheduling. | 14:29 |
| jaypipes | dtantsur: that is *configuration data* | 14:30 |
| jaypipes | dtantsur: that can and should be sent *from nova to ironic before boot*. | 14:30 |
| dtantsur | not entirely nothing. you cannot request RAID-5 from a node that 1. cannot build RAID, 2. does not have it already pre-configured. | 14:30 |
| jaypipes | dtantsur: we are currently woefully lacking in this ability right now, as I mentioned in the email thread. | 14:30 |
| *** edleafe has joined #openstack-ironic | 14:30 | |
| jaypipes | dtantsur: but I'm happy to add this ability to nova. | 14:30 |
| dtantsur | yep, and as I said, this idea was given -2 several times. I'm glad it's changing. | 14:30 |
| jaypipes | dtantsur: the idea that was -2'd was just passing willy-nilly any old information straight through (as a proxy API) from the nova boot command to the virt driver | 14:31 |
| jaypipes | dtantsur: and you are well aware of that. | 14:31 |
| dtantsur | then what is the alternative you're talking about? | 14:31 |
| dtantsur | are you going to build a RAID abstraction in Nova? | 14:32 |
| *** mjturek has joined #openstack-ironic | 14:32 | |
| TheJulia | so, a little late this morning to this discussion because I went and visited the artificial rain simulator... | 14:32 |
| * TheJulia reads | 14:32 | |
| *** gabys has joined #openstack-ironic | 14:32 | |
| jaypipes | dtantsur: what I proposed was allowing a --config-data <SOME_UUID> option to the nova boot command that would be a reference to some metadefs stored in glance that could be passed along to the virt driver. | 14:32 |
| dtantsur | jaypipes: so, passign any old information straight through, but with glance involved? | 14:33 |
| jaypipes | dtantsur: glance metadefs can be fully controlled by the admin, verified by glance's metadefs layer, and have schema(s) and versioning applied to them | 14:33 |
| jaypipes | dtantsur: you're just trying to provoke me... | 14:33 |
| dtantsur | and please pardon my ignorance, how is glance related, especially with booting from volume? | 14:33 |
| dtantsur | Jay, I'm sorry if I sound like that | 14:33 |
| dtantsur | but this has been back-and-forth so many times that please allow me some impacience | 14:33 |
| dtantsur | * impatience? | 14:34 |
| jaypipes | dtantsur: and I'm trying to tell you why glance is involved. the glance metadefs functionality allows an admin (or a service, such as Ironic) to define a document schema called a metadef. this schema would define such fields like, for instance, raid config or disk partitioning, etc. | 14:34 |
| openstackgerrit | Merged openstack/ironic-ui master: Imported Translations from Zanata https://review.openstack.org/603296 | 14:34 |
| openstackgerrit | Merged openstack/ironic-ui master: Add ability to test Ironic UI with Horizon master branch https://review.openstack.org/602184 | 14:34 |
| openstackgerrit | Merged openstack/ironic-ui stable/rocky: Update UPPER_CONSTRAINTS_FILE for stable/rocky https://review.openstack.org/590319 | 14:35 |
| jaypipes | dtantsur: and therefore nova could accept this identifier, pass the identifier to glance which would return the data associated with the particular metadata document | 14:35 |
| jaypipes | dtantsur: so nova doesn't need to be in the business of verifying data for ironic, and Ironic can define the structure of its needed configuration data itself | 14:35 |
| dtantsur | jaypipes: okay, probably, even though the glance schema is (IIRC) quite relaxed. but what about booting from volume? does Cinder pass Glance properties a volume was created from? | 14:35 |
| TheJulia | This really feels like we are struggling in a situation where we as a larger community can't make up our minds, and we're getting conflicting consensuses each time we try to reach consensus :( | 14:36 |
| dtantsur | TheJulia++ | 14:36 |
| *** gabys has quit IRC | 14:37 | |
| TheJulia | The current situation was driven by an explicit desire 1+ years ago to get rid of capabilities, now it seems like we want to keep them, at least from the discussion I've read | 14:37 |
| * TheJulia wonders if it is too early to drink, then remembers she has a gala of sorts to go to tonight | 14:37 | |
| * TheJulia just wants us to all get along, and reach a happy place | 14:38 | |
| jaypipes | TheJulia, dtantsur: please read this. https://docs.openstack.org/glance/pike/user/glancemetadefcatalogapi.html | 14:38 |
| TheJulia | dtantsur: aiui, cinder pulls properties from glance and write an image out | 14:38 |
| jaypipes | TheJulia, dtantsur: along with this: https://docs.openstack.org/glance/pike/user/metadefs-concepts.html | 14:41 |
| TheJulia | Interesting | 14:42 |
| TheJulia | I seem to remember this came up in discussion and largely the problem that seemed to exist was that the same image may need to be able to deployed to multiple different configurations. Granted, I feel like I've only scratched the surface of those links | 14:43 |
| madhuri[m] | TheJulia: By undeployed, do you mean we need to enroll the node back and make it available? | 14:43 |
| jaypipes | TheJulia: there's nothing preventing a user from specifying different image and "config data" IDs. | 14:43 |
| TheJulia | madhuri[m]: no, like openstack baremetal node undeploy | 14:43 |
| TheJulia | jaypipes: Interesting, I think the perception was that there was a resulting hard 1-1 mapping | 14:44 |
| TheJulia | Anyway, we've rabbitholed on raid | 14:44 |
| jaypipes | TheJulia: in fact, I'd love to get out of the flavor and image explosion that occurs because of the (currently very tight) coupling in Ironic and Nova between configuration data and the image bits themselves. | 14:44 |
| dtantsur | with RAID we anyway need some kind of templates on ironic side | 14:45 |
| madhuri[m] | TheJulia: so why do we verify PXE configuration, DHCP options, check images while adoption? | 14:45 |
| dtantsur | it's simply impractical to make Nova, Glance or whatever the hell also responsible for validating our RAID stuff | 14:45 |
| TheJulia | dtantsur: with the links jay provided, I don't think so | 14:45 |
| dtantsur | TheJulia: even we can it does not necessary mean we should | 14:45 |
| NobodyCam | Good Morning Ironic'ers | 14:46 |
| dtantsur | our RAID template is a complex schema, we'd better maintain it ourself | 14:46 |
| NobodyCam | and ofc... | 14:46 |
| dtantsur | morning NobodyCam :) | 14:46 |
| NobodyCam | TGIF! | 14:46 |
| jaypipes | dtantsur: you would define a metadef schema that describes a raid configuration. Ironic would be responsible for defining that schema. | 14:46 |
| NobodyCam | Morning dtantsur | 14:46 |
| TheJulia | dtantsur: s/ourselves/examples ourselves/ right? | 14:46 |
| NobodyCam | wow, hey long time no see jaypipes | 14:46 |
| TheJulia | operational needs can hugely differ from our initial perceptions | 14:46 |
| jaypipes | NobodyCam: hey there Chris :) | 14:46 |
| NobodyCam | :) | 14:46 |
| TheJulia | madhuri[m]: because we want to ensure that the node is possible to be managed | 14:46 |
| dtantsur | jaypipes: do you assume ironic talking to glance to provide schemas or an operator preloading them to glance? | 14:47 |
| jaypipes | dtantsur: preloading. | 14:47 |
| TheJulia | jaypipes: I hate to ask this, but could you elaborate a little more, I'm trying to understand your perception of tight coupling so I can better understand your entire prior statement because my context is different | 14:48 |
| jaypipes | dtantsur: both the schema as well as particular metadata *items*. the schema defines the item structure, but the items themselves would describe different, e.g. RAID configs or disk partitioning configs | 14:48 |
| madhuri[m] | TheJulia: what actions do you mean by manage here? We can't provision it so no deployment | 14:49 |
| dtantsur | okay, technically it will work. logically we're making the Image (!) service responsible for things like RAID or BIOS configuration. | 14:49 |
| dtantsur | it does not seem quite right to me | 14:49 |
| TheJulia | madhuri[m]: but to take over management of the node, which is the purpose of the adoption feature, we need to be able to clean the node and move it to an available state | 14:50 |
| openstackgerrit | yolanda.robla proposed openstack/ironic master: Add automated_clean field to the API https://review.openstack.org/585991 | 14:50 |
| TheJulia | dtantsur: standalone deployments would not be super pleased unless we still translated it so it was directly leveragable | 14:50 |
| *** gabys has joined #openstack-ironic | 14:50 | |
| TheJulia | and we've already offered itup via our api | 14:51 |
| TheJulia | because of all the discussion that a VT trait being required should toggle the VT bios setting if possible | 14:52 |
| jaypipes | TheJulia: right now, the flavor and the image is the only way the nova user has to indicate both scheduling constraints as well as instance configuration goop to the virt driver. This means that, *especially* for Ironic, deployers end up creating dozens of virtually-identical flavors and images that differ only slightly in things like disk partitioning or RAID setup. So, instead of having the ability to say "launch me a 12 CPU, 128G RAM | 14:52 |
| jaypipes | machine, and use this disk partitioning schema", instead we have 43 different flavors that are all pretty much the same but with slightly different configuration details. What I'm suggesting is we remove the instance configuration pieces from the flavor and image extra specs and place them into separate "config data" referenceable metadata items so we can reduce the number and complexity of flavors and also allow this structured configuration | 14:52 |
| jaypipes | data to be sent to the virt driver before launching the instance | 14:52 |
| madhuri[m] | TheJulia: sorry it is bit confusing for me. You said an adopted node is not provisioned or also not made available to Nova. So why an adopted node is made available? | 14:53 |
| madhuri[m] | TheJulia: do you want me to ping later? I can get back later after the ongoing discussion is over. | 14:53 |
| dtantsur | jaypipes: this sounds good, but why do you think the Image service is the right place for it? | 14:53 |
| TheJulia | madhuri[m]: An adopted node is _not_ managed by nova | 14:54 |
| *** munimeha1 has joined #openstack-ironic | 14:54 | |
| TheJulia | it is managed by ironic, and only made available for use by nova after it is deleted, thus putting it into an available state in ironic. | 14:54 |
| TheJulia | madhuri[m]: at which time, nova sees the available node, and can schedule resources to it | 14:54 |
| TheJulia | s/resources/an instance/ | 14:55 |
| *** gabys has quit IRC | 14:55 | |
| jaypipes | dtantsur: it's the "metadata service" in this case. Glance has this API. We should use it, IMHO. | 14:56 |
| TheJulia | jaypipes: oh what the conundrum we've built :( | 14:56 |
| jaypipes | dtantsur: or, you could just rewrite nova to be specific to ironic. but I think you've already started that effort? | 14:56 |
| dtantsur | jaypipes: no, I have not | 14:57 |
| dtantsur | and I'm not sure why you are picking on me. I'm honestly trying to understand your solution and its implication. | 14:57 |
| dtantsur | it is me who will get questions like "I need to put RAID configuration into Glance? Why?" | 14:57 |
| *** e0ne has quit IRC | 14:58 | |
| *** skazi has joined #openstack-ironic | 14:58 | |
| TheJulia | so, if we can ignore raid for a minute, it feels like our deploy templates approach, at least from the handling of traits scheduled upon a node to imply what a user desires in terms of overall high level state of the system. (not finely detailed things like... raid configuration) | 14:58 |
| *** Bhujay has quit IRC | 14:58 | |
| madhuri[m] | TheJulia: ok it is clear now. Thank you :) | 14:59 |
| jaypipes | TheJulia: I'm parsing that last sentence, hold on... | 15:00 |
| TheJulia | jaypipes: no rush, I've got my team's standup for the next half hour to 45 minutes | 15:00 |
| *** e0ne has joined #openstack-ironic | 15:02 | |
| jaypipes | TheJulia: maybe it's better for me to just write up a competing deploy templates spec that uses Glance metadefs instead of the scheduler/placement traits. | 15:04 |
| jaypipes | TheJulia: my issue is I'm not going to do that if there's just built-in resistance in the Ironic contributor team to using Glance for anything. | 15:04 |
| jaypipes | TheJulia: I know traditionally there's been lots of resistance in Ironic to using really any other component in OpenStack as a dependency. | 15:05 |
| *** sthussey has joined #openstack-ironic | 15:05 | |
| TheJulia | jaypipes: If compelling and makes sense, We would either need to provide compatability for not using glance, but I think we could only kind of figure that out once we have a solid understanding of the mechanism your proposing | 15:06 |
| *** jaypipes is now known as leakypipes | 15:06 | |
| leakypipes | TheJulia: ok. I'll type up an ironic-spec. | 15:07 |
| dtantsur | yep, there must be a way to use it without glance, but an optional dependency is totally fine | 15:07 |
| TheJulia | dtantsur: ++ | 15:07 |
| TheJulia | leakypipes: thanks | 15:07 |
| dtantsur | just to clarify: if without glance means without nova too, this is fine | 15:07 |
| rloo | leakypipes, TheJulia, dtantsur: are you talking about what we discussed in rocky ptg: https://etherpad.openstack.org/p/nova-ptg-rocky, line 250ish? | 15:07 |
| leakypipes | dtantsur, TheJulia: see, that's exactly the problem... | 15:08 |
| leakypipes | dtantsur, TheJulia: there's not going to be "a way to use it without glance". | 15:08 |
| dtantsur | leakypipes: you can say, if you want nova, you must use glance. this is fine. | 15:08 |
| TheJulia | there absolutely has to be, it might just be slightly painful and I think that is fine | 15:08 |
| dtantsur | leakypipes: what is not fine is: even for standalone ironic you must use glance | 15:08 |
| rloo | (where glare was proposed) | 15:08 |
| leakypipes | dtantsur, TheJulia: the only thing I can offer is just "if you don't want this functionality, then you don't need to have it, but there is only a Glance implementation for it" | 15:08 |
| leakypipes | dtantsur: if you use nova, you already need to use glance. | 15:09 |
| TheJulia | leakypipes: sounds good to me | 15:09 |
| leakypipes | TheJulia: ok. it sounded like I was being asked to provide some non-Glance implementation for deploy templates. | 15:09 |
| TheJulia | leakypipes: oh absolutely not | 15:09 |
| leakypipes | ok | 15:09 |
| TheJulia | if anything, we can help bridge that gap | 15:10 |
| leakypipes | rloo: and yes, exactly. | 15:10 |
| dtantsur | I think not having deploy templates standalone may be fine. In standalone case people can request e.g. RAID directly via the provisioning API (with deploy steps in place) | 15:10 |
| rloo | leakypipes: then we had decided to do that (although you're saying glance, not glare) | 15:10 |
| *** gabys has joined #openstack-ironic | 15:10 | |
| rloo | glad we all agree :) | 15:11 |
| leakypipes | rloo: there is no more glare, AFAIK. but it's the glance metadefs functionality I was referring to. | 15:11 |
| TheJulia | dtantsur: and through deploy steps, they should be able to epxlicitly state what they want | 15:11 |
| dtantsur | correct | 15:11 |
| rloo | leakypipes: yup | 15:11 |
| TheJulia | they just need step visibility | 15:11 |
| * TheJulia will pickup that spec again today | 15:11 | |
| dtantsur | TheJulia++ | 15:11 |
| * TheJulia will run errands first, then come home, get dressed for the gala, and then hack on the spec with something tasty | 15:12 | |
| dhellmann | dtantsur : I'll take a look at that job failure after the release team meeting | 15:12 |
| dtantsur | thanks dhellmann | 15:12 |
| TheJulia | NobodyCam: I'll be getting a lyft there most likely | 15:12 |
| *** gabys has quit IRC | 15:15 | |
| *** mgoddard has quit IRC | 15:16 | |
| dtantsur | TheJulia: btw! congrats (or condolences? ;) on keeping your TC chair! | 15:18 |
| TheJulia | dtantsur: thanks... I think :) | 15:18 |
| coreywright | any/all: tl;dr are ipa-tempest-dsvm-* tests in zuul known to be unstable/temperamental? my https://review.openstack.org/600139 initially passed, i add a release note, tests start failing, i request a recheck, and now only one test fails. i've reviewed the test logs, but i'm probably overlooking a log because all i can see is that failing test times out waiting for the node to finish cleaning (though i can see | 15:20 |
| coreywright | the custom IPA image with my change being built and being downloaded by the ironic node). | 15:20 |
| patchbot | patch 600139 - ironic-python-agent - Allow erasing metadata from disk partitions - 2 patch sets | 15:20 |
| TheJulia | coreywright: CI has been moderately unstable and the CoreOS jobs run kind of long | 15:20 |
| TheJulia | because of the size of the ramdisk that has to be decompressed | 15:21 |
| TheJulia | We likely need to adjust timings | 15:21 |
| jroll | holy cow it's coreywright \o/ | 15:21 |
| coreywright | yeah, it was coreos (vs tinyipa) | 15:21 |
| coreywright | jroll: :D | 15:21 |
| TheJulia | coreywright: I can try to take a look at timings, but I recently (as in with-in the last 3-6 months that I modified the times | 15:22 |
| coreywright | jroll: your name did not go unnoticed in the channel's membership list, but i didn't figure you wanted to chit-chat on this channel. (i'll PM you) | 15:22 |
| TheJulia | coreywright: people are welcome to chat about unrelated things in channel, fwiw | 15:23 |
| *** gabys has joined #openstack-ironic | 15:23 | |
| * TheJulia notes that by doing so, you become part of the family :) | 15:23 | |
| jroll | but yeah, for timeouts generally a re-check is fine. though you're changing cleaning and it's cleaning that timed out, so I won't 100% it isn't the patch | 15:23 |
| jroll | ++ to both of you :P | 15:23 |
| TheJulia | we need to try and keep the rechecks down, the gate queue has been huge this week. | 15:23 |
| jroll | TheJulia: we may not want to disparage^W talk about how great our employers are in channel, though :) | 15:24 |
| TheJulia | oh, so true | 15:24 |
| TheJulia | :) | 15:24 |
| coreywright | jroll: i wouldn't know what you are talking about / hinting at ;) | 15:24 |
| jroll | :) | 15:25 |
| jroll | coreywright: feel free to PM me at your leisure, if that was an ask and not a tell | 15:25 |
| coreywright | jroll: because of the clean timing out and that being my change, i reviewed the logs, but didn't find anything (eg ipmi console with IPA output), but probably/hopefully just my oversight. | 15:26 |
| jroll | coreywright: yeah, probably just caught a slow CI node. we do have console logs here: http://logs.openstack.org/39/600139/2/check/ipa-tempest-dsvm-partition-bios-ipmi-direct-coreos-src/6a1b2e2/logs/ironic-bm-logs/ | 15:27 |
| *** gabys has quit IRC | 15:27 | |
| jroll | since it's blank it likely never got that far | 15:27 |
| jroll | usually there's one per boot, like so: http://logs.openstack.org/39/600139/2/check/ipa-tempest-dsvm-partition-bios-ipmi-direct-tinyipa-src/2157763/logs/ironic-bm-logs/ | 15:27 |
| jroll | (passing job) | 15:27 |
| coreywright | jroll: yeah, i glanced at ironic-bm-logs but didn't hold out much hope because it says: "ironic-bm-logs/: output from the last successful boot of an ironic "bare metal" VM" and it wasn't a successful, so nothing expected | 15:30 |
| jroll | coreywright: that must be old text, it now captures all boots, up to the point of success or failure | 15:34 |
| jroll | (failure may include a test timeout (hence shutdown) before anything was written to console) | 15:36 |
| coreywright | jroll: gotcha | 15:36 |
| TheJulia | looks like it never heartbeated which means that either coreos didn't boot or IPA never got started in the time window | 15:39 |
| TheJulia | That CoreOS image can take 8-9 minutes to reach the point where services can be started on slow CI nodes | 15:39 |
| jroll | console was blank, so I assume we never even made it to coreos trying to boot | 15:41 |
| TheJulia | not even ipxe attempt which is weird | 15:43 |
| TheJulia | yeah, never even booted | 15:46 |
| TheJulia | the packet counts in the ovs logs are super low | 15:46 |
| TheJulia | s/logs/counters | 15:46 |
| *** lenka has quit IRC | 15:46 | |
| *** mgoddard has joined #openstack-ironic | 15:47 | |
| *** gabys has joined #openstack-ironic | 15:48 | |
| TheJulia | It is using virtio | 15:49 |
| TheJulia | derekh: got a link to the workaround or fix you used? | 15:49 |
| *** JudeCross has joined #openstack-ironic | 15:49 | |
| TheJulia | tl;dr ipxe may have fired up, but the machine never tried to start booting | 15:50 |
| coreywright | TheJulia: ah; thanks | 15:51 |
| derekh | In CI last week we had to downdrade the ipxe package | 15:51 |
| derekh | I'm now hoping to remove the downgrade and change to e1000 https://review.openstack.org/#/c/604506/ | 15:51 |
| patchbot | patch 604506 - ironic - Switch grenade nic driver to e1000 - 2 patch sets | 15:51 |
| *** gabys has quit IRC | 15:53 | |
| derekh | basically ipxe wan't sending out dhcp requests at all | 15:54 |
| derekh | *wasn't | 15:54 |
| TheJulia | woot | 16:01 |
| *** S4ren has joined #openstack-ironic | 16:02 | |
| *** dnuka has joined #openstack-ironic | 16:04 | |
| *** S4ren has quit IRC | 16:08 | |
| *** gabys has joined #openstack-ironic | 16:10 | |
| openstackgerrit | Dhanuka proposed openstack/sushy master: Author: [dnuka] https://review.openstack.org/606125 | 16:11 |
| *** S4ren has joined #openstack-ironic | 16:13 | |
| *** gabys has quit IRC | 16:14 | |
| *** JudeCross has quit IRC | 16:14 | |
| openstackgerrit | Julia Kreger proposed openstack/ironic-python-agent master: Change nic driver to e1000 for CoreOS jobs https://review.openstack.org/606127 | 16:15 |
| TheJulia | coreywright: lets see if ^^^^ works | 16:15 |
| openstackgerrit | Dhanuka proposed openstack/sushy master: Add product property to the root Sushy object https://review.openstack.org/606125 | 16:17 |
| *** gyee has joined #openstack-ironic | 16:24 | |
| *** baha has quit IRC | 16:26 | |
| *** gabys has joined #openstack-ironic | 16:29 | |
| *** gabys has quit IRC | 16:33 | |
| TheJulia | running to the store, bbiab | 16:36 |
| TheJulia | derekh: I'm tempted that we need to re-adjust some of the job timeouts in the ironic repo as well, it just feels like that slight window of variability is hitting a little too often | 16:37 |
| *** ianychoi has quit IRC | 16:39 | |
| derekh | TheJulia: ya, possibly, I wonder could we use logstash to find timeouts, I can take a look monday if you want | 16:39 |
| TheJulia | yeah, that is likely a good idea, but I think a common timeout is the actual nova test timeout, and maybe that should be waiting slightly longer | 16:40 |
| * TheJulia runs into town | 16:41 | |
| derekh | its worth a try | 16:41 |
| *** gkadam has quit IRC | 16:42 | |
| *** munimeha1 has quit IRC | 16:43 | |
| *** ianychoi has joined #openstack-ironic | 16:44 | |
| dtantsur | have a great weekend folks | 16:45 |
| *** dtantsur is now known as dtantsur|afk | 16:45 | |
| *** dnuka has quit IRC | 16:46 | |
| *** gabys has joined #openstack-ironic | 16:53 | |
| *** priteau has joined #openstack-ironic | 16:54 | |
| *** derekh has quit IRC | 16:55 | |
| *** gabys has quit IRC | 16:57 | |
| *** dougsz has quit IRC | 17:00 | |
| mgoddard | leakypipes TheJulia dtantsur|afk: looks like I missed all the fun! I was getting a much needed haircut | 17:03 |
| mgoddard | leakypipes: look forward to reading your spec. In the meantime I'll brush up on glance | 17:04 |
| *** gabys has joined #openstack-ironic | 17:08 | |
| *** trown is now known as trown|lunch | 17:11 | |
| *** gabys has quit IRC | 17:13 | |
| *** baha has joined #openstack-ironic | 17:18 | |
| NobodyCam | morning mgoddard | 17:22 |
| *** jtomasek has joined #openstack-ironic | 17:24 | |
| *** gabys has joined #openstack-ironic | 17:26 | |
| * TheJulia takes a step back and wonders about starting over on the step visibility | 17:31 | |
| *** gabys has quit IRC | 17:31 | |
| *** tssurya has quit IRC | 17:31 | |
| mgoddard | morning NobodyCam | 17:33 |
| TheJulia | /me starts with a blank spec template | 17:37 |
| *** S4ren has quit IRC | 17:38 | |
| *** e0ne has quit IRC | 17:44 | |
| *** dsneddon has joined #openstack-ironic | 17:47 | |
| *** dsneddon has quit IRC | 17:47 | |
| *** dsneddon has joined #openstack-ironic | 17:49 | |
| *** gabys has joined #openstack-ironic | 17:50 | |
| *** JudeCross has joined #openstack-ironic | 17:53 | |
| *** gabys has quit IRC | 17:54 | |
| leakypipes | mgoddard: well, at least you have hair to get a haircut. | 18:01 |
| *** e0ne has joined #openstack-ironic | 18:02 | |
| *** TheJulia is now known as needssleep | 18:02 | |
| * needssleep should have slept in | 18:02 | |
| *** e0ne has quit IRC | 18:05 | |
| *** gabys has joined #openstack-ironic | 18:06 | |
| *** jcoufal has quit IRC | 18:08 | |
| openstackgerrit | Merged openstack/sushy master: Update to use mapped field for Settings Message Severity field https://review.openstack.org/591612 | 18:12 |
| mgoddard | leakypipes: for now. All my male relatives are balding so it's just a matter of time | 18:15 |
| leakypipes | mgoddard: :) it's only your Mum's father that matters. | 18:16 |
| mgoddard | leakypipes: oh dear... | 18:18 |
| needssleep | mgoddard: there is medication to help retain hair though. Some side effects exist though. | 18:19 |
| mgoddard | needssleep: I've accepted it as an inevitability. Would be interested to see solid clinical results of hair loss prevention medication | 18:23 |
| leakypipes | mgoddard: that's the spirit, good man! | 18:24 |
| leakypipes | (accepting the inevitable ;) | 18:24 |
| mgoddard | needssleep: seems like an easy insecurity for pharma to prey on, but perhaps I'm just cynical | 18:24 |
| needssleep | There are many studies on the subject of DHT constricting hair follicles. | 18:25 |
| *** e0ne has joined #openstack-ironic | 18:25 | |
| needssleep | mgoddard: cynical is good! | 18:25 |
| *** gabys has quit IRC | 18:33 | |
| *** trown|lunch is now known as trown | 18:37 | |
| mgoddard | night all | 18:42 |
| needssleep | goodnight | 18:42 |
| AJaeger | woot! python3-first done for ironic - https://review.openstack.org/605271 | 18:46 |
| patchbot | patch 605271 - nova (stable/queens) - Skip ServerShowV247Test.test_update_rebuild_list_s... (MERGED) - 1 patch set | 18:46 |
| AJaeger | patchbot: you're confused ;) | 18:47 |
| *** mgoddard has quit IRC | 18:47 | |
| needssleep | patchbot is... well... patchbot | 18:47 |
| AJaeger | No, I was confused - sorry, https://review.openstack.org/592472 is the final one that just merged | 18:47 |
| patchbot | patch 592472 - openstack-infra/project-config - remove job settings for ironic repositories (MERGED) - 3 patch sets | 18:47 |
| AJaeger | my apologies, patchbot | 18:48 |
| *** AJaeger has left #openstack-ironic | 18:49 | |
| *** gabys has joined #openstack-ironic | 18:49 | |
| *** e0ne has quit IRC | 18:50 | |
| *** gabys has quit IRC | 18:54 | |
| *** gabys has joined #openstack-ironic | 19:12 | |
| *** Chaserjim has joined #openstack-ironic | 19:16 | |
| *** tssurya has joined #openstack-ironic | 19:22 | |
| * needssleep wonders if patchbot is actually a cat | 19:22 | |
| *** e0ne has joined #openstack-ironic | 19:29 | |
| *** e0ne has quit IRC | 19:34 | |
| jroll | I don't think I've seen patchbot treating people like a tree or yelling for no reason | 19:42 |
| * needssleep wonders where yelling came from | 20:01 | |
| needssleep | autocorrect? | 20:01 |
| jroll | my cat yells a lot, dunno about yours :P | 20:01 |
| needssleep | jroll: sanity check needed: which would be better, a list of all steps (cleaning/deploy) or two separate lists? | 20:02 |
| jroll | I AM HUNGRY FEED ME NOW, etc | 20:02 |
| jroll | needssleep: separate imo | 20:02 |
| needssleep | even with the eventual cross-over? or were we going to just dupe the decorators? | 20:02 |
| * needssleep tries to remember | 20:02 | |
| jroll | the latter | 20:02 |
| needssleep | yeah, that sounds right | 20:03 |
| needssleep | yeah, different default priorities in those sorts of cases if present | 20:03 |
| jroll | needssleep: though another option is one API for both, with options to filter. e.g. /.../steps/[?type=clean|deploy] | 20:03 |
| needssleep | jroll: I was thinking all in one api call just to keep the call count down and simplify the interaction | 20:03 |
| jroll | right, the type would be optional | 20:04 |
| needssleep | Because we're going to need to sift through it all conductor side | 20:04 |
| jroll | but if I'm furiously trying to do some manual cleaning, I don't want to filter the steps by hand | 20:04 |
| jroll | so would be nice to be able to get only clean steps | 20:04 |
| needssleep | I'll go ahead and put the type on the endpoint then | 20:06 |
| needssleep | and just default to both | 20:06 |
| jroll | perfect, ty | 20:06 |
| needssleep | I'm putting this all in one spec and trying to be minimalistic, I hope it makes sense | 20:07 |
| *** openstackgerrit has quit IRC | 20:07 | |
| *** Chaserjim has quit IRC | 20:10 | |
| *** mjturek has quit IRC | 20:18 | |
| *** baha has quit IRC | 20:18 | |
| *** lenka has joined #openstack-ironic | 20:26 | |
| *** openstackgerrit has joined #openstack-ironic | 20:53 | |
| openstackgerrit | Julia Kreger proposed openstack/ironic-specs master: A fresh way of looking at step retrieval https://review.openstack.org/606199 | 20:53 |
| needssleep | jroll: ^^^^ | 20:53 |
| jroll | ++ | 20:53 |
| jroll | needssleep: I refuse to read specs this late on a friday though :) | 20:54 |
| needssleep | but all specs should be written on fridays | 20:54 |
| jroll | totes | 20:54 |
| *** mmethot has quit IRC | 21:00 | |
| *** mmethot has joined #openstack-ironic | 21:00 | |
| *** mmethot has quit IRC | 21:05 | |
| *** mgoddard has joined #openstack-ironic | 21:21 | |
| *** dsneddon has quit IRC | 21:22 | |
| *** dsneddon has joined #openstack-ironic | 21:23 | |
| *** tssurya has quit IRC | 21:26 | |
| *** priteau has quit IRC | 21:42 | |
| *** mgoddard has quit IRC | 21:43 | |
| JayF | needssleep: jroll: I /really/ like the idea of a dedicated "just boot the agent and leave it" state. Especially if as part of that you make the code smart enough to use the "warm" agent on a holding->clean or holding->active transition. | 22:31 |
| *** skazi has quit IRC | 22:43 | |
| *** ijw has joined #openstack-ironic | 22:48 | |
| needssleep | Oh yeah, the state machine will allow us to do that easily, the nova code is also kind of trivial | 22:54 |
| needssleep | I think... | 22:54 |
| needssleep | that is always a dangerous thing to say | 22:54 |
| *** lenka has quit IRC | 23:08 | |
| *** sthussey has quit IRC | 23:27 | |
| *** gyee has quit IRC | 23:33 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!