| *** openstack has joined #openstack-ironic | 05:59 | |
| *** mbound has quit IRC | 05:59 | |
| *** z- has joined #openstack-ironic | 06:01 | |
| *** lintan has quit IRC | 06:03 | |
| *** alex_xu has quit IRC | 06:03 | |
| *** lintan has joined #openstack-ironic | 06:03 | |
| *** alex_xu has joined #openstack-ironic | 06:06 | |
| *** bcornec2 has quit IRC | 06:15 | |
| *** yolanda has joined #openstack-ironic | 06:18 | |
| *** yolanda has quit IRC | 06:21 | |
| *** yolanda has joined #openstack-ironic | 06:24 | |
| *** rcernin has joined #openstack-ironic | 06:24 | |
| openstackgerrit | Yuiko Takada Mori proposed openstack/ironic: IPMINative: add NativeIPMISocatConsole class https://review.openstack.org/293874 | 06:27 |
|---|---|---|
| *** jcoufal has quit IRC | 06:36 | |
| *** rama_y has joined #openstack-ironic | 06:38 | |
| *** moshele has joined #openstack-ironic | 06:43 | |
| *** _vsaienko has joined #openstack-ironic | 06:45 | |
| *** bcornec1 has joined #openstack-ironic | 06:47 | |
| *** jaybeale has joined #openstack-ironic | 06:48 | |
| *** Sukhdev has quit IRC | 06:51 | |
| *** hoangcx has quit IRC | 06:51 | |
| *** jaybeale has quit IRC | 06:53 | |
| *** hoangcx has joined #openstack-ironic | 06:54 | |
| *** bcornec1 has quit IRC | 06:56 | |
| *** bcornec1 has joined #openstack-ironic | 07:01 | |
| *** tesseract has joined #openstack-ironic | 07:09 | |
| *** bharathk has quit IRC | 07:16 | |
| *** klindgren has quit IRC | 07:17 | |
| *** bharathk has joined #openstack-ironic | 07:17 | |
| *** pcaruana has joined #openstack-ironic | 07:17 | |
| *** klindgren has joined #openstack-ironic | 07:17 | |
| *** ifarkas has joined #openstack-ironic | 07:19 | |
| *** rama_y has quit IRC | 07:22 | |
| *** bcornec1 has quit IRC | 07:32 | |
| *** ohamada has joined #openstack-ironic | 07:33 | |
| *** ccamacho has joined #openstack-ironic | 07:35 | |
| *** bcornec1 has joined #openstack-ironic | 07:36 | |
| *** praneshp has quit IRC | 07:39 | |
| *** bcornec1 has quit IRC | 07:48 | |
| *** dtantsur|afk is now known as dtantsur | 07:50 | |
| dtantsur | Morning Ironic | 07:50 |
| *** yuriyz|2 has joined #openstack-ironic | 07:51 | |
| openstackgerrit | Yuiko Takada Mori proposed openstack/ironic: IPMITool: add IPMISocatConsole class https://review.openstack.org/293873 | 07:52 |
| *** hoangcx_ has joined #openstack-ironic | 07:54 | |
| *** hoangcx has quit IRC | 07:55 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** zzzeek has joined #openstack-ironic | 08:00 | |
| *** hoangcx_ is now known as hoangcx | 08:04 | |
| *** athomas has joined #openstack-ironic | 08:07 | |
| openstackgerrit | Yuiko Takada Mori proposed openstack/ironic: IPMITool: add IPMISocatConsole class https://review.openstack.org/293873 | 08:09 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/ironic: [WIP] Testing latest u-c https://review.openstack.org/318440 | 08:10 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/ironic: [WIP] Testing latest u-c https://review.openstack.org/318440 | 08:10 |
| *** e0ne has joined #openstack-ironic | 08:13 | |
| *** Nisha_away has joined #openstack-ironic | 08:20 | |
| *** lucas-afk is now known as lucasagomes | 08:28 | |
| lucasagomes | morning all :-) | 08:28 |
| openstackgerrit | Milan Kováčik proposed openstack/ironic-inspector: [WIP]Introducing Inspector Grenade testing https://review.openstack.org/327667 | 08:32 |
| milan | morning dtantsur lucasagomes and all :) | 08:33 |
| lucasagomes | :D | 08:33 |
| dtantsur | morning lucasagomes, welcome back | 08:37 |
| Nisha_away | morning lucasagomes dtantsur and all!!! | 08:40 |
| lucasagomes | o/ | 08:41 |
| dtantsur | hey Nisha_away | 08:43 |
| Nisha_away | dtantsur, i had some queries regarding this spec | 08:45 |
| Nisha_away | https://github.com/openstack/ironic-inspector-specs/blob/master/specs/generic-pci-resource.rst | 08:45 |
| dtantsur | Nisha_away, sure (though it might be better to address the to the author - xek) | 08:45 |
| Nisha_away | hwo is it supposed to work through ironic? ironic and nova-ironic-virt-driver understands only capabilities as extra_spec. How will pci_passthrough filter work for ironic nodes | 08:46 |
| *** derekh has joined #openstack-ironic | 08:47 | |
| Nisha_away | as per spec it populates it as list | 08:47 |
| Nisha_away | while capabilities in ironic are populated as string | 08:47 |
| *** jaybeale has joined #openstack-ironic | 08:48 | |
| Nisha_away | that spec is already merged....so i dont know where to ask or comment | 08:48 |
| dtantsur | Nisha_away, "How will pci_passthrough filter work for ironic nodes" it won't; "as per spec it populates it as list" no quite so | 08:50 |
| dtantsur | Nisha_away, the spec talks about usual capabilities like pci_cpu:4, what's the problem with it | 08:50 |
| Nisha_away | hw_info['pci_devices'] = [{'vendor_id': '8086', 'product_id': '0412'}, | 08:51 |
| Nisha_away | {'vendor_id': '8086', 'product_id': '2955'}] | 08:51 |
| Nisha_away | this is a list, correct | 08:51 |
| *** jaybeale has quit IRC | 08:52 | |
| dtantsur | Nisha_away, yes, but how is it related to ironic, nova or anything around inspector? | 08:53 |
| Nisha_away | inspector isnt going to populate in ironic as this? | 08:54 |
| Nisha_away | or there is one more spec in ironic for this? | 08:54 |
| dtantsur | Nisha_away, 1. no, 2. no | 08:54 |
| dtantsur | iirc (clarify with xek) if you set in inspector.conf: pci_alias={"vendor_id":"8086", "product_id":"1520", "name":"pci_gpu"} | 08:55 |
| dtantsur | then inspector will set this pci_gpu to a number | 08:55 |
| dtantsur | Nisha_away, ^^^ | 08:55 |
| Nisha_away | dtantsur, do u know what timezone is xek available? | 08:56 |
| Nisha_away | and his IRC name | 08:56 |
| Nisha_away | dtantsur, i dont get the details from the spec that how is it suppose to be working, or what inspector is going to populate. | 08:57 |
| dtantsur | Nisha_away, xek is an irc name, I think he'll be around soon | 08:57 |
| Nisha_away | dtantsur, ok, thanks. | 08:57 |
| Nisha_away | dtantsur, could you review https://review.openstack.org/201904 | 08:57 |
| dtantsur | will take a look once I get back (need to relocate to the office) | 08:58 |
| Nisha_away | dtantsur, thanks | 08:58 |
| *** mgould|afk is now known as mgould | 08:59 | |
| Nisha_away | lucasagomes, sambetts|afk and others could you also review https://review.openstack.org/201904 | 08:59 |
| *** dtantsur is now known as dtantsur|brb | 09:00 | |
| *** amotoki has quit IRC | 09:01 | |
| *** Nisha_brb has joined #openstack-ironic | 09:02 | |
| *** dmk0202 has joined #openstack-ironic | 09:02 | |
| lucasagomes | Nisha_away, added to the list, i just came back from holidays so I have a lot to catch up :-/ | 09:04 |
| openstackgerrit | lokesh s proposed openstack/ironic: Add support for the audit middleware https://review.openstack.org/272658 | 09:04 |
| *** Nisha_away has quit IRC | 09:05 | |
| *** ifarkas has quit IRC | 09:06 | |
| xek | Nisha_brb, Hi | 09:14 |
| xek | Nisha_brb, it will work very similar to the capabilities in the iLO driver, but the name will be configurable | 09:15 |
| xek | Nisha_brb, see here http://docs.openstack.org/developer/ironic/deploy/inspection.html#capabilities-discovery | 09:15 |
| xek | Nisha_brb, it will work like the "pci_gpu_devices" capability from the docs | 09:16 |
| *** milan has quit IRC | 09:20 | |
| *** electrofelix has joined #openstack-ironic | 09:27 | |
| Nisha_brb | xek, hi | 09:30 |
| *** tloeb has joined #openstack-ironic | 09:34 | |
| openstackgerrit | Bharath kumar proposed openstack/ironic: adding back the inspect_ports as an optional parameter https://review.openstack.org/325230 | 09:43 |
| Nisha_brb | xek, hello | 09:53 |
| smoriya | morning ironic | 09:55 |
| smoriya | I have a question about ports REST API - 'GET /v1/ports' | 09:56 |
| *** mbound has joined #openstack-ironic | 09:56 | |
| *** ifarkas has joined #openstack-ironic | 09:57 | |
| smoriya | why it take both node and node_uuid parameter? we can pass node=<node_uuid> | 09:58 |
| *** milan has joined #openstack-ironic | 10:00 | |
| lucasagomes | smoriya, hi there, for historical reasons | 10:04 |
| lucasagomes | smoriya, the node parameter came when the support for naming a node was added | 10:04 |
| lucasagomes | prior to that, nodes only used UUID, therefore the parameter was called node_uuid | 10:05 |
| lucasagomes | since it's in the API we can't remove the support for that parameters, so it's a bit duplicated | 10:05 |
| smoriya | lucasgomes, I got it. Thanks! | 10:05 |
| lucasagomes | node accepts UUID or name, and node_uuid only UUID | 10:05 |
| lucasagomes | smoriya, np :-) | 10:05 |
| *** Nisha_brb is now known as Nisha1 | 10:07 | |
| *** mgoddard_ has joined #openstack-ironic | 10:08 | |
| vdrok | morning ironic, dtantsur|brb and lucasagomes ! | 10:08 |
| lucasagomes | vdrok, hey there | 10:08 |
| sambetts|afk | Morning all | 10:09 |
| *** sambetts|afk is now known as sambetts | 10:09 | |
| *** mgoddard has quit IRC | 10:11 | |
| *** dtantsur|brb is now known as dtantsur | 10:13 | |
| vdrok | morning sambetts | 10:14 |
| dtantsur | morning sambetts, vdrok | 10:18 |
| *** sirushti has quit IRC | 10:26 | |
| *** jtomasek_ has joined #openstack-ironic | 10:30 | |
| *** sirushti has joined #openstack-ironic | 10:33 | |
| sambetts | o/ vdrok, dtantsur, lucasagomes | 10:34 |
| *** hoangcx has quit IRC | 10:36 | |
| *** sdake has joined #openstack-ironic | 10:40 | |
| *** e0ne has quit IRC | 10:41 | |
| *** e0ne has joined #openstack-ironic | 10:41 | |
| *** wagiel has joined #openstack-ironic | 10:42 | |
| *** sdake_ has joined #openstack-ironic | 10:42 | |
| openstackgerrit | edan david proposed openstack/ironic-python-agent: Tests should use: self.assertIn(value, list) self.assertNotIn(value, list) https://review.openstack.org/330451 | 10:43 |
| *** sdake has quit IRC | 10:45 | |
| *** Nisha1 has quit IRC | 10:48 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Properly set ephemeral size in agent drivers https://review.openstack.org/330089 | 10:48 |
| *** livelace has joined #openstack-ironic | 10:50 | |
| *** sborkows has joined #openstack-ironic | 10:51 | |
| livelace | Hello. What is mean "ironic node-update $NODE_UUID" ? | 10:52 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Change the logic of selecting image for tests https://review.openstack.org/329625 | 10:52 |
| livelace | Which of nodes ? | 10:52 |
| openstackgerrit | edan david proposed openstack/ironic-python-agent: Tests should use: self.assertIn(value, list) self.assertNotIn(value, list) https://review.openstack.org/330451 | 10:53 |
| lucasagomes | livelace, $NODE_UUID is the variable holding the UUID of the node | 10:53 |
| lucasagomes | e.g ironic node-update 6f68d2bd-513c-4a58-92a0-a978f0a80d04 | 10:53 |
| livelace | lucasagomes, Which node ? I have nova node too :) | 10:53 |
| lucasagomes | livelace, that's just to make scripting easy (not hardcoded) | 10:54 |
| lucasagomes | livelace, node in ironic | 10:54 |
| livelace | lucasagomes, I have to create this node ? | 10:54 |
| lucasagomes | livelace, yes, you have to have your baremetal nodes registered in the ironic database in order to deploy them via Ironic | 10:55 |
| livelace | lucasagomes, Nova node with [ironic] section with settings, right ? | 10:56 |
| *** ccamacho is now known as ccamacho|lunch | 10:56 | |
| lucasagomes | livelace, not sure if I follow... I assume you're trying to configure nova to use Ironic for baremetal deployments, if so, take a look at http://docs.openstack.org/developer/ironic/deploy/install-guide.html#configure-compute-to-use-the-bare-metal-service | 10:58 |
| *** jtomasek_ has quit IRC | 10:59 | |
| livelace | lucasagomes, I did that | 10:59 |
| livelace | After this I should get a baremetal node ? | 11:00 |
| lucasagomes | livelace, right, so I didn't understand your last question. Mind elaborating it? | 11:00 |
| lucasagomes | livelace, after that you should be able to enroll ur baremetal nodes in Ironic | 11:01 |
| lucasagomes | and deploy them via nova | 11:01 |
| lucasagomes | (see http://docs.openstack.org/developer/ironic/deploy/install-guide.html#enrollment) | 11:01 |
| * lucasagomes assumes all other services are also setup correctly | 11:02 | |
| livelace | lucasagomes, After I added [ironic] section to nova node - I should see node in list of "ironic node-list" ? | 11:03 |
| lucasagomes | livelace, no, you've only configured nova to be able talk to the ironic services | 11:04 |
| lucasagomes | I think you are confusing the concept of "node" | 11:04 |
| lucasagomes | "node" in ironic is an abstraction for a baremetal machine | 11:04 |
| hkominos | hello openstackers. quick question. If i run ironic node-show XXX and the state is available does that mean that, that machine is ready for reployment ? because node-validate fails | 11:04 |
| lucasagomes | livelace, you have to register your baremetal machines in Ironic via "ironic node-create" (see the last link I've sent to you) | 11:05 |
| lucasagomes | hkominos, available means the node can be picked for deployment. Some information will be populated at deploy time and that may be the reason why node-validate is failing | 11:06 |
| lucasagomes | hkominos, can you paste the output of node-validate please? | 11:06 |
| lucasagomes | correcting... will be populated by nova at deploy time* | 11:06 |
| livelace | lucasagomes, Thanks, but there aren't any words about "ironic node-create" in the documentation. | 11:07 |
| lucasagomes | livelace, http://docs.openstack.org/developer/ironic/deploy/install-guide.html#enrollment | 11:07 |
| *** moshele has quit IRC | 11:07 | |
| *** moshele has joined #openstack-ironic | 11:08 | |
| hkominos | lucasagomes: http://paste.openstack.org/show/516548/ | 11:09 |
| *** sdake_ has quit IRC | 11:10 | |
| lucasagomes | hkominos, oh right, so it's missing the deploy_ramdisk information, this needs to be populated prior to deploying the node | 11:10 |
| lucasagomes | hkominos, see the same link for enrollment http://docs.openstack.org/developer/ironic/deploy/install-guide.html#enrollment | 11:10 |
| livelace | lucasagomes, I see, it is very strange. First NODE_UUID mentioned in http://docs.openstack.org/developer/ironic/deploy/install-guide.html#flavor-creation, but node creation process below in the page. | 11:11 |
| hkominos | I will look at it again. however i had run root@Controller:~# ironic node-create \ > -d fuel_ipmitool \ > -i deploy_kernel=$(nova image-list |grep ironic-deploy-linux|awk '{print $2}') \ > -i deploy_ramdisk=$(nova image-list |grep ironic-deploy-initramsfs|awk '{print $2}') \ > -i deploy_squashfs=$(nova image-list |grep ironic-deploy-squashfs|awk '{print $2}') \ | 11:11 |
| hkominos | as given in a FUel video and i can see a ramdisk command | 11:12 |
| hkominos | (not sure what these three commnds do though) | 11:12 |
| lucasagomes | livelace, right yeah :-/ the install-guide is not necessarily organized by a set of sequential topics | 11:12 |
| hkominos | btw i just saw that the command is misstyped! | 11:13 |
| hkominos | so lets try again :P | 11:13 |
| lucasagomes | they are usually random because there are many ways to setup an enviromnent and each topic covers a different subject | 11:13 |
| livelace | lucasagomes, Thanks, I will try to understand this. | 11:13 |
| lucasagomes | hkominos, right it may be something to do with the fuel_ipmitool driver then (e.g deloy_squashfs vs deploy_ramdisk seems to be something specific for that driver) | 11:14 |
| lucasagomes | vdrok, yuriyz vsaienko ^ may be able to help you | 11:14 |
| lucasagomes | livelace, cool | 11:14 |
| vdrok | hkominos: right, fuel_ipmitool have 3 required images instead of 2 - kernel, ramdisk and squashfs | 11:17 |
| vdrok | this command adds these images' uuids to a node's driver_info dictionary during creation | 11:18 |
| hkominos | so now in the validate i have Missing 'ipmi_terminal_port' parameter in node's driver_info. | 11:19 |
| hkominos | is that expected? | 11:19 |
| vdrok | hkominos: you need it only if you want console access from ironic | 11:20 |
| *** thrash|g0ne is now known as thrash | 11:20 | |
| vdrok | meaning, deployment will be fine without it :) | 11:21 |
| hkominos | ok then. I have a couple more questions though if you dont mind me asking :) | 11:22 |
| vdrok | yeah sure, if i'm able to answer :) | 11:23 |
| openstackgerrit | edan david proposed openstack/ironic: Use assertIn and assertNotIn https://review.openstack.org/330474 | 11:23 |
| sborkows | Hi, I have a question about deploying ramdisk of modified source code of ironic-python-agent. Is it possible and how to do it? I have read about using diskimage-builder, but as I understand, it will build a non-modified image. | 11:23 |
| vdrok | sborkows: it's described here - https://github.com/openstack/ironic-python-agent/tree/master/imagebuild | 11:24 |
| *** hoangcx has joined #openstack-ironic | 11:24 | |
| vdrok | also I think it's possible to include modified ipa in dib, I'm just not sure how to do it, lucasagomes ^^ | 11:25 |
| *** bcornec1 has joined #openstack-ironic | 11:26 | |
| sborkows | vdrok: thanks, I will check it out | 11:28 |
| *** tobias_fiberdata has joined #openstack-ironic | 11:28 | |
| lucasagomes | vdrok, sborkows not sure if I do it "dib-way" but I usually build it from source and them modify the ramdisk | 11:28 |
| lucasagomes | modify it manually by unpacking, changing the code and packing again | 11:28 |
| vdrok | sborkows: I see this file in ironic-agent element - https://github.com/openstack/diskimage-builder/blob/master/elements/ironic-agent/source-repository-ironic-agent | 11:28 |
| vdrok | possibly you can change the path to the repo to your local one | 11:29 |
| lucasagomes | sborkows, http://docs.openstack.org/developer/ironic/deploy/troubleshooting.html#patching-the-deploy-ramdisk | 11:29 |
| lucasagomes | vdrok, way is better yeah :-) | 11:29 |
| hkominos | Openstack (and FUEL) has more than one Virtual NEtworks (Baremetal,Public,Management etc) all these are isolated and possibly in different nics with different Vlans. I have assumed that an ironic slave will only need the baremetal network. Is that correct? | 11:30 |
| * lucasagomes brb, lunch | 11:31 | |
| *** lucasagomes is now known as lucas-hungry | 11:31 | |
| hkominos | hahahahha | 11:31 |
| vsaienko | hkominos: ironic slave need baremetal and management networks | 11:32 |
| vsaienko | s/need/needs/ | 11:32 |
| hkominos | omg | 11:32 |
| hkominos | just to clarify we are talking about the bare metal machine not the conductor. | 11:32 |
| vsaienko | ah.. yeah baremetal slave needs only baremetal network | 11:33 |
| hkominos | fiouuuuuuuuu | 11:33 |
| hkominos | dude i had like 10 heartattacks | 11:33 |
| vdrok | :D | 11:33 |
| hkominos | (btw that is not clear in the "manual" maybe it should be added") | 11:34 |
| *** e0ne has quit IRC | 11:35 | |
| hkominos | 2)I can also see that when creating a BM flavor one must specify certain metadata. Is that correct? | 11:35 |
| vdrok | hkominos: correct, it will be used for partitioning | 11:36 |
| vsaienko | hkominos: you mean cpu_arch flavor attribute? | 11:37 |
| hkominos | yes | 11:37 |
| vsaienko | it is optional | 11:37 |
| hkominos | i also mean fuel_disk_info, hypervisor_type etc | 11:37 |
| hkominos | (probably that is fuel specific) | 11:37 |
| hkominos | but i cannot find the variable names anywhere | 11:37 |
| vsaienko | right, that attributes are fuel specific | 11:37 |
| hkominos | so what will happen if i dont put metadata? | 11:38 |
| vdrok | hkominos: https://github.com/openstack/fuel-agent/tree/master/contrib/ironic/ironic-fa-deploy | 11:38 |
| vdrok | I think only fuel_disk_info is required apart from the usual vcpu, memory etc | 11:38 |
| vsaienko | if you use fuel_ipmitool driver, fuel_disk_info is required | 11:38 |
| *** hoangcx has quit IRC | 11:39 | |
| *** hoangcx has joined #openstack-ironic | 11:40 | |
| *** yuriyz|2 has quit IRC | 11:40 | |
| vdrok | hkominos: btw, fuel_disk_info should be in image metadata, not flavor. flavor does not contain anything special | 11:41 |
| hkominos | yeah thats what i meant. My problem is that the FUel "manual" has zero information about ironic and i have been struggling for some time for no reason | 11:42 |
| *** tobias_fiberdata has quit IRC | 11:43 | |
| vdrok | yup, I think there should be some links to upstream docs, I'll poke people wrt that :) | 11:43 |
| *** moshele has quit IRC | 11:45 | |
| hkominos | So if a machine has 2 cpus (4cores per cpus) what should i specify in the flavor | 11:45 |
| hkominos | 8Vcpus? | 11:45 |
| vdrok | hkominos: it should be whatever you specify in ironic node properties | 11:46 |
| hkominos | !! ok | 11:46 |
| openstack | hkominos: Error: "!" is not a valid command. | 11:46 |
| vdrok | hkominos: http://docs.openstack.org/developer/ironic/deploy/install-guide.html#enrollment | 11:47 |
| vdrok | point 3 | 11:47 |
| hkominos | damn | 11:51 |
| hkominos | its really really easy to miss small details in the manual | 11:51 |
| *** dprince has joined #openstack-ironic | 11:54 | |
| *** yuriyz|2 has joined #openstack-ironic | 11:55 | |
| * TheJulia_ stumbles through the room as a zombie | 11:56 | |
| *** TheJulia_ is now known as TheJulia | 11:56 | |
| vdrok | morning TheJulia :) | 11:57 |
| TheJulia | good morning vdrok | 11:58 |
| mgould | morning TheJulia | 11:59 |
| *** hoangcx has quit IRC | 11:59 | |
| vdrok | morning mgould | 11:59 |
| *** hoangcx has joined #openstack-ironic | 11:59 | |
| *** bharathk has quit IRC | 12:00 | |
| mgould | morning vdrok | 12:00 |
| hkominos | good morning TheJulia | 12:00 |
| hkominos | Although i asked yesterday , can i use the ubuntu cloud images as my images for a Bare metal node ? | 12:02 |
| *** wajdi has joined #openstack-ironic | 12:02 | |
| hkominos | (RAW format) | 12:02 |
| *** sdake has joined #openstack-ironic | 12:05 | |
| vdrok | hkominos: I think so, yes | 12:05 |
| vdrok | raw image containing only a root partition | 12:06 |
| hkominos | well i think i have added all the variables correctly now. Ironic dployment try #25 in 3 | 12:06 |
| hkominos | 2 | 12:06 |
| hkominos | 1 | 12:06 |
| vdrok | :) | 12:06 |
| dtantsur | morning mgould, TheJulia | 12:06 |
| *** wajdi has quit IRC | 12:07 | |
| hkominos | still has not crashed | 12:07 |
| hkominos | is says spawing | 12:08 |
| hkominos | the fans are cheering | 12:08 |
| hkominos | they are on their seats | 12:08 |
| hkominos | they will stay there cause these HP servers take ages to start | 12:08 |
| hkominos | I hope there is no timeout in the nova | 12:09 |
| hkominos | cause it takes 5 minutes ++ to boot | 12:09 |
| vdrok | hkominos: I think ironic timeout is 30 minutes by default, so you should be fine | 12:09 |
| *** hoangcx has quit IRC | 12:10 | |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Do Not Merge: Canery test for ironic networking changes https://review.openstack.org/330138 | 12:12 |
| *** hoangcx has joined #openstack-ironic | 12:13 | |
| hkominos | vdrok:I must enable PXE boot from the bios for that interface right? | 12:15 |
| TheJulia | hkominos: The cloud image is what diskimage-builder uses to create a functional bootable disk image | 12:15 |
| vdrok | hkominos: umm, I think deploy driver should do that by itself | 12:15 |
| *** slagle_ is now known as slagle | 12:16 | |
| hkominos | TheJulia. It appears that i have made a mistake then. When i specify the image in horizon ,the one that i use is the one that i get from Ubuntu.cloud | 12:16 |
| hkominos | SO you are saying that what i have done is wrong | 12:16 |
| vdrok | hkominos: I think the image should be ok in case of fuel_ipmitool, there are only 2 requirements for it - raw, containing only root partition (also containing grub binary) | 12:18 |
| vdrok | you can create it with dib, or you can use one of them right away, there is an image in ubuntu cloud archive exactly as required for this driver | 12:19 |
| *** hoangcx has quit IRC | 12:19 | |
| *** lucas-hungry is now known as lucasagomes | 12:22 | |
| *** hoangcx has joined #openstack-ironic | 12:23 | |
| *** rcernin has quit IRC | 12:24 | |
| hkominos | trusty-server-cloudimg-amd64-disk1.img | 12:24 |
| hkominos | thats the one that i have | 12:24 |
| hkominos | I think there is something wrong with my image. maybe i should try the dib tool | 12:26 |
| hkominos | I will update on the progress | 12:26 |
| *** ccamacho|lunch is now known as ccamacho | 12:26 | |
| TheJulia | hkominos: have you looked at diskimage-builder? | 12:27 |
| hkominos | I am now | 12:27 |
| TheJulia | ok | 12:27 |
| vdrok | hkominos: that is the whole disk image, the root partition image is this one I think https://cloud-images.ubuntu.com/trusty/current/trusty-server-cloudimg-amd64-root.tar.xz | 12:28 |
| *** trown|outtypewww is now known as trown | 12:30 | |
| *** thiagop has joined #openstack-ironic | 12:30 | |
| *** tobias_fiberdata has joined #openstack-ironic | 12:30 | |
| hkominos | I am confused again. According to the manual the Baremetal services requres two types of images to function. | 12:30 |
| jroll | morning y'all | 12:31 |
| vdrok | morning jroll | 12:31 |
| hkominos | the deploy images and the user images | 12:31 |
| hkominos | I thought that the deploy images are the ones callled ironic-deploy-squashfs etc... | 12:31 |
| hkominos | am i wrong? | 12:31 |
| vdrok | hkominos: correct, deploy images that you specify in driver info - deploy kernel, ramdisk, squashfs, and user image that will be booted after deploy | 12:31 |
| vdrok | deploy images only do the deploy :) | 12:32 |
| hkominos | so they are prebuild and i dont need to worry about them. my only consern is to make-use a user image | 12:32 |
| *** rbudden has joined #openstack-ironic | 12:32 | |
| vdrok | correct | 12:33 |
| dtantsur | our gate is half-broken due to this apache issue... | 12:33 |
| vdrok | this one should be the fix https://review.openstack.org/#/c/329908/ | 12:34 |
| *** jcoufal has joined #openstack-ironic | 12:34 | |
| *** wagiel has quit IRC | 12:35 | |
| hkominos | (btw i am about to itme out. The machine never got to the PXE boot menu it just loaded the previous OS which had in the Hard disk ) | 12:35 |
| hkominos | (priorities are ok i think) | 12:35 |
| dtantsur | vdrok, oh, cool | 12:35 |
| hkominos | However I am sending taged traffic to that machine. Must the traffic be untaged in order to be PXE? | 12:35 |
| hkominos | cause i think there must be native untaged traffic for a pxe boot | 12:36 |
| hkominos | let me check | 12:36 |
| *** moshele has joined #openstack-ironic | 12:37 | |
| thiagop | Good morning, Ironicers | 12:37 |
| vsaienko | hkominos: pxe doesn't work with tagget traffic | 12:37 |
| vdrok | hkominos: btw the previous image link was incorrect, if you still want to try without dib, this is the one - https://cloud-images.ubuntu.com/trusty/current/trusty-server-cloudimg-amd64.tar.gz | 12:37 |
| vdrok | there is trusty-server-cloudimg-amd64.img inside | 12:38 |
| vdrok | morning thiagop | 12:38 |
| *** rcernin has joined #openstack-ironic | 12:39 | |
| hkominos | OK i think that is the problem then. IT never receaves a PXE boot request because its taged | 12:39 |
| hkominos | so that is why the node is not provisioned | 12:39 |
| hkominos | However in the fuel "tutorial" it is said that the bare metalnetwork can be VlanTaged | 12:40 |
| hkominos | how is that possible | 12:40 |
| *** e0ne has joined #openstack-ironic | 12:40 | |
| *** athomas has quit IRC | 12:40 | |
| hkominos | thx for the new image i will try that as well | 12:43 |
| vsaienko | hkminos: yes baremetal network may be tagged. It means that ironic-conductor machine sends tagged packets to network. While ironic slaves - should receive packets as untagged | 12:46 |
| jroll | huh, samsung bought joyent: https://news.samsung.com/global/samsung-to-acquire-joyent-a-leading-public-and-private-cloud-provider | 12:46 |
| hkominos | ok thx. That makes a big difference (also not visible in the manual) | 12:46 |
| *** rbartal has quit IRC | 12:46 | |
| *** athomas has joined #openstack-ironic | 12:46 | |
| *** baoli has joined #openstack-ironic | 12:47 | |
| hkominos | vdrok:in the image you sent me also has some extra files (floppy ,loader,vmzlinuz generic) do i need them? | 12:47 |
| vdrok | hkominos: nope | 12:48 |
| hkominos | vsainko :thx for the help. If you are still at mirantis please tell them to upload a manual :P | 12:50 |
| *** baoli_ has joined #openstack-ironic | 12:53 | |
| *** rloo has joined #openstack-ironic | 12:53 | |
| *** rcernin has quit IRC | 12:54 | |
| vsaienko | hkominos: will do :) | 12:54 |
| *** baoli has quit IRC | 12:55 | |
| moshele | sambetts: hi | 12:57 |
| *** Goneri has joined #openstack-ironic | 12:57 | |
| sambetts | o/ moshele | 12:58 |
| moshele | sambetts: thanks for review my infiniband spec, can you review my respond https://review.openstack.org/#/c/265635/ | 12:59 |
| thiagop | hkominos: If you find the manual missing steps, please patch it ;) | 12:59 |
| moshele | sambetts: basically I ok with putting the "infiniband " mac as address, and I also want to add dhcp option support to the ironic port what do you think | 13:00 |
| thiagop | hkominos: We don't always get the missing points because we're so used to deploy this stuff that the steps seem kinda natural | 13:00 |
| vdrok | thiagop: he's talking about fuel docs :) | 13:00 |
| *** cdearborn has joined #openstack-ironic | 13:00 | |
| thiagop | vdrok: aw, right. Then sorry, hkominos | 13:01 |
| hkominos | I am | 13:01 |
| hkominos | However! | 13:01 |
| hkominos | I get the feeling sometimes that the manual is inteded at the wrong crowd. Maybe a tutorial should be build as a seperate entity. If you are a new user its not possible to even come close to using the manual to do even the simplest tasks.I mean that if I am a new user and i want to learn how to do something in Openstack(fE deploy ironic) the manual is worthless | 13:03 |
| hkominos | there no explanation regarding anything. It is used by people who already know as a reference. | 13:03 |
| hkominos | So i feel something is missing | 13:03 |
| hkominos | that is my view | 13:03 |
| *** jvgrant_ has joined #openstack-ironic | 13:06 | |
| *** rcernin has joined #openstack-ironic | 13:07 | |
| jroll | moshele: I'm not sambetts but I added some comments there | 13:07 |
| jroll | sambetts: great ideas in there, thanks for that :D | 13:07 |
| sambetts | moshele: jroll basically took all the words out of my mouth | 13:08 |
| moshele | jroll: I wanted it to be more generic, if someone need to add more dhcp option, but your understand the direction | 13:09 |
| *** e0ne has quit IRC | 13:09 | |
| jroll | moshele: yeah, I get that, but I don't think I've heard any other requests for that so I'm hesitant to add it | 13:10 |
| moshele | jroll, sambetts: thanks I will update the spec as suggested. it seem that the code change to ironic will be very minimal | 13:10 |
| jroll | moshele: indeed :) | 13:10 |
| sambetts | and hopfully nothing in Nova :) | 13:11 |
| *** e0ne has joined #openstack-ironic | 13:12 | |
| openstackgerrit | edan david proposed openstack/ironic-python-agent: Use assertIn and assertNotIn https://review.openstack.org/330451 | 13:18 |
| *** bharath has joined #openstack-ironic | 13:22 | |
| *** jcoufal has quit IRC | 13:23 | |
| livelace | lucasagomes, Hello again :) Is workflow (pxe_ssh): node up -> pxe booting -> load deploy image -> start ssh inside ? | 13:26 |
| openstackgerrit | garyk proposed openstack/ironic: Fix typo in agent driver https://review.openstack.org/330549 | 13:27 |
| openstackgerrit | edan david proposed openstack/ironic: Use assertIn and assertNotIn https://review.openstack.org/330474 | 13:27 |
| dtantsur | livelace, ssh is working with your hypervisor, not the the node | 13:27 |
| dtantsur | livelace, (i.e. you don't use this driver with real bare metals) | 13:27 |
| lucasagomes | livelace, not really, pxe_ssh is a "way around" for using VMs for testing baremetal. The SSH part is about Ironic ssh'ing into the host machine and issuing commands like virsh start/stop to simulate some sorta of power management (like you would do with IPMI) | 13:28 |
| livelace | :( | 13:28 |
| livelace | How can I provision VMs for testing purpose ? | 13:29 |
| sambetts | livelace: are you using devstack? | 13:29 |
| lucasagomes | livelace, heh I know, that was the only way at the time that we found to be able to have tests in gate | 13:29 |
| livelace | No, its custom | 13:29 |
| lucasagomes | livelace, well, pxe_ssh is a way, another way is using the _ipmitool drivers (or ipminative) + VirtualBMC | 13:29 |
| *** frontrunner has joined #openstack-ironic | 13:30 | |
| lucasagomes | livelace, https://github.com/openstack/virtualbmc (for lack of docs, see the demo gif: https://github.com/umago/virtualbmc/blob/master/images/demo.gif) | 13:30 |
| *** smoriya has quit IRC | 13:32 | |
| livelace | lucasagomes, I have stack inside ovirt and I want to start "baremetal" nodes inside Ovirt, provision them and deploy use images inside these VMs | 13:32 |
| livelace | *user images | 13:32 |
| lucasagomes | livelace, oh, ok that's another can of worms I think because, AFAIK, ovirt do not "pre-create" the VMs | 13:33 |
| lucasagomes | right? | 13:33 |
| livelace | Hm | 13:33 |
| livelace | I can "pre-create" VMs | 13:33 |
| lucasagomes | meaning that, you can't just register them in Ironic because it may be registered in OVirt but the actual VM wasn't yet created | 13:34 |
| livelace | All of VMs are working now | 13:34 |
| lucasagomes | livelace, right | 13:34 |
| * lucasagomes thinks | 13:34 | |
| lucasagomes | I feel I don't have much knowledge as well to actually advise you how to do what you want | 13:34 |
| lucasagomes | but I suspect that you would be better off having Ironic to talk to OVirt CLI/API idk | 13:35 |
| lucasagomes | and managing things that way, but this will required another driver or a translation layer (such as VirtualBMC) | 13:35 |
| *** mtanino has joined #openstack-ironic | 13:35 | |
| jroll | dtantsur: reviewed your lookup/heartbeat spec, it's very close :) | 13:35 |
| *** hoangcx has quit IRC | 13:35 | |
| lucasagomes | that would be a better integration than having 2 systems (OVirt + Ironic/Nova) managing the same VMs | 13:35 |
| lucasagomes | jroll, morning, if you have time https://review.openstack.org/#/c/168799/ :-) | 13:36 |
| lucasagomes | no rush | 13:36 |
| jroll | oooo this looks old | 13:36 |
| * jroll looks | 13:36 | |
| jroll | aha | 13:36 |
| lucasagomes | oh | 13:36 |
| lucasagomes | wrong link | 13:36 |
| jroll | lol | 13:36 |
| *** rloo has quit IRC | 13:36 | |
| jroll | I was just going to say that | 13:36 |
| lucasagomes | jroll, there you go https://review.openstack.org/#/c/323511/ | 13:37 |
| lucasagomes | sorry for that heh | 13:37 |
| jroll | np | 13:37 |
| openstackgerrit | garyk proposed openstack/ironic: No need for 'default=None' in config variable https://review.openstack.org/330557 | 13:37 |
| dtantsur | jroll, re ING states: I was thinking about e.g. if IPA needs rebooting/reboots for any reason during e.g. cleaning | 13:37 |
| *** rloo has joined #openstack-ironic | 13:37 | |
| dtantsur | jroll, the state will be CLEANING and lookup will fail | 13:37 |
| dtantsur | so as a precaution I've included ING states. does it make sense? | 13:37 |
| jroll | dtantsur: no, if a reboot is requested it goes back to CLEANWAIT | 13:37 |
| * jroll finds the code | 13:37 | |
| mat128 | yes clean wait since liberty | 13:38 |
| lucasagomes | +1 on go back to CLEANWAIT | 13:38 |
| jroll | wait, does it? :| | 13:38 |
| dtantsur | also ++ on not requiring exclusive lock right away | 13:38 |
| lucasagomes | that's better because it could be aborted as well (since no system is actually operating on it) | 13:38 |
| jroll | https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent_base_vendor.py#L399 | 13:38 |
| dtantsur | jroll, could you point me please where it sets back CLEAN WAIT? | 13:39 |
| dtantsur | I can't find it | 13:39 |
| livelace | lucasagomes, http://screenshot.net/xgpyea7 | 13:39 |
| jroll | dtantsur: I can't either :( | 13:39 |
| jroll | dtantsur: that's a bug imo though | 13:40 |
| dtantsur | jroll, another case would be IPA crashing, but that's probably not that important | 13:40 |
| jroll | dtantsur: maybe it's fine then | 13:41 |
| livelace | lucasagomes, Can I provision baremetal nodes without Ovirt, only by pxe and ssh ? | 13:41 |
| lucasagomes | if by baremetal nodes you mean "VMs" pretending to be baremetal, then yes | 13:41 |
| dtantsur | jroll, so should I leave a comment that we'll apply more restrictions when we ensure that lookup is not needed in ING states? | 13:42 |
| NobodyCam | Good Morning Ironicers | 13:42 |
| dtantsur | morning NobodyCam | 13:42 |
| livelace | lucasagomes, :) with pxe_ssh ? | 13:42 |
| jroll | dtantsur: that, or we can leave ING just in case | 13:42 |
| mgould | morning NobodyCam | 13:42 |
| lucasagomes | but the VMs needs to be pre-created, since Ironic is about hardware | 13:42 |
| lucasagomes | livelace, yeah | 13:42 |
| NobodyCam | morning dtantsur, mgould | 13:42 |
| NobodyCam | morning lucasagomes :) | 13:42 |
| lucasagomes | NobodyCam, hey there :-) | 13:43 |
| livelace | lucasagomes, Ok, and "baremetals" should automatically take ip/pxe images and start ssh ? | 13:43 |
| NobodyCam | :) | 13:43 |
| *** hoangcx has joined #openstack-ironic | 13:43 | |
| livelace | lucasagomes, I mean coreos deploy images | 13:43 |
| dtantsur | jroll, yuriyz, why does restrict_lookup need to be false for 3rd party drivers? I don't quite get the comment at line 248 | 13:44 |
| jroll | dtantsur: I believe he means "document if the driver requires it to be false" | 13:44 |
| lucasagomes | livelace, yes-ish... I mean, the network bits (IPs/DHCP) comes from Neutron, which is integrated with Ironic/Nova | 13:44 |
| dtantsur | ack | 13:44 |
| NobodyCam | morning jroll :) | 13:45 |
| livelace | lucasagomes, Ok, thanks for your time! | 13:45 |
| lucasagomes | livelace, the pxe_ssh drivers does things like setting up the PXE environment to be able to boot the image and ssh'ing into the host machine to power the VM on or off | 13:45 |
| lucasagomes | now, there are other stuff that also needs to be configured to achieving what you want (neutron, glance, nova, etc...) | 13:46 |
| lucasagomes | livelace, np | 13:46 |
| lucasagomes | livelace, I know OpenStack is a beast to understand/work with | 13:46 |
| dtantsur | jroll, oh, do you think we need a separate node field for heartbeat timestamp? | 13:47 |
| dtantsur | now that it's core API | 13:47 |
| *** n|cotine_ has quit IRC | 13:47 | |
| NobodyCam | gots to reboot. brb | 13:48 |
| jroll | dtantsur: hrm, dunno, maybe let's wait to do that when we do long-running agents? | 13:48 |
| dtantsur | jroll, we already have this timestamp, and iirc it's stored in driver_something which is not really correct for a generic API | 13:48 |
| dtantsur | driver_internal_info['agent_last_heartbeat'] | 13:49 |
| jroll | dtantsur: right, but it isn't super useful other than checking timeouts | 13:49 |
| jroll | but | 13:49 |
| jroll | I'm not super opposed to moving it | 13:49 |
| dtantsur | do we use it at all? | 13:49 |
| jroll | oh guess not | 13:50 |
| jroll | we do use it downstream | 13:50 |
| jroll | but not in ironic, just related apps | 13:50 |
| jroll | monitoring and such | 13:50 |
| jroll | (because long-running agents) | 13:50 |
| dtantsur | jroll, so we should keep compatibility, right? | 13:50 |
| dtantsur | well ok, we can think about it later, and for now just have the agent driver record it as before | 13:51 |
| jroll | dtantsur: meh, it's a json field | 13:51 |
| jroll | yeah, agree | 13:51 |
| vdrok | morning NobodyCam | 13:52 |
| *** jcoufal has joined #openstack-ironic | 13:52 | |
| *** joprovost has joined #openstack-ironic | 13:53 | |
| *** Larion has joined #openstack-ironic | 13:53 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: Promote agent vendor passthru to core API https://review.openstack.org/306418 | 13:54 |
| dtantsur | jroll, yuriyz, NobodyCam ^^^ | 13:54 |
| jroll | dtantsur: thanks | 13:54 |
| NobodyCam | :) | 13:54 |
| NobodyCam | morning vdrok | 13:55 |
| *** tobias_fiberdata has quit IRC | 13:56 | |
| jroll | lucasagomes: some comments on that spec | 13:58 |
| lucasagomes | jroll, gracias! Will take a look | 13:59 |
| jroll | dtantsur: +2 for you, onto the driver comp | 13:59 |
| * jroll fills coffee | 13:59 | |
| *** joprovost1 has joined #openstack-ironic | 14:00 | |
| *** rama_y has joined #openstack-ironic | 14:01 | |
| dtantsur | oh, this is gonna be enjoyable reading | 14:01 |
| *** joprovost has quit IRC | 14:01 | |
| *** joprovost1 is now known as joprovost | 14:01 | |
| *** nicodemos has joined #openstack-ironic | 14:02 | |
| livelace | lucasagomes, Last question, I hope :) /tftpboot directory should change when I change settings for node ? | 14:02 |
| jroll | dtantsur: http://i.imgur.com/ZtMY2my.png | 14:03 |
| dtantsur | LOL | 14:03 |
| lucasagomes | livelace, no worries... The content will be added when a node is being deployed/terminated | 14:03 |
| lucasagomes | Ironic should create/remove the PXE configuration files into that directory | 14:03 |
| lucasagomes | and update the DHCP (neutron) to point to it | 14:04 |
| livelace | lucasagomes, after ironic node-update, for example ? | 14:04 |
| lucasagomes | livelace, no, in that case it won't trigger any update in that directory | 14:05 |
| livelace | lucasagomes, Ok, thanks again | 14:05 |
| lucasagomes | np | 14:05 |
| *** PollyZ has joined #openstack-ironic | 14:07 | |
| *** sacharya has joined #openstack-ironic | 14:08 | |
| *** tobias_fiberdata has joined #openstack-ironic | 14:14 | |
| openstackgerrit | Thiago Paiva Brito proposed openstack/ironic: [WIP] Using Oneview drivers w/ IRONIC_IS_HARDWARE https://review.openstack.org/328451 | 14:15 |
| openstackgerrit | Thiago Paiva Brito proposed openstack/ironic: Using Oneview drivers w/ IRONIC_IS_HARDWARE https://review.openstack.org/328451 | 14:16 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Do Not Merge: another test for ironic networking https://review.openstack.org/330210 | 14:17 |
| NobodyCam | dtantsur: just look at the spec, no need to send current_clean_step if it not empty | 14:17 |
| *** Larion has quit IRC | 14:17 | |
| dtantsur | NobodyCam, sorry, I've switched my context already. What spec are we talking about right now? the agent API one? | 14:18 |
| openstackgerrit | Brad P. Crochet proposed openstack/python-ironicclient: Implementation of baremetal power state commands https://review.openstack.org/172517 | 14:18 |
| openstackgerrit | Brad P. Crochet proposed openstack/python-ironicclient: Add maintenance mode commands https://review.openstack.org/216727 | 14:18 |
| openstackgerrit | Brad P. Crochet proposed openstack/python-ironicclient: Implementation of baremetal port create https://review.openstack.org/172461 | 14:18 |
| openstackgerrit | Brad P. Crochet proposed openstack/python-ironicclient: Add provision state commands https://review.openstack.org/206119 | 14:18 |
| *** sacharya_ has joined #openstack-ironic | 14:18 | |
| NobodyCam | https://review.openstack.org/#/c/306418/12/specs/approved/agent-api.rst@78 | 14:18 |
| openstackgerrit | Vladyslav Drok proposed openstack/python-ironicclient: Add create command to ironic client https://review.openstack.org/328955 | 14:18 |
| NobodyCam | its approved so its just a question from me | 14:19 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-inspector: Add manage_boot parameter to introspection API https://review.openstack.org/316801 | 14:19 |
| NobodyCam | haha | 14:19 |
| NobodyCam | not +a'd yet | 14:19 |
| *** daemontool has joined #openstack-ironic | 14:19 | |
| dtantsur | NobodyCam, not sure still.. do we send this current_clean_step now | 14:19 |
| dtantsur | ? | 14:19 |
| *** joprovost1 has joined #openstack-ironic | 14:20 | |
| *** joprovost has quit IRC | 14:20 | |
| *** joprovost1 is now known as joprovost | 14:20 | |
| dtantsur | NobodyCam, I don't see it in https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent_base_vendor.py#L445 thus I didn't add it, but maybe I'm missing something | 14:21 |
| sambetts | \o/ for promote and standardise agent API :D | 14:21 |
| *** sacharya has quit IRC | 14:21 | |
| *** bharath has quit IRC | 14:21 | |
| *** UForgotten_ has quit IRC | 14:21 | |
| *** UForgotten_ has joined #openstack-ironic | 14:21 | |
| *** UForgotten_ is now known as UForgotten | 14:22 | |
| jroll | dtantsur: NobodyCam: the agent knows it's current step, no need to send it I think | 14:25 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-specs: Collect system logs from IPA https://review.openstack.org/323511 | 14:27 |
| lucasagomes | updated ^ thanks all for the comments :-) | 14:27 |
| *** jistr is now known as jistr|mtg | 14:28 | |
| *** ifarkas has quit IRC | 14:28 | |
| *** amotoki has joined #openstack-ironic | 14:29 | |
| *** sacharya_ has quit IRC | 14:32 | |
| *** sacharya has joined #openstack-ironic | 14:33 | |
| *** bcornec1 has quit IRC | 14:33 | |
| *** ChrisAusten has joined #openstack-ironic | 14:34 | |
| *** xhku has joined #openstack-ironic | 14:34 | |
| *** rama_y has quit IRC | 14:35 | |
| NobodyCam | ahh : https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent_base_vendor.py#L574 | 14:35 |
| livelace | Does anybody have time for stupid questions ? :) | 14:36 |
| jroll | livelace: always feel free to ask, someone will answer if they know and have time :) | 14:36 |
| TheJulia | Plus, no question is a stupid question | 14:36 |
| livelace | Ok. First step. I create node and add port to it. | 14:37 |
| NobodyCam | the only stupid question is the one NOT asked | 14:37 |
| livelace | Second step. I start "baremetal" vm in PXE mode | 14:37 |
| livelace | 3. I see in logs: DHCPDISCOVER(tap70ba5d66-4e) 00:1a:4a:48:1b:9f no address available | 14:38 |
| NobodyCam | livelace: Ironic should be controlling the power of the node | 14:38 |
| TheJulia | are you powering on the VM or is ironic powering up the VM? | 14:38 |
| livelace | I start vm by hand | 14:38 |
| livelace | It's wrong ? | 14:38 |
| *** wagiel has joined #openstack-ironic | 14:39 | |
| livelace | Because ironic doesn't can start it, it's vm | 14:39 |
| NobodyCam | livelace: Ironic should be cntroling the power state of the node | 14:39 |
| *** tloeb has quit IRC | 14:40 | |
| dtantsur | livelace, there is some confusion here, lets try to clarify :) | 14:40 |
| dtantsur | livelace, ironic usually work with bare metals, but we have a special test driver called pxe_ssh which can work with vms | 14:40 |
| openstackgerrit | edan david proposed openstack/ironic: Replace dict.get(key) with dict[key] in unit tests https://review.openstack.org/330617 | 14:40 |
| rloo | hi and goooooood Thursday morning everyone, NobodyCam, livelace, TheJulia, jroll, dtantsur, lucasagomes, sambetts, thiagop, vdrok, and everyone else | 14:40 |
| TheJulia | good morning rloo | 14:41 |
| NobodyCam | morning rloo :) | 14:41 |
| dtantsur | livelace, it works by SSH-ing into your hypervisor and using virsh command (or similar) | 14:41 |
| thiagop | good morning, rloo | 14:41 |
| lucasagomes | rloo, hello there :-) | 14:41 |
| vdrok | morning rloo :) | 14:41 |
| dtantsur | livelace, if you want to play with vms you should be using this driver. booting vms manually will never work, as our networking service won't recognize them | 14:41 |
| *** pcaruana has quit IRC | 14:41 | |
| dtantsur | morning rloo, thiagop | 14:41 |
| livelace | dtantsur, I can't provision VM with any plugin ? | 14:41 |
| rloo | folks, there are two +2 for active node creation. I'm going to +A unless people want to review it today: https://review.openstack.org/#/c/275766/ | 14:41 |
| wagiel | Hi openstackers, need some guidence here, I have devstack ironic on real hardware, instance is spawning, but dont get the ip addr from dhcp, i see dhcp requests coming with tcpdump and dnsmasq process running | 14:42 |
| dtantsur | livelace, pxe_ssh is a way to go. but note that it does not ssh into the vm itself, it sshes into its hypervisor | 14:42 |
| wagiel | but dont see anything listening on port 53 | 14:42 |
| livelace | dtantsur, agent_ssh ? | 14:42 |
| dtantsur | livelace, pxe_ssh or agent_ssh would work. they slightly differ in how exactly they flash the image | 14:42 |
| *** tloeb has joined #openstack-ironic | 14:43 | |
| dtantsur | livelace, maybe somebody already asked, but why aren't you using devstack? | 14:43 |
| jroll | dtantsur: reviewed driver comp, this is awesome :D | 14:43 |
| jroll | dtantsur: if you finish the TODOs I'm +2 | 14:43 |
| jroll | rloo: morning :) | 14:43 |
| livelace | dtantsur, But why Ironic can't wait for baremetal nodes and provision their by their availability ? | 14:43 |
| dtantsur | jroll, cool! will try to finish it tomorrow | 14:44 |
| livelace | dtantsur, For some reasons | 14:44 |
| *** wajdi has joined #openstack-ironic | 14:44 | |
| dtantsur | livelace, it does not have all required information to provision them | 14:44 |
| *** trown is now known as trown|brb | 14:44 | |
| lucasagomes | wagiel, the port (mac address) registered in Ironic is the same mac address in the DHCP request? | 14:44 |
| dtantsur | it is possible to implement, it's just not how ironic works | 14:45 |
| livelace | dtantsur, Very strange, I set port with mac address. It isn't enough ? All thinstation/vdi works similar | 14:45 |
| livelace | lucasagomes, Yes | 14:46 |
| lucasagomes | livelace, that answer was to another question | 14:46 |
| dtantsur | livelace, well, our primary use case is to respond to requests like "please provision this machine". we haven't previously seen a use case for provisioning everything that shows up on network with a valid MAC | 14:47 |
| wagiel | lucasagomes, shoud I see dhcp listening on 53 for incoming reqs? | 14:47 |
| lucasagomes | livelace, but no, registering the mac address in Ironic shouldn't and is not enough. Unless Ironic have total control over the network bits (which its doesn't because we delegate it to Neutron) | 14:47 |
| dtantsur | also while it would work smoothly with vms, for bare metals it would require several manual and vendor-specific actions | 14:47 |
| lucasagomes | wagiel, yup, I thought you said you've seem the requests already coming via tcpdump | 14:48 |
| livelace | dtantsur, More useful will be, if: 1. Provision bunch of nodes by MAC mask/list 2. Provision bunch of nodes by hardware template | 14:49 |
| dtantsur | livelace, there seem to be a valid use case, but on the other hand I'm not sure how it's going to play with e.g. nova scheduler (essentially any scheduler) | 14:50 |
| sambetts | livelace: Ironic is designed to be a baremetal hypervisor under Nova, we respond to a nova boot my an instance, we aren't dictated my new hardware just appearing | 14:50 |
| sambetts | s/my/by | 14:50 |
| sambetts | livelace: are you using Ironic with OpenStack or standalone? | 14:51 |
| livelace | For example, in Openshift we label hosts with some labels (like: ssd, hpc etc.) It's very useful. | 14:51 |
| livelace | sambetts, Openstack | 14:51 |
| *** trown|brb is now known as trown | 14:52 | |
| dtantsur | I've seen such approach with SpinalStack btw | 14:52 |
| sambetts | livelace: then once a baremetal node is registered in Ironic you should be booting instances through nova | 14:52 |
| dtantsur | livelace, similarly, you can label ironic nodes (and we do it in tripleo) with a so called capability, for example profile:ssd | 14:53 |
| dtantsur | livelace, then you can create a nova flavor specifying the same capability. then when a user requests a node of this flavor, a node with profile:ssd will get chosen | 14:53 |
| mgould | livelace: it sounds like you're trying to use Ironic for something it wasn't designed for | 14:53 |
| mgould | what ultimate problem are you trying to solve? | 14:53 |
| *** sdake has quit IRC | 14:54 | |
| livelace | mgould, I'm trying to test Ironic :) But use case is - provision some types of linux into baremetal nodes with testing purpose | 14:54 |
| mgould | OK | 14:55 |
| mgould | so each test run starts with "give me a bunch of clean Linux installs on bare-metal nodes so I can run my test scripts on them"? | 14:56 |
| livelace | Deploy bunch of version, then start testing on these nodes | 14:56 |
| livelace | mgould, Yes | 14:56 |
| mgould | right, good | 14:56 |
| mgould | this is indeed the problem that Ironic solves :-) | 14:56 |
| livelace | Cool. But I can't work with it on my testing server :) | 14:56 |
| *** amotoki_ has joined #openstack-ironic | 14:57 | |
| dtantsur | folks, I have to disappear for a couple of hours, will continue in the evening | 14:58 |
| lucasagomes | dtantsur, see ya | 14:58 |
| *** dtantsur is now known as dtantsur|bbl | 14:58 | |
| NobodyCam | have a good day dtantsur|bbl | 14:59 |
| *** tloeb_ has joined #openstack-ironic | 14:59 | |
| *** sdake has joined #openstack-ironic | 14:59 | |
| vdrok | rloo: I'd take a look in an hour or so | 14:59 |
| *** amotoki has quit IRC | 15:00 | |
| rloo | vdrok: a look at what? | 15:00 |
| vdrok | at active node creation | 15:00 |
| *** tloeb__ has joined #openstack-ironic | 15:00 | |
| rloo | vdrok: oh yeah. forgot i had asked. (my head is in driver composition reformation. ha ha.) thx! | 15:00 |
| *** tloeb has quit IRC | 15:00 | |
| *** tloeb_ has quit IRC | 15:00 | |
| *** wagiel has quit IRC | 15:02 | |
| dtantsur|bbl | rloo, I feel sorry for you :) | 15:02 |
| rloo | dtantsur|bbl: i think i feel sorry-er? for the people that are going to review the code :) | 15:02 |
| dtantsur|bbl | and what about someone who is going to write this code? | 15:03 |
| * dtantsur|bbl now goes for real | 15:03 | |
| *** tloeb__ has quit IRC | 15:04 | |
| *** amotoki_ has quit IRC | 15:06 | |
| JayF | I'm going through some of the IPA bugs right now, updating status | 15:09 |
| *** edand has joined #openstack-ironic | 15:14 | |
| JayF | Any objection to me marking https://bugs.launchpad.net/ironic-python-agent/+bug/1367915 (IPA is missing developer documentation) as resolved? When this was filed there were no docs, now almost all of IPA has docstrings and is reasonably well documented | 15:14 |
| openstack | Launchpad bug 1367915 in ironic-python-agent "ironic-python-agent is missing developer documentation" [Medium,In progress] - Assigned to Clif Houck (f-te-u) | 15:14 |
| jroll | +1 | 15:15 |
| *** piet has joined #openstack-ironic | 15:16 | |
| *** jistr|mtg is now known as jistr | 15:19 | |
| JayF | sambetts: didn't you fix https://bugs.launchpad.net/ironic-python-agent/+bug/1580932 | 15:19 |
| openstack | Launchpad bug 1580932 in ironic-python-agent "The 'logs' inspection collector does not work with TinyIPA" [Low,Triaged] | 15:19 |
| JayF | sambetts: also congrats | 15:19 |
| sambetts | jroll: wow thanks! | 15:21 |
| jroll | :D | 15:21 |
| sambetts | JayF: you spotted the email before I did :-P | 15:21 |
| jroll | JayF used to work on email systems, he's addicted to reading email | 15:22 |
| * jroll ducks | 15:22 | |
| sambetts | haha | 15:22 |
| NobodyCam | lol | 15:22 |
| jroll | NobodyCam: btw, great success on the cruze with hammering the socket on :P | 15:23 |
| jroll | also fixed the exhaust leak I had left on the audi | 15:23 |
| jroll | productive night it was | 15:23 |
| NobodyCam | jroll: lol... I might have ground oil drain bolts in to squares with a grinder to get tem off before | 15:24 |
| *** ifarkas has joined #openstack-ironic | 15:24 | |
| NobodyCam | :) nice | 15:24 |
| jroll | NobodyCam: yeah, that was the backup plan, or an air chisel :P | 15:25 |
| NobodyCam | hehehehe | 15:25 |
| lucasagomes | JayF, sambetts I think the bug about the logs is not fixed yet the log collector still relies on journalctl | 15:25 |
| * NobodyCam does like the "bigger" hammer methods thou :) | 15:25 | |
| lucasagomes | the collect logs for deploy indirectly fixes that problem | 15:26 |
| lucasagomes | JayF, sambetts and congrats for the nomination :-) | 15:26 |
| openstackgerrit | edan david proposed openstack/ironic: Use assertIsNotNone instead of assertTrue https://review.openstack.org/330641 | 15:26 |
| NobodyCam | :) ++ | 15:26 |
| sambetts | thanks livelace | 15:26 |
| JayF | lucasagomes: ah, the collect logs for deploy is what I was thinking of | 15:27 |
| sambetts | thankks lucasagomes | 15:27 |
| sambetts | sorry for the noise livelace | 15:27 |
| sambetts | JayF, lucasagomes: there is work to make it pick up syslog etc too right? | 15:28 |
| *** sacharya_ has joined #openstack-ironic | 15:28 | |
| lucasagomes | sambetts, right, waiting on the spec ( https://review.openstack.org/#/c/323511/ ) prior to submit the code | 15:28 |
| lucasagomes | should be close/done | 15:28 |
| livelace | lucasagomes, virtualbmc isn't working for me, because vm doesn't exist in ovirt/libvirt. I can't start VM which doesn't exist in libvirt. | 15:28 |
| sambetts | lucasagomes: awesome :) | 15:28 |
| lucasagomes | livelace, exactly, told ya that :-) | 15:28 |
| mgould | NobodyCam, jroll: it's always satisfying when a problem can be solved by hitting things with hammers | 15:29 |
| sambetts | livelace: what is your hypervisor? | 15:29 |
| livelace | sambetts, qemu/kvm | 15:29 |
| jroll | mgould: :D | 15:29 |
| lucasagomes | livelace, if you want to use OVirt I would recommend having a driver in Ironic or an translation layer that talks to the OVirt CLI/API | 15:29 |
| livelace | sambetts, management - ovirt | 15:29 |
| livelace | lucasagomes, I understood that. I will try variants ... | 15:30 |
| *** sacharya has quit IRC | 15:30 | |
| *** sacharya has joined #openstack-ironic | 15:30 | |
| livelace | Maybe I switch to my gentoo and build qemu with bmc support :) | 15:31 |
| livelace | https://www.youtube.com/watch?v=oUv2RZa9uM0 | 15:31 |
| lucasagomes | livelace, fwiw https://review.openstack.org/#/c/301226/ | 15:32 |
| lucasagomes | someone already came up with some ovirt driver, I haven't tested it tho | 15:32 |
| *** edand has quit IRC | 15:32 | |
| *** sacharya_ has quit IRC | 15:33 | |
| *** openstackgerrit has quit IRC | 15:34 | |
| *** openstackgerrit has joined #openstack-ironic | 15:34 | |
| livelace | lucasagomes, Therefore, I like Open Source :) | 15:35 |
| lucasagomes | :-) | 15:35 |
| *** milan has quit IRC | 15:37 | |
| *** bcornec1 has joined #openstack-ironic | 15:37 | |
| JayF | vdrok: https://bugs.launchpad.net/ironic-python-agent/+bug/1590493 we def need to get that parition image job on IPA :( | 15:38 |
| openstack | Launchpad bug 1590493 in ironic-python-agent "IPA fails to deploy partition image with CoreOS based ramdisk" [Undecided,New] | 15:38 |
| vdrok | JayF: morning, yep, I'll propose this to project config today | 15:38 |
| NobodyCam | jroll: I see you have +2 on 306418 (Agent pass thru promote spec) I don't see any thing about a api version bump. are we expectiong to bump the version?" | 15:44 |
| jroll | NobodyCam: oh, is there not? there should be, it's even in the POC code already | 15:44 |
| jroll | dtantsur|bbl: ^ | 15:44 |
| *** sacharya has quit IRC | 15:44 | |
| NobodyCam | I'll add a note | 15:44 |
| *** sacharya has joined #openstack-ironic | 15:45 | |
| thiagop | Folks, I'm having a bit of trouble trying to find out why this happens -> http://logs.ironic-oneview-ci.lsd.ufcg.edu.br/51/328451/7/silent/dsvm-tempest-ironic-agent_pxe_oneview/159d302/logs/screen-ir-cond.txt.gz#_2016-06-16_13_12_20_159 | 15:45 |
| thiagop | Probably because I don't know how the port creation workflow actually works (apparently, at some time, Ironic expects that Neutron has a port related to the port on Ironic) | 15:46 |
| thiagop | is there some docs that you can point me to understand how this works? Walking through code has proven ineffective... | 15:47 |
| jroll | that's a 10. address, I can't reach that | 15:47 |
| jlvillal | Congrats JayF and sambetts (in order of who spoke last) on your nominations! :) | 15:47 |
| JayF | ty | 15:47 |
| sambetts | thiagop: Nova creates a neutron port, and it does an ironic port update setting a field "port_vif_id" on the ironci port.extra, the dhcp provider in Ironic then looks at that field | 15:48 |
| mgould | yes, congrats JayF and sambetts! | 15:48 |
| mgould | I'd vote for you both if I could :-) | 15:48 |
| * sambetts just realised he forgot to congratulate JayF ... feels bad :-p | 15:48 | |
| thiagop | Me too, congratz for the nomination JayF and sambetts | 15:49 |
| sambetts | congrats JayF :D | 15:49 |
| thiagop | sambetts: so, since the info isn't on the node, I should be looking an error on n-cpu, right? | 15:50 |
| openstackgerrit | Vladyslav Drok proposed openstack/python-ironicclient: Add create command to ironic client https://review.openstack.org/328955 | 15:51 |
| *** tobias_fiberdata has quit IRC | 15:52 | |
| sambetts | thiagop: 2016-06-16 13:12:18.893 16111 DEBUG nova.virt.ironic.driver [req-675a8e40-c0a5-4c02-bb28-907f9819e647 tempest-BaremetalBasicOps-688855481 tempest-BaremetalBasicOps-688855481] plug: instance_uuid=feaf146a-b961-4e2f-9a80-5f9377cdb32f vif=[] _plug_vifs /opt/stack/new/nova/nova/virt/ironic/driver.py:999 | 15:52 |
| sambetts | thiagop: it appears plug vifs is being called without any vifs | 15:53 |
| *** tobias_fiberdata has joined #openstack-ironic | 15:53 | |
| sambetts | thiagop: 2016-06-16 13:12:17.130 16111 DEBUG nova.network.neutronv2.api [req-675a8e40-c0a5-4c02-bb28-907f9819e647 tempest-BaremetalBasicOps-688855481 tempest-BaremetalBasicOps-688855481] [instance: feaf146a-b961-4e2f-9a80-5f9377cdb32f] No network configured allocate_for_instance /opt/stack/new/nova/nova/network/neutronv2/api.py:544 | 15:53 |
| thiagop | sambetts: which log is that? | 15:53 |
| sambetts | thiagop: n-cpu | 15:54 |
| sambetts | it appears that nova is not creating a port for your instance | 15:54 |
| *** Larion has joined #openstack-ironic | 15:56 | |
| thiagop | sambetts: I'll see if I missed some configuration... | 15:56 |
| *** e0ne has quit IRC | 15:56 | |
| *** z- has quit IRC | 15:57 | |
| *** z- has joined #openstack-ironic | 15:57 | |
| *** sacharya_ has joined #openstack-ironic | 15:57 | |
| *** garthb has joined #openstack-ironic | 15:58 | |
| *** ipukha has left #openstack-ironic | 15:59 | |
| *** tobias_fiberdata has quit IRC | 15:59 | |
| sambetts | thiagop: 13:12:52 2016-06-16 13:12:15,809 17489 INFO [tempest.lib.common.rest_client] Request (BaremetalBasicOps:test_baremetal_server_ops): 202 POST http://127.0.0.1:8774/v2.1/servers 0.636s | 15:59 |
| sambetts | 13:12:52 2016-06-16 13:12:15,810 17489 DEBUG [tempest.lib.common.rest_client] Request - Headers: {'Content-Type': 'application/json', 'Accept': 'application/json', 'X-Auth-Token': '<omitted>'} | 15:59 |
| sambetts | 13:12:52 Body: {"server": {"imageRef": "9bd87a36-af1e-42f6-bf8f-d3dd90cfadee", "flavorRef": "94edacbd-d0fd-4ac8-a474-7d9dd611ad59", "name": "tempest.common.compute-instance-1828306964", "key_name": "tempest-BaremetalBasicOps-758411598"}} | 15:59 |
| sambetts | I don't see any network info being passed as part of the instance boot | 15:59 |
| sambetts | I do see this: 13:12:52 2016-06-16 13:12:15,171 17489 WARNING [tempest.common.fixed_network] Network with name: private not found | 16:00 |
| thiagop | sambetts: I thought it was automagic | 16:00 |
| *** sacharya has quit IRC | 16:00 | |
| thiagop | This last line seems very strange | 16:00 |
| *** tesseract has quit IRC | 16:00 | |
| thiagop | devstack always creates this 'private' network, wth? | 16:00 |
| sambetts | your BM network in neutron needs to be called 'private' for the Ironic devstack script and the tempest to pick it up | 16:00 |
| sambetts | thiagop: are you running flat provider networking> | 16:01 |
| sambetts | ? | 16:01 |
| *** dprince has quit IRC | 16:01 | |
| *** dprince has joined #openstack-ironic | 16:02 | |
| thiagop | sambetts: on the CI cloud, yes. On devstack, don't know. Not by manual configuration though. | 16:02 |
| sambetts | thiagop: if you are running flat networking for your CI via devstack, e.g. Q_USE_PROVIDER_NETWORKING=True and PROVIDER_NETWORK_TYPE="flat", then the network created is called "public" by default, and you need to add PHYSICAL_NETWORK=private to make it be called "private" so that Ironic devstack/tempest can use it | 16:03 |
| *** jaybeale has joined #openstack-ironic | 16:04 | |
| *** Larion has quit IRC | 16:04 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-specs: Collect system logs from IPA https://review.openstack.org/323511 | 16:04 |
| lucasagomes | folks I'm calling it a day | 16:05 |
| *** jvgrant_ has quit IRC | 16:05 | |
| lucasagomes | have a great evening all, talk to you all tomorrow | 16:06 |
| thiagop | sambetts: OK, that makes sense | 16:06 |
| sambetts | cya lucasagomes | 16:06 |
| * lucasagomes gotta run to catch his train back home | 16:06 | |
| lucasagomes | see ya | 16:06 |
| thiagop | good night lucasagomes | 16:06 |
| *** lucasagomes is now known as lucas-afk | 16:06 | |
| NobodyCam | night lucas | 16:06 |
| *** milan has joined #openstack-ironic | 16:06 | |
| * thiagop doesn't handle neutron very well | 16:07 | |
| *** nicodemos has quit IRC | 16:07 | |
| sambetts | thiagop: ;) its a difficult dance to learn | 16:07 |
| *** sacharya has joined #openstack-ironic | 16:07 | |
| *** nicodemos has joined #openstack-ironic | 16:07 | |
| rloo | TheJulia, NobodyCam: what do you think about Satoru's comment about the volume target info being associated with the node and the potential security risk, line 261: https://review.openstack.org/#/c/200496/38/specs/approved/volume-connection-information.rst | 16:09 |
| *** sacharya_ has quit IRC | 16:10 | |
| TheJulia | rloo: let me get some coffee and I'll take a look | 16:10 |
| NobodyCam | rloo: looking at ANC now will look in couple minutes | 16:10 |
| rloo | thx TheJulia, NobodyCam. | 16:10 |
| *** mgoddard_ has quit IRC | 16:11 | |
| *** sabeen has joined #openstack-ironic | 16:11 | |
| rloo | so, for folks that want a little break, this is interesting. removing osc from requirements: https://review.openstack.org/#/c/297730/ | 16:15 |
| jlvillal | rloo: I thought we were going the other way (says that without looking at the patch) | 16:15 |
| rloo | pavlo is right, osc isn't needed to run the ironic CLI part. but it is needed to run openstack baremetal. | 16:15 |
| rloo | jlvillal: our client has an osc plugin. but it won't work w/o osc. | 16:15 |
| rloo | i'm not sure what the right answer is. | 16:16 |
| jlvillal | rloo: Ah, now reading the patch makes more sense. | 16:16 |
| rloo | if we remove it we should at least document that osc needs to be installed. | 16:16 |
| rloo | but if we deprecate/remove ironic CLI in the future, wouldn't osc be needed, or would we still leave it as is, w/o osc being installed. | 16:17 |
| rloo | i wonder what the other osc plugins do. | 16:17 |
| *** sborkows has quit IRC | 16:18 | |
| *** mgoddard has joined #openstack-ironic | 16:18 | |
| *** dprince has quit IRC | 16:20 | |
| *** sdake has quit IRC | 16:20 | |
| sambetts | I guess it depends how you view drivers/plugins and dependencies, the fact that ironic client needs OSC is like Ironic needing nova, neutron, and glance etc but we don't make them dependecies | 16:21 |
| *** dprince has joined #openstack-ironic | 16:21 | |
| rloo | sambetts: so ironic doesn't need nova. it can be standalone. | 16:22 |
| sambetts | rloo: the same as ironic client can run standalone without OSC | 16:22 |
| rloo | sambetts: and right now, the client has both CLI and OSC plugin. So I can see removing osc. but we WILL (heh) remove the CLI at some point. | 16:22 |
| rloo | sambetts: so we remove osc now, after we remove CLI, we add osc back. or maybe by then, things will have changed wrt osc/plugins, dunno. | 16:23 |
| rloo | sambetts: i think the reason i don't want to remove osc, is because i want to encourage folks to use osc instead of cli, cuz we're going to deprecate that sooner rather than later I hope. | 16:24 |
| rloo | sambetts: but yeah, it is an osc plugin. so I'm back to 'i don't know what the best solution is'. so going to ignore it i guess. | 16:24 |
| sambetts | I don't even think we need to then, because its a plugin, if we don't import any code from it, technically another something like OSC might come along which is OSC plugin compatible | 16:24 |
| sambetts | which could use our plugin, do we make ironicclient depend on that too | 16:25 |
| rloo | sambetts: true. the thing is, we've implemented it as a plugin cuz of the way osc is done. but for a user, should they know htat it is a plugin and that it won't work if they don't also install osc? | 16:26 |
| rloo | sambetts: i think the problem is that osc should have made some basic package w/o the actual commands from keystone/nova/etc. then we'd all be in an even playing field. | 16:27 |
| sambetts | rloo: ++, I think that was brought up in the OSC session at the summit | 16:27 |
| rloo | sambetts: well, that may not be the problem. maybe the problem is that we should put the osc plugin in a separate package from python-ironicclient. | 16:27 |
| *** vmud213 has joined #openstack-ironic | 16:27 | |
| *** moshele has quit IRC | 16:27 | |
| stevemar | rloo: sambetts so we haven't publicly advertised it yet, but we have an osc-lib package that's almost ready for public consumption | 16:28 |
| stevemar | rloo: sambetts it'll contain all the base OSC bits without the keystone/nova/etc commands | 16:28 |
| *** sdake has joined #openstack-ironic | 16:28 | |
| rloo | stevemar to the rescue! :) | 16:28 |
| sambetts | stevemar: whats your stance on OSC plugins importing OSC or not? | 16:29 |
| stevemar | rloo: sambetts so folks don't have to depend on python-openstackclient, but rather osc-lib (MUCH smaller) instead | 16:29 |
| rloo | stevemar: when do you think the package will be avail? | 16:29 |
| sambetts | stevemar: even if we don't use code from osc-lib/ | 16:29 |
| sambetts | ? | 16:29 |
| stevemar | #link https://github.com/openstack/osc-lib | 16:29 |
| TheJulia | rloo: replied, hopefully that thought makes sense | 16:29 |
| stevemar | sambetts: so i think most folks didn't want it in requirements cause it brought in so many dependencies (especially if folks just wanted to use it as a python lib and not a CLI) | 16:30 |
| stevemar | sambetts: using osc-lib should clean that mess up | 16:30 |
| sambetts | stevemar: I guess my question is if we require on osc-lib will we still need to install main OSC to actually use the command client? | 16:31 |
| stevemar | sambetts: nope! | 16:31 |
| sambetts | stevemar: oh awesome, so main OSC repo basically becomes a repo that just hold plugins for the main projects? | 16:32 |
| *** dtantsur|bbl is now known as dtantsur | 16:32 | |
| stevemar | sambetts: at least... i hope not | 16:32 |
| stevemar | let me double check | 16:32 |
| stevemar | sambetts: want to jump over to #openstack-sdks | 16:32 |
| sambetts | sure | 16:32 |
| stevemar | sambetts: dtroyer isn't here | 16:32 |
| dtantsur | NobodyCam, could you please provide an example where I can NOT bump API version when introducing new API? | 16:33 |
| *** bharath has joined #openstack-ironic | 16:34 | |
| dtantsur | jroll, I didn't put the API version bumping to the spec for the same reason I didn't say that I'll be following i18n guidelines, writing proper unit tests and using creating release notes :) because it's a requirement for all API changes | 16:35 |
| rloo | jlvillal: did you see david lyle's comment wrt the apache bug: https://bugs.launchpad.net/ironic/+bug/1590139 | 16:36 |
| openstack | Launchpad bug 1590139 in Ironic "gate job fails when unable to start apache service during horizon setup" [Critical,Triaged] | 16:36 |
| *** appu103_ has joined #openstack-ironic | 16:38 | |
| sambetts | rloo: I've got a conversation going in #openstack-sdks about that patch | 16:38 |
| jlvillal | rloo: I did. It makes sense. | 16:38 |
| rloo | sambetts: let me know what is decided :) | 16:38 |
| rloo | jlvillal: sigh. | 16:38 |
| jlvillal | rloo: I am looking at logstash to figure out approximately when it started | 16:39 |
| rloo | thx jlvillal | 16:39 |
| jlvillal | My logstash kung fu is not very strong... | 16:39 |
| *** jvgrant_ has joined #openstack-ironic | 16:40 | |
| sambetts | rloo: will do :) | 16:41 |
| sambetts | rloo: left a comment on that patch about waiting until we hear back | 16:42 |
| NobodyCam | dtantsur: replied and vote changed you now have 2 +2's ... any one else what to look or I'm going to land it :) | 16:42 |
| *** dmk0202 has quit IRC | 16:42 | |
| dtantsur | NobodyCam, thanks | 16:43 |
| NobodyCam | no no thank you :) | 16:43 |
| JayF | clif_h: "done" is a relative thing. When that bug was filed there were 0 docs, now there are tons and tons. I got a +1 in here before closing it, but I don't think that means we can't keep working on it | 16:46 |
| JayF | clif_h: if you have an outstanding docs patch tho I'd love to review it, going to look for it now | 16:47 |
| rloo | sambetts: so they don't know yet, when osc-lib will be avail? | 16:47 |
| NobodyCam | JayF: did you want to take a look at https://review.openstack.org/#/c/306418 you had a +2 before | 16:47 |
| JayF | already open | 16:48 |
| JayF | in my tabs | 16:48 |
| sambetts | rloo: I hoping to find that out, and if osc-lib actually does what we think it does | 16:48 |
| NobodyCam | :) | 16:48 |
| *** jaybeale has quit IRC | 16:49 | |
| rloo | sambetts: ok, thx. | 16:49 |
| *** david-lyle has joined #openstack-ironic | 16:50 | |
| jlvillal | rloo: Unfortunately the logstash data seems to have been dropped for anything before the 9th. And I filed the bug on the 7th. | 16:50 |
| rloo | jlvillal: hmm. i wonder what changed on 6th or 7th. | 16:51 |
| dtantsur | TIL https://github.com/openstack/osc-lib thanks sambetts | 16:52 |
| sambetts | rloo, jlvillal: was that the day logs.openstack.org blew up and broke the gate for the day? | 16:52 |
| jlvillal | sambetts: I don't recall | 16:53 |
| rloo | sambetts: i don't think so. | 16:53 |
| vdrok | sambetts: but still I will be pulling this osc-lib even if I only need ironic commands? | 16:53 |
| dtantsur | wow, so many comments on the driver composition spec, thanks jroll, rloo, vdrok. I'll slowly start addressing them tomorrow :) | 16:53 |
| vdrok | :) | 16:53 |
| *** jvgrant_ has quit IRC | 16:53 | |
| rloo | dtantsur: you asked and you received :) | 16:53 |
| vdrok | sambetts: or if I only need python api | 16:54 |
| dtantsur | heh, true | 16:54 |
| JayF | dtantsur: re: agent api spec; is the intention to *ever* deprecate the passthru endpoints/ | 16:54 |
| rloo | dtantsur: i'm not done but wanted to make sure i understood it before i continued. | 16:54 |
| *** mbound has quit IRC | 16:54 | |
| vdrok | sambetts: I'd expect that if someone wants to use openstack command, he would install openstack client :) | 16:54 |
| dtantsur | JayF, yes. and the whole AgentVendorPassthru | 16:54 |
| *** bcornec1 has quit IRC | 16:55 | |
| dtantsur | JayF, the reason is driver composition: I don't want to bring vendor mixins there | 16:55 |
| JayF | dtantsur: just making sure my assumption was good; it occurred to me something api-ish could be different | 16:55 |
| *** rbrud has joined #openstack-ironic | 16:55 | |
| rloo | vdrok: sambetts and i had a discussion earlier in irc, did you see that? | 16:55 |
| dtantsur | rloo, I'll try to answer what's easily answerable right now then | 16:55 |
| vdrok | looking | 16:56 |
| rloo | dtantsur: don't answer here. going into a meeting soon. | 16:56 |
| sambetts | vdrok: we plan to remove the ironic commands, and only have openstack client in the future, and we want to encrourage people to use the openstack client over the ironic ones | 16:56 |
| vdrok | yeah, but there is the case of only python api still. tho if it is small, maybe that's OK | 16:56 |
| sambetts | there is an interesting case of CLI vs python language hook | 16:57 |
| dtantsur | rloo, yeah, on the spec ofc | 16:57 |
| rloo | dtantsur: thx! (although not sure i have time to review it again today) | 16:57 |
| rloo | vdrok: would it make sense to put our osc plugin in a separate package? | 16:58 |
| dtantsur | no worries | 16:58 |
| dtantsur | rloo, we don't do it now with our "ironic" command, right? though it would probably cut some dependencies | 16:58 |
| sambetts | that package is called ironicclient so I expect it to be a client, maybe we need to split the language hook ] | 16:58 |
| vdrok | hm | 16:59 |
| *** derekh has quit IRC | 16:59 | |
| vdrok | that would be a breakage for lots of people I think | 16:59 |
| *** PollyZ_ has joined #openstack-ironic | 16:59 | |
| sambetts | so would removing the current requirement, if people are only installing python-ironiclient and then using the openstack... commands | 17:00 |
| vdrok | yeah, maybe you're right. but I personally would not expect openstackclient being installed when I install ironic client :) will hold +2 for now then | 17:01 |
| *** PollyZ has quit IRC | 17:02 | |
| *** nicodemos has quit IRC | 17:03 | |
| *** nicodemos has joined #openstack-ironic | 17:03 | |
| sambetts | vdrok: maybe we don't advertise our OSC intergration well enough | 17:04 |
| clif_h | JayF: I don't have an outstanding docs patch atm | 17:04 |
| clif_h | just wondering if closing that was premature | 17:04 |
| vdrok | sambetts: heh, we don't do it at all :) http://docs.openstack.org/developer/python-ironicclient/ | 17:05 |
| JayF | clif_h: I put a comment in the bug responding as well; tl;dr we have enough docs now that probably a catchall bug is not super valuable. However we should (I should, probably as docs liason) file smaller bugs for each module and mark low-hanging-fruit | 17:05 |
| *** mbound has joined #openstack-ironic | 17:05 | |
| *** daemontool has quit IRC | 17:08 | |
| *** trown is now known as trown|lunch | 17:08 | |
| *** PollyZ_ has quit IRC | 17:08 | |
| *** PollyZ has joined #openstack-ironic | 17:08 | |
| *** ifarkas has quit IRC | 17:08 | |
| dtantsur | clif_h, as a person managing our bug list of time to time, I prefer bugs with a clear definition of done | 17:09 |
| dtantsur | :) | 17:09 |
| *** wagiel has joined #openstack-ironic | 17:09 | |
| livelace | Incredible useful article http://apahim.livejournal.com/2395.html | 17:11 |
| stevemar | dtantsur: we plan on advertising osc-lib soon :) | 17:11 |
| wagiel | lucasagomes, I see in logfile, DHCP, sockets bound exclusively to interface tap75c17816-33, tap is port from br-int, would that means i should be expecting dhcp request w/ tcpdump on that interface?~> | 17:11 |
| wagiel | ~.~> | 17:11 |
| stevemar | dtantsur: it's not ready for public consumption just yet, but i saw the convo happening and had to chime in | 17:11 |
| clif_h | dtantsur: makes sense | 17:11 |
| dtantsur | stevemar, awesome, thanks! will be looking forward to it | 17:12 |
| stevemar | dtantsur: we'll post on the -dev ML when it's ready | 17:12 |
| stevemar | shouldn't be too long now | 17:12 |
| *** bharath has quit IRC | 17:12 | |
| dtantsur | vdrok, as a data point: ironic-inspector-client depends on OSC, but it's our only CLI, we don't have another one | 17:12 |
| stevemar | I think we started it on May 10th :) | 17:12 |
| *** bharath has joined #openstack-ironic | 17:13 | |
| *** sabeen has quit IRC | 17:13 | |
| *** wagiel has quit IRC | 17:14 | |
| TheJulia | vdrok: seems like your -1 on 275766 is purely over doc and I need to push a follow-up revision anyway, would you be okay if we did that instead of revising the base revision, and spending another week or so getting reviews? | 17:14 |
| TheJulia | s/base revision/changeset/ | 17:14 |
| vdrok | dtantsur: I see. so you think we should leave it? | 17:15 |
| dtantsur | vdrok, I'm a bit undecided, need to think more | 17:15 |
| dtantsur | I'm pretty sure we'll have to depend on it as soon as we deprecate the 'ironic' command | 17:16 |
| vdrok | TheJulia: well, it is, kind of. the main reason for -1 is this inconsistency - boot.validate and deploy.prepare calls | 17:16 |
| dtantsur | vdrok, also as a packager I probably won't remove this dependency from our RPM | 17:16 |
| TheJulia | vdrok: that deploy.prepare does not check the checksum? | 17:16 |
| TheJulia | vdrok: that would be a bug oustide of ANC | 17:16 |
| dtantsur | vdrok, so yeah, I'm slightly towards leaving it there | 17:16 |
| TheJulia | and truthfully, it might be intended | 17:16 |
| vdrok | TheJulia: what I mean is - there is a possibility that deploy.validate does check something that deploy.prepare relies on, and we do only boot.validate. but then, afterwards it means that it will fail anyway I think... | 17:18 |
| vdrok | yeah, I think I'm ok with that, thank you TheJulia :) | 17:18 |
| *** ohamada has quit IRC | 17:18 | |
| dtantsur | devananda, when you're here please take a look at https://review.openstack.org/#/c/306418/ it's collected 3x +2 already | 17:19 |
| jroll | TheJulia: we don't check checksums in validate because we download the image to the agent, not to the conductor :) | 17:19 |
| TheJulia | vdrok: that is a good point, at the same time there is no guarentee that the user even care about providing real information that needs to be validated :) but deploy.prepare does actually attempt to check if a file exists and fails appropriately | 17:20 |
| TheJulia | jroll: yup, that is what I meant by intended, for the most part at least | 17:20 |
| jroll | yeah | 17:20 |
| dtantsur | ok, calling it a day now, see you tomorrow | 17:20 |
| vdrok | yep, I just cannot come up with something that will break because of that | 17:20 |
| * sambetts -> home | 17:20 | |
| *** dtantsur is now known as dtantsur|afk | 17:20 | |
| sambetts | night all o/ | 17:20 |
| vdrok | night dtantsur and sambetts | 17:20 |
| *** sambetts is now known as sambetts|afk | 17:21 | |
| sambetts|afk | cya vdrok | 17:21 |
| *** rcernin has quit IRC | 17:21 | |
| *** athomas has quit IRC | 17:22 | |
| *** PollyZ has quit IRC | 17:27 | |
| *** moshele has joined #openstack-ironic | 17:28 | |
| *** ccamacho has quit IRC | 17:30 | |
| *** MrWorldWide has joined #openstack-ironic | 17:33 | |
| *** MrWorldWide has quit IRC | 17:33 | |
| TheJulia | vdrok: deploy.validate does actually have a checksum presence check, but truthfully I think that is the only thing that might be useful to call out of what deploy.validate consists of since the goal for the user is to just create a node in active state. Granted, some of those things might matter upon a redeployment, but from ANC's perspective, the node is already deployed. | 17:34 |
| vdrok | TheJulia: yeah, I agree | 17:37 |
| *** ChubYann has joined #openstack-ironic | 17:38 | |
| *** aparnav has joined #openstack-ironic | 17:41 | |
| *** appu103_ has quit IRC | 17:41 | |
| *** purgat0ry has joined #openstack-ironic | 17:45 | |
| purgat0ry | Hey, I have a question. Is Ironic for you to put OpenStack onto bare metal? | 17:46 |
| mgould | purgat0ry: that's not its core use case, but it can be used for that | 17:46 |
| *** aparnav has quit IRC | 17:47 | |
| clif_h | purgat0ry: more like use ironic to manage bare-metal nodes | 17:47 |
| mgould | the core use case is "let users of my cloud provision bare metal machines as if they were VMs" | 17:47 |
| mgould | but the TripleO project does use Ironic internally to put OpenStack services onto bare metal machines | 17:48 |
| purgat0ry | Why is it so difficult to install OpenStack onto bare metal? It seems like doing it all in a VM is doing too much. it seems like more work to do it through a VM. Why do it that way? | 17:48 |
| TheJulia | purgat0ry: For testing maybe? | 17:49 |
| krtaylor | hm, yes, for all the same reasons that you would us a VM in the first place, privacy, other applications resources, isolation, etc | 17:51 |
| TheJulia | purgat0ry: For most, the difficulty seems to be choosing a method, or configuring all the details to meet your needs, which vary from deployment to deployment... at that point, there doesn't seem to be much of a difference, except deploying everything in VMs seems like it would only really be for testing. | 17:51 |
| krtaylor | btw, I pushed the WIP patch last night to add all remaining drivers to stackalytics driver page, reviews appreciated -> https://review.openstack.org/330270 | 17:52 |
| JayF | purgat0ry: if you're interested in using openstack to deploy openstack onto bare metal, there's a whole project around that called tripleo | 17:52 |
| JayF | purgat0ry: also aiui the new OpenStack Ansible stuff is pretty good, but that's more a pure install automation piece that direct integration with openstack | 17:52 |
| * mgould -> home | 17:52 | |
| *** mgould is now known as mgould|afk | 17:52 | |
| jlvillal | purgat0ry: Using Bifrost/Ironic and something else is a way to install OpenStack onto bare-metal. The something else could be Kolla or OpenStack Ansible or ????? | 17:55 |
| cinerama | jlvillal, kolla are doing stuff to use bifrost aiui | 17:55 |
| JayF | aiui kolla is basically forked openstack ansible designed to use docker instead of lxc directly | 17:56 |
| JayF | I usually reccomend openstack ansible because it's the OG :P | 17:56 |
| *** vmud213 has quit IRC | 17:56 | |
| TheJulia | IBM and Fujitsu CI systems failed on ANC, but in neutron network setup.... | 17:56 |
| *** PollyZ has joined #openstack-ironic | 17:57 | |
| *** electrofelix has quit IRC | 17:58 | |
| *** mtanino has quit IRC | 17:59 | |
| *** baoli_ has quit IRC | 18:00 | |
| *** PollyZ has quit IRC | 18:00 | |
| *** Sukhdev has joined #openstack-ironic | 18:03 | |
| *** moshele has quit IRC | 18:05 | |
| vdrok | JayF: https://review.openstack.org/330738 I've added a depends-on on the bugfix, otherwise both will be just failing | 18:06 |
| *** rcernin has joined #openstack-ironic | 18:07 | |
| vdrok | good night everyone! | 18:09 |
| *** trown|lunch is now known as trown | 18:09 | |
| *** sabeen has joined #openstack-ironic | 18:10 | |
| JayF | vdrok: catch you in time? | 18:10 |
| vdrok | yep | 18:10 |
| JayF | vdrok: had a thought. There's also a bug open about testing localboot in the agent | 18:10 |
| JayF | vdrok: maybe we should have parition images localboot in that job, and kill two bugs with one job? | 18:11 |
| *** mtanino has joined #openstack-ironic | 18:11 | |
| JayF | I'm unsure if localboot support exists in devstack at all today though, so that may be a deep rabbithole | 18:11 |
| *** PollyZ has joined #openstack-ironic | 18:11 | |
| vdrok | JayF: yeah, I don't know too | 18:11 |
| vdrok | I'll look at it tomorrow then, should I wip it for now? | 18:12 |
| JayF | nah | 18:12 |
| JayF | this is beter than nothing was just a thought | 18:12 |
| JayF | I always favor better sooner :) we can always change it later if we want | 18:12 |
| vdrok | ok, will look into it, thanks! | 18:12 |
| *** purgat0ry has quit IRC | 18:14 | |
| *** baoli has joined #openstack-ironic | 18:16 | |
| *** bharath has quit IRC | 18:19 | |
| *** PollyZ has quit IRC | 18:27 | |
| *** rama_y has joined #openstack-ironic | 18:30 | |
| *** jvgrant_ has joined #openstack-ironic | 18:30 | |
| *** jvgrant_ has quit IRC | 18:32 | |
| *** sacharya has quit IRC | 18:32 | |
| *** sacharya_ has joined #openstack-ironic | 18:32 | |
| *** sacharya has joined #openstack-ironic | 18:33 | |
| *** baoli has quit IRC | 18:34 | |
| *** aarefiev_ has joined #openstack-ironic | 18:34 | |
| *** baoli has joined #openstack-ironic | 18:37 | |
| *** sacharya_ has quit IRC | 18:37 | |
| *** PollyZ has joined #openstack-ironic | 18:37 | |
| *** mgoddard has quit IRC | 18:42 | |
| *** dmk0202 has joined #openstack-ironic | 18:44 | |
| *** dmk0202 has quit IRC | 18:45 | |
| *** amit213 has joined #openstack-ironic | 18:51 | |
| *** baoli has quit IRC | 18:56 | |
| *** e0ne has joined #openstack-ironic | 18:58 | |
| *** yolanda has quit IRC | 19:00 | |
| *** ijw has joined #openstack-ironic | 19:03 | |
| *** ijw has quit IRC | 19:06 | |
| openstackgerrit | Merged openstack/ironic: No need for 'default=None' in config variable https://review.openstack.org/330557 | 19:06 |
| *** ijw has joined #openstack-ironic | 19:06 | |
| openstackgerrit | Merged openstack/ironic: Fix typo in agent driver https://review.openstack.org/330549 | 19:09 |
| *** sacharya has quit IRC | 19:13 | |
| *** sacharya has joined #openstack-ironic | 19:13 | |
| *** e0ne has quit IRC | 19:22 | |
| *** dprince has quit IRC | 19:27 | |
| *** SiRiuS__ has joined #openstack-ironic | 19:28 | |
| *** MattMan has quit IRC | 19:32 | |
| *** MattMan has joined #openstack-ironic | 19:32 | |
| *** jvgrant_ has joined #openstack-ironic | 19:33 | |
| *** dmk0202 has joined #openstack-ironic | 19:40 | |
| *** PollyZ has quit IRC | 19:42 | |
| *** mbound has quit IRC | 19:43 | |
| *** dmk0202 has quit IRC | 19:43 | |
| *** sabeen has quit IRC | 19:46 | |
| *** jvgrant_ has quit IRC | 19:47 | |
| *** fragatin_ has quit IRC | 19:51 | |
| *** sabeen has joined #openstack-ironic | 19:58 | |
| JayF | https://review.openstack.org/#/c/330089/3 is pretty simple (+17,-8) and fixes a pretty big bug (partition images + agent driver) if anyone has a little review bandwidth to throw at it | 20:03 |
| *** PollyZ has joined #openstack-ironic | 20:03 | |
| *** baoli has joined #openstack-ironic | 20:07 | |
| *** fragatina has joined #openstack-ironic | 20:10 | |
| *** fragatina has quit IRC | 20:10 | |
| *** fragatina has joined #openstack-ironic | 20:11 | |
| *** jcoufal has quit IRC | 20:12 | |
| openstackgerrit | Merged openstack/python-ironicclient: Implementation of baremetal port create https://review.openstack.org/172461 | 20:13 |
| *** rbrud has quit IRC | 20:16 | |
| *** ayoung has quit IRC | 20:18 | |
| *** openstackstatus has joined #openstack-ironic | 20:19 | |
| *** ChanServ sets mode: +v openstackstatus | 20:19 | |
| *** jvgrant_ has joined #openstack-ironic | 20:19 | |
| *** aarefiev_ has quit IRC | 20:20 | |
| *** lucas-afk has quit IRC | 20:21 | |
| rloo | JayF: wrt that patch, what does the '(+17,-8)' mean? | 20:22 |
| JayF | lines changed | 20:22 |
| JayF | +17 -8 | 20:22 |
| JayF | so small diff | 20:22 |
| rloo | JayF: small is -1 :) | 20:22 |
| JayF | lol | 20:22 |
| rloo | JayF: is it a high priority thing? | 20:22 |
| JayF | It completely breaks partition images with agent driver | 20:22 |
| JayF | it was discovered when vdrok setup the partition image job | 20:23 |
| rloo | JayF: ok, then it is worth looking at. Looking... | 20:23 |
| JayF | so we want to fix it, and get partition image jobs running so it doesn't happe nagain | 20:23 |
| mat128 | JayF: ephemeral_mb vs gb | 20:23 |
| mat128 | omg | 20:23 |
| JayF | I used to think the 'if it isn't tested, it doesn't work' opinions were overkill. After this week, I'm completely sold | 20:23 |
| *** dmk0202 has joined #openstack-ironic | 20:23 | |
| rloo | JayF: did this used to work? | 20:23 |
| JayF | I honestly have never personally tested it, I don't use partition images downstream at all | 20:24 |
| JayF | but it's supposed to work | 20:24 |
| rloo | JayF: "supposed to"... | 20:24 |
| JayF | Meaning support was added, but it was never tested in the gate | 20:24 |
| JayF | so I can't say personally with 100% certainty it did | 20:24 |
| JayF | because I didn't do it, and tempest never tested it | 20:24 |
| JayF | that's the only reason I hedged there :) | 20:25 |
| rloo | JayF: i was just amused. All the stuff we put in ironic are supposed to work ;) | 20:25 |
| JayF | Well like I said, I'm being rapidly sold on CI | 20:25 |
| JayF | after all the talk of DIB images working and then losing a day because they don't + this being broken, etc | 20:26 |
| rloo | ++ | 20:26 |
| JayF | just lots of examples of nontested stuff being broken have been making my life bad this week | 20:26 |
| JayF | lol | 20:26 |
| *** jvgrant_ has quit IRC | 20:26 | |
| * rloo thinks it must be raining for JayF | 20:27 | |
| *** piet has quit IRC | 20:28 | |
| *** ChrisAusten has quit IRC | 20:28 | |
| *** lucasagomes has joined #openstack-ironic | 20:28 | |
| *** mbound has joined #openstack-ironic | 20:33 | |
| *** jvgrant_ has joined #openstack-ironic | 20:41 | |
| *** rbrud has joined #openstack-ironic | 20:41 | |
| NobodyCam | rloo: quick question on the comment on https://review.openstack.org/#/c/272658/53/ironic/api/app.py Line#124 | 20:43 |
| NobodyCam | the PycadfAuditApiConfigError would have a message like 'Error parsing audit map file: %s' % err | 20:43 |
| rloo | NobodyCam: so if you saw 'File ... cannot be read. Error parsing audit map file: ..' that would be OK? | 20:44 |
| rloo | NobodyCam: maybe we should just change that exception string. | 20:44 |
| NobodyCam | ya kinda makes me want a new error type... /me adds | 20:44 |
| rloo | NobodyCam: what happens if there is an EnvironmentError, OSError. | 20:45 |
| rloo | would it show eg 'File .. cannot be read. File cannot be read'? | 20:45 |
| *** PollyZ has quit IRC | 20:45 | |
| rloo | Maybe it can be something like 'Error with file x. Reason: blah'. | 20:45 |
| rloo | do we already have some sort of exception with files? | 20:46 |
| NobodyCam | lol... good point ... the error is a general InputFileError I can adjust the test to remove can not be read | 20:46 |
| NobodyCam | i'll take a look | 20:46 |
| wajdi | "blah" would still be the exception though right? You don't want to lose that information. | 20:47 |
| *** [1]cdearborn has joined #openstack-ironic | 20:47 | |
| rloo | wajdi: yeah, we all want to see blah still :) | 20:47 |
| NobodyCam | oh ya | 20:47 |
| rloo | wajdi: it is the hardcoded part that may not make sense. | 20:48 |
| rloo | wajdi: for that particular exception. i don't have the string handy to show you. | 20:48 |
| wajdi | seems to be: _msg_fmt = _("File %(file_name)s cannot be read. Reason: %(reason)s") | 20:49 |
| NobodyCam | wajdi: that's what is in the patch now | 20:51 |
| NobodyCam | I kinda like "Error with file %(file_name)s. Reason: %(reason)s" | 20:52 |
| rloo | wajdi: yeah, that's it. the problem is the 'cannot be read' part. could be confusing depending on the actual error | 20:52 |
| wajdi | Right. Yeah. Error with file.... works. | 20:52 |
| *** cdearborn has quit IRC | 20:53 | |
| rloo | NobodyCam: or 'Excuse me, we seem to have a problem with your file .' | 20:53 |
| *** baoli has quit IRC | 20:53 | |
| * NobodyCam thinks rloo is too nice :p | 20:53 | |
| wajdi | Pardon me, sorry to disturb your work. But your file seems to be out of sorts. :P | 20:54 |
| NobodyCam | lol | 20:54 |
| NobodyCam | ++ | 20:54 |
| wajdi | I think I would have a good laugh if I saw that come up in a log | 20:54 |
| TheJulia | ++ | 20:54 |
| *** sacharya has quit IRC | 20:54 | |
| *** [1]cdearborn has quit IRC | 20:54 | |
| rloo | sounds like one of those easter egg things :) | 20:54 |
| *** sacharya has joined #openstack-ironic | 20:55 | |
| rloo | JayF: I reviewed that patch you wanted. Go chase someone else now :) | 20:55 |
| *** jvgrant_ has quit IRC | 20:55 | |
| *** PollyZ has joined #openstack-ironic | 20:55 | |
| *** sacharya_ has joined #openstack-ironic | 20:56 | |
| JayF | rloo: maybe if I wait long enough I can land it :P | 20:56 |
| rloo | JayF: i think we usually wait a few days, not necessarily a week. Maybe it depends on how many cores reply, can't remember. | 20:56 |
| rloo | JayF: but yeah. | 20:56 |
| JayF | I know, it's mostly a joke | 20:58 |
| rloo | NobodyCam: for that audit middleware one, will you/someone add documentation as to how it works? cuz it doesn't explain it now. | 20:58 |
| NobodyCam | rloo: as part of the patch or is a follow up okay? | 20:59 |
| *** sacharya has quit IRC | 20:59 | |
| rloo | NobodyCam: a follow up is ok if you are sure it works as desired. although i am not sure i can +2 it w/o actually knowing how it is meant to work. | 21:00 |
| *** trown is now known as trown|outtypewww | 21:00 | |
| NobodyCam | ack. let push up these changes and see what we can do | 21:00 |
| rloo | NobodyCam: but I would be really happy if others like JayF would review it and +2 it. I regret a bit, that I looked at it. Too many reviews this week... | 21:00 |
| *** rbrud has quit IRC | 21:01 | |
| NobodyCam | :) I do thank you for the reviews :) | 21:01 |
| JayF | if you want me to take a look at something, add me as a reviewer | 21:01 |
| * rloo thinks maybe we should have some initiation for new cores. | 21:01 | |
| JayF | If I'm not initiated by now, I never will be | 21:01 |
| rloo | JayF: EVERYTHING! | 21:01 |
| JayF | lol | 21:01 |
| *** wajdi_ has joined #openstack-ironic | 21:01 | |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Add support for the audit middleware https://review.openstack.org/272658 | 21:02 |
| NobodyCam | JayF: ^^^^^ :p | 21:02 |
| *** baoli has joined #openstack-ironic | 21:02 | |
| JayF | I'm in the middle of something but if you add me to reviewers list I'll get to it | 21:02 |
| NobodyCam | :) awesome TY | 21:02 |
| rloo | TheJulia: have you confirmed that this breaks the standalone case? https://review.openstack.org/#/c/285852/ | 21:03 |
| *** rbrud has joined #openstack-ironic | 21:04 | |
| *** wajdi has quit IRC | 21:05 | |
| *** wajdi has joined #openstack-ironic | 21:05 | |
| *** wajdi_ has quit IRC | 21:05 | |
| jroll | rloo: I think you mean a hazing, right? :D | 21:07 |
| *** wajdi_ has joined #openstack-ironic | 21:07 | |
| rloo | jroll: ahhhh, which is more politically correct? ;) | 21:07 |
| jroll | I prefer actually correct to politically :) | 21:07 |
| *** praneshp has joined #openstack-ironic | 21:08 | |
| rloo | jroll: :) | 21:09 |
| TheJulia | rloo: it does break stand-alone | 21:09 |
| *** sacharya has joined #openstack-ironic | 21:10 | |
| rloo | TheJulia: ok, then I'm not going to review it. | 21:10 |
| mat128 | JayF: wajdi told me about that OEM bug | 21:10 |
| JayF | mat128: just thought you'd wanna know, we'll have the patch up for it shortly | 21:10 |
| *** wajdi has quit IRC | 21:11 | |
| *** sachary__ has joined #openstack-ironic | 21:11 | |
| *** wajdi_ has quit IRC | 21:11 | |
| JayF | mat128: also, you all might care to look at my example hardware managers stuff I pushed up to the ml earlier, has some good skeleton code should you choose to make some custom managers | 21:11 |
| openstackgerrit | Brad Morgan proposed openstack/ironic-python-agent: Fix unwanted unmounting of /usr/share/oem https://review.openstack.org/330789 | 21:12 |
| mat128 | JayF: yeah I've seen it in your github account | 21:12 |
| JayF | jroll: devananda ^ if you'd like to vote | 21:12 |
| *** sacharya_ has quit IRC | 21:13 | |
| jroll | +2 | 21:13 |
| mat128 | JayF: you mentioned looking at other labels, have you read https://coreos.com/os/docs/latest/sdk-disk-partitions.html ? | 21:13 |
| JayF | mat128: we did lots of research aruond that over mon/tues | 21:14 |
| *** sacharya has quit IRC | 21:14 | |
| mat128 | JayF: ok, was just making sure we didnt miss "ROOT" or something ;) | 21:14 |
| JayF | Yeah, feel free to do addl research on your own but this is where I disappeared off to early in the week | 21:15 |
| JayF | because logs for this are misleading as well, it'll tell you it mounted stuff when it clearly didn't | 21:15 |
| NobodyCam | rloo: something like this: http://docs.openstack.org/developer/keystonemiddleware/audit.html | 21:15 |
| *** krtaylor has quit IRC | 21:15 | |
| rloo | NobodyCam: Nope. as an operator, i don't see how that helps at all. | 21:15 |
| rloo | NobodyCam: i commented about that in the patch. I don't think the install guide shoudl have a link to that. | 21:16 |
| NobodyCam | ack ok... just started to look for whats out there | 21:16 |
| NobodyCam | I can write up something too | 21:16 |
| rloo | NobodyCam: why does the operator care that it is an 'audit middleware'? and then the rest would be very confusing cuz it isn't being enabled/configured in ironic, as described in that page. | 21:17 |
| rloo | NobodyCam: I mean, I read that, and I looked at the patch, and I thought huh? | 21:17 |
| NobodyCam | ++ ahh that helps me!! | 21:17 |
| rloo | NobodyCam: take a look at the patch, comments from me and (I forgot their name). | 21:18 |
| *** sachary__ has quit IRC | 21:20 | |
| *** sacharya has joined #openstack-ironic | 21:20 | |
| *** baoli has quit IRC | 21:21 | |
| *** sacharya_ has joined #openstack-ironic | 21:21 | |
| *** baoli has joined #openstack-ironic | 21:21 | |
| *** ChrisAusten has joined #openstack-ironic | 21:23 | |
| *** sacharya has quit IRC | 21:25 | |
| *** livelace has quit IRC | 21:30 | |
| *** bcornec1 has joined #openstack-ironic | 21:35 | |
| *** rcernin has quit IRC | 21:36 | |
| *** dmk0202 has quit IRC | 21:38 | |
| *** Goneri has quit IRC | 21:45 | |
| *** wagiel has joined #openstack-ironic | 21:47 | |
| mrda | Morning Ironic | 21:52 |
| NobodyCam | morning mrda | 21:53 |
| mrda | o/ | 21:53 |
| jlvillal | Good morning mrda | 21:54 |
| mrda | hey jlvillal | 21:58 |
| thiagop | morning mrda , night folks! | 22:01 |
| * thiagop out | 22:01 | |
| *** thiagop has quit IRC | 22:01 | |
| *** wagiel has quit IRC | 22:02 | |
| *** amotoki has joined #openstack-ironic | 22:03 | |
| *** dmk0202 has joined #openstack-ironic | 22:04 | |
| *** ayoung has joined #openstack-ironic | 22:07 | |
| *** rbrud has quit IRC | 22:11 | |
| *** baoli has quit IRC | 22:11 | |
| *** PollyZ has quit IRC | 22:12 | |
| *** baoli has joined #openstack-ironic | 22:12 | |
| *** PollyZ has joined #openstack-ironic | 22:12 | |
| *** ChrisAusten has quit IRC | 22:20 | |
| *** PollyZ has quit IRC | 22:23 | |
| *** PollyZ has joined #openstack-ironic | 22:24 | |
| *** baoli has quit IRC | 22:27 | |
| *** baoli has joined #openstack-ironic | 22:27 | |
| openstackgerrit | John L. Villalovos proposed openstack/ironic: Trying to debug apache restart issue https://review.openstack.org/330812 | 22:28 |
| *** dmk0202 has quit IRC | 22:39 | |
| *** baoli has quit IRC | 22:40 | |
| *** PollyZ_ has joined #openstack-ironic | 22:45 | |
| *** PollyZ has quit IRC | 22:48 | |
| *** sabeen has quit IRC | 22:51 | |
| *** rbudden has quit IRC | 22:52 | |
| klindgren | so how are people working around: https://bugs.launchpad.net/neutron/+bug/1527061 | 22:55 |
| openstack | Launchpad bug 1527061 in neutron "Nova should not throw exception when port binding fails for Ironic" [Undecided,Confirmed] | 22:55 |
| klindgren | installing whatever L2 agent on the compute node that fronts-ironic? | 22:56 |
| klindgren | because even without that bug in nova - their is seperate thing in liberty apparently - the dhcp_opts are not updated in dnsmasq on a port with vif_type: biniding_failed | 22:57 |
| *** piet has joined #openstack-ironic | 23:01 | |
| *** joprovost has quit IRC | 23:02 | |
| *** alaski_ is now known as alaski | 23:02 | |
| *** PollyZ_ has quit IRC | 23:03 | |
| *** ayoung has quit IRC | 23:07 | |
| *** garthb has quit IRC | 23:20 | |
| *** praneshp has quit IRC | 23:29 | |
| *** hoangcx has quit IRC | 23:37 | |
| *** chlong has quit IRC | 23:42 | |
| *** ayoung has joined #openstack-ironic | 23:44 | |
| *** piet has quit IRC | 23:45 | |
| *** piet has joined #openstack-ironic | 23:46 | |
| *** joprovost has joined #openstack-ironic | 23:50 | |
| openstackgerrit | Ramamani Yeleswarapu proposed openstack/ironic: Centralize config options - [cimc] https://review.openstack.org/301376 | 23:56 |
| openstackgerrit | Ramamani Yeleswarapu proposed openstack/ironic: Centralize config options - [conductor] https://review.openstack.org/301399 | 23:56 |
| openstackgerrit | Ramamani Yeleswarapu proposed openstack/ironic: Centralize config options - [cisco_ucs] https://review.openstack.org/301382 | 23:56 |
| openstackgerrit | Ramamani Yeleswarapu proposed openstack/ironic: Centralize config options - [console] https://review.openstack.org/295946 | 23:56 |
| *** sdake has quit IRC | 23:57 | |
| *** mbound has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
