| *** rama_y has quit IRC | 00:05 | |
| *** jaypipes has quit IRC | 00:14 | |
| *** jdwidari has joined #openstack-ironic | 00:21 | |
| *** hw_wutianwei has joined #openstack-ironic | 00:38 | |
| *** vmud213 has joined #openstack-ironic | 00:53 | |
| *** vmud213 has quit IRC | 00:55 | |
| *** vmud213 has joined #openstack-ironic | 00:56 | |
| *** tuanluong has joined #openstack-ironic | 00:56 | |
| *** bfournie has joined #openstack-ironic | 00:57 | |
| *** vmud213 has quit IRC | 00:57 | |
| *** vmud213 has joined #openstack-ironic | 00:57 | |
| *** vmud213 has quit IRC | 00:58 | |
| *** vmud213 has joined #openstack-ironic | 00:58 | |
| *** bfournie has quit IRC | 00:58 | |
| *** vmud213 has quit IRC | 00:58 | |
| *** bfournie has joined #openstack-ironic | 00:59 | |
| *** vmud213 has joined #openstack-ironic | 01:00 | |
| *** phuongnh has joined #openstack-ironic | 01:04 | |
| *** vmud213 has quit IRC | 01:08 | |
| *** vmud213 has joined #openstack-ironic | 01:08 | |
| openstackgerrit | Michael Turek proposed openstack/ironic master: Generate iPXE template on deploy for boot from volume https://review.openstack.org/478690 | 01:09 |
|---|---|---|
| *** jaganathan has quit IRC | 01:15 | |
| *** litao__ has joined #openstack-ironic | 01:20 | |
| TheJulia | rloo: no problem, thank you for raising that. I swear I looked that up before too... :(. Anyway, noted. My preference is to get the check before landk f, although if I have to toss up a quick patch to address it after the fact then so be it :\ | 01:25 |
| *** krtaylor has joined #openstack-ironic | 01:30 | |
| *** crushil_ has joined #openstack-ironic | 01:35 | |
| *** vmud213 has quit IRC | 01:37 | |
| *** vmud213 has joined #openstack-ironic | 01:39 | |
| *** vmud213 has quit IRC | 01:39 | |
| *** vmud213 has joined #openstack-ironic | 01:40 | |
| *** crushil_ has quit IRC | 01:56 | |
| *** vmud213 has quit IRC | 02:11 | |
| *** crushil_ has joined #openstack-ironic | 02:11 | |
| *** vmud213 has joined #openstack-ironic | 02:11 | |
| *** Nisha_Agarwal has joined #openstack-ironic | 02:13 | |
| *** crushil_ has quit IRC | 02:15 | |
| *** vmud213 has quit IRC | 02:19 | |
| *** vmud213 has joined #openstack-ironic | 02:20 | |
| *** gcb has joined #openstack-ironic | 02:24 | |
| *** rloo has quit IRC | 02:41 | |
| openstackgerrit | Dao Cong Tien proposed openstack/ironic master: WIP/DNM:Temp patch to test boot-from-volume job https://review.openstack.org/476854 | 02:43 |
| *** vmud213 has quit IRC | 02:56 | |
| openstackgerrit | Nisha Agarwal proposed openstack/proliantutils master: i[WIP] Adds nvdimm_n, logical_nvdimm_n, persistent_memory attributes https://review.openstack.org/478702 | 02:57 |
| *** shobhar has joined #openstack-ironic | 02:59 | |
| *** Nisha_Agarwal has quit IRC | 03:04 | |
| *** joe817 has quit IRC | 03:06 | |
| *** joe817 has joined #openstack-ironic | 03:06 | |
| openstackgerrit | Dao Cong Tien proposed openstack/ironic master: Tempest scenario test for boot-from-volume https://review.openstack.org/472740 | 03:16 |
| openstackgerrit | Dao Cong Tien proposed openstack/ironic master: WIP/DNM:Temp patch to test boot-from-volume job https://review.openstack.org/476854 | 03:16 |
| *** baoli has joined #openstack-ironic | 03:27 | |
| *** udesale has joined #openstack-ironic | 03:36 | |
| *** adisky__ has joined #openstack-ironic | 03:56 | |
| *** amotoki_away is now known as amotoki | 03:58 | |
| *** vmud213 has joined #openstack-ironic | 04:04 | |
| *** yuanying has joined #openstack-ironic | 04:18 | |
| *** yuanying_ has quit IRC | 04:19 | |
| *** namnh has joined #openstack-ironic | 04:23 | |
| openstackgerrit | Ngo Quoc Cuong proposed openstack/ironic master: Delete log translation functions and add hacking rules https://review.openstack.org/478421 | 04:27 |
| *** vmud213 has quit IRC | 04:46 | |
| *** links has joined #openstack-ironic | 04:53 | |
| *** jaianshu has joined #openstack-ironic | 05:00 | |
| *** links has quit IRC | 05:03 | |
| *** yuanying has quit IRC | 05:04 | |
| *** jaianshu_ has joined #openstack-ironic | 05:04 | |
| *** yuanying has joined #openstack-ironic | 05:06 | |
| *** marios has joined #openstack-ironic | 05:11 | |
| *** jaianshu_ has quit IRC | 05:12 | |
| *** pcaruana has joined #openstack-ironic | 05:14 | |
| *** links has joined #openstack-ironic | 05:15 | |
| *** mvpnitesh has joined #openstack-ironic | 05:19 | |
| *** pcaruana has quit IRC | 05:30 | |
| *** pcaruana has joined #openstack-ironic | 05:33 | |
| *** joe817 has quit IRC | 05:33 | |
| *** marios has quit IRC | 05:34 | |
| *** ChubYann has quit IRC | 05:34 | |
| *** vmud213 has joined #openstack-ironic | 05:39 | |
| *** pcaruana has quit IRC | 05:39 | |
| openstackgerrit | Debayan Ray proposed openstack/ironic master: Update new proliantutils version to 2.3.0 https://review.openstack.org/478741 | 05:42 |
| *** mjura has joined #openstack-ironic | 05:46 | |
| *** rcernin_ has joined #openstack-ironic | 05:48 | |
| openstackgerrit | Anshul Jain proposed openstack/proliantutils master: Redfish: Adds manager support for redfish API's. https://review.openstack.org/473719 | 05:49 |
| *** mvpnitesh has quit IRC | 05:55 | |
| *** karts has quit IRC | 05:56 | |
| *** rcernin_ is now known as rcernin | 06:04 | |
| openstackgerrit | Hironori Shiina proposed openstack/ironic master: [DNM] Test grenade job with fleetifying nova conductor patch https://review.openstack.org/477738 | 06:04 |
| openstackgerrit | Dao Cong Tien proposed openstack/ironic master: WIP/DNM:Temp patch to test boot-from-volume job https://review.openstack.org/476854 | 06:24 |
| *** rbartal has joined #openstack-ironic | 06:27 | |
| pas-ha | morning ironic :) | 06:32 |
| openstackgerrit | Hironori Shiina proposed openstack/ironic master: Set nomulticell flag for starting nova-compute in upgrade https://review.openstack.org/477313 | 06:34 |
| *** e0ne has joined #openstack-ironic | 06:36 | |
| openstackgerrit | Dao Cong Tien proposed openstack/ironic master: WIP/DNM:Temp patch to test boot-from-volume job https://review.openstack.org/476854 | 06:42 |
| openstackgerrit | Shu Muto proposed openstack/ironic-ui master: Switch from oslosphinx to openstackdocstheme https://review.openstack.org/478365 | 06:46 |
| *** hshiina has joined #openstack-ironic | 06:46 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic master: [DNM] Test n-g-s upgrade https://review.openstack.org/477655 | 06:47 |
| vdrok | good morning ironic and pas-ha | 06:48 |
| *** makowals has joined #openstack-ironic | 06:57 | |
| *** Goneri has joined #openstack-ironic | 06:59 | |
| *** Goneri has quit IRC | 07:04 | |
| openstackgerrit | Dao Cong Tien proposed openstack/ironic master: Adds clean step 'restore_irmc_bios_config' to iRMC drivers https://review.openstack.org/416403 | 07:21 |
| openstackgerrit | Anshul Jain proposed openstack/proliantutils master: Redfish: Adds 'update_firmware' APIs support through redfish https://review.openstack.org/472532 | 07:21 |
| *** marios has joined #openstack-ironic | 07:24 | |
| *** Nisha_Agarwal has joined #openstack-ironic | 07:28 | |
| openstackgerrit | Ngo Quoc Cuong proposed openstack/ironic master: Delete log translation functions and add hacking rules https://review.openstack.org/478421 | 07:29 |
| *** mvpnitesh has joined #openstack-ironic | 07:30 | |
| *** tesseract has joined #openstack-ironic | 07:31 | |
| *** jkilpatr has quit IRC | 07:33 | |
| *** thaj has joined #openstack-ironic | 07:35 | |
| thaj | Good morning all | 07:35 |
| *** pcaruana has joined #openstack-ironic | 07:35 | |
| *** hjensas has joined #openstack-ironic | 07:35 | |
| *** hjensas has quit IRC | 07:35 | |
| *** hjensas has joined #openstack-ironic | 07:35 | |
| *** ccamacho has quit IRC | 07:40 | |
| *** ccamacho has joined #openstack-ironic | 07:40 | |
| *** tesseract has quit IRC | 07:47 | |
| *** openstackgerrit has quit IRC | 07:47 | |
| *** tesseract has joined #openstack-ironic | 07:48 | |
| *** ralonsoh has joined #openstack-ironic | 07:50 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** zzzeek has joined #openstack-ironic | 08:00 | |
| *** junbo has quit IRC | 08:03 | |
| *** jaganathan has joined #openstack-ironic | 08:06 | |
| *** junbo has joined #openstack-ironic | 08:06 | |
| thaj | Hello Dmitry | 08:09 |
| *** dtantsur|afk is now known as dtantsur | 08:11 | |
| dtantsur | morning ironic | 08:11 |
| *** derekh has joined #openstack-ironic | 08:15 | |
| *** tesseract has quit IRC | 08:30 | |
| *** amarao has joined #openstack-ironic | 08:31 | |
| *** tesseract has joined #openstack-ironic | 08:32 | |
| *** openstackgerrit has joined #openstack-ironic | 08:34 | |
| openstackgerrit | Hironori Shiina proposed openstack/ironic master: [WIP] Add API for volume resources to api-ref https://review.openstack.org/478798 | 08:34 |
| *** racedo has joined #openstack-ironic | 08:38 | |
| openstackgerrit | Hironori Shiina proposed openstack/ironic master: [WIP] Add API for volume resources to api-ref https://review.openstack.org/478798 | 08:39 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic master: [DNM] Test n-g-s upgrade https://review.openstack.org/477655 | 08:43 |
| openstackgerrit | Dao Cong Tien proposed openstack/ironic master: WIP/DNM:Temp patch to test boot-from-volume job https://review.openstack.org/476854 | 08:49 |
| *** snecklifter has joined #openstack-ironic | 08:51 | |
| openstackgerrit | Anshul Jain proposed openstack/proliantutils master: Redfish: Adds 'update_firmware' APIs support through redfish https://review.openstack.org/472532 | 08:52 |
| *** athomas has joined #openstack-ironic | 08:55 | |
| *** mgoddard has joined #openstack-ironic | 09:03 | |
| *** vmud213 has quit IRC | 09:03 | |
| *** sambetts|afk is now known as sambetts | 09:17 | |
| sambetts | Morning all | 09:19 |
| dtantsur | hi sambetts | 09:22 |
| *** hshiina has quit IRC | 09:23 | |
| *** Nisha_ has joined #openstack-ironic | 09:25 | |
| *** paramite has quit IRC | 09:26 | |
| *** Nisha_Agarwal has quit IRC | 09:26 | |
| openstackgerrit | Shu Muto proposed openstack/ironic-ui master: Switch from oslosphinx to openstackdocstheme https://review.openstack.org/478365 | 09:30 |
| *** Nisha_ has quit IRC | 09:30 | |
| sambetts | o/ dtantsur | 09:31 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Use adapters for cinderclient https://review.openstack.org/476171 | 09:33 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Use adapters for neutronclient https://review.openstack.org/476170 | 09:33 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Refactor common keystone methods https://review.openstack.org/476974 | 09:33 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Rework keystone auth for glance https://review.openstack.org/467728 | 09:33 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Fix swiftclient creation https://review.openstack.org/476173 | 09:33 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Use adapters for inspectorclient https://review.openstack.org/476172 | 09:33 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Introduce keystoneauth adapters for clients https://review.openstack.org/469627 | 09:33 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Finalize migration to keystoneauth adapters https://review.openstack.org/478825 | 09:33 |
| *** tiendc has joined #openstack-ironic | 09:36 | |
| *** tesseract has quit IRC | 09:38 | |
| openstackgerrit | Anshul Jain proposed openstack/proliantutils master: Redfish: Adds 'update_firmware' APIs support through redfish https://review.openstack.org/472532 | 09:39 |
| *** tesseract has joined #openstack-ironic | 09:40 | |
| *** shu-mutou-AWAY is now known as shu-mutou | 09:41 | |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Allow to set image_url to glance API file https://review.openstack.org/474996 | 09:47 |
| dtantsur | sambetts: did you have a chance to check TheJulia's comments on https://review.openstack.org/#/c/454243/ ? | 09:48 |
| patchbot | patch 454243 - ironic - Logic for skipping deployment with BFV | 09:48 |
| sambetts | dtantsur: was just working through my review queue, so I'm nearly there :-P | 09:50 |
| dtantsur | awesome :) | 09:50 |
| *** b00tcat has quit IRC | 09:59 | |
| *** b00tcat has joined #openstack-ironic | 10:00 | |
| sambetts | dtantsur, TheJulia: I don't think this actually touches vlad's spec to refactor the orcastration into the conductor which so far I don't agree with, I have major concerns about having to include these conditional statements all over deploy interfaces to basically make them noop, where as if we don't even call into the deploy driver when we're not writing an image, then there isn't a need to | 10:02 |
| sambetts | touch the deploy interfaces at all... | 10:03 |
| dtantsur | ah, hmm, I see now | 10:03 |
| dtantsur | unfortunately, I do belive that it touches the vdrok | 10:03 |
| dtantsur | damn | 10:03 |
| dtantsur | the vdrok's spec | 10:03 |
| dtantsur | sambetts: either the deploy driver is an orchestrator, or the conductor. | 10:04 |
| vdrok | hey dtantsur sambetts morning | 10:04 |
| dtantsur | if the former, then we have to keep this logic there; if the latter, then we should go with vdrok's spec | 10:04 |
| dtantsur | morning vdrok :) | 10:04 |
| *** ralonsoh has quit IRC | 10:04 | |
| sambetts | dtantsur: the deploy interface should be the orchestrator if your writing an image, if your not writing an image then it shouldn't be | 10:04 |
| dtantsur | sambetts: well, the orchestator knows what you have to do | 10:05 |
| dtantsur | I don't feel easy about spreading this logic all over the code base, to be honest | 10:05 |
| dtantsur | like, conductor also knows if we should call networking interfaces or not, but we still keep it in deploy drivers | 10:05 |
| dtantsur | now we need to keep in mind, that deploy drivers are NOT always called | 10:05 |
| openstackgerrit | Dao Cong Tien proposed openstack/ironic master: Adds clean step 'restore_irmc_bios_config' to iRMC drivers https://review.openstack.org/416403 | 10:06 |
| openstackgerrit | Ankit Kumar proposed openstack/proliantutils master: Redfish: Adds support to get ilo firmware version https://review.openstack.org/478478 | 10:06 |
| vdrok | yeah, I suppose we need to decide this once and do the same thing for all interfaces | 10:06 |
| dtantsur | sambetts: e.g. this code https://review.openstack.org/#/c/454243/17/ironic/drivers/modules/iscsi_deploy.py@458 where should it go then? | 10:06 |
| patchbot | patch 454243 - ironic - Logic for skipping deployment with BFV | 10:06 |
| dtantsur | lines 458 - 468 | 10:06 |
| sambetts | in the conductor manager, conditionallised on if we need to write an image here: https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L2791-L2808, IMO the conductor is the core orchestrator, it just delegates the orchestration to the interfaces that know what they have to do, if we need to write an image then the deploy interface understands how to do it, if we're not | 10:08 |
| sambetts | writing an image the | 10:08 |
| sambetts | we don't need to call into the deploy driver | 10:08 |
| dtantsur | sambetts: so you actually suggest moving the orchestration to the conductor, right? :) which is what vdrok proposes IIUC | 10:09 |
| vdrok | dtantsur: right, but sambetts concerns is mostly about network interfaces AFAIU | 10:10 |
| vdrok | because some of them might be tightly coupled with deploy interfaces | 10:10 |
| dtantsur | well, the code in question touches network interfaces | 10:10 |
| sambetts | the conductor already orchestrates stuff, it calls task.driver.depoloy.prepare and task.driver.deploy.deploy | 10:10 |
| sambetts | what I'm saying is that it doesn't need to do that if we don't need to write an image | 10:10 |
| dtantsur | sambetts: yep, and now it will call "task.driver.network.remove_provisioning_network(task)", right? | 10:10 |
| dtantsur | isn't it what vdrok proposes? | 10:11 |
| *** shu-mutou is now known as shu-mutou-AWAY | 10:11 | |
| dtantsur | if we don't call deploy(), then the code on lines 462 - 466 must go *somewhere*. and this code includes calling into a network interface. | 10:12 |
| sambetts | the difference is that vdrok is proposing doing it in all cases, when we need to write an image that is a complex process and requires delegated control and might be different per deploy driver, in the case of BFV its just attach the volume, atttach the network and power on in every case | 10:12 |
| dtantsur | sambetts: so you propose duplicating this logic between conductor (in case of BFV) and deploy (in case of normal deploy)? | 10:13 |
| vdrok | yeah, not pretty :/ | 10:13 |
| dtantsur | maybe then we need a method on the base DeployInterface, something like finish_deploy with this code? | 10:14 |
| dtantsur | otherwise we'll surely forget to update one of the copies one day | 10:14 |
| *** ralonsoh has joined #openstack-ironic | 10:15 | |
| sambetts | dtantsur: so in the conductor it'd be something like: if storage.needs_to_write_image: call deploy driver like normal, else call deploy driver finish deploy? | 10:15 |
| *** mat128 has joined #openstack-ironic | 10:15 | |
| dtantsur | sambetts: yep, maybe | 10:17 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Serve boot configs from Ironic API https://review.openstack.org/369438 | 10:17 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Implement Driver API for dynamic boot config https://review.openstack.org/394398 | 10:17 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Add RPC API to get the boot config https://review.openstack.org/394399 | 10:17 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: DNM test enabled dynamic iPXE https://review.openstack.org/391943 | 10:17 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Allow not writing pxe configs to disk https://review.openstack.org/394397 | 10:17 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Convert iPXE boot config to proper Jinja template https://review.openstack.org/393462 | 10:17 |
| sambetts | dtantsur: IMO thats what boot.prepare_instnace is for | 10:18 |
| dtantsur | maybe? as long as we don't duplicate this code | 10:19 |
| dtantsur | sambetts: to be honest, I'd prefer to do such refactorings async to the whole BFV work.. | 10:22 |
| dtantsur | otherwise we'll never finish it | 10:22 |
| sambetts | dtantsur: there is a concern in my head that the current way of doing it will result in all out of tree deploy interfaces being broken if they don't update to be aware of adding these conditionals to make themselves effectivy noop | 10:23 |
| dtantsur | sambetts: the same concern applies to a potential boot interface change | 10:24 |
| dtantsur | and my suggestion with finish_deploy() will only work, if the out-of-tree deploy drivers don't require additional finish steps | 10:24 |
| * dtantsur has a feeling that writing out-of-tree deploy and boot interfaces is not an easy task | 10:26 | |
| sambetts | so we have functions that exist today that are actually kind of like the finish_deploy function you've proposed: https://github.com/openstack/ironic/blob/0424e1a075132b40909b2d41500581524935c706/ironic/drivers/modules/agent.py#L262 | 10:28 |
| dtantsur | I think it's actually reboot_and_finish_deploy | 10:29 |
| dtantsur | yeah, kinda https://github.com/openstack/ironic/blob/31d7eb0e05e496cb226ae8f8bc73100e8d6d4486/ironic/drivers/modules/agent_base_vendor.py#L586-L589 | 10:30 |
| dtantsur | but we cannot just call it from the conductor, as it's specific to IPA | 10:30 |
| sambetts | dtantsur: yeah because it does its soft power off | 10:31 |
| openstackgerrit | Debayan Ray proposed openstack/ironic master: Refresh firmware versions on firmware update https://review.openstack.org/346683 | 10:31 |
| sambetts | dtantsur: but if we made it a interface abstract function then the conductor could call it right? | 10:31 |
| dtantsur | sambetts: yes, though it won't fix deploy interfaces that don't have it, or that rely on additional actions post-deploy | 10:32 |
| dtantsur | which brings us back to my point: we break them anyway (with BFV) | 10:33 |
| sambetts | dtantsur: acutually that wouldn't work for BFV anyway because soft power off | 10:33 |
| dtantsur | yep | 10:33 |
| *** dsneddon has quit IRC | 10:34 | |
| *** dsneddon has joined #openstack-ironic | 10:35 | |
| *** dsneddon has quit IRC | 10:36 | |
| *** udesale has quit IRC | 10:44 | |
| sambetts | hmm... ok... I think this is going to need more thought and design... | 10:46 |
| *** jaianshu has quit IRC | 10:47 | |
| *** phuongnh has quit IRC | 10:47 | |
| *** dsneddon has joined #openstack-ironic | 10:50 | |
| dtantsur | sambetts: yeah. Do you think we should delay BFV until then? | 10:50 |
| sambetts | no, I think we shgould just get this landed if it works and refactor the whole conductor deploy interface interacted as part of a larger picture | 10:51 |
| dtantsur | ++ | 10:51 |
| sambetts | part of me wonders if the storage interfaces themseleves should have deploy function or not, but thats a conversation for another day | 10:52 |
| dtantsur | yeah, I'm also not sure if the storage interface is perfectly designed.. but I put quite a few thoughts in it back then, and was unable to come up with a perfect solution | 10:52 |
| dtantsur | our deploy interface is kinda weird.. | 10:53 |
| *** jkilpatr has joined #openstack-ironic | 10:56 | |
| *** jkilpatr has quit IRC | 10:56 | |
| *** jkilpatr has joined #openstack-ironic | 10:57 | |
| sambetts | yeah I agree... there is a mix of things specific to how to deploy the image to the server e.g. talking to IPA or setting up ISCSI, and also how to boot the server etc :/ | 10:58 |
| sambetts | maybe deploy needs to be split, into deploy orcestrator interface and image writer interface | 10:59 |
| *** vmud213 has joined #openstack-ironic | 11:00 | |
| dtantsur | this is what I'm thinking too.. need to talk on the PTG (or bring it to vdrok's spec) | 11:04 |
| dtantsur | sambetts: mind re-review the patch then please? | 11:04 |
| sambetts | dtantsur: just did | 11:04 |
| dtantsur | thanks! | 11:05 |
| openstackgerrit | bhavani proposed openstack/ironic-specs master: Remove pbr warnerrors in favor of sphinx check https://review.openstack.org/478879 | 11:25 |
| openstackgerrit | M V P Nitesh proposed openstack/python-ironicclient master: Using internal Resource.to_dict() https://review.openstack.org/478834 | 11:25 |
| *** hjensas is now known as hjensas|lunch | 11:27 | |
| dtantsur | vdrok: mind checking https://review.openstack.org/#/c/454243/ as well? maybe we can merge it today | 11:30 |
| patchbot | patch 454243 - ironic - Logic for skipping deployment with BFV | 11:31 |
| vdrok | dtantsur: will try to do today | 11:31 |
| *** vmud213 has quit IRC | 11:36 | |
| *** mvpnitesh has quit IRC | 11:44 | |
| *** mvpnitesh has joined #openstack-ironic | 11:44 | |
| *** fellypefca has joined #openstack-ironic | 11:44 | |
| TheJulia | I think the existing deploy interface solves our needs with-in our context as a singular workflow driving engine of a machine through that lifecycle. Changing that model is definitely a larger discussion because otherwise we will just create more headaches for ourselves. For BFV, the conundrum that we will also eventually have to deal with is someone who has BFV with dedicated storage fabrics, and then wants to | 11:47 |
| TheJulia | deploy via the agent because their san is loosely connected to cinder or whatever. I suspect I might write a POC storage interface driver at some point, if I have time to kind of do the same thing, but for local iscsi offering or something. If anyone has an FC san or FCoE san with a decent API they would like to donate to the cause.... ;) | 11:47 |
| dtantsur | we're trying to get our hands on hardware like that, but without too much luck so far | 11:50 |
| dtantsur | also morning TheJulia | 11:50 |
| TheJulia | Good morning dtantsur and everyone else! | 11:51 |
| fellypefca | Good morning TheJulia, dtantsur and ironic :) | 11:51 |
| dtantsur | hi fellypefca | 11:51 |
| TheJulia | I did mention during the project update that hardware access was one of our hurdles, and there was some interest in trying to help solve that problem from the back of the room. We should put together a wish list during the PTG of things that could help various efforts, and how they could constructively help. That could then become an avenue to get time in a lab with said hardware if we ask the right people. :) | 11:52 |
| *** mvpnitesh has quit IRC | 11:53 | |
| *** mvpnitesh has joined #openstack-ironic | 11:53 | |
| *** litao__ has quit IRC | 11:57 | |
| *** tuanluong has quit IRC | 11:59 | |
| fellypefca | dtantsur I have already updated the description of RFE, explaining in detail the problem and our proposed solution. | 12:00 |
| dtantsur | thanks. link handy? | 12:00 |
| *** namnh has quit IRC | 12:01 | |
| fellypefca | dtantsur https://bugs.launchpad.net/ironic/+bug/1666497 | 12:01 |
| openstack | Launchpad bug 1666497 in Ironic "[RFE] Onetime boot feature in OneView drivers" [Low,In progress] - Assigned to Fellype Cavalcante (fellypefca) | 12:01 |
| dtantsur | thnx, I'll get to it when I finish one email | 12:01 |
| *** bfournie has quit IRC | 12:01 | |
| fellypefca | dtantsur ok! :) | 12:02 |
| *** tiendc has quit IRC | 12:04 | |
| *** rbudden has joined #openstack-ironic | 12:05 | |
| *** vmud213 has joined #openstack-ironic | 12:07 | |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Finalize migration to keystoneauth adapters https://review.openstack.org/478825 | 12:09 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Use adapters for cinderclient https://review.openstack.org/476171 | 12:09 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Use adapters for neutronclient https://review.openstack.org/476170 | 12:09 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Refactor common keystone methods https://review.openstack.org/476974 | 12:09 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Rework keystone auth for glance https://review.openstack.org/467728 | 12:09 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Fix swiftclient creation https://review.openstack.org/476173 | 12:09 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Use adapters for inspectorclient https://review.openstack.org/476172 | 12:09 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic master: Introduce keystoneauth adapters for clients https://review.openstack.org/469627 | 12:09 |
| fxpester | hi all, finally able to boot node, but it boots to kernel panic | 12:16 |
| fxpester | 'no root= passed to kernel' | 12:16 |
| fxpester | in generated config I can see 'root={{ ROOT }}' - something is broken or this rcord is correct ? | 12:16 |
| fxpester | that config path is - /tftpboot/4ca1c9a5-dd51-41e7-b67a-7b953287fc25/config | 12:17 |
| thaj | Hi Dmitry | 12:18 |
| *** baoli has quit IRC | 12:19 | |
| thaj | while ironic deployment, i am getting the the following error in the baremetal server | 12:19 |
| thaj | "PXE-E51: No DHCP or proxyDHCP offers were received" | 12:19 |
| *** hjensas|lunch has quit IRC | 12:19 | |
| thaj | Anything need to be set in the neutron config files or in ironic config files. | 12:20 |
| fxpester | thaj: use tcpdump to check if ironic recieves dhcp requests and sends response | 12:20 |
| thaj | It is basically devestack ironic | 12:20 |
| dtantsur | fellypefca: I really wonder why we even supply persistent=True when booting an IPA.. | 12:20 |
| fxpester | tcpdump -an -i any port 67 or port 68 | 12:20 |
| thaj | Hi fxpester. Using wireshark, we tried | 12:21 |
| thaj | we get only "DHCP discover". After that, nothing happens | 12:21 |
| sambetts | dtantsur: that was one of my thoughts too, also every call to set_boot_device goes through this IPMI tool specific function https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/deploy_utils.py#L670 | 12:21 |
| fellypefca | dtantsur yeah, me too :/ | 12:22 |
| *** cdearborn has joined #openstack-ironic | 12:22 | |
| dtantsur | fellypefca: maybe we should change https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/pxe.py#L488 to pass persistent=False? with an optional flag to make it always persistent? | 12:22 |
| dtantsur | I think it would solve your problem in a much cleaner way | 12:22 |
| dtantsur | sambetts: ^^ | 12:22 |
| sambetts | dtantsur: +1 that was my suggestion, and I'd like to get rid of the deploy_utils function too | 12:23 |
| fxpester | thaj: ironic will answer dhcp only if correct setuped AND expecting a node (so it`s like 30 minutes after you done 'nova boot') | 12:23 |
| dtantsur | oh, catching an IPMI exception is awesome as well. but that's a different story | 12:23 |
| fxpester | guys can you clean for me driver fake_pxe will work with IPA or with other way with iSCSI or both ? | 12:24 |
| dtantsur | fxpester: it will, as long as you actually manage the power at the right point in time | 12:25 |
| dtantsur | I'm aware of people who did it | 12:25 |
| thaj | fxpester: after nova boot, i get ironic status as "Clean failed". I am able to power on/off the baremetal from controller node. But, problem occurs only, while 'nova boot' | 12:26 |
| dtantsur | and btw thanks for the detailed explanation fellypefca, I finally understand what you're trying to solve :) | 12:26 |
| * dtantsur can be slow at times | 12:26 | |
| dtantsur | fellypefca: e.g. ironic-inspector uses non-persistent boot: https://github.com/openstack/ironic-inspector/blob/master/ironic_inspector/introspect.py#L118-L119 | 12:27 |
| fellypefca | dtantsur But it wouldn't be a better idea that the solution we proposed be accepted for our driver and than we create a spec to modify that in Ironic? We commit ourselves to stay ahead of this. | 12:28 |
| dtantsur | fellypefca: it does not require a spec or even an RFE. it's a bug IMO. | 12:28 |
| fxpester | thaj: so status 'clean failed' - ironic will not answer dhcp, you need to disable 'cleaning' in config | 12:28 |
| dtantsur | fellypefca: we only need to double-check that we set the boot device each time we reboot during cleaning | 12:29 |
| fellypefca | dtantsur oh, ok | 12:30 |
| fellypefca | dtantsur But can this general solution be in another patch or does it have to be in ours? | 12:30 |
| dtantsur | fellypefca: I don't see a reason to work around a bug in one driver, if we can just fix it | 12:30 |
| dtantsur | fellypefca: yeah, I think we need a non-persistent set_boot_device in https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent_base_vendor.py#L132 as well | 12:30 |
| thaj | fxpenster: i have ironic.conf file created after running stack.sh . Here, we have a field called 'automated_clean = True'. Should i make this false. Then how can i restart to make into effect? | 12:31 |
| dtantsur | fellypefca: bonus point: a bug fix is potentially backporteable, while your RFE is not | 12:32 |
| *** bfournie has joined #openstack-ironic | 12:33 | |
| thaj | In which file i should make 'cleaning' disable?. local.conf or ironic.conf. Do i need to build devstack again? | 12:33 |
| dtantsur | fxpester: why do you think cleaning has to be disabled? | 12:34 |
| sambetts | thaj: if you want to change a setting without restarting the whole of devstack you can edit the /etc/ironic/ironic.conf, and the go into screen -x and restart ironic-api and ironic-conductor | 12:34 |
| dtantsur | this is a dangerous thing to recommend, are you sure you're aware of the consequences? | 12:34 |
| dtantsur | and are you sure that deployment will magically succeed, given that it uses the same DHCP mechanism? | 12:35 |
| *** chlong_ has joined #openstack-ironic | 12:35 | |
| fxpester | dtantsur: it is his current problem, ofc I would recomend just to put his machine in 'ready to deployment' state | 12:36 |
| dtantsur | well, I suspect the next step will be disabling deployment :) | 12:36 |
| * dtantsur gets back to whatever his was doing then | 12:36 | |
| fxpester | if I know ironic states so well) | 12:36 |
| *** vmud213 has quit IRC | 12:37 | |
| dtantsur | thaj: the DHCP problem is probably related to the node not being able to access neutron | 12:37 |
| dtantsur | it may be wrong wiring (you're using devstack + baremetal, right? it can be strange at times) or e.g. firewall | 12:37 |
| dtantsur | it can be also that ports are not enrolled correctly (wrong MACs) | 12:37 |
| thaj | Dmitry. How to fix the neutron issue. What should i do for neutron config? | 12:37 |
| dtantsur | thaj: it's hard to tell from a distance. try using tcpdump on the interface you expect your requests to come, make sure they actually arrive | 12:38 |
| thaj | Dmitry. we tried using wireshark | 12:38 |
| dtantsur | (also please use nicks when asking something, just "Dmitry" will not highlight a message for me, and I'll easily miss it) | 12:39 |
| thaj | ok , nicks | 12:39 |
| *** dmellado has quit IRC | 12:39 | |
| *** trown|outtypewww is now known as trown | 12:40 | |
| dtantsur | so, wireshark. do you see DHCP arriving where you expect it to arrive? | 12:40 |
| thaj | in hardware info file, i have included the BMC ip , MAC, userid, passworld of baremetal node | 12:40 |
| thaj | nicks, no progress of DHCP after DISCOVER | 12:40 |
| sambetts | thaj: how did you plumb the physical NIC into the neutron network bridge? | 12:41 |
| *** dmellado has joined #openstack-ironic | 12:42 | |
| thaj | Not clear with your question, sambetts. Can you please ask little bit in lower level? May be i am a stranger to Ironic! | 12:42 |
| fellypefca | dtantsur But if you look at the first discussions of the patch, our first solution was to change that in the general Ironic. And after discussions with the community staff, they asked that this behavior be internal to our driver. | 12:42 |
| dtantsur | fellypefca: have a link to the discussion handy? I wonder what the concerns were.. | 12:43 |
| sambetts | dtantsur: https://bugs.launchpad.net/ironic/+bug/1666497 | 12:44 |
| openstack | Launchpad bug 1666497 in Ironic "[RFE] Onetime boot feature in OneView drivers" [Low,In progress] - Assigned to Fellype Cavalcante (fellypefca) | 12:44 |
| fellypefca | dtantsur https://review.openstack.org/#/c/436469/ | 12:44 |
| patchbot | patch 436469 - ironic - Adds onetime boot in oneview drivers | 12:44 |
| dtantsur | fellypefca: at least, which patchset? :) | 12:44 |
| fellypefca | The first discussions with vdrok, julia and yuriy | 12:45 |
| fellypefca | patchset | 12:45 |
| fellypefca | 2 | 12:45 |
| fellypefca | sorry | 12:45 |
| *** sasha2 has joined #openstack-ironic | 12:46 | |
| thaj | Hi Sambetts | 12:46 |
| *** jcoufal has joined #openstack-ironic | 12:46 | |
| dtantsur | TheJulia, vdrok, what was the reasoning against using non-persistent boot for booting IPA? that's the only reason we even have non-persistent boot.. | 12:47 |
| thaj | can you pls re-tell your question in simple way " how did you plumb the physical NIC into the neutron network bridge?" | 12:47 |
| dtantsur | if we never going to use it, we can well remove it | 12:47 |
| *** dprince has joined #openstack-ironic | 12:47 | |
| *** baoli has joined #openstack-ironic | 12:48 | |
| *** baoli has quit IRC | 12:48 | |
| *** jkilpatr has quit IRC | 12:49 | |
| dtantsur | for me it looks this way: 1. we accept persistent=False to be used in one-shot boots like IPA; 2. we never use though; 3. we ask vendors to work around this by hacking their management interfaces. | 12:49 |
| *** baoli has joined #openstack-ironic | 12:49 | |
| *** jkilpatr has joined #openstack-ironic | 12:49 | |
| dtantsur | TheJulia, vdrok, yuriyz: I really wonder if this ^^ is the right thing to do.. | 12:49 |
| TheJulia | by 1, you mean expect instead of accept? | 12:50 |
| dtantsur | TheJulia: no, I mean accept. Introduce this flag in the API. | 12:50 |
| TheJulia | For one-shots, I don't know. For normal power operations, it has to do with buggy bmcs, like ones that if you set netboot, they only ever netboot once | 12:51 |
| dtantsur | fellypefca: btw, part of the concern was around changing the default value for the parameter in that function. this is not what I propose - I propose changing where it's used. | 12:51 |
| sambetts | IMO we should kill deploy_utils try_set_boot_device, PXEboot etc should call manager_utils.node_set_boot_device directly, and should pass persisent=False | 12:52 |
| dtantsur | sambetts++ | 12:52 |
| TheJulia | oh | 12:52 |
| TheJulia | Well, Agent boots are not always one time boots | 12:53 |
| TheJulia | the agent can reboot the node for cleaning | 12:53 |
| dtantsur | TheJulia: yes, and we can request a temporary boot again | 12:53 |
| dtantsur | TheJulia: and anyway, the OneView change in question does exactly this: forces a temporary boot device, even if persistent is requested | 12:54 |
| TheJulia | but only if the conductor orchestrates it. i.e. I can't have a firmware script for cleaning that just reboots the node after updating firmware. | 12:54 |
| dtantsur | TheJulia: this is precisely why we have "reboot_required" flag in clean steps, I think | 12:55 |
| TheJulia | for just deploy... | 12:55 |
| * TheJulia looks up the rackspace cleaning example because she seems to remember a shutdown -r in the script | 12:55 | |
| dtantsur | TheJulia: and as I mentioned above, the OneView change in question breaks it anyway :) | 12:55 |
| TheJulia | only if wired in for deployment which I think their change was | 12:56 |
| dtantsur | well, this reduces usefulness of the change though. | 12:56 |
| sambetts | TheJulia: they do it in cleaning too | 12:57 |
| dtantsur | also the RFE explicitly mentions cleaning | 12:57 |
| sambetts | TheJulia: https://review.openstack.org/#/c/436469/19/ironic/drivers/modules/oneview/deploy_utils.py L106 | 12:57 |
| patchbot | patch 436469 - ironic - Adds onetime boot in oneview drivers | 12:57 |
| fellypefca | We're worried if this get not merged in Pike release. Our solution was close to be accepted and with this change of approach, it might not come in time. :( | 12:58 |
| TheJulia | whew, the published examples don't have shutdown -r | 12:58 |
| dtantsur | fellypefca: the solution we're talking about is much closer to being accepted IMO | 12:58 |
| dtantsur | and it's potentially backportable, as I mention above | 12:58 |
| TheJulia | ugh, yeah, that does break cleaning, even if reboot_required I think | 12:58 |
| * TheJulia double checks | 12:59 | |
| dtantsur | fellypefca: also, as it seems now, your solution may break cleaning when reboot_required is used in clean steps | 12:59 |
| fxpester | can someone post working config ? I doubt about mine the path is - /tftpboot/4ca1c9a5-dd51-41e7-b67a-7b953287fc25/config data - http://paste.openstack.org/show/614051/ | 13:00 |
| fxpester | so I ask for your working config to compare them ^^ | 13:00 |
| fxpester | mine leads to kernel panic | 13:00 |
| dtantsur | fxpester: which release is that? have you changed anything? ({{ DISK_IvimT FIER }} looks suspicious) | 13:02 |
| fxpester | 2015.01 kilo | 13:02 |
| TheJulia | yeah, prepare_cleaning gets called once, to start, then after reboot it is heartbeats that drive cleaning | 13:02 |
| TheJulia | so if there is a reboot... rutro. | 13:02 |
| fellypefca | Oh :/ | 13:03 |
| TheJulia | then again | 13:04 |
| TheJulia | most of such things should be handled by the server profile template in terms of firmware and everything else | 13:05 |
| TheJulia | pure cleaning would be it I think, unless someone implemented their own hardware manager with the oneview driver | 13:05 |
| vdrok | sorry, was on lunch | 13:05 |
| fellypefca | dtantsur TheJulia So, the best solution is to pass persistent=False here(https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/pxe.py#L488)? with an optional flag to make it always persistent? | 13:06 |
| dtantsur | fellypefca: I think so + fix cleaning to also do the same | 13:06 |
| TheJulia | fellypefca: I _think_, however for cleaning you would need to pass persistent=True | 13:07 |
| dtantsur | TheJulia: or call set_boot_device on every reboot | 13:07 |
| TheJulia | yup | 13:07 |
| dtantsur | I mean, cleaning reboot, of course | 13:07 |
| TheJulia | yeah, I'm just worried about possible tribal knowledge issues with expectation of having the conductor manage the reboot, but I'm also not sure how we document that effectively as a "don't do X in your own hardware managers!" | 13:09 |
| TheJulia | stepping away for a little bit | 13:10 |
| vdrok | dtantsur: hrm, so out-of-tree drivers will need to set some flag if they need persistent boot? | 13:11 |
| vdrok | during deployment | 13:11 |
| sambetts | vdrok: what type of driver? | 13:12 |
| sambetts | management? deploy? | 13:12 |
| vdrok | if we're talking about https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/pxe.py#L488 I guess it is management interface that will be looking if this flag is present? | 13:12 |
| sambetts | deploy drivers would call set_boot_device themselves so could pass persistent=True if they need it | 13:13 |
| sambetts | if a management interface knows that it can't do one time boot, then it can just ignore the flag | 13:13 |
| vdrok | OK, but this means a change is needed to deploy interfaces? | 13:13 |
| vdrok | to explicitly say persistent=True | 13:14 |
| dtantsur | vdrok: I don't get why.. our PXE interface has a bug that it requests a persistent boot, when a temporary is needed. and our Agent deploy interface has a bug, that it does not request any boot device, when doing a reboot. | 13:14 |
| sambetts | IMO we can leave the deploy_utils function there with a deprcation warning for out of tree drivers that are using it | 13:14 |
| fellypefca | dtantsur: besides call set_boot_device on every reboot, what would be a other solution to fix the cleaning? I understand the problem of the cleaning (reboot required as a cleaning step) but a need a hand on the way where to place the change... :) | 13:14 |
| sambetts | and then the in-tree ones should be calling the set_boot_device directly with the flag set as expected | 13:15 |
| dtantsur | fellypefca: I think calling set_boot_device isthe right thing to do. the only alternative I see it to say with persistent boot.. | 13:15 |
| fellypefca | dtantsur: gotcha | 13:16 |
| vdrok | dtantsur: "Agent deploy interface has a bug, that it does not request any boot device, when doing a reboot" here you're saying about cleaning reboots? | 13:17 |
| fellypefca | So we leave the onetime boot patch and propose this general solution in another one? | 13:19 |
| *** hjensas|lunch has joined #openstack-ironic | 13:19 | |
| dtantsur | vdrok: yep | 13:21 |
| dtantsur | fellypefca: that's what I (and sambetts apparently) suggest | 13:21 |
| sambetts | makes sense to me | 13:22 |
| *** mvpnitesh has quit IRC | 13:25 | |
| vdrok | ugh, honestly I'm not sure without seeing the patch :( we have 3 functions for setting boot device, deploy_utils.try_set_boot_device (with persistent=True by default), conductor.utils.node_set_boot_device (with persistent=False by default) and management interfaces method set_boot_device (which also defaults to False). we just need to make sure the changes don't break anyone | 13:25 |
| *** mvpnitesh has joined #openstack-ironic | 13:25 | |
| *** jkilpatr has quit IRC | 13:26 | |
| vdrok | eg agent driver configure_local_boot calls deploy_utils.try_set_boot_device relying on default value of True, while reboot_to_instance requests persistent device explicitly | 13:27 |
| vdrok | and in first case deploy_utils.try_set_boot_device is called, in the second - conductor.utils.node_set_boot_device | 13:28 |
| sambetts | deploy_utils.try_set_boot_device needs to die | 13:28 |
| *** jkilpatr has joined #openstack-ironic | 13:28 | |
| vdrok | right, but it can be used by someone, that's the reason I was opposed to changing its defaults initially (IIRC :)) | 13:29 |
| vdrok | *someone out of tree | 13:29 |
| vdrok | if we deprecate it, then I guess it should be OK | 13:29 |
| fxpester | Folowed standart deploy images way with coreos, but at deploy stage it boot to this and get frozen - https://pasteboard.co/30lmvfLZc.png | 13:31 |
| fxpester | any ideas what is it ? | 13:31 |
| fxpester | oh no, it is fedora images | 13:31 |
| fxpester | and looks ok... hm | 13:31 |
| fxpester | there is no default login/passwords ? | 13:31 |
| *** baoli has quit IRC | 13:32 | |
| *** jkilpatr_ has joined #openstack-ironic | 13:34 | |
| *** jkilpatr_ has quit IRC | 13:36 | |
| *** jkilpatr_ has joined #openstack-ironic | 13:37 | |
| dtantsur | fxpester: nope. you need to use nova keypairs, or embed logins in the image | 13:37 |
| *** jkilpatr has quit IRC | 13:37 | |
| *** mjturek has joined #openstack-ironic | 13:43 | |
| *** mtanino__ has joined #openstack-ironic | 13:44 | |
| *** rama_y has joined #openstack-ironic | 13:44 | |
| *** livelace has joined #openstack-ironic | 13:46 | |
| *** mvpnitesh has quit IRC | 13:48 | |
| *** mvpnitesh has joined #openstack-ironic | 13:48 | |
| *** mvpnitesh has quit IRC | 13:51 | |
| *** mvpnitesh has joined #openstack-ironic | 13:51 | |
| openstackgerrit | Dao Cong Tien proposed openstack/ironic master: WIP/DNM:Temp patch to test boot-from-volume job https://review.openstack.org/476854 | 13:53 |
| openstackgerrit | Merged openstack/ironic master: Don't retry power status if power action fails https://review.openstack.org/449454 | 13:54 |
| *** rama_y has quit IRC | 13:55 | |
| *** makowals has quit IRC | 13:58 | |
| *** joe817 has joined #openstack-ironic | 14:01 | |
| *** makowals has joined #openstack-ironic | 14:02 | |
| *** skazi__ has joined #openstack-ironic | 14:04 | |
| *** racedo_ has joined #openstack-ironic | 14:09 | |
| jlvillal | TheJulia, I did a -1 on https://review.openstack.org/466333 as I think we want to use the latest API version to make sure we don't break it later. WDYT? | 14:11 |
| patchbot | patch 466333 - ironic - DevStack: Add configuration for boot-from-volume | 14:11 |
| *** racedo has quit IRC | 14:11 | |
| TheJulia | jlvillal: latest as in explicitly request latest? | 14:12 |
| jlvillal | TheJulia, Yes. Like all the other uses in devstack/lib/ironic | 14:12 |
| jlvillal | TheJulia, But I could be wrong. | 14:12 |
| TheJulia | I thought we did explicit versions, but it could have been changed while I was not looking :) | 14:13 |
| jlvillal | TheJulia, I only see latest used in current devstack/lib/ironic | 14:14 |
| TheJulia | works for me then :) | 14:14 |
| *** shobhar has quit IRC | 14:14 | |
| jlvillal | TheJulia, Which makes sense to me to make sure we don't break something in a future API change. | 14:14 |
| TheJulia | it does | 14:14 |
| TheJulia | jlvillal: thanks! | 14:15 |
| jlvillal | :) | 14:15 |
| *** rodrigofm has joined #openstack-ironic | 14:21 | |
| *** jaypipes has joined #openstack-ironic | 14:24 | |
| *** baoli has joined #openstack-ironic | 14:25 | |
| TheJulia | fxpester: vmxnet3 ? Driving a machine testing on vmware? The question may also be what version of the CoreOS image you have. The latest Ironic Python Agent code is not going to be able to talk to a kilo ironic installation. | 14:25 |
| *** jkilpatr_ has quit IRC | 14:26 | |
| fxpester | so, old way without IPA is not supported in current images ? | 14:27 |
| *** jkilpatr has joined #openstack-ironic | 14:27 | |
| *** baoli has quit IRC | 14:28 | |
| *** mvpnitesh has quit IRC | 14:30 | |
| *** e0ne has quit IRC | 14:30 | |
| *** zhurong has joined #openstack-ironic | 14:31 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic master: [DNM] Test n-g-s upgrade https://review.openstack.org/477655 | 14:32 |
| *** ralonsoh has quit IRC | 14:32 | |
| *** rbartal has quit IRC | 14:36 | |
| *** makowals has quit IRC | 14:37 | |
| *** baoli has joined #openstack-ironic | 14:39 | |
| *** zhurong has quit IRC | 14:40 | |
| *** racedo__ has joined #openstack-ironic | 14:47 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Make redfish power interface wait for the power state change https://review.openstack.org/478958 | 14:48 |
| *** racedo_ has quit IRC | 14:49 | |
| TheJulia | fxpester: We changed the heartbeating method in one of the more recent releases and only maintained that support for current-1 support. You might be able to build an IPA image using stable/mitaka, or liberty. Best bet may be to resurect the bash ramdisk from Diskimage-Builder, but then you will be limited to the iscsi based deployment drivers. | 14:51 |
| TheJulia | fxpester: tl;dr, a lot has changed since kilo. | 14:52 |
| *** tuanla_fujitsu has joined #openstack-ironic | 14:53 | |
| *** tuanla__ has joined #openstack-ironic | 14:54 | |
| openstackgerrit | Xavier proposed openstack/ironic master: [WIP] Migrate oneview driver to use python-hpOneView https://review.openstack.org/466872 | 14:56 |
| ricardoas | dtantsur TheJulia fellypefca sambetts: sorry for losing the persistent boot party :( i´ll take a look at your comments, submit the bug and see if there is something oneview related i can add, ok? | 14:56 |
| dtantsur | sure | 14:57 |
| dtantsur | it's never late for a persistent boot party :) | 14:57 |
| TheJulia | it is also never late for a review party, or beer :) | 14:57 |
| ricardoas | \o/ | 14:57 |
| *** makowals has joined #openstack-ironic | 14:59 | |
| *** joe817 has quit IRC | 15:01 | |
| *** joe817 has joined #openstack-ironic | 15:01 | |
| *** vmud213 has joined #openstack-ironic | 15:02 | |
| *** keerthivasan has joined #openstack-ironic | 15:06 | |
| keerthivasan | facing few issues on node scheduling on ironic | 15:06 |
| keerthivasan | in ironic spinning 20 nodes using ironic...but scheduling happens for only 10 at a time ? | 15:08 |
| keerthivasan | any clue ? | 15:08 |
| *** cdearborn has quit IRC | 15:10 | |
| *** rpioso has joined #openstack-ironic | 15:10 | |
| keerthivasan | hi rpioso any idea ? | 15:11 |
| dtantsur | keerthivasan: I think it's configurable in nova | 15:15 |
| *** jkilpatr has quit IRC | 15:16 | |
| TheJulia | keerthivasan: As dtantsur said, that is inside nova's configuration. I think concurrency or something like that. | 15:16 |
| keerthivasan | checking the same configuration... | 15:17 |
| keerthivasan | thanks dtantsur | 15:18 |
| *** rama_y has joined #openstack-ironic | 15:19 | |
| keerthivasan | I am right max_concurrent_builds ? | 15:20 |
| NobodyCam | Good Morning Ironic'ers | 15:20 |
| *** jistr is now known as jistr|afk | 15:20 | |
| rama_y | Good morning NobodyCam | 15:21 |
| *** cdearborn has joined #openstack-ironic | 15:22 | |
| NobodyCam | Morning rama_y :) | 15:22 |
| TheJulia | keerthivasan: looks like it :) I seem to remember there is another thing that might need to be tuned from there | 15:23 |
| *** skazi__ has quit IRC | 15:23 | |
| TheJulia | but I just don't remember anymore | 15:23 |
| TheJulia | Good morning rama_y, NobodyCam | 15:23 |
| rama_y | Good morning TheJulia | 15:23 |
| NobodyCam | Morning TheJulia :) did you get power back? | 15:23 |
| TheJulia | Oh yes, of course :) | 15:24 |
| NobodyCam | :) | 15:24 |
| *** cdearborn_ has joined #openstack-ironic | 15:24 | |
| TheJulia | my high amperage extension cable is on the delivery truck now \o/ | 15:25 |
| *** shobhar has joined #openstack-ironic | 15:25 | |
| *** mjura has quit IRC | 15:25 | |
| NobodyCam | w00t | 15:25 |
| openstackgerrit | Merged openstack/ironic master: Logic for skipping deployment with BFV https://review.openstack.org/454243 | 15:26 |
| NobodyCam | :) | 15:26 |
| NobodyCam | nice | 15:26 |
| * TheJulia feels a slight amount of stress disappear | 15:26 | |
| *** cdearborn__ has joined #openstack-ironic | 15:26 | |
| keerthivasan | TheJulia: that configuration change is not working....looking for other config | 15:27 |
| *** rcernin has quit IRC | 15:28 | |
| TheJulia | keerthivasan: that might be the other thing I'm remembering. :\ | 15:28 |
| openstackgerrit | Merged openstack/ironic master: Add CRUD notification objects for volume connector and volume target https://review.openstack.org/463930 | 15:28 |
| rama_y | TheJulia: I am testing BFV on BM node; getting fault 'no valid hostname found' from nova when I create an instance. | 15:28 |
| rama_y | TheJulia: here is my etherpad: https://etherpad.openstack.org/p/Testing_'boot_from_volume'_patches around line #694 | 15:29 |
| TheJulia | rama_y: Anything in the nova logs? | 15:29 |
| dtantsur | TheJulia: two BFV patches at once, w00t :) | 15:29 |
| rama_y | Same info | 15:29 |
| *** jkilpatr has joined #openstack-ironic | 15:29 | |
| TheJulia | rama_y: Okay, can you by chance provide the nova logs, about a thousand lines before and a hundred lines after the error that your seeing? I'm assuming your seeing this in the nova-compute log output? | 15:30 |
| TheJulia | dtantsur: That just leaves the API, and it is being worked on from the note update I saw about an hour ago | 15:30 |
| TheJulia | well, just leaves the API before we are out of a single stack of patches | 15:30 |
| TheJulia | mjturek: btw, I added the fix you uploaded to the list | 15:31 |
| rama_y | TheJulia: ok; will do so shortly; thanks. | 15:31 |
| mjturek | thx TheJulia, I'm just about to push a rev on that | 15:31 |
| TheJulia | rama_y: if your not seeing that error in nova-compute's log, then it is something likely scheduling related, or the compute service just was not running maybe | 15:32 |
| TheJulia | mjturek: \o/ | 15:32 |
| TheJulia | rama_y: out of curiosity, do you mean the error is "No valid host was found. There are not enough hosts available." ? | 15:36 |
| rama_y | TheJulia: yes | 15:37 |
| rama_y | TheJulia: That is the error I am seeing; I added a BM node as the 4th node in addition to the 3 VM nodes. | 15:38 |
| TheJulia | So that is totally a scheduling issue then | 15:39 |
| TheJulia | node-4? | 15:39 |
| TheJulia | err, node-3? | 15:39 |
| rama_y | TheJulia: I put 2 nodes in maintenance; another VM node is active; so assumed nova would pick node-3 | 15:39 |
| rama_y | node-3 is my BM node | 15:39 |
| TheJulia | rama_y: possibly | 15:39 |
| *** cdearborn has quit IRC | 15:40 | |
| TheJulia | can we get the ironic node-show output for node-3 into that etherpad? | 15:40 |
| TheJulia | This might be easy to spot | 15:40 |
| rama_y | I did; at the very bottom | 15:40 |
| * TheJulia reloads the etherpad and looks | 15:40 | |
| TheJulia | I just see the openstack server show output | 15:41 |
| TheJulia | which is nova | 15:41 |
| rama_y | Oh sorry; let me do so | 15:41 |
| TheJulia | openstack baremetal node show (i think) | 15:41 |
| TheJulia | no worries | 15:41 |
| rama_y | Yes; now it is there at the bottom | 15:42 |
| *** cdearborn_ has quit IRC | 15:42 | |
| TheJulia | your missing the properties for nova to use to schedule | 15:43 |
| TheJulia | An example would be like line 441 | 15:43 |
| TheJulia | You may want to consult the devstack plugin for ironic on examples of adding the properties, in terms of the way the data is structured | 15:46 |
| TheJulia | you'll also need to wait 2-3 minutes at a minimum before trying to re-deploy. I think nova updates its scheduling data every minute by default | 15:47 |
| *** Nisha_Agarwal has joined #openstack-ironic | 15:48 | |
| *** hshiina has joined #openstack-ironic | 15:56 | |
| keerthivasan | TheJulia are you able to recollect the parameter ? | 15:59 |
| * sambetts ----> home | 16:00 | |
| sambetts | night all, cya'll tomorrow o/ | 16:00 |
| *** sambetts is now known as sambetts|afk | 16:00 | |
| TheJulia | goodnight sambetts|afk | 16:00 |
| rama_y | TheJulia: great, thanks for the pointer. | 16:00 |
| sambetts|afk | o/ TheJulia | 16:00 |
| *** amarao has quit IRC | 16:04 | |
| *** mat128_ has joined #openstack-ironic | 16:04 | |
| *** mat128 has quit IRC | 16:06 | |
| *** jistr|afk is now known as jistr | 16:07 | |
| *** makowals has quit IRC | 16:09 | |
| *** jaganathan has quit IRC | 16:19 | |
| *** e0ne has joined #openstack-ironic | 16:22 | |
| *** keerthivasan has quit IRC | 16:25 | |
| *** snecklifter has quit IRC | 16:28 | |
| *** Nisha_Agarwal has quit IRC | 16:30 | |
| *** Nisha_Agarwal has joined #openstack-ironic | 16:30 | |
| *** hjensas|lunch is now known as hjensas | 16:34 | |
| *** marios has quit IRC | 16:37 | |
| *** shobhar has quit IRC | 16:39 | |
| *** trown is now known as trown|lunch | 16:41 | |
| *** pcaruana has quit IRC | 16:48 | |
| *** jdwidari has quit IRC | 16:49 | |
| dtantsur | g'night | 16:50 |
| *** dtantsur is now known as dtantsur|afk | 16:50 | |
| *** derekh has quit IRC | 16:50 | |
| TheJulia | goodnight dtantsur|afk | 16:50 |
| mjturek | adios dtantsur|afk | 16:50 |
| dtantsur|afk | o/ | 16:51 |
| TheJulia | dtantsur|afk: minimum basic creates the instance first and then attaches the volume, not applicable in our case due to physical limitations of the universe at this time. | 16:52 |
| *** tesseract has quit IRC | 16:58 | |
| *** baoli has quit IRC | 16:59 | |
| *** baoli has joined #openstack-ironic | 17:00 | |
| *** mgoddard has quit IRC | 17:02 | |
| *** shobhar has joined #openstack-ironic | 17:09 | |
| *** hshiina has quit IRC | 17:11 | |
| jlvillal | TheJulia: What's next for BFV? I see the Whiteboard has the two items crossed out as they merged. | 17:16 |
| TheJulia | I'll post the next two patches | 17:17 |
| jlvillal | TheJulia: Thanks | 17:17 |
| *** racedo__ has quit IRC | 17:18 | |
| *** vmud213 has quit IRC | 17:19 | |
| TheJulia | jlvillal: updated | 17:19 |
| jlvillal | TheJulia: Thanks | 17:20 |
| TheJulia | https://etherpad.openstack.org/p/Ironic-BFV is also the etherpad tracking all the BFV stuff right now | 17:20 |
| *** tuanla_fujitsu has quit IRC | 17:24 | |
| *** tuanla__ has quit IRC | 17:24 | |
| *** skazi has quit IRC | 17:25 | |
| *** fellypefca has quit IRC | 17:35 | |
| *** baoli has quit IRC | 17:37 | |
| *** baoli has joined #openstack-ironic | 17:38 | |
| *** mgoddard has joined #openstack-ironic | 17:41 | |
| *** trown|lunch is now known as trown | 17:49 | |
| *** jaganathan has joined #openstack-ironic | 17:55 | |
| *** mgoddard has quit IRC | 18:02 | |
| *** rodrigofm has quit IRC | 18:04 | |
| *** cdearborn has joined #openstack-ironic | 18:05 | |
| *** rodrigofm has joined #openstack-ironic | 18:05 | |
| *** cdearborn__ has quit IRC | 18:18 | |
| *** dsneddon has quit IRC | 18:20 | |
| *** dsneddon has joined #openstack-ironic | 18:20 | |
| *** mtanino__ has quit IRC | 18:25 | |
| *** e0ne has quit IRC | 18:31 | |
| *** shobhar has quit IRC | 18:32 | |
| *** dprince has quit IRC | 18:32 | |
| *** skazi has joined #openstack-ironic | 18:36 | |
| *** Nisha_Agarwal has quit IRC | 18:37 | |
| *** ChubYann has joined #openstack-ironic | 18:41 | |
| *** dprince has joined #openstack-ironic | 18:48 | |
| *** mjturek has quit IRC | 18:50 | |
| *** mjturek has joined #openstack-ironic | 18:51 | |
| *** racedo has joined #openstack-ironic | 18:53 | |
| *** adisky__ has quit IRC | 18:54 | |
| *** mjturek has quit IRC | 18:55 | |
| *** livelace has quit IRC | 18:59 | |
| *** rodrigofm has quit IRC | 19:00 | |
| *** rodrigofm has joined #openstack-ironic | 19:01 | |
| *** racedo has quit IRC | 19:04 | |
| *** rodrigofm has quit IRC | 19:06 | |
| *** rodrigofm has joined #openstack-ironic | 19:06 | |
| *** racedo has joined #openstack-ironic | 19:07 | |
| *** racedo has quit IRC | 19:10 | |
| *** mat128_ has quit IRC | 19:11 | |
| *** mjturek has joined #openstack-ironic | 19:11 | |
| *** e0ne has joined #openstack-ironic | 19:25 | |
| *** e0ne has quit IRC | 19:35 | |
| openstackgerrit | Hugo Nicodemos proposed openstack/ironic master: Introduce hpOneView and ilorest to OneView https://review.openstack.org/468129 | 19:54 |
| *** e0ne has joined #openstack-ironic | 19:59 | |
| openstackgerrit | Hugo Nicodemos proposed openstack/ironic master: Migrate python-oneviewclient validations to Ironic OneView drivers https://review.openstack.org/468428 | 20:05 |
| *** racedo has joined #openstack-ironic | 20:14 | |
| *** jcoufal has quit IRC | 20:34 | |
| *** cdearborn has quit IRC | 20:38 | |
| *** racedo has quit IRC | 20:40 | |
| *** trown is now known as trown|pto | 20:43 | |
| *** dprince has quit IRC | 20:44 | |
| *** hamzy has quit IRC | 20:44 | |
| *** baoli has quit IRC | 21:02 | |
| *** e0ne has quit IRC | 21:16 | |
| *** racedo has joined #openstack-ironic | 21:21 | |
| *** paramite has joined #openstack-ironic | 21:22 | |
| *** pooja has joined #openstack-ironic | 21:22 | |
| *** mjturek has quit IRC | 21:27 | |
| *** racedo has quit IRC | 21:36 | |
| *** e0ne has joined #openstack-ironic | 21:38 | |
| *** athomas has quit IRC | 21:38 | |
| *** jaganathan has quit IRC | 21:39 | |
| *** e0ne has quit IRC | 21:43 | |
| *** paramite has quit IRC | 21:43 | |
| *** bfournie has quit IRC | 22:04 | |
| *** MattMan_1 has quit IRC | 22:18 | |
| *** MattMan_1 has joined #openstack-ironic | 22:18 | |
| openstackgerrit | John L. Villalovos proposed openstack/ironic master: Deprecate DHCP providers https://review.openstack.org/456127 | 22:21 |
| *** cdearborn has joined #openstack-ironic | 22:23 | |
| NobodyCam | lol.. with that I guess I can abandon this review I started like a year ago. https://review.openstack.org/#/c/343947/ | 22:31 |
| patchbot | patch 343947 - bifrost - Initial work to add ISC-DHCP | 22:31 |
| *** cdearborn has quit IRC | 22:40 | |
| *** bfournie has joined #openstack-ironic | 23:05 | |
| *** johnthetubaguy has quit IRC | 23:18 | |
| *** mjturek has joined #openstack-ironic | 23:19 | |
| *** jaypipes has quit IRC | 23:20 | |
| *** mtanino_ has joined #openstack-ironic | 23:25 | |
| *** johnthetubaguy has joined #openstack-ironic | 23:28 | |
| *** baoli has joined #openstack-ironic | 23:48 | |
| *** baoli has quit IRC | 23:53 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
