Monday, 2017-01-23

*** sacharya has joined #openstack-ironic00:03
*** sacharya has quit IRC00:27
*** sacharya has joined #openstack-ironic00:27
*** zackf has joined #openstack-ironic00:39
*** mtanino has quit IRC00:43
*** zackf has quit IRC00:44
*** hoangcx has joined #openstack-ironic00:54
*** causten has joined #openstack-ironic00:56
*** sacharya has quit IRC00:57
*** alexpilotti has joined #openstack-ironic01:00
*** causten has quit IRC01:01
*** hw_wutianwei1 has joined #openstack-ironic01:02
*** alexpilotti has quit IRC01:04
*** hw_wutianwei has quit IRC01:05
*** hw_wutianwei1 has quit IRC01:07
*** phuongnh has joined #openstack-ironic01:07
*** ijw has joined #openstack-ironic01:09
*** hw_wutianwei has joined #openstack-ironic01:13
*** ppiela has quit IRC01:18
*** causten has joined #openstack-ironic01:21
openstackgerritCuong Nguyen proposed openstack/ironic: Use context manager for better file handling
*** ijw_ has joined #openstack-ironic01:56
*** ijw has quit IRC01:59
*** baoli has quit IRC02:05
*** tiendc has joined #openstack-ironic02:16
*** hecliunyx has quit IRC02:21
*** hecliunyx has joined #openstack-ironic02:27
*** soliosg has quit IRC02:32
*** soliosg_ is now known as soliosg02:32
*** sacharya has joined #openstack-ironic02:43
*** causten has quit IRC02:49
*** soliosg has left #openstack-ironic02:50
*** jerrygb has joined #openstack-ironic02:55
*** amotoki has quit IRC02:59
openstackgerritzhouya proposed openstack/ironic-python-agent: Get numa_node info when collecting pci devices info
*** rbudden has quit IRC03:02
*** zackf has joined #openstack-ironic03:04
*** amotoki has joined #openstack-ironic03:10
*** zackf has quit IRC03:11
*** amotoki has quit IRC03:15
*** amotoki has joined #openstack-ironic03:20
*** SerenaFeng has joined #openstack-ironic03:24
*** SerenaFeng has quit IRC03:24
*** SerenaFeng has joined #openstack-ironic03:24
*** SerenaFeng has quit IRC03:24
*** SerenaFeng has joined #openstack-ironic03:24
*** baoli has joined #openstack-ironic03:26
*** SerenaFeng has quit IRC03:27
*** SerenaFeng has joined #openstack-ironic03:29
*** baoli has quit IRC03:30
*** amotoki_ has joined #openstack-ironic03:33
*** amotoki has quit IRC03:35
*** early has joined #openstack-ironic03:35
*** early` has quit IRC03:36
*** SerenaFeng has quit IRC03:39
*** SerenaFeng has joined #openstack-ironic03:39
*** amotoki_ has quit IRC03:42
*** hecliunyx has quit IRC03:42
*** csmart has quit IRC03:43
*** SerenaFeng has quit IRC03:48
*** baoli has joined #openstack-ironic03:53
*** soliosg has joined #openstack-ironic03:53
openstackgerritDao Cong Tien proposed openstack/python-ironicclient: Extends driver-list, driver-show supporting new hardware types
openstackgerritBob Fournier proposed openstack/ironic-inspector: Add plugin to process basic management LLDP TLVs
*** baoli has quit IRC04:00
*** jerrygb has quit IRC04:03
*** Sukhdev has joined #openstack-ironic04:03
openstackgerritBob Fournier proposed openstack/ironic-inspector: Add plugin to process basic management LLDP TLVs
*** aparnav has joined #openstack-ironic04:10
*** amotoki has joined #openstack-ironic04:16
*** csmart has joined #openstack-ironic04:17
*** lindycoder has joined #openstack-ironic04:21
openstackgerritAparna proposed openstack/ironic: Doc for sanitize erase support in iLO drivers
openstackgerritDao Cong Tien proposed openstack/python-ironicclient: Add args to CLI 'node-create' for selecting hardware interfaces
openstackgerritNisha Agarwal proposed openstack/proliantutils: snmp inspection for disk size
*** hecliunyx has joined #openstack-ironic04:44
*** causten has joined #openstack-ironic04:50
*** causten has quit IRC04:55
*** lindycoder has quit IRC05:00
*** hecliunyx has quit IRC05:00
*** udesale has joined #openstack-ironic05:02
*** sayantani01 has quit IRC05:03
*** hecliunyx has joined #openstack-ironic05:08
*** sayantani01 has joined #openstack-ironic05:13
*** hecliunyx has quit IRC05:14
openstackgerritNisha Agarwal proposed openstack/ironic: SNMP agent support for OOB inspection for iLO Drivers
*** causten has joined #openstack-ironic05:27
*** hecliunyx has joined #openstack-ironic05:32
*** ijw_ has quit IRC05:32
*** nmathew has joined #openstack-ironic05:34
*** dyasny has quit IRC05:49
*** jerrygb has joined #openstack-ironic06:04
*** jerrygb has quit IRC06:08
*** sacharya has quit IRC06:10
openstackgerritAnh Tran proposed openstack/bifrost: Typo fix: enviroment => environment
*** causten has quit IRC06:24
*** moshele has joined #openstack-ironic06:26
openstackgerritNaohiro Tamura proposed openstack/ironic: Add Inject NMI to api-ref
*** dyasny has joined #openstack-ironic06:46
openstackgerritNisha Agarwal proposed openstack/proliantutils: snmp inspection for disk size
*** sacharya has joined #openstack-ironic07:11
*** Sukhdev has quit IRC07:15
*** sacharya has quit IRC07:15
*** ChubYann has quit IRC07:18
openstackgerritGrzegorz Grasza (xek) proposed openstack/ironic: Add RPC and object version pinning
openstackgerritGrzegorz Grasza (xek) proposed openstack/ironic: Add version column
openstackgerritGrzegorz Grasza (xek) proposed openstack/ironic: Add version column
*** amarao has quit IRC07:30
*** rbartal has joined #openstack-ironic07:30
*** tesseract has joined #openstack-ironic07:31
openstackgerritGrzegorz Grasza (xek) proposed openstack/ironic: Add RPC and object version pinning
openstackgerritGrzegorz Grasza (xek) proposed openstack/ironic: Add version column
xekmorning ironic :)07:34
pas-hamorning xek :) and all others :)07:38
openstackgerritDao Cong Tien proposed openstack/python-ironicclient: Extends driver-list, driver-show supporting new hardware types
*** sayantani01 has quit IRC07:44
*** yonglihe has joined #openstack-ironic07:45
*** e0ne has joined #openstack-ironic07:46
openstackgerritVasyl Saienko proposed openstack/ironic: WIP: ironic standalone tests
*** e0ne has quit IRC08:09
*** amoralej|off is now known as amoralej08:12
*** jtomasek has quit IRC08:15
*** jtomasek has joined #openstack-ironic08:17
*** pcaruana has joined #openstack-ironic08:17
*** ccamacho has joined #openstack-ironic08:22
*** causten has joined #openstack-ironic08:24
*** causten has quit IRC08:29
openstackgerritYuiko Takada Mori proposed openstack/ironic: Add previous config name to Multitenancy doc
*** amarao has joined #openstack-ironic08:33
openstackgerritNaohiro Tamura proposed openstack/ironic: Add soft reboot, soft power off and power timeout to api-ref
*** mgoddard has joined #openstack-ironic08:36
*** ohamada has joined #openstack-ironic08:38
openstackgerritPavlo Shchelokovskyy proposed openstack/ironic: Serve boot configs from Ironic API
openstackgerritPavlo Shchelokovskyy proposed openstack/ironic: Implement Driver API for dynamic boot config
openstackgerritPavlo Shchelokovskyy proposed openstack/ironic: Add RPC API to get the boot config
openstackgerritPavlo Shchelokovskyy proposed openstack/ironic: Allow not writing pxe configs to disk
openstackgerritPavlo Shchelokovskyy proposed openstack/ironic: Convert iPXE boot config to proper Jinja template
openstackgerritPavlo Shchelokovskyy proposed openstack/ironic: DNM test enabled dynamic iPXE
*** ralonsoh has joined #openstack-ironic08:54
*** zzzeek has quit IRC09:00
*** zzzeek has joined #openstack-ironic09:01
*** yolanda has quit IRC09:02
*** mjura has joined #openstack-ironic09:05
*** yolanda has joined #openstack-ironic09:07
*** e0ne has joined #openstack-ironic09:08
*** nmathew has quit IRC09:08
*** mgoddard has quit IRC09:14
*** yolanda has quit IRC09:15
*** yolanda has joined #openstack-ironic09:24
*** Nisha_Agarwal has joined #openstack-ironic09:26
openstackgerritAnh Tran proposed openstack/python-ironicclient: Typo fix: prefered => preferred
*** yarkot has quit IRC09:29
*** hw_wutianwei has quit IRC09:33
*** hw_wutianwei has joined #openstack-ironic09:33
*** lennyb has joined #openstack-ironic09:37
*** nmathew has joined #openstack-ironic09:38
*** amotoki has quit IRC09:44
*** derekh has joined #openstack-ironic09:47
*** MattMan has quit IRC09:54
*** MattMan has joined #openstack-ironic09:54
*** mgoddard has joined #openstack-ironic09:56
*** lindycoder has joined #openstack-ironic10:03
vdrokgood morning ironic, xek and pas-ha !10:03
*** sambetts|afk is now known as sambetts10:06
*** priteau has joined #openstack-ironic10:07
phuongnhmorning vdrok :)10:15
*** sasha2 has quit IRC10:17
*** phuongnh has quit IRC10:20
openstackgerritGalyna Zholtkevych proposed openstack/ironic: Migrate to session for glance client
openstackgerritGalyna Zholtkevych proposed openstack/ironic: Make _send_sensors_data concurrent
*** hoangcx has quit IRC10:27
openstackgerritDebayan Ray proposed openstack/proliantutils: Change Log as part of proliantutils 2.2.0 release
*** zhugaoxiao has quit IRC10:34
*** udesale has quit IRC10:55
*** dtantsur|afk is now known as dtantsur11:05
dtantsurMorning Ironic11:05
vdrokgood morning dtantsur !11:06
dtantsurpas-ha, hey! do you plan on backporting your PXE takeover patch?11:10
*** hw_wutianwei has quit IRC11:14
*** zackf has joined #openstack-ironic11:21
*** nmathew has quit IRC11:24
*** zackf has quit IRC11:26
*** ccamacho1 has joined #openstack-ironic11:29
*** ccamacho has quit IRC11:29
*** krtaylor has quit IRC11:31
*** moshele has quit IRC11:31
*** krtaylor has joined #openstack-ironic11:33
mrtenioMorning dtantsur vdrok!11:39
*** jkilpatr has quit IRC11:41
vdrokmorning mrtenio11:44
yolandahi, good morning11:45
yolandadtantsur, so i'm working in the custom deployment steps spec11:45
yolandamy initial thought was to expose all public methods from the base DeployInterface class, what do you think?11:46
dtantsuryolanda, I think we should get something similar to clean steps, which we defined by
dtantsurnote that actually other interfaces may provide deploy steps, e.g. RaidInterface11:47
yolandaand do we have a common denominator of the steps that all interfaces provide?11:48
dtantsursorry, what do you mean by common denominator here?11:48
yolandato create the default deployment steps11:49
yolandaor each one shall expose their own, similar as the get_clean_steps() there? get_deployment_steps() ?11:51
*** tiendc has quit IRC11:52
*** nmathew has joined #openstack-ironic11:59
yolandaanyway, going to update the spec trying to resemble the cleanup part, that makes logic, then we can iterate from there12:00
openstackgerritDmitry Tantsur proposed openstack/ironic: Add hardware type for IPMI using ipmitool
dtantsuryolanda, yep. the default deployment steps should, I guess, be defined in the DeployInterface base class12:05
TheJuliaGood morning12:05
dtantsurmorning TheJulia!12:08
* dtantsur wonders why he doesn't see EU folks today12:08
TheJuliaNo clue... I'm only up early because I'm about to jump on a flight12:09
yolandadtantsur, and how could that help with custom vendor config, such as customizing an image, adding boot parameters, or touch config of files?12:09
dtantsurTheJulia, I see. safe flight!12:09
dtantsuryolanda, drivers overriding get_deploy_steps12:10
yolandadtantsur, so we shall override this driver in tripleo, let's say?12:10
dtantsuryolanda, we may extend it even in ironic, we're just laying out the basis for now12:11
dtantsurbut yes, for clearly non-ironic things like touching files, it's probably going to be a custom driver or something like that12:11
dtantsurwe can do the same thing as in clean steps, and allow IPA to provide deploy steps12:12
TheJuliadtantsur: thanks!  I just hope inflight wifi works so I don't miss the weekly meeting12:12
yolandaok just using class inheritance, override it, etc12:12
*** jkilpatr has joined #openstack-ironic12:13
*** amotoki has joined #openstack-ironic12:15
TheJuliaAnyone seen hshiina recently on irc?12:22
openstackgerritGalyna Zholtkevych proposed openstack/ironic: Migrate to session for glance client
dtantsurTheJulia, not me.. is it about overwriting your changes?12:25
TheJuliaEh, yeah.  Hoping to get a little better communication going12:29
*** amotoki has quit IRC12:33
cdanielmorning all12:38
*** baoli has joined #openstack-ironic12:38
*** baoli has quit IRC12:39
TheJuliaGood morning12:40
*** Nisha_Agarwal has quit IRC12:45
*** patriciadomin_ is now known as patriciadomin12:49
*** rbartal has quit IRC12:52
*** dprince has joined #openstack-ironic13:00
*** rbartal has joined #openstack-ironic13:04
jrollmorning y'all13:07
*** trown|outtypewww is now known as trown13:12
dtantsurmorning jroll13:12
vdrokmorning TheJulia cdaniel and jroll13:22
openstackgerritAparna proposed openstack/ironic: Support `update_firmware` to hpsum based firmware update
*** amoralej is now known as amoralej|luch13:26
*** awiddersheim has quit IRC13:27
openstackgerritAndrey Shestakov proposed openstack/bifrost: Fixes to allow install Ironic on remote node
openstackgerritAndrey Shestakov proposed openstack/bifrost: Add support of collecting logs from baremetal nodes
*** awiddersheim has joined #openstack-ironic13:29
dtantsurjroll, wdyt about weren't we supposed to kill this code?13:29
openstackLaunchpad bug 1654190 in Ironic "Got 'ImportError' when build doc" [Undecided,Incomplete]13:29
mrtenioHi, I am trying to change the boot order using python-ironicclient, but I keep getting timeout. How could I increase the timeout limit? It is happening in one of our CI's jobs. Thanks13:30
jrolldtantsur: yep, I believe we were waiting for deprecation period things13:30
* jroll takes a look13:31
openstackgerritKyrylo Romanenko proposed openstack/python-ironicclient: Add usage documentation for Baremetal OSC Plugin
*** aparnav has quit IRC13:31
jrollaha, we've been clear to delete that13:32
openstackgerritJim Rollenhagen proposed openstack/ironic: Revert "Revert "Remove ClusteredComputeManager""
jrolldtantsur: we've already done it once, too, ^^ there you are :)13:34
*** mrtenio has quit IRC13:37
*** mrtenio has joined #openstack-ironic13:38
*** mrtenio has quit IRC13:39
*** noama has joined #openstack-ironic13:46
*** jheroux has joined #openstack-ironic13:50
*** noama has quit IRC13:51
*** noama has joined #openstack-ironic13:52
*** amoralej|luch is now known as amoralej13:58
*** sasha2 has joined #openstack-ironic14:02
*** jcoufal has joined #openstack-ironic14:02
dtantsurjroll, also I've updated the IPMI hw type patch this morning14:03
jrolldtantsur: thanks14:04
jrollI kind of hit a wedge with vendor passthru friday night14:04
jrolllooking again to see if I'm missing something, though14:04
jrolltl;dr vendor passthru stuff assumes we have the mapping mixin thing, which we don't in dynamic drivers14:04
*** rloo has joined #openstack-ironic14:05
dtantsurjroll, where do we have mixins beyond old agent passthru?14:05
dtantsuranyway, I think this is why we need to keep vendor_interface changeable ;)14:05
*** Goneri has joined #openstack-ironic14:05
jrolldtantsur: these are all ready for code review, just needs docs/release notes afaik:
patchbotpatch 417970 - ironic - WIP: API changes for driver-composition14:05
patchbotpatch 423536 - ironic - WIP: support dynamic drivers in driver properties API14:05
jrolllet me look14:05
patchbot(let <variable> = <value> in <command>) -- Defines <variable> to be equal to <value> in the <command> and runs the <command>. '=' and 'in' can be omitted.14:05
dtantsurjroll, we get to them today then! /me is looking forward to play with API finally14:06
jrolloh huh, we do just use them directly14:06
jrollwtf was I thinking14:06
jrollfriday night coding isn't my best work :P14:06
dtantsurwe used to have mixins all over the place due to old agent API :)14:06
dtantsurhah, I can imagine :D14:06
jrolldtantsur: yeah, I can finish them this afternoon, if you want to review so I can get your comments fixed up14:07
jrolloh here's the only one left in the project14:07
jrollI guess I could probably kill that14:07
dtantsurI may or may not have a meeting soon. depending on it, I'll maybe get to it soon or late14:07
jrollthat's what confused me, though14:07
jrollok, no worries14:08
dtantsurjroll, ignore this one, it's fake anyway14:08
openstackgerritAndrey Shestakov proposed openstack/bifrost: Fix "Create SSH directory for ironic user" task
jrollwell, I was writing tests against that driver :)14:08
dtantsurouch, I see :D14:08
jrollbut yeah, I should be able to finish everything today14:09
*** ppiela has joined #openstack-ironic14:09
jrolland hopefully set up a CI job tomorrow14:09
dtantsuryeah.. I'll try to also get to the clients' review queue14:09
dtantsurwe have to do final releases of them by Thu14:09
jrolloooo yeah14:10
dtantsurspeaking of which14:10
dtantsursambetts, around?14:10
jrolldtantsur: are we just going to do driver composition client stuff in pike, then?14:10
dtantsurjroll, unless we're super fast, yes.14:10
jrollyeah, fair enough14:11
* jroll seems to recall someone wanting to pick that work up14:11
dtantsuryeah, there are patches up14:11
sambettsdtantsur: hi whats up?14:11
jrollah cook14:11
dtantsursambetts, mind checking please? we need to wrap up everything around client soon14:11
patchbotpatch 409789 - python-ironic-inspector-client - Clarify that node names can be used in addition to...14:11
*** amotoki has joined #openstack-ironic14:11
sambettsdtantsur: I'll take a look :)14:12
dtantsursambetts, would be completely awesome if you manage to check bfournie's LLDP stuff too, as it has a client part14:12
jrollthese should be straightforward, maybe we can land them, hrm14:12
* jroll will look today14:13
*** Nisha_Agarwal has joined #openstack-ironic14:13
dtantsurmaybe I should not have asked for a Depends-On on these patches, as they're anyway coded according to the spec.. dunno14:13
*** moshele has joined #openstack-ironic14:13
*** jerrygb has joined #openstack-ironic14:14
jrollmeh, we can remove them if we want to land it14:14
jrollthere's functional tests, so it's needed14:14
jrolllooks like check-osc-commands might be broken14:15
jrollgood week for that14:15
*** jerrygb_ has joined #openstack-ironic14:15
jrollit's non-voting, ignore me14:15
openstackgerritGalyna Zholtkevych proposed openstack/ironic: Make _send_sensors_data concurrent
*** jerrygb__ has joined #openstack-ironic14:18
openstackgerritKyrylo Romanenko proposed openstack/python-ironicclient: [trivial] Fix typo in documentation
*** jerrygb has quit IRC14:19
*** baoli has joined #openstack-ironic14:19
*** jerrygb_ has quit IRC14:20
*** amotoki has quit IRC14:25
*** jcoufal has quit IRC14:29
openstackgerrityolanda.robla proposed openstack/ironic-specs: Ironic deployment steps
openstackgerritMiles Gould proposed openstack/ironic: Allow setting UEFI boot devices in DRAC driver
dtantsurmgould|afk seems not so afk ^^^14:30
openstackgerritMiles Gould proposed openstack/ironic: Allow setting UEFI boot devices in DRAC driver
*** jcoufal has joined #openstack-ironic14:31
openstackgerritMerged openstack/ironic: Updated from global requirements
openstackgerrityolanda.robla proposed openstack/ironic-specs: Ironic deployment steps
yolandahi dtantsur , new version uploaded... when you have time, can you check that i have not said anything crazy?14:35
dtantsursure, though "when I have time" is not easy this time of the cycle :(14:35
openstackgerritOksana Voshchana proposed openstack/ironic: WIP: TEST GRENADE DEV
dtantsuryolanda, if you don't mind, please keep the status whiteboard updated for this work, see around line 17014:36
*** moshele has quit IRC14:36
*** mgould|afk is now known as mgould14:37
mgoulddtantsur: merely away from IRC :-)14:37
dtantsurI see :) morning mgould14:37
mgould^^^ patch now ready for review, I think14:38
* mgould curses whichever DRAC engineer decided to use "Bios" and "Uefi" for setting boot modes, but "IPL" and "UEFI" for setting boot *devices*14:38
*** jcoufal_ has joined #openstack-ironic14:38
mgouldconsistency is the hobgoblin of little minds14:39
jrolldtantsur: see open discussion re: resource classes:
jrollso I guess that's toward the top of my list this week14:39
mgouldAFAICT "IPL" is a term from mainframe bootloaders?14:39
openstackgerritOksana Voshchana proposed openstack/ironic: WIP: TEST GRENADE DEV
*** nmathew has quit IRC14:40
*** jcoufal has quit IRC14:41
*** jcoufal has joined #openstack-ironic14:42
yolandaof course i failed on linters, need to tidy it14:42
dtantsurjroll, ugh! thanks for asking. Do we have some docs to refer to?14:43
jrolldtantsur: I wrote some last cycle, people didn't want to merge them because they thought plans might change >.>14:44
patchbotpatch 345081 - ironic - Add docs for node.resource_class and flavor creation14:44
* jroll will rebase14:44
openstackgerritMiles Gould proposed openstack/ironic: Allow setting UEFI boot devices in DRAC driver
*** jcoufal_ has quit IRC14:45
jrolldtantsur: fairly accurate as is thuogh14:45
openstackgerritMerged openstack/python-ironic-inspector-client: Clarify that node names can be used in addition to UUIDs
*** vinsh has quit IRC14:46
*** vinsh has joined #openstack-ironic14:47
dtantsurjroll, let's try to land it asap (maybe get someone from nova review it too)14:47
jrolldtantsur: yeah, working on things now14:47
jrollflavor is still unclear14:47
dtantsurjroll, what's the consequence of NOT doing it in Ocata?14:48
jrolldtantsur: there will be a delay in scheduling availability until after the first resource tracker run post-upgrade14:49
*** Nisha_Agarwal has quit IRC14:50
openstackgerritSofiia Andriichenko proposed openstack/python-ironicclient: Negative tests for testing actions with port.
dtantsurupgrade to Pike final, right?14:51
jrolldtantsur: yeah14:51
jrollstill talking some things through in -nova, btw14:51
dtantsurah, ok14:51
jkilpatrok finally got statsd getting metrics, does anyone know what the units are for these?14:51
openstackgerritSofiia Andriichenko proposed openstack/python-ironicclient: Add negative test-cases for openstack node create command
jkilpatrnode_sync_power_state has data points that are like 1k, I assume its micro or milli seconds14:51
jrolljkilpatr: I believe it's milli14:52
*** udesale has joined #openstack-ironic14:52
*** srobert has joined #openstack-ironic14:53
jkilpatrjroll, thanks. any idea hat a node stats controller is?14:53
*** srobert has quit IRC14:53
jrolljkilpatr: controllers are the classes that are used for API request handling, e.g. ironic/api/v1/controllers/ would have a NodeStatesController14:55
openstackgerritSofiia Andriichenko proposed openstack/python-ironicclient: Negative tests for testing actions with node-get command
jrolljkilpatr: so that would have a metric of ironic.api.v1.controllers.node.NodeStatesController.foo_method14:55
jkilpatrjralph, yup seeing that, so provision would be the time it took the api to respond to a provisioning request in milli seconds?14:56
jrolljkilpatr: I believe so, yes14:56
*** jaosorior has joined #openstack-ironic14:56
jrolloh god there's three jr<tab> people here now14:56
jrollwe're ruined14:57
*** causten has joined #openstack-ironic14:57
* jroll /nick trollenhagen14:57
dtantsurlet's kick jrist out :D14:57
jkilpatrevery time  that happens I want to go and write an irc plugin that is context sensitive (tab completes to the last person who talked)14:57
* dtantsur dreams about the same thing14:57
jrollweechat does this :P14:58
jkilpatrjroll, so boot device controller would be how long it took to swap the boot device?14:58
openstackgerritMiles Gould proposed openstack/ironic: Allow setting UEFI boot devices in DRAC driver
jkilpatris that time for the full operation of just to register the request. Looks like the full operation because it takes like 2 seconds14:58
jrolljkilpatr: probably, yes. some of these api calls send a message to the conductor and that will fire a thread and return immediately14:58
jrollright, it depends, but for BMC things it's usually async14:58
openstackgerritSofiia Andriichenko proposed openstack/python-ironicclient: Negative tests for testing actions with node-port-list.
jrolljkilpatr: in which case you likely want to look for ironic.conductor.manager.ConductorManager._do_node_set_power_state or something like that14:59
*** rbartal has quit IRC15:00
mgould^^^ /me learns to run tox *before* pushing, *especially* for trivial changes15:00
openstackgerritZhenguo Niu proposed openstack/ironic: Add nodes tagging support - objs
openstackgerritZhenguo Niu proposed openstack/ironic: Add DB layer support for node tags filter
openstackgerrityolanda.robla proposed openstack/ironic-specs: Ironic deployment steps
*** Syed__ has joined #openstack-ironic15:01
*** rloo has quit IRC15:07
*** sayantani01 has joined #openstack-ironic15:08
*** rloo has joined #openstack-ironic15:08
*** sayantani01 has quit IRC15:08
dtantsurjroll, I'm pressing the button on fyi15:09
patchbotpatch 424113 - ironic - Revert "Revert "Remove ClusteredComputeManager""15:09
jrolldtantsur: +115:09
*** sayantani01 has joined #openstack-ironic15:11
*** Nisha_Agarwal has joined #openstack-ironic15:13
TheJuliaI <3 ATC flight delays15:14
*** srobert has joined #openstack-ironic15:15
rloohi and good morning to the start of another glorious week (TM) :)15:17
rlooTheJulia: ?15:17
*** zackf has joined #openstack-ironic15:18
rloojroll, xek, jlvillal: are you all available to discuss status of rolling upgrades in about an hour or so? (before weekly meeting?)15:18
*** mtanino has joined #openstack-ironic15:19
jrollrloo: I can be, yeah15:19
openstackgerritZhenguo Niu proposed openstack/ironic: Add DB layer support for node tags filter
jlvillalrloo, Sure. Maybe vsaienk0 would like to join.15:19
openstackgerritMerged openstack/ironic: Turn NOTE into docstring
rloojlvillal: sure, anyone is welcome to join! just wanted a few key folks there :)15:19
openstackgerritMerged openstack/ironic: Fix typo in the metrics.rst file
TheJuliarloo: traveling today, got a boarding notice for my next flight as we were taxiing to the arriving gate for my layover, got to the gate and air traffic control put a 1.5 hour hold on the flight.15:20
rloojlvillal: did we convert you to east coast time? :)15:20
rlooTheJulia: OH :-(15:20
jlvillalrloo, No. I just woke up about 10 minutes ago.15:21
*** vinsh_ has joined #openstack-ironic15:21
rlooTheJulia: I thought ATC = active tech. contributor in openstack, but there's another world out there :D15:21
rloojlvillal: good morning :)15:22
openstackgerritJohn L. Villalovos proposed openstack/ironic: WIP/DNM: Testing multinode job
*** vinsh has quit IRC15:24
TheJuliarloo: I realized that might be how it was interpreted right after you raised the question mark.15:26
rlooTheJulia: :)15:27
openstackgerritJim Rollenhagen proposed openstack/ironic: Add docs for node.resource_class and flavor creation
jrollrloo: ^ it returns from the grave, would love your review there if you have time15:27
dtantsursambetts, hey, have you seen ? mgould and I can't wrap our heads around it15:27
openstackLaunchpad bug 1656010 in neutron (Ubuntu) "Incorrect notification to nova about ironic baremetall port (for nodes in 'cleaning' state)" [Undecided,New]15:27
dtantsuralso morning jlvillal, rloo15:27
rloojroll: heh. sure, what's the urgency on that?15:28
jrollrloo: this week?15:28
jrollor like15:28
jrollthis release, I guess15:28
rloojroll: ah, ok, will look this afternoon then or tomorrow.15:28
sambettsdtantsur: /me hates launchpads limited width for bug reports...15:28
openstackgerritMerged openstack/python-ironicclient: Typo fix: prefered => preferred
sambettsdtantsur: I think I know whats happening15:29
*** baoli has quit IRC15:30
dtantsurmind triaging this one please? I barely understand a word :(15:30
jrollis that even a bug?15:30
sambettsdtantsur: basically in neutron they have a notifyer thing which when a new port is created it sends a message to nova so that nova can bind it without a user having to go to nova and do a vif-attach15:30
sambettsdtantsur: this is saying that a notification is sent for the ports we create for cleaning15:31
*** ccamacho1 has quit IRC15:32
vdrokmorning rloo jlvillal and sambetts15:33
jrolldtantsur: do you know offhand if our osc plugin has a --wait for e.g. waiting for enroll->manageable->available?15:33
dtantsurjroll, the patch is on review15:33
jrollI'm updating devstack to do that thing15:33
*** beekneemech is now known as bnemec15:33
mgouldsambetts: so the fix is "don't send that notification if we're not going to create an instance"?15:33
dtantsuryou can just merge it ;)15:33
rloodtantsur, jroll: you mean
patchbotpatch 354806 - python-ironicclient - Add --wait to OSC provisioning commands15:34
sambettsdtantsur: nova ignores it because there is no device_id AKA instance id with a id == ironic so :/ I guess we probably need to add code to the nova notify to ignore ports with the baremetal vnic type15:34
sambettsmgould:  ^15:34
openstackgerritMerged openstack/bifrost: Typo fix: enviroment => environment
rloojroll, dtantsur: was sitting in my inbox still. will look at it today ...15:34
*** yolanda has quit IRC15:36
*** yolanda has joined #openstack-ironic15:36
*** sayantani01 has quit IRC15:36
*** sayantani01 has joined #openstack-ironic15:37
*** yolanda has quit IRC15:37
*** udesale has quit IRC15:37
*** yolanda has joined #openstack-ironic15:38
jrolloh, we have it for ironic cli already, woohoo15:40
dtantsuryep :)15:40
*** baoli has joined #openstack-ironic15:41
jrollthe downside being now I have no excuse to fix a bunch of crap15:41
jroller, not to fix15:42
TheJuliarloo: I'm going to review the rpc pin and patches up to the db online upgrade stuff.  I still have the -2 on the first patch, and kind of feel like I should pull it so we can get that stuff landed15:42
rlooTheJulia: thx for reviewing! we can discuss today about whether/when to remove the -2. I want to discuss with xek and see where testing is. i am hoping xek will be around in 30 min or so to discuss. you can join too!15:43
rlooTheJulia: I'm currently 'stuck' on the version column patch. i think that is the most crucial one so i want to make sure i totally grok it.15:43
TheJuliarloo: I'll try, they are telling us we should begin boarding in about 20-25 minutes and might be taking off during the beginning of the meeting15:44
TheJuliarloo: 10-415:44
rlooTheJulia: OH, don't miss your flight. guess you may not be at ironic meeting too?15:44
rlooTheJulia: oh yeah, just read what you said (again). heh.15:44
openstackgerritJesse J. Cook proposed openstack/ironic: Add rescuewait timeout periodic task
openstackgerritJesse J. Cook proposed openstack/ironic: Add API and conductor methods for [un]rescue
TheJuliaI'm hoping to, but wifi was also refusing to work on my last flight.15:45
*** jerrygb has joined #openstack-ironic15:45
rlooTheJulia: for boot from volume, is there anything that must/want to get into ocata? I saw the note about not being able to get client stuff. in that case, is there anything else we should try to get in?15:46
rlooTheJulia: if you could update etherpad with that, so we know whether there are any priorities there...15:46
*** ccamacho has joined #openstack-ironic15:46
*** jerrygb__ has quit IRC15:47
TheJuliaI'd like to try get more stuff under the hood, but I also wf-1'ed the current first patch because it got reverted, and I won't be in front of a git client for a few hours15:47
rlooTheJulia: ok, maybe add a note/link to those and mention that they will be ready to review later today/tomorrow, whenever.15:48
TheJuliarloo: but at the same time, if we can get it all lined up, first thing in pike would be awesome.15:48
rlooTheJulia: definitely!15:48
TheJuliarloo: I put a note on the first one like four hours ago :)15:48
* TheJulia has been up for a while15:49
* rloo looks and yes, it is there. Thx TheJulia!15:49
*** rbudden has joined #openstack-ironic15:51
openstackgerritMichael Turek proposed openstack/ironic: [WIP] OpenBMC Driver
*** jcoufal_ has joined #openstack-ironic15:56
sambettsjroll: should the nova notifier be notifying nova of baremeta:none owned ports??
jrollhuh, we use capabilities in devstack now? O_o15:58
openstackgerritJim Rollenhagen proposed openstack/ironic: Allow using resource classes
openstackgerritJim Rollenhagen proposed openstack/ironic: DNM: turn on resource classes
jrollsambetts: I know absolutely nothing about neutron notifications15:58
*** jcoufal has quit IRC15:59
mgouldsambetts: makes sense, thanks16:00
mgoulddoes that happen in code that we control?16:01
* mgould continues to be awed at wifi on planes16:02
mgouldwe're living in The Future, people16:02
sambettsjroll: I think the problem is that we seem to be overriding the device_owner field for ports created/passed through nova for Ironic changing them to baremetal:none16:02
sambettsjroll: which broke the nova notifications for those ports16:02
jrollsambetts: I don't even understand what the problem is, or if it's a real problem (vs superficial scary looking things in logs), to be clear16:03
sambettsjroll: and someone fixed it by changing the notification code in neutron instead of fixing the device_owner in Ironic16:03
*** aparnav has joined #openstack-ironic16:03
sambettsjroll: this is the bug report that lead to that neutron change:
openstackLaunchpad bug 1606229 in neutron "vif_port_id of ironic port is not updating after neutron port-delete" [Medium,In progress] - Assigned to Andrey Shestakov (ashestakov)16:04
jrollI see16:04
jrollso the problem is that we don't handle the case where a port we're using is deleted?16:04
sambettsjroll: if we left the device_owner alone for ports that came through nova then our cleaning ports wouldn't have resulted in nova notifications16:04
*** dyasny has quit IRC16:04
*** dyasny has joined #openstack-ironic16:05
jrollsambetts: which means... what exactly16:05
sambettsjroll: yeah, we don't support interface detach in nova driver16:05
* jroll doesn't understand any of this16:05
jrollmaybe you can summarize in a way that my 5-year-old brain can understand on the ML? :)16:05
*** yolanda has quit IRC16:06
*** amotoki has joined #openstack-ironic16:06
*** yolanda has joined #openstack-ironic16:06
sambettsjroll: heh, yeah I'll try to put something togther16:07
jrollthanks :D16:08
aparnavHi jroll16:09
sambettsjroll: TL;DR we change a nova set value on the port, which broke nova notifications, and they "fixed" it in neutron instead of ironic, which now results in notifications being sent to nova for neutron ports owned by Ironic16:09
jrollhi aparnav16:09
TheJuliarloo: Just looked at the version rev.  Yeah, I totally get your comment earlier regarding it :)16:09
jrollsambetts: ahhhh, even if it's totally unrelated, I see16:09
sambettsjroll: yeah16:09
rlooTheJulia: it may be short, but is important!16:10
aparnavI have reframed the rfe you have approved previously, I wanted to confirm whether change looks good or not16:10
openstackLaunchpad bug 1648448 in Ironic "[RFE]iLO driver `update_firmware` to support hpsum based firmware update" [Wishlist,In progress] - Assigned to Aparna (aparnavtce)16:10
sambettsjroll: I suggest we revert the neutron change, and fix the device_owner thing on the Ironic side16:10
sambettsjroll: tbh I think we do some weird stuff with device_id in the neutron net-interface which I'm not sure we should be doing16:11
jrollaparnav: what did it say before?16:11
TheJuliarloo: I feel like it is missing something, I just can't put my finger on it.  Likely contextual with patches beyond the the database upgrade command rev16:11
rlooTheJulia: some stuff related to that got merged a cycle or two ago...16:12
openstackgerritSam Betts proposed openstack/ironic: Rework Ironic devstack baremetal network simulation
jrollaparnav: is the update mode going to be a config, or how is that set?16:12
jrollsambetts: I trust you to do the right thing, I don't have room in my brain for this, honestly16:12
sambettsjroll: heh, I'll dig into it16:13
aparnavjroll, I have proposed to add it as a new manual clean step. Now I decided to leverage the existing update_firmware availble for iLO drives.16:13
*** pcaruana has quit IRC16:13
aparnavjroll, It is not a config, It uses the existing `update_firmware` with some changes added to support hpsum mode of firmware update16:14
jrollaparnav: ok cool, sounds great to me :)16:14
aparnavjroll, thanks :)16:15
xekrloo, Hi16:17
rlooxek: hi!16:17
rlooxek: you avail to talk about rolling upgrades now?16:17
xekrloo, yes :)16:17
* jlvillal is here16:17
rlooxek: sweet. jroll, jlvillal, vsaienk0 ^^16:17
rlooxek: so where are you wrt coding?16:18
rlooTheJulia: ^^ if you're around16:18
* jroll will be listening in 2 minutes, carry on and I'll read back16:18
xekrloo, I think we are finished on the implementation side16:18
xekrloo, we tested the upgradability by running tempest16:18
xekrloo, on partialy upgraded and upgraded installation16:19
xekrloo, with pinning and whatnot, but without creating and testing resources between the upgrade steps16:19
rlooxek: so wrt coding. is this the last code patch in the series:
patchbotpatch 420728 - ironic - Add compatibility with Newton when creating a node16:20
TheJuliarloo: perfect timing, just boarded the aircraft16:20
xekrloo, this specifically fixes one of the incompatibilities we found16:21
*** moshele has joined #openstack-ironic16:21
openstackgerritMerged openstack/ironic: Revert "Revert "Remove ClusteredComputeManager""
rlooxek: is it needed for rolling upgrades to work?16:21
xekrloo, it is needed for the upgrade to ocata16:21
rlooxek: ok, so we need that patch then.16:22
rlooxek: the one after is a PoC:
patchbotpatch 415462 - ironic - [PoC] Add fake fields in Node object and RPC16:22
rlooxek: is that one needed?16:22
jrollwell, we upgrade the api side first, right?16:22
rloojroll: no conductor side16:22
jroller, sorry, conductor16:22
rloojroll: yeah, conductor side first :)16:22
xekrloo, no that one we are using for the upgrade test, it contains various test changes in the rpc and the database16:23
rlooxek: ok.16:23
jrollso node-create is only broken if we upgrade with pins16:23
rlooxek: so coding wise, i think if 'we' can all agree on the version column patch, we're probably good.16:23
*** moshele has quit IRC16:23
jrollI'm wondering, with driver composition, if changing node-create like that will cause some awkwardness16:23
xekjroll, yes16:24
rlooat least for me, that seems the most complex:
patchbotpatch 412397 - ironic - Add version column16:24
jlvillalI just updated the Whiteboard lines 110-117 with current progress of testing:
jrolldtantsur: how bad would running ocata like this for some amount of time be?
patchbotpatch 420728 - ironic - Add compatibility with Newton when creating a node16:24
* jroll thinks that's a rabbit hole16:24
*** stendulker has joined #openstack-ironic16:25
TheJuliaPossibly multiple rabbit holes16:26
rloooh, interesting. so rolling upgrades can be a problem if we switch how we create stuff.16:26
rloowhat about disallowing node creation during upgrades?16:27
jrollrloo: well, it's unclear if that causes problems, it's probably fine16:27
xekrloo, jroll it only changes the place where we create it (to be the api service instead of the conductor)16:27
rloojroll, xek: ok, we should review/comment on that patch itself then.16:27
jrollrloo: the issue would be, if someone creates a node with e.g. --deploy-interface foo, that wouldn't be validated, It hink16:27
xekso for the pinned rolling upgrade timeframe, the behavior follows that of newton16:28
jrolloh actually16:28
jrollmaybe you're right16:28
jrollmaybe not though16:28
jrollxek: the problem is driver_factory.check_and_update_node_interfaces() may work differently on the conductor16:28
rloojroll: let's confirm. we can also add code in to prevent the newton code from handling --deploy-interface.16:28
jrollactually yes, this would cause API services to load drivers up, which is bad16:29
xekrloo, yes16:29
rloojroll: err, ocata code from handling newton-api request with --deploy-interface16:29
dtantsurjroll, it's going to work, in some cases, I guess16:32
rloowhatever we do to solve this, we may have a similar problem in the future if we move creation of eg chassis to conductor.16:32
*** dsneddon_afk is now known as dsneddon16:33
openstackgerritMerged openstack/python-ironicclient: Add a new command "ironic node-inject-nmi"
jrollrloo: yes, but we'll know in advance, so we can make sure we plan things out so that it doesn't completely fall apart if we fall back to old way16:34
rloojroll: right16:34
jrollit's too far gone for this case, I think16:34
jrollbut I'm not sure if it will actually break16:34
TheJuliaDropping off until we are in the air16:34
rlooi think from coding pov, if we can grok the version column patch and solve this create-node issue, the feature is probably good to go in.16:34
rlooi'll take a look at the create-node patch later today/this week.16:35
rloowhat about testing side? jlvillal said he updated etherpad with that status.16:35
jrollrloo: sure, I agree with that. I'm curious if we'll get CI running, though, and if we won't then is it worth landing anyway?16:35
rloojroll: my current question!16:36
jlvillalrloo, We aren't done yet for sure. But did make a lot of progress last week. Which I am very happy about :)16:36
rlooline 11116:36
rlooi don't recall what CI we need for this. need to look at spec...16:36
rlooxek: all your testing has been with your local environment, right?16:37
xekrloo, yes16:37
rloojroll: xek had mentioned to me that in other projects the CI wasn't avail when the code was, so they put it in as 'experimental'. something to consider.16:38
jrollrloo: yeah, I suppose16:38
rloojlvillal, xek: in english, what of those do we think can be done in ocata?16:38
*** sacharya has joined #openstack-ironic16:39
jlvillalrloo, I doubt that the multi-node grenade will be done in Ocata.16:39
rloojlvillal: ok. so no testing of rolling upgrades.16:40
rloojlvillal: your status has 'multi-tenant', not multi-node.16:40
xekjlvillal, rloo, jroll, I remember when discussing this list, that we accepted that the scope of the testing could be more narrow, or not ideal at first16:40
jlvillalrloo, Correct. I have been working on getting grenade to work with multi-tenant. Will add multi-node after that.16:40
rloojlvillal: does multi-node need multi-tenant, or was that the order you decided to take? (just wondering)16:41
jlvillalrloo, It will be multi-node, multi-tenant, and grenade.16:41
jrollok so with this rpcapi thing, we won't be setting the calculated defaults for nodes created without specific interfaces16:41
jrolldtantsur: ^16:41
jrollI think that's the main point16:41
jlvillalrloo, The current just multi-node job is tested with multi-tenant. But it was mostly just an order I decided.16:42
*** hshiina has joined #openstack-ironic16:42
rloojlvillal: ok.16:42
dtantsursorry, I'm distracted with other things, what's the tl;dr of the problem?16:42
*** hoonetorg has joined #openstack-ironic16:42
openstackgerritAndrey Shestakov proposed openstack/bifrost: Fix endpoint service name for inspector
jlvillalrloo, We are close on the multi-tenant with grenade. So once that is done, add multi-node and see what breaks...16:43
rloojroll: i'm not a testing person. what do you think the minimum CI is needed for rolling upgrades, and/or for 'experimental' rolling upgrades?16:43
jrolldtantsur: just trying to track down if this patch, when using the "old" node create stuff, will break us:
patchbotpatch 420728 - ironic - Add compatibility with Newton when creating a node16:43
openstackgerritZhenguo Niu proposed openstack/ironic: Add nodes tagging support - API
jrollrloo: the first two sub-bullets in IMO16:44
*** amarao has quit IRC16:44
jrolldtantsur: if you're busy just ignore me :)16:44
rloojlvillal: first bullet can be done?16:44
jrollfirst bullet is already done16:44
jlvillalrloo, I think that would be a question for xek. If you mean unit tests.16:44
dtantsurjroll, this rolling upgrade thing, are we starting with Newton?16:44
dtantsurI thought the first upgrade was Ocata->Pike..16:45
rloojlvillal: no, grenade. but it is done.16:45
openstackgerritMerged openstack/ironic: Ipmitool management driver for Inject NMI
rloodtantsur: we're shooting for newton->ocata16:45
jrolldtantsur: yeah, that's what current code does, newton->ocata16:45
*** hoonetorg has quit IRC16:45
jrolldtantsur: so the question is, does this break down enough that ocata-pike should be the first16:45
jlvillalrloo, We have 'grenade' but not grenade with multi-node. So I think first bullet for multi-node grenade is not done. Since we don't have any multi-node grenade yet.16:45
dtantsurso, is the concern that we won't have hardware types enabled on API?16:46
rloojroll: so most likely, 2nd bullet will not be done. if it is done and code is done, then upgrades can be in ocata.16:46
jrollrloo: jlvillal: oh, I was wrong, the first sub-bullet is *not* done, jlvillal is right16:46
rloojroll: if 2nd bullet is not done and code is done, then ??16:46
rloojroll: sigh.16:46
xekjroll, dtantsur, also, keep in mind, that it is recomended in the upgrade procedure to upgrade the ironic client after the ironic service upgrade is finished, so would the --deploy-interface be somehow available in the old client when the new services are runnin pinned?16:46
jrolldtantsur: right, we so far haven't loaded drivers in the api services16:46
openstackgerritKyrylo Romanenko proposed openstack/python-ironicclient: Add usage documentation for Baremetal OSC Plugin
jrollxek: people just pip install ironicclient whenever they want, we must assume old/new clients need to always work16:47
jrollxek: (except when used in services like nova)16:47
*** hoonetorg has joined #openstack-ironic16:47
rloojroll, xek: i think we should spend some time looking into that create-node patch to see what the issues are.16:47
jrollrloo: I'm half talking here and half doing that16:48
dtantsurjroll, xek, right. we can take one more step back, and read the configuration options the way we used to; and forbid hardware types in the fallback case16:48
rloojroll: yeah, i figured so :) but i think we need to clarify the ci part.16:48
rloojroll: cuz i believe we can solve the coding part. i am not familiar enough about the testing part to have any confidence in that16:49
jrolldtantsur: we never read the configs, we used rpcapi.get_topic_for(), which now includes hardware types :/16:50
jrollnope, it will, damn16:50
dtantsurwe did read the configs to determine the default network interface16:50
jrollah, yeah16:50
jrollhrm, idk if I love that16:51
jrollso here's where I'm at:16:51
jroll1) we'll need to take a step back (or two) in driver composition16:51
jroll2) we won't have CI at all16:51
jroll3) we're a week or so from our proposed soft feature freeze, and ~2 weeks from final release16:52
jrollI'm not sure I see much benefit in landing this now, as opposed to the first week of pike, personally16:52
dtantsuralso, aren't we supposed to update all conductors first?16:52
jrollother than to say "we did it!"16:52
jrolldtantsur: we are, this is about the time between "everything is upgraded" and "rpc versions are unpinned"16:53
dtantsurhmm, yes16:53
jrollit would work if we didn't pin rpc16:53
vdrokdtantsur: I think we won't be able to send ver 1.36 message if the rpc client is capped16:53
dtantsurright, right16:53
xekjroll, well, the patch is a fix for a bug we found doring testing, there may be others16:54
patchbotpatch 420728 - ironic - Add compatibility with Newton when creating a node16:54
jrollxek: right, because we don't have CI, we don't even know if this works yet16:54
xekjroll, shouldn't we treat all of them like bugs, and fix them after soft feature freeze?16:54
jrollxek: right, but we can't say this works until we have CI anyway, so we won't actually know if the bug fixes actually fix it16:55
rlooxek: what happens if there are bugs we don't find, and someone tries to do a rolling upgrade? if they could get into a bad state... i'm not sure we want to risk that.16:55
rloojroll: is there any worth in having out there as experimental, so people can 'test' it?16:56
rloojroll: even with CI, there's only so much we'll be able to test.16:56
jrollrloo: idk, people can just as easily test it after the branch. for example in a 7.1.0 release 3 weeks after PTG :)16:56
rloojroll: with the node-create, we could add code that just returns an error if they try with old api and conductor is pinned.16:57
*** david-lyle_ has joined #openstack-ironic16:57
rloojroll: true, if/when there is a 7.1.0 release.16:57
jrollrloo: we already did that (inadvertently)16:57
jrollthey can also test master, you get my point16:57
rloojroll: true16:58
rloojlvillal: the likelihood of getting multinode grenade is very low, right?16:58
rloojlvillal: in the next 2 weeks16:58
jlvillalrloo, Yes.16:59
rloooh, meeting in one minute16:59
* jroll does the thing16:59
rloowe can discuss more at meeting or after. but i think we are punting to ocata.16:59
*** mjturek has joined #openstack-ironic16:59
jrolls/ocata/pike/ :)17:00
* jroll wonders out loud if a version column in ocata would be useful for ocata->pike17:00
rloojroll: ++ to version column and yeah, pike. (pike, makes me think of star trek, i'm old...)17:01
*** baoli has quit IRC17:01
*** Sukhdev has joined #openstack-ironic17:01
*** rpioso has joined #openstack-ironic17:03
*** david-lyle_ has quit IRC17:03
*** stendulker has quit IRC17:05
*** rajinir has joined #openstack-ironic17:06
*** stendulker has joined #openstack-ironic17:08
*** stendulker has quit IRC17:10
openstackgerritAndrey Shestakov proposed openstack/bifrost: Add wait_for_node_deploy parameter to deploy task with custom instance_info
*** e0ne has quit IRC17:13
*** stendulker has joined #openstack-ironic17:13
*** stendulker has quit IRC17:13
*** stendulker has joined #openstack-ironic17:14
*** ccamacho has quit IRC17:14
openstackgerritSam Betts proposed openstack/ironic: [WIP] Don't override device_owner for tenant network ports
*** tesseract has quit IRC17:15
openstackgerritAndrey Shestakov proposed openstack/bifrost: Allow enroll node with empty properties.
sambettsjroll: ^ /me thinks that should solve the issue, but I'm waiting to see the consequences17:16
*** baoli has joined #openstack-ironic17:18
*** causten has quit IRC17:21
*** causten has joined #openstack-ironic17:23
*** causten has quit IRC17:26
*** stendulker has quit IRC17:28
*** mat128 is now known as mat128|afk17:28
*** mat128_ has joined #openstack-ironic17:29
*** stendulker has joined #openstack-ironic17:29
openstackgerritAndrey Shestakov proposed openstack/bifrost: Add missing packages
*** stendulker has quit IRC17:30
*** stendulker has joined #openstack-ironic17:31
*** causten has joined #openstack-ironic17:31
*** stendulker has quit IRC17:32
*** derekh has quit IRC17:33
*** stendulker has joined #openstack-ironic17:33
*** causten has quit IRC17:35
*** dyasny has quit IRC17:36
*** sayantani01 has quit IRC17:36
*** aNuposic has joined #openstack-ironic17:36
*** sacharya_ has joined #openstack-ironic17:37
*** sacharya has quit IRC17:38
*** stendulker has quit IRC17:39
*** stendulker has joined #openstack-ironic17:41
rookdtantsur: hey - curious wrt nodes going into maintenance -- could we put in an interval to check if the node should _still_ be in maintenance mode?17:41
*** ohamada has quit IRC17:42
dtantsurrook, we've been planning on it for some time17:42
dtantsurjust never got to it. I think JayF was planning on something like that17:42
rookdtantsur ok17:42
JayFthere is a spec under review17:42
jrollrook: dtantsur
patchbotpatch 334113 - ironic-specs - Add spec for specific faults support17:42
JayFthanks for saving me a google jroll17:43
rookdtantsur: we are seeing if there is another service hitting nodes via IPMI there is some sort of locking that causes nodes to be put into maintenance.17:43
rookdtantsur: ie, Sat617:43
*** stendulker has quit IRC17:44
*** stendulker has joined #openstack-ironic17:45
rlooxek: is there any merit in adding the 'version' column and getting dbsync online migrations in, in ocata? (w/o the rolling upgrades code)17:45
rlooxek: guess it'd need the mappings too.17:45
*** aparnav has quit IRC17:45
jrollwe could skip this null bit: "This version column will be null at first and will be filled with the appropriate versions by a data migration script. If there is a change in Ocata that requires migration of data, we will check for null in the new version column."17:45
xekrloo, I think the online migration scripts will be easier for developers17:46
TheJuliaI think there is always merit in getting the database structural changes in early, which is a behavior we will need to enforce moving forward.17:46
dtantsurrook, why would Sat6 hit nodes via IPMI? this is a know problem indeed17:46
TheJuliajroll: or s/a data migration script/`the command goes here`17:46
rloojroll: so my only concert wrt adding version column is that the feature itself that uses it, won't be avail. we have to be really sure we will use this...17:46
rookdtantsur it captures the state of the node.17:46
jrollrloo: I think we've agreed by now that it is useful17:47
TheJuliarloo: if we suppress the version column from being returned by the API, which we should likely do anyway, then we can always remove it without it being a big deal17:47
rloojroll: yup, just wanted to point it out cuz i'm always hesitant to add things too early :) but i'm good with it.17:47
jrollrloo: you? hesitant?17:47
dtantsurrook, why can't it use ironic API fo rthat?17:47
rloojroll: heh, like the resource classes stuff :)17:47
jrollrloo: :)17:48
*** dyasny has joined #openstack-ironic17:48
rloojroll, TheJulia, xek: so if we add new version column, jroll sez just use existing VERSION values, xek sez add dbsync migrations. ?17:48
jrollI guess we'd need a migration to add existing values17:49
*** stendulker_ has joined #openstack-ironic17:49
TheJuliaThe reason not to populate them now is not to churn the database17:50
rloojroll: right. i was thinking we could code, but that won't migrate any objects that aren't actually touched/modified.17:50
*** trown is now known as trown|lunch17:50
rlooit would be a good test/use of the new dbsync command.17:50
jrollwell, we wouldn't need the new command, though we could17:50
TheJuliaOur documentation would need to be a bit more specific in the upgrade process then, but I think that would work...17:51
*** stendulker_ has quit IRC17:52
jrollwe don't have to add the new data migration command yet17:52
jrollcould just add 'update nodes set version=foo' in the migration to add the column17:52
TheJuliaWell, the spec details that it would just need to be run before, once we have it, we would need to be hyper detailed on the order of operations17:52
jrollagree, but not for the ocata upgrade17:53
*** stendulker has quit IRC17:53
TheJuliajroll: exactly17:53
TheJuliaSorry, my brain tends to think off in the future17:53
* jroll feels like there's a couple threads here17:53
*** mgoddard has quit IRC17:54
jrollyeah my question right now is "how much of this, if any, should we land in the next 1.5 weeks"17:54
TheJuliaI think update nodes set version=blah could work, but I would be worried about how database locking would occur with all rows being updated, then again, it might not matter, I doubt there are many super large installs where that column update wouldn't take that long.17:55
aslezilaareifiev: jayF: jroll: sambetts:  ( This is the spec for collecting biosdevname of NICs. re uploaded a patch as suggested by the community. Have a +2 on it. I would be great if someone could review it.17:55
patchbotpatch 404372 - ironic-inspector-specs - Collect NIC name given by BIOS17:55
TheJuliajroll: I think if we add the version field and the versions, then we can add the command later17:55
vdrokas it seems to me, the rpc version cap is pretty straightforward, and the behaviour does not change compared to the current situation17:55
TheJuliaDocs will need to be very specific on the upgrade happy path regardless17:55
dtantsursee you tomorrow folks!17:56
TheJuliagoodnight dtantsur17:56
rloobye dtantsur!17:56
*** sayantani01 has joined #openstack-ironic17:56
*** dtantsur is now known as dtantsur|afk17:56
xekTheJulia, they are updated using a limit, so this won't have a big impact17:56
rloovdrok: without CI, that part of it seems risky17:56
*** david-lyle_ has joined #openstack-ironic17:56
*** david-lyle_ has quit IRC17:56
rlooxek: TheJulia was talking about not using the dbsync command17:56
rlooxek: but doing a db update in code17:56
vdrokgood night dtta17:57
*** david-lyle_ has joined #openstack-ironic17:57
vdrokouch, dtantsur|afk17:57
xekrloo, in that case, the service will be down during the upgrade, so it doesn't matter17:57
rlooxek: oh yeah, forgot about that!17:57
jrollyeah, I think I'd be fine adding the version column/mappings this cycle, and hold the new dbsync command and rpc cap for pike17:57
rlooxek: so what do you think? it'll make the code a bit easier in pike. but up to you i think.17:58
TheJuliaxek: I agree with the idea of that being helpful, I'm just not sure someone would actually do that out of the gate until they realize that their database is locked up for x amount of time and do something awful like control-c dbsync.  Actually, that is the perfect argument for the online db upgrade command landing17:58
rlooxek: cuz then, the .version will all be available for the objects17:58
rlooTheJulia: if we assume a table with 10k nodes, any idea how long it'd take to add a new column and populate it?17:59
TheJuliarloo: entirely dependent upon the disks, and db version at this point17:59
TheJuliarloo: and any existing io contention17:59
jrollI mean, even at a million rows, it should be fast18:00
xekrloo, I can change the patch to go before the rpc one and add an (old style) offline migration in alembic18:00
jrolland there's no way someone has a million rows in any of these tables without bragging about it18:00
rloojroll: yeah, that's what I'd think but I have no practical db experience18:00
TheJuliajroll: That would be if the index is entirely in ram, if not, then it could be a not fun day for whoever the dab is18:00
TheJuliaerr, s/dab/dba/18:00
*** ralonsoh has quit IRC18:01
* TheJulia has had indexes > physical ram and it is not fun18:01
vdrokwill the value of version be empty then?18:01
jrollTheJulia: yeah, maybe you're right, dunno18:01
jrollit'd be fine to populate it after the upgrade in the background somewhere18:01
rloovdrok: no, we're talking about coding it so the value of version == the object's VERSION value18:01
jrolljust... more code18:01
rloovdrok: and hence the lock on the db18:02
rloovdrok: and amoutn of time that would take18:02
openstackgerritBob Fournier proposed openstack/ironic-inspector: Add plugin to process basic management LLDP TLVs
jrollTheJulia: if the alembic migration sets a default for the new column, such that it is the right one, that shouldn't be any heavier than just adding the column, right?18:02
TheJuliaOf course, to have a db that size, someone would have to be operating a very, exceptionally large, ironic install18:02
jrollor an exceptionally small db machine :P18:02
TheJuliaThat does happen as well18:03
* TheJulia wonders what the airplane internet lag actually is at the moment18:04
jrolldoesn't seem too bad18:04
*** mgould is now known as mgould|afk18:04
TheJuliajroll: Oh, that is right, setting a default, however if we set a table default, we would later have to remove the default18:04
* TheJulia hopes nobody interacts with ironic directly via the DB18:04
jrollTheJulia: yeah, I'm thinking if it can just write it and not set the table default18:05
rlooTheJulia: the default value would be the object's class version. wouldn't need to remove it i think.18:05
jrollNote that this statement uses the Column construct as is from the SQLAlchemy library. In particular, default values to be created on the database side are specified using the server_default parameter, and not default which only specifies Python-side defaults18:05
*** david-lyle_ is now known as david-lyle18:05
jrollI think default=foo is what we would want18:06
TheJuliarloo: I need to think about that since we've been bitten by using current state with database upgrades before18:06
TheJuliaerr, current state in code18:06
TheJuliaSpecifically testing though... It should work if we just keep updating the default.....18:07
TheJulia(for new records that is)18:07
rlooTheJulia: ok, good to think about it. if we don't explicitly set the default value via code, what is the (if any) value of adding version column + dbsync-background-command-to-set in ocata, and rest of code in pike?18:07
xekjroll, rloo, TheJulia, there wouldn't be an internal database default, the default value will be filled based on the object's version18:07
vdrokyeah, having default=object.VERSION should be OK18:08
TheJuliaoh, that might bite us, old row, doesn't get touched, db upgrade, new default, no default on row... it might represent the new default automatically18:08
rlooTheJulia: ? we want the old row to have the new default18:08
vdrokTheJulia: if there were no updates, that means it should be OK18:09
xekvdrok, it will be more complicated than that, because the version column is on the base mode and versioned objects are not that tightly coupled with sqlalchemy models18:09
TheJuliabut if there are migration activities, we need to undertake them, I would think18:09
xekvdrok, *base model18:09
rlooxek: if we add a version column in ocata (master), and set that value to be the object's VERSION, why won't that work?18:10
rlooxek: this is during the cold upgrade to ocata.18:10
rlooif this isn't that simple, we should just punt it all to pike.18:11
xekrloo, because the version column is on the base model, so there is only one default we could set for all objects18:11
TheJuliaWe would have to touch each row, I think we need the command + the version column field to at least build the pattern moving forward, since we can't use server side defaults for asserting the object version in the db18:11
xekrloo, I think we should keep the version column on the base model, for consistency18:11
vdrokxek: hmm, the create_node gets the values from object.get_changes right? And it could just include version=object.VERSION18:11
xekrloo, so we can assume it's always present18:11
jrollsounds like we should just punt the whole thing :P18:12
TheJuliaxek: except we need to scrub it from api replies18:12
TheJulias/api/all api/18:12
rloojroll: yeah, that's what i'm thinking. i thought it was straightforward but am not sure now.18:12
xekrloo, vdrok, TheJulia the current implementation in my patch should work fine18:12
rloounless someone xek? can convince us otherwise.18:13
TheJuliaI think I'm onboard with eek as long as we not setting any server side table defaults18:13
* TheJulia tosses a wet cat at autocorrect18:13
rlooxek: what part of the current implementation? we don't want the rolling upgrades part. and unless we can guarantee that version column is populated in ocata (for upgrading to pike), i see no use in adding it in ocata18:13
patchbotpatch 412397 - ironic - Add version column18:14
rlooxek: i don't want to add obj_get_changes(). i don't think i do anyway.18:14
xekrloo, the target_version would be always equal to obj.VERSION in this case18:14
TheJuliaBut given we've got a shed, and we're trying to figure out how to modify the foundation without moving the shed, punting does sound like the ideal action project wise18:14
openstackgerritNisha Agarwal proposed openstack/proliantutils: snmp inspection for disk size
* TheJulia doesn't like the idea of punting18:15
vdrokxek: unless someone sets a pin :) We'd have to remove that option too for now18:15
rlooxek: so we'd add a version column with null value, and then it would get set if an object gets modified. which means we'd need the offline dbsync command.18:15
xekvdrok, agreed, so this line would just be changes['version'] = self.VERSION18:16
rloovdrok: we wouldnt' have a pin; we're talking about no rolling upgrades code, just the version column.18:16
xekrloo, no, I will also add a data migration to the alembic script, to fill the columns with proper versions18:16
vdrokrloo: I mean that, current patches don't do just set the version, they do more. Some things should be dropped from there18:16
rloovdrok: right, we wouldn't want all that code.18:17
xekrloo, so that any scripts that are implemented later on can assume that there is always a version there18:17
jrollyeah, this seems not worth it to me18:17
jrollas much as I want it to be18:17
rlooxek: by data migration, you mean using the dbsync command, right?18:17
*** amotoki has quit IRC18:18
xekxek, no, the old alembic upgrade migration18:18
rlooxek: this is where i am confused then. that was what we were talking about i think, adding to the alembic migration to set version to the object's VERSION. why would we need the code you pointed out?18:19
xekrloo, to set the version properly for objects created after the upgrade18:19
rlooxek: right, in pike. we'd need that code before we change the version for any object, right?18:20
xekrloo, yes, so that when doing a rolling upgrade to pike, the ocata version, that can still be running, writes it's object versions when it is writing data18:21
*** sambetts is now known as sambetts|afk18:21
rlooxek: gotcha. so it is needed as part of the rolling upgrades code that will go in pike (or as i mentioned, before we change the version for any object)18:22
rlooxek: so maybe it isn't worth doing in ocata then. cuz i'm not sure i want to put in that code w/o the rest of the rolling upgrades code.18:23
xekrloo, yes, that is if we want to assume that the we can always count on the version being there18:23
rlooxek: exactly. so i think for now, if we punt to pike, we'll know exactly what state/value that new column will have. it might be safer to deal with then.18:23
TheJuliaI need to drop off because we're landing shortly.  I think we should not make that assumption for pike, but that is just me. :)18:23
*** causten has joined #openstack-ironic18:24
rlooI agree with TheJulia. so lets punt this feature to pike. Sorry xek.18:24
xekrloo, yeah, we can always get the patches as they are in pike, the upgrades from ocata will be possible just like upgrades from newton if we get it in in ocata18:24
*** ChubYann has joined #openstack-ironic18:25
rlooxek: i'll continue to review the patches, but probably not as frequently until we get ocata out.18:25
TheJuliaxek: alternatively, first thing we add is the version column, from there we cut a release, and version from there.18:25
TheJuliaor, we back-version the data18:25
*** causten has quit IRC18:25
rlooTheJulia: i think it is risky. we talk about getting rolling upgrades code in early pike, but it will still depend on CI.18:25
* TheJulia is thinking so Ocata -> Pike online upgrades would work18:25
TheJuliarloo: agreed18:26
TheJuliarloo: then again, we should just get the command in that does a noop first18:26
rlooTheJulia: i think we're good with Ocata -> Pike upgrade working, with xek's stuff. (just like we would have been with newton->ocata, and w/o the node-create issue)18:26
TheJuliaokay, going now since they are now going to yell at me for having a laptop out :)18:26
rloobye TheJulia!18:26
xekI also have to go, it's my girlfriends birthday today :P18:27
vdrokand /me leaves too :)18:27
joannaxek: wish her best!18:27
vdrokgood night everyone!18:27
rlooxek: enjoy! thx for staying late!18:27
joannanight, vdrok :)18:28
rloobye vdrok18:29
openstackgerritAndrey Shestakov proposed openstack/bifrost: Add socat package
*** mjura has quit IRC18:38
* jroll lunches18:39
*** jcoufal has joined #openstack-ironic18:39
*** jcoufal__ has joined #openstack-ironic18:42
*** jcoufal_ has quit IRC18:42
*** jcoufal_ has joined #openstack-ironic18:45
*** mgoddard has joined #openstack-ironic18:45
*** jcoufal has quit IRC18:45
*** e0ne has joined #openstack-ironic18:46
*** mrtenio has joined #openstack-ironic18:46
mrtenioHi, sorry, today morning I asked something and left. Is there a timeout parameter when setting the boot device? Or is it always 60 seconds?18:48
*** jcoufal__ has quit IRC18:48
*** jcoufal has joined #openstack-ironic18:58
openstackgerritSolio Sarabia proposed openstack/ironic-tempest-plugin: WIP/DNM: Migrate ironic-tempest-plugin to new repo
openstackgerritHironori Shiina proposed openstack/python-ironicclient: Fix node-inject-nmi to pass an empty body
*** jcoufal_ has quit IRC19:01
hshiinahi, a small fix is necessary for nmi in ironicclient19:03
hshiinaplease review :
patchbotpatch 424293 - python-ironicclient - Fix node-inject-nmi to pass an empty body19:04
*** jcoufal_ has joined #openstack-ironic19:07
*** trown|lunch is now known as trown19:09
*** jcoufal has quit IRC19:10
*** jcoufal__ has joined #openstack-ironic19:10
*** hshiina has quit IRC19:11
*** amarao has joined #openstack-ironic19:11
*** jcoufal_ has quit IRC19:12
aNuposicmrtenio: This might be something helpful
*** amoralej is now known as amoralej|off19:13
aNuposicmrtenio: not sure though if that's exactly what you are looking19:14
*** amarao has quit IRC19:16
*** amotoki has joined #openstack-ironic19:18
*** Sukhdev_ has joined #openstack-ironic19:21
*** hshiina has joined #openstack-ironic19:22
*** mgoddard has quit IRC19:29
*** strigazi is now known as strigazi_AFK19:30
*** jkilpatr has quit IRC19:34
*** causten has joined #openstack-ironic19:40
openstackgerritHironori Shiina proposed openstack/python-ironicclient: Fix node-inject-nmi to pass an empty body
*** jkilpatr has joined #openstack-ironic19:47
*** Nisha_Agarwal has quit IRC19:48
*** jaosorior has quit IRC19:50
*** bcornec1 has joined #openstack-ironic19:52
*** sayantani01 has quit IRC19:53
*** sayantani01 has joined #openstack-ironic19:57
*** causten has quit IRC19:58
*** sayantan_ has joined #openstack-ironic19:58
*** sayantani01 has quit IRC19:58
*** hshiina has quit IRC19:59
rlooJayF: a while ago, we discussed some doc, was it admin? and how we have some of that stuff in our tree and others in the openstack admin guide. or whatever guide it was. did you ever remove the content from the openstack <whatever> guide, and/or point it to our intree stuff?19:59
*** Sukhdev has quit IRC20:01
*** sayantan_ has quit IRC20:01
*** jcoufal__ is now known as jcoufal20:04
openstackgerritJim Rollenhagen proposed openstack/ironic: Allow using resource classes
openstackgerritJim Rollenhagen proposed openstack/ironic: DNM: turn on resource classes
*** dprince has quit IRC20:10
*** dyasny has quit IRC20:14
*** causten has joined #openstack-ironic20:28
*** aNuposic has quit IRC20:31
*** mat128_ has quit IRC20:36
csmart\o morning20:38
*** moshele has joined #openstack-ironic20:43
*** bcornec1 has quit IRC20:48
*** aNuposic has joined #openstack-ironic20:48
*** jcoufal_ has joined #openstack-ironic20:50
*** Sukhdev has joined #openstack-ironic20:51
*** cfarquhar has joined #openstack-ironic20:52
*** jcoufal has quit IRC20:53
*** jcoufal has joined #openstack-ironic20:53
*** causten has quit IRC20:56
*** jcoufal_ has quit IRC20:56
*** mjturek has quit IRC20:58
*** causten has joined #openstack-ironic21:00
*** hoangcx has joined #openstack-ironic21:01
* TheJulia re-appears having had lunch \o/21:04
mrteniocsmart, morning :)21:05
*** mjturek has joined #openstack-ironic21:11
openstackgerritMario Villaplana proposed openstack/ironic-python-agent: WIP: IPA rescue extension
openstackgerritJulia Kreger proposed openstack/ironic: Add base cinder common interface
*** jerrygb has quit IRC21:12
mariojvJayF: i have not tested this out yet, but that ^ has the bash, network config template, and cloud-config for coreos rescue w/ ipv4 dhcp tenant networks21:13
openstackgerritJulia Kreger proposed openstack/ironic: Add base cinder common interface
mariojvi suspect tinyipa will be easier, going to focus on reviews the rest of the day and try to get the remainder of that working tomorrow21:14
mariojvi think the agent driver patch will need a small change to make configdrive optional though21:14
*** rama_y has joined #openstack-ironic21:15
openstackgerritRamamani Yeleswarapu proposed openstack/ironic-inspector-specs: Collect Bios, Memory and Processor Info from Dmidecode
*** mrtenio is now known as mrtenio-afk21:20
*** ijw has joined #openstack-ironic21:23
*** mrtenio-afk has quit IRC21:24
openstackgerritMerged openstack/ironic: Add tests for Payloads with SCHEMAs
*** aNuposic has quit IRC21:28
jrollrloo: how terrible would it be to add an api version in one patch, and use it for a thing in the next patch? should I squash those instead?21:30
*** causten has quit IRC21:30
rloojroll: depends. do you mean it 'only' adds an api version?21:30
rloojroll: if it only does that, it makes no sense?21:31
*** jerrygb has joined #openstack-ironic21:31
jrollrloo: no, this api version does many things21:31
jrollrloo: so would you rather those many things be in one patch or a couple21:31
*** causten has joined #openstack-ironic21:31
rloojroll: oh. crap. in one patch. otherwise i think it messes things. ie, you'd need to bump the version with each patch.21:32
jrolloh I'm not even sure I can version this, hm21:32
*** jerrygb_ has joined #openstack-ironic21:32
jrollrloo: alright, I will squash these things21:33
rloojroll: thx. i prefer bigger patches myself but i am the odd person out in that regard i think.21:33
jrollrloo: huh, interesting21:33
*** jkilpatr has quit IRC21:34
rloojroll: well, within reason. but 'reason' is interpreted differently so...21:34
*** Goneri has quit IRC21:35
*** jerrygb has quit IRC21:36
*** srobert_ has joined #openstack-ironic21:36
*** srobert has quit IRC21:39
*** srobert_ has quit IRC21:41
openstackgerritRamamani Yeleswarapu proposed openstack/ironic: Enable OSProfiler support in Ironic
*** moshele has quit IRC21:47
*** jkilpatr has joined #openstack-ironic21:47
*** anush has quit IRC21:48
openstackgerritMario Villaplana proposed openstack/ironic: Agent driver rescue implementation
*** priteau has quit IRC21:55
openstackgerritMario Villaplana proposed openstack/ironic-python-agent: WIP: IPA rescue extension
*** causten has quit IRC21:56
*** causten has joined #openstack-ironic21:57
*** causten has quit IRC22:01
*** harlowja has quit IRC22:03
* jroll is fairly certain all vendor passthru endpoints are broken, lol22:03
*** ElCoyote_ has quit IRC22:03
jrolldtantsur|afk: ^^22:03
*** baoli has quit IRC22:04
*** e0ne has quit IRC22:04
*** hoangcx has quit IRC22:05
*** e0ne has joined #openstack-ironic22:06
*** e0ne has quit IRC22:06
*** e0ne has joined #openstack-ironic22:06
openstackgerritJim Rollenhagen proposed openstack/ironic: Add dynamic driver functionality to driver list/show APIs
openstackgerritJim Rollenhagen proposed openstack/ironic: WIP: vendor passthru for hw types
*** e0ne has quit IRC22:07
*** e0ne has joined #openstack-ironic22:07
*** dprince has joined #openstack-ironic22:07
*** e0ne has quit IRC22:08
*** e0ne has joined #openstack-ironic22:08
*** e0ne has quit IRC22:08
*** aNuposic has joined #openstack-ironic22:08
*** e0ne has joined #openstack-ironic22:09
*** e0ne has quit IRC22:09
*** jheroux has quit IRC22:12
*** dprince has quit IRC22:15
openstackgerritJulia Kreger proposed openstack/ironic: Add Cinder storage driver
*** trown is now known as trown|outtypewww22:22
*** mjturek has quit IRC22:24
* jroll out for the night, see y'all tomorrow22:26
*** mgoddard has joined #openstack-ironic22:26
*** rama_y has quit IRC22:27
jlvillalNot sure if people know. But there is a known issue with all grenade jobs. Waiting for fix to land. So likely our gate is broken too.22:33
jlvillalNeed this to land:
patchbotpatch 424323 - requirements (stable/newton) - Update packaging to 16.822:33
*** dyasny has joined #openstack-ironic22:36
rloojlvillal: thx for the heads up. that might explain why i couldn't get a newton ironic to work earlier today.22:40
jlvillalrloo: Yeah. I was trying to figure out why I kept dying locally, after fixing the new Python 3 pip3 issue in devstack for my local build. And then ran into that issue.22:42
jlvillalI feel like I've wasted most of my day because other stuff has broken me...22:42
rloojlvillal:  :-(. I took the easy way out and didn't look into it.22:42
jlvillalrloo: Smart! I wish I had that option. But since my work is make grenade work, I kind of had to look into it :)22:43
rloojlvillal: well, it is good that someone looks into the issue...22:43
jlvillalrloo: Thanks22:44
rloojlvillal: so thank you! :D22:44
rloojlvillal: i'm already almost screaming cuz i wanted to check someone's patch, so didn't want to get side-tracked...22:44
jlvillalI do see I have 210 unread threads in openstack-dev. maybe I should go read them...22:45
rloojlvillal:  :)22:46
*** mtanino has quit IRC22:49
*** mgoddard has quit IRC22:55
*** rama_y has joined #openstack-ironic23:01
*** dsneddon is now known as dsneddon_afk23:07
*** dsneddon has joined #openstack-ironic23:11
*** alexpilotti has joined #openstack-ironic23:19
jlvillalrloo: FYI: The fix for grenade has landed. So hopefully are grenade jobs are working again.23:23
*** jkilpatr has quit IRC23:29
rloojlvillal: thx. we can only hope :)23:33
*** harlowja has joined #openstack-ironic23:34
*** harlowja has quit IRC23:35
*** mtanino has joined #openstack-ironic23:40
*** Sukhdev has quit IRC23:41
openstackgerritJoanna Taryma proposed openstack/ironic: WIP host ipxe static assets in devstack on :80
*** awiddersheim has quit IRC23:45
*** awiddersheim has joined #openstack-ironic23:45
*** harlowja has joined #openstack-ironic23:48
openstackgerritOpenStack Proposal Bot proposed openstack/python-ironic-inspector-client: Updated from global requirements
openstackgerritOpenStack Proposal Bot proposed openstack/python-ironicclient: Updated from global requirements

Generated by 2.14.0 by Marius Gedminas - find it at!