Thursday, 2015-04-16

« Wednesday, 2015-04-15 Index Friday, 2015-04-17 »
*** kbs has joined #openstack-ironic00:05
*** achanda_ has joined #openstack-ironic00:11
*** achanda_ has quit IRC00:16
*** achanda has quit IRC00:16
*** linggao has joined #openstack-ironic00:21
*** linggao has quit IRC00:21
openstackgerritRuby Loo proposed openstack/ironic: Update some config option's help strings  https://review.openstack.org/17416400:28
*** r-daneel has quit IRC00:37
*** mdbooth has quit IRC00:46
*** mdbooth has joined #openstack-ironic00:52
openstackgerritJohn L. Villalovos proposed openstack/ironic: ironic/tests/drivers/ilo: Add spec= & autospec=True to mocks  https://review.openstack.org/17417401:01
*** kkoski has joined #openstack-ironic01:02
*** kkoski has quit IRC01:03
*** kkoski has joined #openstack-ironic01:06
*** kkoski has quit IRC01:07
*** davideagnello has quit IRC01:08
*** thrash is now known as thrash|g0ne01:09
openstackgerritZhai, Edwin proposed openstack/ironic: Take place of metrics with meters in option  https://review.openstack.org/17191501:30
*** jerryz has joined #openstack-ironic01:35
*** mtanino has quit IRC01:37
*** rloo has quit IRC01:38
*** chenglch has joined #openstack-ironic01:40
*** dttocs_ has quit IRC01:40
*** harlowja is now known as harlowja_away02:05
openstackgerritJohn L. Villalovos proposed openstack/ironic: ironic/tests/drivers/ilo: Add spec= & autospec=True to mocks  https://review.openstack.org/17417402:07
*** kbs has quit IRC02:35
*** chenglch|2 has joined #openstack-ironic02:58
*** chenglch has quit IRC02:59
*** jerryz has quit IRC03:00
openstackgerritHaomeng,Wang proposed openstack/python-ironicclient: validate the command empty arguments  https://review.openstack.org/17261803:13
*** achanda has joined #openstack-ironic03:33
*** chenglch|2 has quit IRC03:34
*** Marga_ has joined #openstack-ironic03:49
*** Marga_ has quit IRC03:49
*** Marga_ has joined #openstack-ironic03:50
*** dttocs has joined #openstack-ironic04:07
openstackgerritMerged openstack/ironic: document "scheduler_use_baremetal_filters" option in nova.conf  https://review.openstack.org/17414104:10
*** Guest16106 has quit IRC04:25
*** naohirot has joined #openstack-ironic04:27
*** saripurigopi has joined #openstack-ironic04:28
*** achanda has quit IRC04:30
*** achanda has joined #openstack-ironic05:00
*** Marga_ has quit IRC05:07
*** yog__ has joined #openstack-ironic05:23
openstackgerritMichael Davies proposed openstack/python-ironicclient: Cache negotiated api microversion for this server  https://review.openstack.org/17367405:26
*** bnemec has quit IRC06:18
*** jcoufal has joined #openstack-ironic06:18
*** jcoufal has quit IRC06:30
openstackgerritsandhya proposed openstack/ironic:  Implement Chassis Level Discovery Tool  https://review.openstack.org/15857706:35
*** ukalifon1 has joined #openstack-ironic06:39
*** sandhya has joined #openstack-ironic06:43
*** rameshg87 has joined #openstack-ironic06:48
*** viktors|afk is now known as viktors06:53
openstackgerritHaomeng,Wang proposed openstack/python-ironicclient: Ensure command line arguments are not empty  https://review.openstack.org/17261806:55
openstackgerritHaomeng,Wang proposed openstack/python-ironicclient: Ensure command line arguments are not empty  https://review.openstack.org/17261806:58
*** bnemec has joined #openstack-ironic07:00
*** achanda has quit IRC07:02
*** achanda has joined #openstack-ironic07:04
*** a1exhughe5 has joined #openstack-ironic07:10
*** athomas has joined #openstack-ironic07:19
*** dtantsur|afk is now known as dtantsur07:19
dtantsurMorning!07:19
pshigedtantsur: morning :)07:19
rameshg87dtantsur pshige o/07:21
dtantsuro/07:21
pshigerameshg87: o/07:21
*** achanda has quit IRC07:25
*** foexle has joined #openstack-ironic07:26
*** ifarkas has joined #openstack-ironic07:39
*** chlong has quit IRC07:40
openstackgerritYuiko Takada proposed stackforge/ironic-discoverd: Load authenticate token from HTTP header  https://review.openstack.org/17423107:41
*** jistr has joined #openstack-ironic07:43
*** rameshg87 is now known as rameshg87-lunch07:57
openstackgerritYuiko Takada proposed stackforge/ironic-discoverd: Use keystonemiddleware auth credentials  https://review.openstack.org/17078208:02
*** andreykurilin has joined #openstack-ironic08:05
openstackgerritYuiko Takada proposed stackforge/ironic-discoverd: Load authenticate token from HTTP header  https://review.openstack.org/17423108:07
*** lucasagomes has joined #openstack-ironic08:14
*** derekh has joined #openstack-ironic08:19
openstackgerritYuiko Takada proposed stackforge/ironic-discoverd: Load authenticate token from HTTP header  https://review.openstack.org/17423108:20
*** andreykurilin has quit IRC08:26
*** derekh is now known as derekh_afk08:30
*** ndipanov has joined #openstack-ironic08:33
*** jmankov has quit IRC08:39
*** jmank has joined #openstack-ironic08:40
*** pas-ha has joined #openstack-ironic08:42
*** agordeev has quit IRC08:56
openstackgerritNaohiro Tamura proposed openstack/ironic-specs: iRMC Virtual Media Deploy Driver for Ironic  https://review.openstack.org/13486508:58
*** edcragg has joined #openstack-ironic09:01
*** dtantsur is now known as dtantsur|brb09:06
openstackgerritRamakrishnan G proposed openstack/ironic-specs: Add generic RAID configuration spec for liberty  https://review.openstack.org/17321409:13
*** agordeev has joined #openstack-ironic09:17
*** kbs has joined #openstack-ironic09:17
kejiahumorning all09:25
pshigekejiahu: morning :)09:26
kejiahumorning pshige :)09:29
*** saripurigopi_ has joined #openstack-ironic09:31
*** athomas has quit IRC09:32
*** saripurigopi has quit IRC09:32
*** saripurigopi_ is now known as saripurigopi09:32
*** jmank has quit IRC09:39
*** saripurigopi has quit IRC09:43
*** athomas has joined #openstack-ironic09:43
*** athomas_ has joined #openstack-ironic09:48
*** athomas has quit IRC09:50
*** Marga_ has joined #openstack-ironic09:55
*** naohirot has quit IRC09:55
*** romcheg has joined #openstack-ironic09:55
*** pas-ha has quit IRC10:02
*** a1exhughe5 has quit IRC10:04
*** pas-ha has joined #openstack-ironic10:15
*** a1exhughe5 has joined #openstack-ironic10:18
*** pelix has quit IRC10:19
*** yuanying has quit IRC10:20
*** yuanying has joined #openstack-ironic10:21
*** pelix has joined #openstack-ironic10:22
*** yuanying has quit IRC10:32
openstackgerritMerged openstack/ironic: Fix heartbeat when clean step in progress  https://review.openstack.org/17411510:32
*** sandhya has quit IRC10:36
openstackgerritShivanand Tendulker proposed openstack/ironic-specs: UEFI secure boot support for pxe_ilo driver  https://review.openstack.org/17429510:47
openstackgerritShivanand Tendulker proposed openstack/ironic: grub2 bootloader support for uefi boot mode  https://review.openstack.org/16619210:48
openstackgerritShivanand Tendulker proposed openstack/ironic: Secure boot support for pxe_ilo driver  https://review.openstack.org/15480810:49
openstackgerritShivanand Tendulker proposed openstack/ironic: Secure boot support for pxe_ilo driver  https://review.openstack.org/15480810:52
*** derekh_afk is now known as derekh10:54
*** dtantsur|brb is now known as dtantsur10:56
*** Marga_ has quit IRC10:59
*** pas-ha has quit IRC11:18
*** pas-ha has joined #openstack-ironic11:22
*** rameshg87-lunch has left #openstack-ironic11:23
*** thrash|g0ne is now known as thrash11:38
*** saripurigopi has joined #openstack-ironic11:39
*** lucasagomes is now known as lucas-hungry11:45
*** jamielennox is now known as jamielennox|away11:47
*** saripurigopi has quit IRC11:49
*** trown|outttypeww is now known as trown12:02
*** Marga_ has joined #openstack-ironic12:06
*** saripurigopi has joined #openstack-ironic12:07
*** zhenguo has quit IRC12:17
*** saripurigopi has quit IRC12:18
*** dprince has joined #openstack-ironic12:20
openstackgerritVictor Sergeyev proposed openstack/ironic: Run tests in py34 environment  https://review.openstack.org/15619212:20
*** kbs1 has joined #openstack-ironic12:46
*** kbs has quit IRC12:47
openstackgerritDmitry Tantsur proposed openstack/python-ironicclient: Implement retries on Conflict  https://review.openstack.org/17435912:47
dtantsur^^^ \o/12:48
openstackgerritDmitry Tantsur proposed openstack/python-ironicclient: Implement retries on Conflict  https://review.openstack.org/17435912:49
*** lucas-hungry is now known as lucasagomes12:49
*** Marga_ has quit IRC12:51
*** Marga_ has joined #openstack-ironic12:58
*** zhidong has joined #openstack-ironic13:00
*** zyu4_ has joined #openstack-ironic13:00
dtantsurlucasagomes, ^^^ mind having a look?13:01
lucasagomesdtantsur, just did :-)13:02
lucasagomesleft some comments already13:02
dtantsurthanks13:04
dtantsurlucasagomes, what do you feel about adding configuration options in the follow-up patch?13:05
lucasagomesdtantsur, hmm wouldn't do any harm really... tho, it sounds something cool to have in that patch13:05
lucasagomesbut I'm ok with 1 or 2 patches to be honest13:06
dtantsurlucasagomes, ack will do13:06
dtantsur(in case w'll have to cherry-pick it)13:06
*** kkoski has joined #openstack-ironic13:07
BadCubmornin folks13:13
dtantsurBadCub, o/13:13
*** rameshg87 has joined #openstack-ironic13:14
*** Marga_ has quit IRC13:15
*** rloo has joined #openstack-ironic13:16
BadCubneed coffee13:19
lucasagomesBadCub, morning ! +1 coffee13:23
* BadCub returns13:29
BadCubheya dtantsur and lucasagomes13:29
*** trown is now known as trown|brb13:30
*** trown|brb is now known as trown13:34
*** afaranha_ has joined #openstack-ironic13:37
*** dttocs has quit IRC13:38
*** afaranha_ has quit IRC13:38
*** afaranha has joined #openstack-ironic13:39
*** MattMan has quit IRC13:42
*** stevemar has joined #openstack-ironic13:43
afaranhaHello, did anyone get this error recently: "set_provision_state() got an unexpected keyword argument 'configdrive'" ( http://paste.openstack.org/show/204100/ ) ? I tried to set in nova.conf the force_config_drive to False but I get the same error13:43
rloohi afaranha, there's an ordering issue. Does http://docs.openstack.org/developer/ironic/deploy/upgrade-guide.html help?13:46
NobodyCammorning Ironicers13:47
BadCubmornin NobodyCam13:47
NobodyCam:)13:47
*** MattMan has joined #openstack-ironic13:50
openstackgerritDmitry Tantsur proposed openstack/python-ironicclient: Implement retries on Conflict  https://review.openstack.org/17435913:51
dtantsurlucasagomes, ^^^13:51
dtantsurNobodyCam, rloo, morning13:51
afaranharloo, thanks, it worked13:51
NobodyCammorning dtantsur rloo BadCub lucasagomes13:51
lucasagomesrloo, afaranha NobodyCam morning13:51
lucasagomesdtantsur, will take a look very soon!13:52
*** Sukhdev has joined #openstack-ironic13:52
stevemardoes ironicclient (or something similar) have an openstackclient plugin?13:52
NobodyCammorning afaranha :)13:52
dtantsurstevemar, it's in progress, lemme find a link13:52
NobodyCamstevemar: there are patches up for that now13:52
BadCubmornin rloo13:52
rlooafaranha: phew :-)13:52
stevemarlinks please :)13:52
dtantsurstevemar, https://review.openstack.org/#/c/172963/13:53
stevemari was looking at python-ironicclient on github, but didn't see any osc related code13:53
afaranhamorning folks13:53
stevemaroh it's still in review13:53
rloohiya NobodyCam, BadCub, lucasagomes, dtantsur, afaranha, stevemar13:53
dtantsuryep13:53
stevemari think someone opened a bug for code in review? https://bugs.launchpad.net/python-openstackclient/+bug/144498313:54
openstackLaunchpad bug 1444983 in python-openstackclient "Usage is not described accurately for 99% of openstack CLI sub commands" [Undecided,New]13:54
stevemarthe example that is cited is baremetal related13:54
dtantsurthat's the result of people not distinguishing between downstream and upstream...13:55
stevemarI also don't think it's a legit bug, but i wanted to see if there was a live version of ironicclient running somewhere13:55
stevemarglad to see i'm not crazy :)13:55
dtantsurstevemar, sorry for confusing, we're already trying these patches at Red Hat internally, so that might be the cause13:55
stevemarAhhhh13:55
stevemaryeah, that would do it13:55
stevemarit's cool!13:55
dtantsurstevemar, if you're interested in more details, you can drop by #rdo or ping thrash13:56
stevemarjust had me confused this early in the morning13:56
dtantsur:)13:56
lucasagomeshelp baremetal power [--on | --off] <node> x.x13:56
lucasagomes--on OR --off13:56
lucasagomesdamn13:56
*** jgrimm is now known as zz_jgrimm13:56
* lucasagomes thinks more about a positional argument there13:56
dtantsur++13:56
dtantsurlucasagomes, as our folks are using it, I think we should review the whole patch chain. wdyt?13:57
stevemarmake it <state>13:57
lucasagomesdtantsur, +113:57
lucasagomeswill do13:57
stevemaras an option13:57
dtantsurdefinitely13:57
lucasagomesI was reviewing the client queue just now13:57
lucasagomeswill look at that patch series13:57
dtantsurlucasagomes, me too, after I test the retry thingy in real life13:58
stevemari reviewed a bunch of the earlier patches, haven't caught up to the newer ones yet o_O13:58
BadCubdtantsur: lucasagomes is that the patch series from Brad?13:58
stevemarit is13:58
dtantsuryep13:58
BadCubcool. I have them all listed on the L pad as well. :)13:59
dtantsurgood!13:59
BadCubhttps://etherpad.openstack.org/p/ironic-liberty-priorities under "Patches of Interest"14:00
openstackgerritZhenguo Niu proposed openstack/ironic: Rearrange some code in PXEDeploy.prepare  https://review.openstack.org/17392514:00
*** zhenguo has joined #openstack-ironic14:01
thrashstevemar: i've noticed it with others14:03
lucasagomesBadCub, cool!14:03
BadCublucasagomes: :)14:03
thrashdtantsur: lucasagomes if you have suggestiong on arguments, let me know. I'm trying to follow the OSC guidelines, but it's not always clear.14:03
lucasagomesthrash, sure :-) thanks14:04
thrashstevemar: for the power part... you suggest 'baremetal power <state>'14:05
thrash?14:05
lucasagomesBtw folks, I'd like to call your attention to http://lists.openstack.org/pipermail/openstack-dev/2015-April/061673.html14:05
*** zhidong has quit IRC14:05
thrashstevemar: however, it would be 'baremetal power <uuid> <state>'14:05
lucasagomesthis concerns me quite a bit14:05
*** zyu4_ has quit IRC14:06
*** Sukhdev has quit IRC14:06
NobodyCamlucasagomes: ugggh :(14:06
NobodyCamis there another package that is currentlly being maintained?14:07
lucasagomesnop :-/14:08
* lucasagomes was thinking about forking it14:08
stevemarthrash, i actually like --off | --on14:08
stevemari was just giving options14:08
stevemarthrash, but i didn't understand the bug =\14:08
lucasagomesNobodyCam, but I sent the email before, see what people think14:08
rloolucasagomes: doesn't doug hellmann have something to do with wsme?14:09
thrashstevemar: oh... ok..14:09
lucasagomesrloo, yes, he's core14:09
thrashstevemar: so, the bug... I can explain that.14:09
stevemarshoot14:09
rloolucasagomes: have you asked him about it?14:09
lucasagomesbut I believe he's pretty busy and not having time to look at the project14:09
thrashstevemar: try 'openstack help user list'14:09
lucasagomesrloo, yup, tried to ping him for the last week at least14:09
thrashnotice that '-h, --help' is listed there.14:09
* lucasagomes which there was evasdrop for the wsme channel14:09
thrashstevemar: and that one may be a bad example...14:09
thrash:P14:10
rloolucasagomes: oh. maybe he's too busy being excited, thrilled, i forgot the verbs, announcing new packages. (just joking.)14:10
lucasagomesheh14:10
thrashstevemar: well.. hmm.. let me see how she wrote it up.14:10
jlvillallucasagomes, The cores: https://review.openstack.org/#/admin/groups/202,members  which you probably already know :)14:10
lucasagomesjlvillal, yeah :-)14:10
rloolucasagomes: (they have some script that is used to generate those email announcements)14:11
openstackgerritMerged openstack/ironic: Update ilo drivers documentation for inspection  https://review.openstack.org/17006514:11
lucasagomesrloo, do they?14:11
lucasagomesI don't know14:11
thrashstevemar: well... this may not be a bug after all...14:11
NobodyCamlucasagomes: what about django14:11
thrashlet me talk to ola.14:11
jlvillallucasagomes, So will you make wsme_lag ;)  lag = lucas a gomes14:11
*** HenryG_ is now known as HenryG14:11
lucasagomesNobodyCam, wsme works on top of turbogears and flask too14:11
lucasagomesNobodyCam, but, I don't know it works well for us14:12
rloolucasagomes: yeah. harlowja_away showed it to me the other day.14:12
lucasagomesis just that we need some community around it to improve it and fix bugs from time to time14:12
* rloo votes for lucasagomes to become core in wsme.14:12
lucasagomesjlvillal, :D WSMJO, web service made just OK14:12
thrashstevemar: I think the misunderstanding comes from the help text of -h... 'show this help message and exit'14:12
lucasagomesrloo, I don't know much the code, but yeah I would try to pay some more attention to it14:13
thrashif that is shown when you do 'openstack help <command>', you would expect -h to give you the same output...14:13
*** saripurigopi has joined #openstack-ironic14:15
stevemarthrash, oh... 'openstack help user set' and 'openstack -h user set' should have the same14:15
stevemaroutput14:15
thrashstevemar: definnitely does not.14:15
thrashwhether it is 'openstack -h user set' or 'openstack user set -h'14:16
stevemarthrash, same for --help14:16
thrash-h gives the top-level help14:16
thrashstevemar: yep14:16
stevemaroookay, fun times, but i get it now. i'll re-word the bug14:16
*** yog__ has quit IRC14:17
thrashstevemar: thanks!14:20
stevemarthrash, better now? https://bugs.launchpad.net/python-openstackclient/+bug/144498314:21
openstackLaunchpad bug 1444983 in python-openstackclient "-h and --help are not recognized and revert to top level help" [Undecided,New]14:21
thrashMuch better. :)14:21
*** mtanino has joined #openstack-ironic14:27
rloohi lucasagomes, wrt https://review.openstack.org/#/c/172618/, i see you commented on it. i don't think we want to ensure all arguments are empty. Or do we?14:34
lucasagomesrloo, not sure to be honest14:34
lucasagomesmaybe we should jsut make sure that "" is seem as non passed argument14:35
lucasagomeswhere " " actually has some value14:35
rloolucasagomes: see my example, "ironic chassis-create -d ''"14:35
lucasagomesoh14:35
rloolucasagomes: I don't see why we should disallow that.14:35
lucasagomeshaving an empty description14:35
lucasagomesindeed14:35
lucasagomessounds like a valid case14:36
lucasagomesmaybe that bug is invalid them14:36
lucasagomesthen*14:36
rloolucasagomes: ok thx, wanted to get confirmation. the bug is valid, the solution isn't.14:36
lucasagomesyeah, depends on the argument14:36
lucasagomesbut it makes pretty complicate to filter by argument :/14:36
rloolucasagomes: the bug is "node-show ''". The code ends up doing a node-list cuz the node-ident isn't specified.14:36
lucasagomesyeah14:37
lucasagomesrloo, one way to solve that would be to pass the node-ident in the body message instead of the URI14:37
lucasagomesGET {'node_ident': ''} /v1/nodes14:38
lucasagomesinstead of GET /v1/nodes/<node ident>14:38
lucasagomeswhich empty will return node list14:38
* lucasagomes thinks, but need tests14:38
rloolucasagomes: oh, didn't realize/remember that you could do that. i'd leave it for Haomeng|2 to fix ;)14:38
lucasagomesor just check in the show command if node_ident is being passed with some value :-)14:38
rloolucasagomes: i was thinking the latter would be easier :-)14:39
lucasagomes+114:39
rloolucasagomes: ok thx. on to the next patch to review :-)14:39
lucasagomescool thank u14:39
*** edcragg has quit IRC14:40
*** a1exhughe5 has quit IRC14:42
*** tiagogomes has quit IRC14:42
rameshg87JoshNang: are you around ?14:45
*** mgoddard1 has joined #openstack-ironic14:46
JoshNangrameshg87: o/14:46
rameshg87JoshNang: hello14:46
rameshg87JoshNang: sorry to have just pinged and leaving14:46
rameshg87JoshNang: wanted to discuss about https://review.openstack.org/#/c/168698/14:46
openstackgerritRuby Loo proposed openstack/ironic: Replace 'metrics' with 'meters' in option  https://review.openstack.org/17191514:46
rameshg87i will be back in a short while, can we discuss on it ?14:47
JoshNangyup! i'm just starting my day, so i'll be around for a while14:47
rameshg87okay14:47
*** rameshg87 is now known as rameshg87-dinner14:47
*** mgoddard has quit IRC14:47
jrollmorning all :)14:50
jrolllucasagomes: maybe we should just ditch wsme and use pecan directly14:50
jrollit's really wsmaPITA14:50
*** subscope has quit IRC14:50
*** zz_jgrimm is now known as jgrimm14:51
lucasagomesjroll, yeah... I like the custom types of wsme tho14:51
lucasagomesI think it's handy14:51
lucasagomesbut yeah the idea of only pecan is not bad too :-)14:51
lucasagomesor flask! since it at elast has some community around14:51
dtantsurjroll, morning14:51
dtantsurlucasagomes, flask is pretty call, I'm using it for discoverd14:51
jrolllucasagomes: I mean, the custom types are really just a hack on top of all of it14:52
dtantsur* call = cool14:52
jrollhey dtantsur :)14:52
jrollI'm fine with flask, fine with falcon, fine with pecan14:52
jrollafaik pecan is the "default" for openstack thing14:52
lucasagomesyeah14:52
jrolls14:52
lucasagomeswell... it was suppose to be pecan + wsme14:52
lucasagomesbut I doubt it's near it14:52
jrollI seem to remember TC trying to make barbican switch from falcon->pecan during incubation14:52
jrollwhich is really silly IMO14:52
lucasagomesreally? ouch14:52
lucasagomestotally14:52
lucasagomesyeah so there are alternatives, plenty we need to try to make the best decision here14:54
lucasagomessince our projects depends on wsme now14:54
*** dttocs has joined #openstack-ironic14:55
lucasagomesironic & ipa14:55
pshigejroll: morning :)14:56
jrollheya pshige :)14:56
jrolllucasagomes: I switched ipa to pecan/wsme to be common :(14:57
jrolllucasagomes: used to be https://github.com/rackerlabs/teeth-rest14:57
lucasagomesjroll, right, and that's good actually14:57
NobodyCammorning pshige jroll JoshNang14:57
lucasagomeswe can share code etc, yet more with the ironic-libs project now14:57
pshigeNobodyCam: morning :)14:57
JoshNanggood morning!14:58
jrolllucasagomes: meh, IPA doesn't need a large rest framework14:58
NobodyCam:)14:58
lucasagomesjroll, heh true14:58
jrolltranslate a python dict to a json response14:59
*** Marga_ has joined #openstack-ironic14:59
* NobodyCam git clone's a fresh cup of coffee..brb15:00
*** a1exhughe5 has joined #openstack-ironic15:01
jrolllucasagomes: looks like mordred has a similar opinion15:01
mordredhonestly - I have NO skin in this game15:02
mordredbut of flask, falcon and pecan15:02
mordredflask is the one that's a 'real' thing outside of openstack15:02
lucasagomesyeah15:02
mordredand falcon and pecan are basically both only used by openstack15:02
mordredso I'd probably suggest we use flask if we're going to move to something15:03
lucasagomesit has some community around it and that's a good sight15:03
mordredyup15:03
mordredFlask is already in global-requirements too15:04
jroll+115:05
lucasagomesnice15:06
*** zhenguo has quit IRC15:06
*** ukalifon1 has quit IRC15:07
pshigeJoshNang: morning :)15:09
JoshNango/15:09
dtantsurso I'd like to ask here out loud: any objections to our client retrying 409 by default?15:09
NobodyCamhttp://flask-restful.readthedocs.org/15:09
dtantsurFWIW I believe it's a sane thing to do15:09
dtantsurJoshNang, o/15:10
NobodyCamdtantsur: +1 from /me15:10
*** a1exhughe5 has quit IRC15:10
lucasagomesdtantsur, +1 (as stated in my review)15:10
*** tiagogomes_ has joined #openstack-ironic15:11
jrolldtantsur: it seems fine, everywhere we've used the client we write logic for that anyway15:11
pshigeno objection15:11
*** edcragg has joined #openstack-ironic15:13
NobodyCam509?15:14
jroll509?15:15
dtantsur409 Conflict15:15
NobodyCamlineline 235 of shell.py15:15
dtantsuroh good catch :D15:16
NobodyCam:D15:16
*** jcoufal has joined #openstack-ironic15:17
*** thrash is now known as thrash|f00dz15:18
*** saripurigopi has quit IRC15:18
jrollI don't get it :/15:18
NobodyCamtypo15:19
jrollwhat's line 235 have to do with it :P15:19
NobodyCamhttps://review.openstack.org/#/c/174359/3/ironicclient/shell.py15:20
dtantsurlol15:20
NobodyCam:)15:21
jrolloh in the change15:21
jrollheh15:21
NobodyCamhehehee15:21
*** saripurigopi has joined #openstack-ironic15:22
*** mgoddard1 has quit IRC15:24
*** alexpilotti has joined #openstack-ironic15:24
*** mgoddard has joined #openstack-ironic15:27
*** a1exhughe5 has joined #openstack-ironic15:31
openstackgerritDmitry Tantsur proposed openstack/python-ironicclient: Implement and enable retries on Conflict  https://review.openstack.org/17435915:35
dtantsurlucasagomes, NobodyCam ^^^15:35
openstackgerritDmitry Tantsur proposed openstack/python-ironicclient: Implement and enable retries on Conflict  https://review.openstack.org/17435915:39
*** Marga_ has quit IRC15:39
*** jcoufal has quit IRC15:43
dtantsursee you tomorrow15:46
*** dtantsur is now known as dtantsur|afk15:46
NobodyCamwow I thought there was a bug for that15:46
NobodyCamhave a good night dtantsur|afk15:47
*** kozhukalov has joined #openstack-ironic15:54
*** ifarkas has quit IRC15:55
openstackgerritLucas Alvares Gomes proposed openstack/ironic: Troubleshoot: Do not power off node if deployment fails  https://review.openstack.org/17293215:55
*** a1exhughe5 has quit IRC15:58
*** Sukhdev has joined #openstack-ironic16:02
openstackgerritJohn L. Villalovos proposed openstack/ironic: Create a 3rd party mock specs file  https://review.openstack.org/17444416:04
NobodyCammorning jlvillal :)16:05
jlvillalNobodyCam, Good morning :)16:05
NobodyCam:)16:05
jlvillalNobodyCam, I'm just mocking the world today.  Or at least mocking Ironic ;)16:05
NobodyCam:) Ironiclly Mocking you say ... lol16:06
jlvillal:D16:06
*** Marga_ has joined #openstack-ironic16:07
*** jistr has quit IRC16:08
*** athomas_ has quit IRC16:14
*** Marga_ has quit IRC16:16
openstackgerritJohn L. Villalovos proposed openstack/ironic: ironic/tests/drivers/drac: Add spec= & autospec=True  https://review.openstack.org/17414516:17
*** krtaylor has quit IRC16:28
*** saripurigopi has quit IRC16:29
openstackgerritJohn L. Villalovos proposed openstack/ironic: ironic/tests/drivers/amt: Add autospec=True to mocks  https://review.openstack.org/17411316:31
*** Marga_ has joined #openstack-ironic16:32
NobodyCamdo we translate client log messages?16:33
openstackgerritJohn L. Villalovos proposed openstack/ironic: ironic/tests/drivers/ilo: Add spec= & autospec=True to mocks  https://review.openstack.org/17417416:35
openstackgerritJohn L. Villalovos proposed openstack/ironic: ironic/tests/drivers/ilo: Add spec= & autospec=True to mocks  https://review.openstack.org/17417416:36
*** derekh has quit IRC16:39
*** EmilienM is now known as EmilienM|afk16:39
*** davideagnello has joined #openstack-ironic16:44
*** andreykurilin__ has joined #openstack-ironic16:45
*** pelix has quit IRC16:45
*** rameshg87-dinner is now known as rameshg8716:45
rameshg87JoshNang: hello16:45
JoshNangrameshg87: o/16:45
rameshg87JoshNang: okay to discuss now ?16:46
rameshg87JoshNang: if so, wdyt about my reply comment on the review ?16:46
JoshNangone sec, let me bring it up again16:48
rameshg87yeah okay16:48
*** thrash|f00dz is now known as thrash16:49
JoshNangnot sure i understand your comment. it is a separate interface, and making it a 'helper for the deploy driver' seems like we're coupling it too tightly16:50
NobodyCamgah why did I add -r to me tox command16:50
NobodyCambrb16:51
rameshg87yeah JoshNang but the problem statement itself was to decouple the boot logic from the deploy drivers16:51
rameshg87idk, but logically the cleaning related stuffs doesn't seem to depend on the mechanism in which bare metal is brought up16:52
rameshg87i thought we were just concerned on different mechanisms in bring up bare metal nodes to either "do some operation" or "boot user image"16:52
jrollso really the boot interface is there to allow other things to say "I want to boot $image via {disk,pxe} on $network"16:53
jroll(imho)16:53
*** krtaylor has joined #openstack-ironic16:53
rameshg87jroll: yes16:53
rameshg87in my opinion too..16:53
jrollso deploy.prepare_cleaning() should call boot_image($ipa_ramdisk, 'pxe', $cleaning_network)16:54
jrollor something16:54
JoshNang++16:54
jrollthat's how I've always imagined it16:54
JayFjroll: that'd make a pxe-pxe driver much easier, wouldn't it?16:55
rameshg87jroll: but why does it have to provide $cleaning_network16:55
jrollJayF: yes16:55
jrollrameshg87: to pxe boot, you need to create ports in neutron, which you need a network for16:55
rameshg87the cleaning_network thing was specific to pxe driver16:55
jrollalternatively could pass neutron ports instead, idk16:56
rameshg87some other boot mechanism might not need that16:56
*** harlowja_away is now known as harlowja16:58
*** andreykurilin__ has quit IRC16:58
rameshg87oh i missed to address.  jroll, ^^^16:58
harlowjarloo what did u show u :-P16:59
*** mtanino has quit IRC16:59
jrollrameshg87: right, maybe it's optional or something16:59
rlooharlowja: did u show u something? that oslo script that generates the release announcement emails with all those exciting verbs in it :)17:00
harlowjaoh17:00
harlowjahaha17:00
harlowjaya17:00
rameshg87yeah JoshNang, but why do you feel boot interface should participate in cleaning ?17:00
JayFrameshg87: cleaning requires us to boot a thing, so cleaning would be a consumer of the interface17:00
JayFI think that's what he was getting at?17:00
JoshNangrameshg87: so my concern is that we've got a bunch of drivers duplicating code to deal with ipa, and i'd like to do that less17:01
NobodyCamanyone else seeing this, or did I break something? http://paste.openstack.org/show/NUc3zh62ecrHJ9uf97cB/17:01
rameshg87JayF: yeah of course17:01
JoshNang^ that's part of it and will definitely make cleaning better17:01
*** Marga__ has joined #openstack-ironic17:02
*** Marga_ has quit IRC17:02
*** Sukhdev has quit IRC17:02
jrollJoshNang: I'm also curious what the -1 is about specifically (though I'll admit I haven't read the whole thing yet)17:02
rameshg87JoshNang: yeah i hope you mean the same things that we did for cleaning in agent_ilo and agent_ipmitool17:03
rameshg87JoshNang: that should get addressed as part of this17:03
JoshNangrameshg87: right17:03
JoshNangwhich, \o/17:03
rameshg87jroll: JoshNang: JayF: what i meant in comment on L157 of https://review.openstack.org/#/c/168698/2/specs/liberty/new-boot-interface.rst17:05
rameshg87is that 'boot' will not be here: https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L192-L20017:05
jrollah, I see17:06
rameshg87but (a refactored version) of https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent.py#L391-L414 would use boot interface to boot the ramdisk17:06
jrollI think I agree17:06
JoshNangyeah i think this makes sense, i'll remove my -117:07
rameshg87and https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/ilo/deploy.py#L587-L601 would be merged to the above17:07
JoshNangand look elsewhere for deduping the IPA code all over17:07
rameshg87i mean in the end there would be just one piece of code17:08
rameshg87okay17:08
rameshg87JoshNang: jroll: JayF: please have a look at the spec and see if that proposal is missing something still and comment on review when you get some time17:08
rameshg87i think i would goto sleep then :)17:09
jrollwill do :)17:09
rameshg87have a great day folks17:09
NobodyCamhave a good night rameshg8717:09
rameshg87good night ironic :)17:09
rloonight rameshg8717:09
BadCubg'night rameshg8717:09
*** rameshg87 has quit IRC17:09
NobodyCammorning rloo :)17:09
NobodyCam(if I didn;t already say as such)17:09
rlooNobodyCam: thx. Didn't I say hi already? or was that yesterday, or the day before or ... ;)17:10
NobodyCamlol ++17:10
*** pas-ha has quit IRC17:11
slaglethrash: fyi, filed a bug too: https://bugs.launchpad.net/python-ironicclient/+bug/144512817:13
openstackLaunchpad bug 1445128 in python-ironicclient "osc plugin doesn't define API_VERSIONS" [Undecided,New]17:13
slaglei couldnt find you on lp btw17:13
thrashslagle: that code hasn't been merged upstream yet.17:14
slagleoh, so my bug was premature :)17:14
thrasha bit, yes. :)17:14
slaglei'm just being...hopeful17:15
slaglei'll close it, thx17:15
lucasagomesfolks I will call it a day!17:15
lucasagomeshave a good night everyone17:15
*** lucasagomes is now known as lucas-dinner17:15
NobodyCamhave a great night lucas-dinner17:15
*** achanda has joined #openstack-ironic17:18
*** foexle has quit IRC17:20
NobodyCamummm waffle ...17:23
*** Nisha has joined #openstack-ironic17:29
*** Marga__ has quit IRC17:30
NobodyCambrb17:35
*** kkoski has quit IRC17:35
*** kkoski1 has joined #openstack-ironic17:35
*** Marga_ has joined #openstack-ironic17:38
*** EmilienM|afk is now known as EmilienM17:43
*** alexpilotti has quit IRC17:48
*** vipul has quit IRC17:50
*** edcragg has quit IRC17:51
*** EmilienM is now known as EmilienM|afk17:55
openstackgerritMerged openstack/ironic: Replace 'metrics' with 'meters' in option  https://review.openstack.org/17191517:57
*** trown is now known as trown|lunch17:58
*** vipul has joined #openstack-ironic17:59
*** Marga_ has quit IRC18:02
*** Nisha has quit IRC18:02
*** ijw has joined #openstack-ironic18:06
*** Marga_ has joined #openstack-ironic18:06
openstackgerritJohn L. Villalovos proposed openstack/ironic: ironic/tests/drivers/irmc: Add spec_set & autospec=True  https://review.openstack.org/17451418:06
openstackgerritRuby Loo proposed openstack/ironic: Remove ironic compute driver and sched manager  https://review.openstack.org/17451918:08
*** davideagnello has quit IRC18:09
*** davideagnello has joined #openstack-ironic18:09
openstackgerritDoug Hellmann proposed openstack/python-ironicclient: Uncap library requirements for liberty  https://review.openstack.org/17453318:13
NobodyCambrb18:15
openstackgerritBrad P. Crochet proposed openstack/python-ironicclient: Conform to new OSC plugin requirements  https://review.openstack.org/17455118:15
*** andreykurilin__ has joined #openstack-ironic18:17
openstackgerritRuby Loo proposed openstack/ironic: Remove ironic compute driver and sched manager  https://review.openstack.org/17451918:17
*** achanda has quit IRC18:20
*** Sukhdev has joined #openstack-ironic18:24
*** achanda has joined #openstack-ironic18:28
rlooadam_g, devananda: we still have migrate_nova stuff (added by https://review.openstack.org/#/c/112402/). We should delete it?18:29
*** EmilienM|afk is now known as EmilienM18:33
*** dprince has quit IRC18:35
*** Marga_ has quit IRC18:49
*** Marga_ has joined #openstack-ironic18:49
*** achanda has quit IRC18:52
*** Marga_ has quit IRC18:54
*** Marga_ has joined #openstack-ironic18:54
*** trown|lunch is now known as trown18:54
*** ijw has quit IRC19:01
*** achanda has joined #openstack-ironic19:12
*** dttocs has quit IRC19:15
jrollNobodyCam: http://imgur.com/gallery/QD5735j19:15
BadCubLOL19:16
*** kozhukalov has quit IRC19:20
NobodyCamjroll: lol19:26
NobodyCami'm sure he'll eat thru the door on the clip19:26
jrolloh ya19:27
rlooNobodyCam: I want to add something to the meeting agenda for next monday. The agenda still has stuff from last week's meeting though. Should i just add it in?19:33
NobodyCamrloo: please delete the old stuff..19:34
NobodyCamI have not gotten to it yet19:35
rlooNobodyCam: ok.19:35
NobodyCamhttp://scooterbbs.net/board/DCForumID3/8858.html :(19:35
rlooNobodyCam: oh no.19:35
NobodyCam:(19:35
BadCub:(19:35
*** Marga_ has quit IRC19:35
NobodyCamrloo: or rut-row19:36
rlooNobodyCam: whatever you bow-wow19:36
NobodyCamoh what I typo'd that: http://www.internetslang.com/RUH_20ROH-meaning-definition.asp19:37
rlooNobodyCam: that explains why I didn't understand it ;)19:39
NobodyCamlol :)19:39
jrollhey, this is something we should be aware of for installations not using glance: https://github.com/openstack/glance-specs/blob/master/specs/kilo/deactivate-image.rst19:40
jrollglance images with a status of "deactivated" won't return any image data19:40
jrollI haven't verified that it's a bug, but it's probably a bug19:41
*** lucas-dinner has quit IRC19:41
*** andreykurilin__ has quit IRC19:45
openstackgerritChris Krelle proposed openstack/ironic: Check temp dir is usable for ipmitool driver  https://review.openstack.org/16038319:46
-openstackstatus- NOTICE: gerrit has been restarted to clear a problem with its event stream. any gerrit changes updated or approved between 19:14 and 19:46 utc will need to be rechecked or have their approval reapplied for zuul to pick them up19:47
openstackgerritChris Krelle proposed openstack/ironic: Add check for usable tftp directory to PXE driver  https://review.openstack.org/16745119:49
NobodyCamgah I've ended up with a bad dependicy in my commit... (see line #10 of http://paste.openstack.org/show/O1gLrUQ5rf2zVgV9BhPF ) ... any one got a slick trick to remove it with out a hard reset?19:53
jrollNobodyCam: git rebase -i master, remove the commit you don't want in there19:53
jrollassuming you just don't want that commit in that chain at all19:54
*** lucas-dinner has joined #openstack-ironic19:54
jroll"remove the commit" meaning delete that line in the editor when selecting commits for the rebase19:54
openstackgerritChris Krelle proposed openstack/ironic: Add a check to pxe driver for TFTP service  https://review.openstack.org/16991819:55
NobodyCamjroll: awesome TY :)19:55
rlooNobodyCam: I updated the agenda and added a Next Meeting section so I can easily find the date for the next meeting. Not sure if that works for others or not.19:55
NobodyCamrloo: :)19:56
NobodyCamwill look in a minute19:56
jrollnp :)19:56
rlooNobodyCam: thx, feel free to undo. I think ideally, I'd like 'Agenda for <date/time> meeting' but not sure if that makes it harder to update.19:57
NobodyCamrloo: looks good to me20:03
NobodyCam:) Thank you20:03
NobodyCam:) woo hoo my glasses are ready for pick up. and :( boo hoo I need glasses...lol20:06
*** Marga_ has joined #openstack-ironic20:06
*** Marga_ has quit IRC20:11
rlooNobodyCam: if this is the first time in your life that you're wearing glasses, that isn't too bad.20:12
NobodyCam:) it is!20:13
NobodyCamand I only have to ware them for work!20:13
NobodyCamso my vanity will remain intact20:13
NobodyCam:-p20:13
rlooNobodyCam: wow, so if you didn't have to work... ;)20:13
NobodyCamlol20:14
*** andreykurilin__ has joined #openstack-ironic20:16
*** dttocs has joined #openstack-ironic20:19
*** Sukhdev has quit IRC20:24
*** dprince has joined #openstack-ironic20:24
* BadCub has to get eyes checked today too *pouts*20:25
*** romcheg has quit IRC20:25
openstackgerritRuby Loo proposed openstack/ironic: Remove scripts for migrating nova baremetal  https://review.openstack.org/17459320:29
rlooadam_g, devananda: wrt my migrate_nova question earlier, I decided to delete them. feel free to -2 it if I shouldn't have ^^20:33
*** romcheg has joined #openstack-ironic20:34
*** achanda has quit IRC20:35
*** achanda has joined #openstack-ironic20:40
TheJuliaBadCub: I did that last week :(20:41
NobodyCamTheJulia: :(20:44
BadCubTheJulia: so much fun! :p20:44
BadCubI know I am going to need new script *frowns*20:45
* BadCub has to run out for a few.20:46
TheJuliaEnjoy!20:50
NobodyCam:/20:50
*** dprince has quit IRC20:54
openstackgerritJarrod Johnson proposed stackforge/pyghmi: Implement FRU inventory (WIP)  https://review.openstack.org/17460120:55
openstackgerritBrad P. Crochet proposed openstack/python-ironicclient: Implementation of baremetal reboot command  https://review.openstack.org/17251720:56
openstackgerritBrad P. Crochet proposed openstack/python-ironicclient: Add support for 'baremetal show' command  https://review.openstack.org/17173920:56
openstackgerritBrad P. Crochet proposed openstack/python-ironicclient: Conform to new OSC plugin requirements  https://review.openstack.org/17455120:56
openstackgerritBrad P. Crochet proposed openstack/python-ironicclient: Implementation of baremetal power on and off  https://review.openstack.org/17294620:56
openstackgerritBrad P. Crochet proposed openstack/python-ironicclient: Implementation of set and unset  https://review.openstack.org/17296320:56
openstackgerritBrad P. Crochet proposed openstack/python-ironicclient: Introduce openstackclient plugin  https://review.openstack.org/17167220:56
openstackgerritBrad P. Crochet proposed openstack/python-ironicclient: Implementation of baremetal create  https://review.openstack.org/17217120:56
openstackgerritBrad P. Crochet proposed openstack/python-ironicclient: Add support for microversion  https://review.openstack.org/17217020:56
openstackgerritBrad P. Crochet proposed openstack/python-ironicclient: Implementation of baremetal port create  https://review.openstack.org/17246120:56
openstackgerritBrad P. Crochet proposed openstack/python-ironicclient: Implement the rest of the args for baremetal list  https://review.openstack.org/17169920:56
openstackgerritBrad P. Crochet proposed openstack/python-ironicclient: Implement baremetal delete command  https://review.openstack.org/17175120:56
NobodyCamwow20:56
NobodyCam:)20:56
thrashNobodyCam: rebase... :)20:56
NobodyCam:)20:57
NobodyCamlol: /me luvs section #4 https://tools.ietf.org/html/rfc751420:58
*** andreykurilin__ has quit IRC21:04
*** Marga_ has joined #openstack-ironic21:07
*** harlowja is now known as harlowja_away21:11
*** Marga_ has quit IRC21:12
*** trown is now known as trown|outttypeww21:13
*** EmilienM is now known as EmilienM|afk21:13
openstackgerritJames Slagle proposed stackforge/ironic-discoverd: Define API_VERSIONS in shell.py  https://review.openstack.org/17457521:18
*** Marga_ has joined #openstack-ironic21:18
*** ijw has joined #openstack-ironic21:22
*** Marga_ has quit IRC21:23
*** Marga_ has joined #openstack-ironic21:24
*** romcheg has quit IRC21:30
*** Marga_ has quit IRC21:34
*** ijw has quit IRC21:34
*** Marga_ has joined #openstack-ironic21:35
*** kkoski1 has quit IRC21:36
*** lucas-dinner has quit IRC21:38
*** jamielennox|away is now known as jamielennox21:39
*** stevemar has quit IRC21:39
NobodyCamlol I think one of my pre-flight check found a config error in the xcat third party tests: http://xcat.sourceforge.net/OpenStack/CI/169918/6/check/check-ironic-xcat-third-party/2382/console.html21:47
jrollNobodyCam: left a review on that :P21:49
NobodyCamjroll: :) TY21:49
jrollI was going to say xcat might need to set that option :P21:49
mgagneI would like to unit test that the conductor is doing proper driver validation on my driver. I know you can unit test a driver by calling validate() directly but I would like to also make sure that from the POV of the conductor, validate() is properly called on my driver. Which test should I look at to get an example?21:50
mgagneI think I'm looking for validate_driver_interfaces() in conductor.manager21:52
jrollmgagne: I think you want https://github.com/openstack/ironic/blob/master/ironic/tests/conductor/test_manager.py#L199821:52
jrolland I think in general we rely on conductor tests21:52
jrollyou shouldn't need to worry about this as a driver author21:53
openstackgerritJosh Gachnang proposed openstack/ironic: Add cleaning network docs  https://review.openstack.org/17462021:53
NobodyCamjroll: oh yes... good catch!! TY!21:53
jroll:)21:53
jrollNobodyCam: though that might be related to agent driver only, not sure21:54
mgagnesure but lets say I'm an ass and forget to inherit from VendorInterface or whatever interface with validate(). conductor will never call validate() and I'll only find out in integration/prod21:54
mgagnejroll: ^21:54
jrollah21:54
* jroll looks21:54
rloo? all interfaces have a validate()21:55
mgagnebecause that's what happened in code I inherited from =)21:55
NobodyCamjroll: oh it is only defined in agent section: https://github.com/openstack/ironic/blob/master/etc/ironic/ironic.conf.sample#L36721:55
*** EmilienM|afk is now known as EmilienM21:55
jrollmgagne: maybe something like https://github.com/openstack/ironic/blob/master/ironic/tests/drivers/test_pxe.py#L49821:55
mgagneobject doesn't have validate() =)21:55
jrollmgagne: so set the node to have your driver and do something like that21:56
jrollso it's going through the task manager the same way the conductor does it21:56
mgagnejroll: alright, I'll check the links you sent me21:56
jroll:)21:56
mgagnejroll: because I did like test_validate_good did21:56
mgagnejroll: and removed inheritance from VendorInterface, tests are still passing21:57
jrolloh?21:57
mgagnejroll: and it's bad mmmkkk21:57
jrollare you asserting specific things your vendor validate method does?21:57
mgagnefor sure21:57
mgagneI'm validating that an exception is raised from example21:57
mgagnefor*21:58
jrollthat's odd21:58
mgagnecome to think of it, I'm sure you don't HAVE to inherit from VendorInterface21:58
mgagneas long as the methods are there, I guess conductor will be happy21:59
jrolloh right, it'll still call the method21:59
jrollyou just can't call the superclass if the inheritance is wrong21:59
mgagneyep21:59
mgagneI think I got my answer21:59
jroll:D21:59
mgagnenow lets say I'm super clever and introduce a new interface to BaseDriver22:00
NobodyCamI'd ask for a spec22:00
mgagnein my own super driver. how can I make sure conductor will check my new super interface validity? :D22:00
mgagneI guess I'll gave to write my own version of https://github.com/openstack/ironic/blob/master/ironic/tests/conductor/test_manager.py#L199822:01
jrollmgagne: write tests for the conductor method that calls it :)22:01
jrollright22:01
NobodyCam:)22:01
mgagnejroll: super, I think I got all I needed to know, thanks!22:01
devanandamgagne: I don't, offhand, think we're asserting that the interface is a subclass of the base class -- but we could.22:01
NobodyCamhey morning devananda22:02
devanandalike the conductor might test that each interface on the node's driver is an instance of the right type of interface22:02
NobodyCam:-p22:02
devanandag'morning ;)22:02
mgagnedevananda: the driver I'm developing is in his own repository, I'm not hacking Ironic directly22:02
mgagnedevananda: so I would still have to have a test adding coverage for the conductor part22:03
NobodyCammgagne: can I ask what the driver for?22:03
mgagneNobodyCam: our internal driver =)22:04
NobodyCam:)22:04
mgagnehooking to existing systems22:04
mgagnejust wish to make sure my new interface gets picked up by this validation: https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L1231-L123722:04
mgagneI could put everything in vendor instead but why make it simple when it can be complex :D22:06
jrollI mean, if you're adding new base interfaces, you're basically forking ironic :/22:07
NobodyCamthats really what we had in mind with Vendor passthru22:07
devanandamgagne: eh? wait - are you adding a new driver that implements the existing interface contracts, or are you adding a NEW interface contract?22:07
devanandamgagne: it sounds like the latter, and I'd really like to understand why you think that's required22:08
*** thrash is now known as thrash|g0ne22:10
BadCubwell folks, I am heading out to get the eyes tormented/tested. See y'all tomorrow!22:12
mgagnedevananda: both22:12
mgagnedevananda: I'm fine with moving stuff to vendor too22:13
*** BadCub is now known as BadCub_Hiding22:13
*** harlowja_away is now known as harlowja22:13
devanandamgagne: right. so I'm curious why you need to add a new interface, and would like to second what jroll said above -- that that would be forking the project and in no way guaranteed to be supportable on any future release22:14
devanandamgagne: whereas if you stay with the driver API as we've defined it, having an out of tree driver is just fine22:14
mgagnedevananda: I understand what you are saying. Previouss dev thought it was a great idea to split some part of what should be in vendor in its own driver22:15
devanandaheh22:16
mgagneI already refactor a lot of the code, I don't mind adding a new commit with "come on guys, vendor exists for that purpose"22:17
devananda:)22:17
NobodyCam:)22:21
jxiaobinhi devananda22:21
* NobodyCam is going to pick up his glasses so ya'll stop being fuzzy blobs22:21
devanandaNobodyCam: what if I like being a fuzzy blob?22:22
* mrda find it an interesting time to join the conversation22:22
NobodyCamlol I can always take them off :)22:22
NobodyCammorning mrda22:22
mrdaMornin' (I'm very far away, so I'll remain a fuzzy blob)22:23
NobodyCamlol... the glasses fix upclose things far away thing look fine to /me22:23
NobodyCamlol22:23
jxiaobin@devananda, in nova host_manager.py, if "force_hosts" is identified, the whole filter chain will be skipped. what if the "host" which hosting ironic service has nodes with different hardware configuration?22:24
jxiaobin@devananda, in my test env, I have 2 flavors, one has 128GB ram, another 256GB, correspondingly there are 2 ironic nodes having 128GB & 256 GB ram respectively. If I specify the ironic host to provision, the result is not what I desired22:28
devanandajxiaobin: what result do you expect // what are you getting?22:29
*** Marga_ has quit IRC22:29
jxiaobin@devananda, say flavor bm.large has 128GB ram, I want bm.large hit the node with 128GB ram, but actually the node with 256GB ram was hit22:30
*** Marga_ has joined #openstack-ironic22:30
jxiaobin@devananda, the root cause is I specify availability-zone:host in request, and the filters were skipped22:30
*** Marga_ has quit IRC22:30
*** krtaylor has quit IRC22:31
*** Marga_ has joined #openstack-ironic22:31
devanandajxiaobin: so nova did exactly what you asked for22:31
devanandajxiaobin: if you tell nova to force a host, it forces it. what's the problem?22:31
jxiaobin@devananda for Ironic, there are multiple nodes are the host, and the nodes have different configurations22:32
*** Marga_ has quit IRC22:32
*** Marga_ has joined #openstack-ironic22:32
*** Haomeng has joined #openstack-ironic22:34
jxiaobin@devananda, for BM case, the host I specified is actually the Ironic service node, which actually hosting several "Ironic Nodes" with different hardware configurations22:35
devanandajxiaobin: you can specify the hypervisor host as well, eg, availability-zone:host:node-uuid22:35
jxiaobin@devananda, yeah, that's possible, but what if node is missing22:36
jxiaobin@devananda, the result seems not correct for users22:36
*** Haomeng|2 has quit IRC22:36
jxiaobin@devananda I asked bm.large, but ended up with bm.xlarge22:37
jxiaobin@devananda do you agree this is a defect need fix?22:38
devanandajxiaobin: in the bare metal case, you should use the exact-match scheduler filters, which addresses that most of the time22:38
devanandajxiaobin: however, since you're cutting out the scheduler by saying "force host", no, I don't think that's a bug22:39
devanandawell - to be more clear, perhaps forcehost should refuse to work with ironic if you don't specify the node22:39
devanandathat is, if you just say "--force-host availability-zone:host" it should fail.22:40
devanandayou must specify the node uuid as well22:40
*** EmilienM is now known as EmilienM|afk22:40
jxiaobin@devananda, why not go through baremetal filters instead of return directly?22:40
jxiaobin@devananda, to me, this way makes more sense22:41
devanandajxiaobin: because force-hosts is a means to bypass the scheduler filters. that's the whole point of that function22:41
jxiaobin@devananda, that makes sense for vm, but not bm, due to the bm architecture22:42
jxiaobin@devananda, specifying node id is not convenient22:44
jxiaobin@devananda, not cloudy22:44
jxiaobin@devananda just my 2 cents22:45
devanandajxiaobin: eh? --force-hosts is not cloudy ...22:45
devanandamost users will not (and should not) know what hypervisor host their VM is on22:46
jxiaobin@devananda yeah, true for vm22:46
devanandatrue for bare metal too22:46
jxiaobin@devananda depends on how bm/vm is deployed22:46
devanandauser shouldn't care which particular server they get, so long as it is of the right flavor, has the right network hooked in, etc...22:46
jxiaobin@devananda if vm/bm are mixed in the same cell, the ironic host actually manages all the ironic nodes22:48
jxiaobin@devananda to avoid landing bm into vm HV, I have to specify the ironic host22:49
devanandajxiaobin: oh. nope. you should be using a flavor to differentiate that22:49
devanandathese flavors for VMs. those flavors for bare metal.22:49
jxiaobin@devananda, how to differ vm flavors from bm flavors? any flags?22:50
jxiaobin@devananda my understanding is there is no way to say that this flavor is for vm, that one for bm, am I right?22:51
devanandahttp://docs.openstack.org/developer/ironic/deploy/install-guide.html#configure-compute-service-to-use-the-bare-metal-service22:51
devanandahttp://docs.openstack.org/developer/ironic/deploy/install-guide.html#flavor-creation22:52
jxiaobin@devananda what's the difference? the hardware configuration? cpu arch?22:53
jxiaobin@devananda vm flavors can have these attributes as well22:54
*** Marga_ has quit IRC22:54
*** Marga_ has joined #openstack-ironic22:55
devanandajxiaobin: I need to run ... yes, you can use separate flavors for bare metal that won't get mapped to VM hypervisor hosts.22:55
*** davideagnello has quit IRC22:57
jxiaobin@devananda I did so based on the doc, but actually, if I don't specify az/host, ironic node lands on vm HV22:57
* TheJulia reads22:58
*** yuanying has joined #openstack-ironic23:01
*** chlong has joined #openstack-ironic23:02
*** jgrimm is now known as zz_jgrimm23:04
*** Marga_ has quit IRC23:05
*** dttocs has quit IRC23:06
*** EmilienM|afk is now known as EmilienM23:10
*** stevemar has joined #openstack-ironic23:12
*** stevemar has quit IRC23:17
TheJuliajxiaobin: that honestly doesn't sound quite right, have you verified your nova.conf settings?23:23
*** dguerri is now known as _dguerri23:23
*** _dguerri is now known as dguerri23:24
Haomengjxiaobin: check if you apply this patch - https://review.openstack.org/#/c/174141/23:25
Haomengjxiaobin: option scheduler_use_baremetal_filters should be set to true for ironic.conf23:26
jxiaobin@TheJulia this is the patch I committed23:26
jxiaobinsorry @Haomeng23:26
Haomengjxiaobin: and make sure your nova code is latest which include this nova patch - https://review.openstack.org/#/c/170299/23:27
jxiaobinthe key is I specify az:host23:27
jxiaobin@Haomeng, you're right, if the exact filters are configured, theoretically the bm will not land on vm HV23:28
jxiaobinunless the vm flavor has identical configurations with bm flavor23:29
*** davideagnello has joined #openstack-ironic23:31
*** Sukhdev has joined #openstack-ironic23:34
*** Marga_ has joined #openstack-ironic23:36
*** Marga_ has quit IRC23:40
*** saripurigopi has joined #openstack-ironic23:47
Haomengjxiaobin: yes23:55
*** saripurigopi has quit IRC23:58
« Wednesday, 2015-04-15 Index Friday, 2015-04-17 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!