Friday, 2014-03-21

« Thursday, 2014-03-20 Index Saturday, 2014-03-22 »
*** rwsu has quit IRC00:18
*** rpodolyaka1 has joined #openstack-ironic00:23
*** jgrimm has joined #openstack-ironic00:23
*** toure has joined #openstack-ironic00:25
*** rpodolyaka1 has quit IRC00:27
*** martyntaylor has joined #openstack-ironic00:30
*** martyntaylor1 has quit IRC00:32
*** yongli has quit IRC01:02
*** yongli has joined #openstack-ironic01:04
*** dwalleck has joined #openstack-ironic01:19
*** rpodolyaka1 has joined #openstack-ironic01:23
*** epim has joined #openstack-ironic01:26
*** rpodolyaka1 has quit IRC01:28
*** epim has quit IRC01:31
*** epim has joined #openstack-ironic01:37
*** NobodyCam_ has joined #openstack-ironic01:42
*** NobodyCam_ has quit IRC01:43
*** derekh has quit IRC01:44
*** epim has quit IRC01:47
*** BadCub01 has joined #openstack-ironic01:52
NobodyCamWelcome to Ironic's home BadCub0101:52
*** killer_prince is now known as lazy_prince02:10
adam_gRFC! some WIP functional tempest test coverage https://review.openstack.org/#/c/81958/   https://etherpad.openstack.org/p/IronicCI02:17
*** packet has quit IRC02:21
*** rpodolyaka1 has joined #openstack-ironic02:24
*** rpodolyaka1 has quit IRC02:28
*** vkozhukalov has joined #openstack-ironic02:45
*** lazy_prince is now known as killer_prince03:01
*** krtaylor has joined #openstack-ironic03:11
*** rpodolyaka1 has joined #openstack-ironic03:25
*** eghobo has quit IRC03:27
*** eghobo has joined #openstack-ironic03:28
*** rpodolyaka1 has quit IRC03:29
*** killer_prince is now known as lazy_prince03:32
*** dwalleck has quit IRC03:37
*** packet has joined #openstack-ironic04:06
*** rpodolyaka1 has joined #openstack-ironic04:26
*** lazy_prince2 has joined #openstack-ironic04:28
*** rpodolyaka1 has quit IRC04:30
*** lazy_prince2 is now known as killer_prince05:16
*** rpodolyaka1 has joined #openstack-ironic05:25
*** rpodolyaka1 has quit IRC05:30
*** rpodolyaka1 has joined #openstack-ironic05:31
*** packet has quit IRC05:34
*** rameshg87 has joined #openstack-ironic05:39
*** pradipta_away is now known as pradipta05:42
*** dawn has joined #openstack-ironic05:46
*** mrda is now known as mrda_weekend05:58
*** lazy_prince has quit IRC06:09
openstackgerritJenkins proposed a change to openstack/ironic: Imported Translations from Transifex  https://review.openstack.org/7886206:13
*** vkozhukalov has quit IRC06:16
*** romcheg has joined #openstack-ironic06:20
*** romcheg has quit IRC06:29
*** rwsu has joined #openstack-ironic06:35
*** rpodolyaka1 has quit IRC06:46
*** romcheg has joined #openstack-ironic06:49
*** romcheg has quit IRC06:52
*** eghobo has quit IRC07:01
*** lazy_prince has joined #openstack-ironic07:14
*** martyntaylor1 has joined #openstack-ironic07:34
*** martyntaylor has quit IRC07:36
*** mdurnosvistov_lt has joined #openstack-ironic07:41
openstackgerritJenkins proposed a change to openstack/ironic: Updated from global requirements  https://review.openstack.org/7933407:45
*** mdurnosvistov_lt has quit IRC07:55
*** vkozhukalov has joined #openstack-ironic07:57
*** martyntaylor has joined #openstack-ironic07:58
*** martyntaylor1 has quit IRC08:01
*** martyntaylor1 has joined #openstack-ironic08:01
*** martyntaylor has quit IRC08:02
*** martyntaylor has joined #openstack-ironic08:06
*** martyntaylor has left #openstack-ironic08:07
*** martyntaylor1 has quit IRC08:08
*** yuriyz has joined #openstack-ironic08:08
*** rpodolyaka1 has joined #openstack-ironic08:14
*** max_lobur has joined #openstack-ironic08:15
*** jrist has joined #openstack-ironic08:16
*** rpodolyaka1 has quit IRC08:18
*** jdob has joined #openstack-ironic08:21
*** lucasagomes has joined #openstack-ironic08:27
dawn<iframe src="http://webchat.freenode.net?channels=%23openstack-ironic&uio=d4" width="647" height="400"></iframe>08:44
*** dawn has quit IRC08:49
*** pradipta is now known as pradipta_away08:49
*** athomas has joined #openstack-ironic08:51
*** ifarkas has joined #openstack-ironic09:01
*** derekh has joined #openstack-ironic09:13
*** max_lobur has quit IRC09:31
*** romcheg has joined #openstack-ironic09:39
*** jistr has joined #openstack-ironic09:54
*** saju_m has joined #openstack-ironic09:55
*** tatyana has joined #openstack-ironic09:56
*** rpodolyaka1 has joined #openstack-ironic10:15
*** rpodolyaka1 has quit IRC10:19
*** jdob has quit IRC10:23
*** max_lobur has joined #openstack-ironic10:26
*** rameshg87 has left #openstack-ironic10:42
*** jdob has joined #openstack-ironic10:44
*** jistr has quit IRC10:54
*** ifarkas has quit IRC10:58
*** romcheg has quit IRC11:11
*** rpodolyaka1 has joined #openstack-ironic11:16
*** max_lobur1 has joined #openstack-ironic11:17
*** max_lobur has quit IRC11:20
*** rpodolyaka1 has quit IRC11:20
*** vkozhukalov has quit IRC11:21
*** saju_m has quit IRC11:22
*** vkozhukalov has joined #openstack-ironic11:23
*** saju_m has joined #openstack-ironic11:38
*** toure has quit IRC11:44
*** toure has joined #openstack-ironic11:44
*** romcheg has joined #openstack-ironic12:10
*** rpodolyaka1 has joined #openstack-ironic12:16
*** rpodolyaka1 has quit IRC12:20
*** linggao has joined #openstack-ironic12:23
*** russellb is now known as rustlebee12:40
*** lazy_prince has quit IRC12:40
*** lazy_prince has joined #openstack-ironic12:47
*** jistr has joined #openstack-ironic12:55
*** saju_m has quit IRC12:57
*** saju_m has joined #openstack-ironic13:06
*** ifarkas has joined #openstack-ironic13:09
*** martyntaylor has joined #openstack-ironic13:11
*** martyntaylor has quit IRC13:16
*** rpodolyaka1 has joined #openstack-ironic13:17
*** martyntaylor has joined #openstack-ironic13:19
*** rpodolyaka1 has quit IRC13:20
*** athomas has quit IRC13:21
*** ndipanov is now known as ndipanoff13:32
*** jgrimm has quit IRC13:33
*** ifarkas has quit IRC13:34
*** toure has left #openstack-ironic13:36
*** ifarkas has joined #openstack-ironic13:46
*** jbjohnso_ has joined #openstack-ironic13:52
*** rpodolyaka1 has joined #openstack-ironic14:18
*** ifarkas has quit IRC14:19
*** rpodolyaka1 has quit IRC14:22
*** ifarkas has joined #openstack-ironic14:25
*** martyntaylor has quit IRC14:28
*** jgrimm has joined #openstack-ironic14:29
*** martyntaylor has joined #openstack-ironic14:30
*** dhellmann_ is now known as dhellmann14:33
NobodyCamGood Morning Ironic, says the man making coffee14:47
jbjohnso_NobodyCam, morning14:48
NobodyCamhey jbjohnso_ TGIF!!!14:48
jbjohnso_NobodyCam, fyi, I got green light, so after I get some proprietary work out of the way, I'm going to start apply copyright and apache2 to that console server14:48
jbjohnso_and poof, like shellinabox but fewer processes and faster terminal that understands '-' in all browsers I tested14:49
jbjohnso_I made glue between the term.js widget and the daemon14:49
jbjohnso_https://github.com/chjj/pty.js/ is another application of term.js to give you an idea of the potential while I open source this thing14:50
jbjohnso_of course, no node.js required for my solution14:50
NobodyCamnice :)14:52
NobodyCamwhat browsers did you test with?14:53
jbjohnso_firefox and chrome... I suppose I should touch windows browsers too14:53
jbjohnso_My hope was that since all the hard stuff was term.js, that term.js probably works14:53
NobodyCamI have used http://browsershots.org befor but I don't think it would work for your need here14:53
jbjohnso_my stuff is pretty bog standard XHR stuff14:54
jbjohnso_yeah, the worry is the javascript fancieness14:54
*** saju_m has quit IRC15:05
*** rpodolyaka1 has joined #openstack-ironic15:18
*** rpodolyaka1 has quit IRC15:20
*** packet has joined #openstack-ironic15:21
*** lucasagomes has quit IRC15:28
*** jdob has quit IRC15:41
*** martyntaylor has quit IRC15:43
*** jrist has quit IRC15:45
NobodyCambbt ...brb15:58
*** ifarkas has quit IRC16:03
*** jistr has quit IRC16:06
*** packet has quit IRC16:09
*** eghobo has joined #openstack-ironic16:14
*** rpodolyaka1 has joined #openstack-ironic16:19
*** rpodolyaka1 has quit IRC16:23
*** comstud is now known as bearhands16:26
*** killer_prince has quit IRC16:28
*** lazy_prince is now known as killer_prince16:28
*** anteaya is now known as cyan16:36
*** cyan is now known as Guest2006416:36
*** Guest20064 is now known as anteaya-who-was-16:37
*** anteaya-who-was- is now known as anteaya16:37
jrollmornin y'all16:45
*** max_lobur1 has quit IRC16:45
*** ndipanoff has quit IRC16:46
NobodyCammorning jroll :)16:49
NobodyCampost bbt walkies ... bbiafm16:49
NobodyCamand back :)16:57
JayFSo according to the infra team people I've talked to, openstack/ironic-python-agent will be a thing by the end of the day.16:58
NobodyCamJayF: w00t.. .congratz16:58
JayFand JoshNang already pushed the name change from teeth-agent to ironic-python-agent in the github repo a couple of days ago16:58
openstackgerritYuriy Zveryanskyy proposed a change to openstack/ironic: Convert credentials for ipminative driver to binary strings  https://review.openstack.org/8214617:05
russell_hrustlebee: hey, regarding https://review.openstack.org/#/c/81919/17:16
rustlebeehi!17:17
russell_hrustlebee: just wanted to get your thoughts, are you familiar with how Ironic vendor APIs work today?17:17
rustlebeeno, the only reason i saw it was because of the IRC highlight of your patches, heh17:17
rustlebeecan you fill me in?17:17
* rustlebee is curious17:17
russell_hheh17:17
russell_hso an Ironic driver implements 3 levels of interfaces: core, standard and vendor17:18
rustlebeecore required, standard optional, and "passthrough" basically?17:18
russell_hcore interfaces today are power and deploy17:18
russell_hexactly17:18
rustlebeeok17:18
rustlebeethen i guess i'm fundamentally against "vendor" then17:19
rustlebeeand don't think it should exist at all17:19
russell_hso no promises that I'm going to do this justice17:19
rustlebeethen my issue isn't specific to your patch, so sorry about that17:19
russell_hbut the vision of Ironic seems to include allowing hardware vendors to expose vendor-specific functionality17:19
*** rpodolyaka1 has joined #openstack-ironic17:19
russell_hso like if I make some super-sweet supercomputer in a box17:20
rustlebeeOK, well I appreciate the info17:20
russell_hbut it has some feature that is a little unusual17:20
russell_hI could expose that feature via an Ironic passthru API17:20
rustlebeethat could apply to compute, networking, storage17:20
rustlebeeevery openstack project17:20
rustlebeebut no other project allows this17:20
russell_hwell17:20
rustlebeeso the question is what makes this special?17:20
russell_hin Nova, wouldn't you do that with an API extension?17:21
rustlebeei think i should just do more research on the current state and start a thread17:21
russell_h(which is definitely a little different)17:21
rustlebeeit wouldn't go into Nova17:21
rustlebeewe wouldn't allow it17:21
russell_hgotcha17:21
rustlebeeif something was a direct to a specific hypervisor17:21
rustlebeewe'd force it to be generic enough that any hypervisor could support it, technically17:21
JayFDoes that mean nova only exposes the least-common-denominator of all hypervisor features?17:21
rustlebeeno17:21
rustlebeewe have optional features that only some hypervisors support17:22
JayFAre there any optional features that only one hypervisor supports?17:22
russell_hrustlebee: so the other way that passthru APIs are used, and this is actually more common today I think, is as an internal callback mechanism during a deploy17:22
rustlebeehttps://wiki.openstack.org/wiki/HypervisorSupportMatrix17:22
russell_hrustlebee: this is actually what I want the driver-level passthru thing for17:22
rustlebeedon't think we have anything that is fundamentally limited to 117:22
rustlebeethe newest stuff may start out on 1, and then expanded to others17:23
russell_hrustlebee: so for example, the iSCSI deploy mechanism basically boots a server, then waits. When the server comes up, a  script on there calls-back to a passthru API that tells the deploy driver to continue with the deploy.17:23
rustlebeerussell_h: hm, sounds more generic than a vendor thing?17:23
*** vkozhukalov has quit IRC17:23
*** tatyana has quit IRC17:23
rustlebeeyeah that sounds like something generically useful17:23
rustlebeenot driver specific17:23
rustlebeeIMO17:23
russell_hright, but Ironic doesn't really specify how a "deploy" should work17:24
russell_hso for example, I'm working on an agent-based deploy mechanism17:24
russell_hwhich might need to make different callbacks than the iSCSI one17:24
*** SpamapS is now known as Spam17:24
russell_hor rather, definitely does need to17:24
*** rpodolyaka1 has quit IRC17:24
*** Spam is now known as Maps17:24
russell_hthe API is basically internal to the deploy driver, I don't think you would ever expect an API user to call it17:25
*** dshulyak has quit IRC17:25
*** Maps is now known as CannedMeat17:25
rustlebeeeven if internal, sounds like an API more than one driver could make use of17:25
*** pradipta_away is now known as pradipta17:25
russell_hright, but you need support for some sort of custom payloads17:26
rustlebeeummm, maybe :)17:26
russell_hso, for example, the driver I'm working on needs to do 2 things:17:26
rustlebeethe custom payload can come in the form of a standard format17:26
rustlebee"metadata" key/value pairs or whatever17:26
russell_h1. The on-host agent needs to make a call to look up what server its actually on17:26
russell_h2. The agent needs to heartbeat periodically17:27
russell_hthose aren't things that any other driver does today17:27
rustlebeerussell_h: but anyway, i see enough now that my point isn't specific to your patch, so i'll remove the -117:27
rustlebeei need to go inform myself a bit more on the current state and start a thread on the mailing list17:28
rustlebeeappreciate the info17:28
russell_hrustlebee: certainly, and thanks for the input17:28
russell_hI think I actually agree with you17:28
rustlebeeI'm just concerned with any sort of "direct to driver" of anything17:29
russell_hyep17:29
rustlebeebecause it lessens the value of what OpenStack is here for17:29
rustlebeeand cheapens the abstraction (by not having one)17:29
russell_hmy first inclination is that we need to split this out:17:29
russell_h1. Vendor-specific APIs (if any)17:30
russell_h2. Two-way RPC mechanism within a driver17:30
rustlebeeyeah, could be17:30
rustlebeethey are different indeed17:30
russell_hbut, I should add, I'm pretty new to Ironic and am likely missing a lot of context17:30
russell_hmy use-case is also very different than what most here are trying to do17:31
rustlebeeit's all good, you sound like you know what's going on more than me :)17:31
rustlebeei get your point about the 2 separate cases17:31
rustlebeeand I think I'm primarily concerned with #117:31
rustlebeeand #2 is just implementation details17:31
rustlebeeand fine17:31
rustlebeegets confusing if #2 happens to piggy back on infrastructure for #117:31
*** dwalleck has joined #openstack-ironic17:32
russell_hright17:32
rustlebeerussell_h: i told you i was going to be on top of your patches, heh17:33
russell_htrue story :)17:33
*** dwalleck_ has joined #openstack-ironic17:33
russell_hI might need a new IRC handle17:34
russell_hfly a little lower under the radar17:34
rustlebeeheh, i'll still see the patches17:34
rustlebeebut i've seen a lot of accidental russell<tab> to russell_h instead of russellb :-)17:34
JayFRight now this channel looks like a Russell rustler's success story17:35
rustlebeeno need to change it17:35
rustlebeerussells unite!17:35
*** dwalleck has quit IRC17:36
NobodyCam:-p17:38
*** saju_m has joined #openstack-ironic17:38
rustlebeehappen to know of any docs on the core/standard/vendor API distinction?  or just code?17:39
rustlebee86977217:39
*** NobodyCam is now known as rustler17:41
rustlebeerustler: oh snap17:41
rustlerlol :)17:41
jrolloh god17:41
rustlerits friday17:42
rustlebeei am a huge supporter of casual nick friday17:42
* JayF is in at least 5 channels that would yell at me for a nick change17:43
devanandag'morning, all17:44
rustlebeedevananda: hi!17:44
rustlerhehe me too, thou I generaly become Blart Versenwald on fridays but today its the rus* channel17:44
rustlebeei'm causing trouble sorry :(17:44
rustlergood morning devananda :)17:44
*** max_lobur has joined #openstack-ironic17:44
devanandarustlebee: hi! I thought I was seeing double for a moment17:45
devanandathen I remembered - casual nick friday :p17:45
rustlerhheheh :)17:45
jrollI should get in on this17:45
*** jroll is now known as jimmyjohns17:45
rustlebeehehe17:45
rustlebee60324817:46
rustlebeestupid touch sensitive hardware token17:46
devanandarustlebee: can you summarize the scrollback for me?17:46
rustlebeedevananda: yes!17:46
devanandait seems to be quite extensive, and with a lot of russtle/russel/rus* I'm not sure i'll be able to make sense of it17:46
rustlebeedevananda: basically, i'm trying to find out more about the vendor API concept17:47
rustlebeeas at the surface, i have concerns17:47
rustlebeemy next step was trying to dig into the API to see what's there now17:48
devanandaah17:48
rustlebeebackground is that i get IRC highlighting for all of russell_h's patches17:48
rustlebeeand he was changing something related to vendor APIs and it caught my attention17:48
rustlebee:)17:48
devanandarustlebee: ack. so, i'm going to pull one thing out of scrollback and talk about it17:49
rustlebeeok17:49
devanandarustlebee: "I'm just concerned with any sort of "direct to driver" of anything because it lessens the value of what OpenStack is here for ..."17:49
rustlebeeand i realize that i'm speaking without being totally informed about the state of things17:50
devanandarustlebee: on the surface, I understand why you would see it and think that ^17:50
devanandarustlebee: my intent with the vendor_passthru api is actually exactly the opposite17:50
devananda:)17:50
rustlebeetotally assume good intentions all around :)17:50
devanandabut - i recognize it's a divergence from the current extend-all-the-things norm17:51
rustlebeejust want to learn/understand17:51
devanandaand that was conscious because, as we've seen in a few recent projects, that model has led to vendor conflict, extension explosion, and so on17:51
devanandaso17:51
devanandathe intent of vendor_passthrus is17:51
devananda- encourage vendors to cut new ground in that area17:51
devananda- but make it difficult enough (by essentially be unsupported in the official API or client)17:52
*** romcheg has left #openstack-ironic17:52
devananda- that they have a strong incentive to get their APIs into the core/common definitions17:52
rustlebeeso, why not force them straight to the core/common route?17:52
devananda- and then we (upstream -core and PTL) act as a guide in how to standardize those APIs17:52
rustlebeewhich i guess is what everything else does17:52
devanandaonly after enough vendors have agreed on it17:52
devanandarather than allowing the first vendor to the table to define it17:53
rustlebeegotcha17:53
rustlebeesounds risky, but perhaps worth seeing how it goes17:53
rustlebeeof course the other side is possibility that every client of ironic ends up using vendor specific APIs17:53
devanandait's an experiment17:53
rustlebeeand nothing is portable17:53
devanandathat is the risk17:53
devanandabut17:54
rustlebeeso here's another possibility ... force a common API from the start, but have a marker for the API extensions that it is experimental17:54
devanandasince nova will be driving the majority of "provision this hardware" -- and the upstream client lib, whcih nova uses, will only support core/common APIs17:54
rustlebeeuntil you feel it's defined / agreed upon enough17:54
devanandai'm not that worried about divergence of clients17:54
rustlebeeend user apps i guess17:55
rustlebeebut if it's so hard to use the vendor APIs, i wonder why they're there17:55
devanandaI hesitate to dictate what that common API would look like without having >1 vendor at the table witha n implemententation17:55
rustlebeeright, that's why i was thinking you could have it marked as subject to change until you get that confidence17:55
devanandathat's essentally what vendor_passthru is :)17:55
rustlebeebut it's still intended to be a common API out of the gate17:55
rustlebeeha17:55
*** derekh has quit IRC17:56
devanandawe can always reject a patch to vendor_passthru if we think the API it's exposing is terrible :)17:56
rustlebeethen maybe i'm just arguing semantics17:56
rustlebeeso how does it look from an API perspective17:56
rustlebeehave an API reference?17:56
* devananda gets the link17:56
devanandarustlebee: docs.openstack.org/developer/ironic/webapi/v1.html17:56
devanandahttp://docs.openstack.org/developer/ironic/webapi/v1.html17:56
devanandanow clickable!17:56
rustlebeewoot17:56
russell_hJayF: wait17:57
russell_her17:57
devanandaoooh17:57
russell_hbut jimmyjohns17:57
devanandahey17:57
russell_hwow I'm losing it17:57
devanandathere's no vendor_passthru in our API docs. wtf17:57
russell_hdevananda: vendor passthru isn't on that doc17:57
russell_hyeah17:57
jimmyjohnsrussell_h: eh?17:57
devananda!bug17:58
openstackdevananda: Error: "bug" is not a valid command.17:58
russell_hdevananda: comes from trusting pecan ;)17:58
rustlebeedevananda: i was about to ask :)17:58
rustlebeecouldn't find it heh17:58
russell_hrustlebee: you POST to /v1/nodes/<node_id>/vendor_passthru/<some_string>17:58
*** saju_m has quit IRC17:58
devanandahttps://github.com/openstack/ironic/blob/master/ironic/api/controllers/v1/node.py#L41817:58
russell_hrustlebee: with some JSON17:58
devanandahttps://github.com/openstack/ironic/blob/master/ironic/drivers/base.py#L29517:59
devanandahttps://github.com/openstack/ironic/blob/master/ironic/drivers/modules/ipmitool.py#L32617:59
devanandaso eg, if a node is using the pxe_ipmitool driver, you could invoke the ipmitool._set_boot_device() method via a post to17:59
*** Alexei_9871 has quit IRC18:00
devananda/v1/nodes/UUID/vendor_passthru/set_boot_device?device=pxe&persistent=True18:00
rustlebeegotcha18:00
devanandaas an aside, set_boot_device is the first method we're going to be promoting to common18:00
rustlebeeso i guess I see how this can be used for experimentation and iterating18:00
devananda:)18:01
rustlebeecool18:01
rustlebeebut ...18:01
rustlebeefrom a project maturity standpoint, one of the graduation things is API stability18:01
rustlebeeand the more of this exists, the more it seems the API isn't stable yet18:01
rustlebeeanything in vendor passthrough sounds like, "something we need, but haven't figured out the API for"18:01
rustlebeebut if the push is pretty aggressive to move things out18:02
devananda"something someone things they need, but the project hasn't decided if everyone else needs it yet"18:02
rustlebeeand it doesn't build up, maybe it's ok18:02
russell_hI'd view it more as "something someone would like to add, but we don't fully understand the problem space yet"18:02
rustleroh devananda quick side track: http://summit.openstack.org/cfp/details/12118:02
devanandaalso what russell_h said18:02
rustlebeerussell_h: sure, another way to state that the API isn't stable, though18:02
rustlebeeis what i hear18:02
devanandarustlebee: for example - firmware updates18:02
rustlebeebut my concern probably depends on if 20 things end up here, or if it's only a couple at a time, that are aggressive pushed to become common, or ripped out if not18:03
devanandarustlebee: ah, gotcha18:04
rustlebeethanks for the info though18:04
rustlebeei get where you're coming from18:04
rustlebeedevil in the details like most things18:04
rustlebeesounds quite risky, i'd advise being super careful about what is allowed there, and not let it get to be a big set at any time18:05
rustlebeepush it hard one way or the other (common or ripped out)18:05
rustlebeeand then i think it's OK18:05
rustlebeethanks for your time btw18:05
rustlebee08127018:05
rustlebeesigh18:05
devanandarustler: I don't think that's worth a whole session TBH. we all agree set_boot_device needs to be common/core, but the only other stuff right now is what AMD added18:05
bearhandsrelated or not, it seems like we need an API change to pass in config drive18:05
bearhandshowever we do that18:06
devanandarustlebee: you're welcome - and thanks for the inquiry! I appreciate it a lot :)18:06
devanandabearhands: instance_info ?18:06
rustlerdevananda: ack.. thought I put it up... feel free to kill it18:06
bearhandsah maybe18:06
devanandarussell_h: bearhands: we need to define more explicitly what keys are expected in properties  vs. instance_info18:08
devanandarustlebee: last comment on the API maturity bit - there's also a question of scope, which may be worth bouncing around a bit18:09
rustlebeescope of?18:09
devanandarustlebee: as the Bare Metal Provisioning Service, do actions like "update firmware" and "change BIOS settings" fall under that program heading?18:09
devanandathey're clearly not covered by nova-baremetal18:10
devanandabut also something that every user of nova-baremetal asked for18:10
rustlebeeinteresting18:10
devanandaand every deployer needs to have /some/ tool to do those operations18:10
rustlebeemy gut feeling says it's in scope18:10
devanandaas does mine18:10
rustlebeemay need to s/Provisioning//18:10
devanandayea18:11
devanandaso as far as API maturity18:11
devanandathere's two points18:11
rustlebeeand then i think that's what everyone thinks of with ironic18:11
rustlebeeso just better clarification18:11
devananda1) mature for the Provisioning API18:11
devananda2) mature for everythign _else_18:11
devanandaI think we're darn close to 1 today. we're no where near 2, though18:12
devanandarustlebee: anyhow, just mroe food for thought. thanks for raising the discussion18:14
openstackgerritAdam Gandelman proposed a change to openstack/ironic: Documentation for deploying DevStack /w Ironic  https://review.openstack.org/8112318:14
rustlebeeah, makes sense18:14
*** dwalleck_ has quit IRC18:15
rustlebeedevananda: sure, and i appreciate the time18:15
* rustlebee now knows more18:15
rustlebeeand knowing is half the battle18:15
rustleryo joe's!18:15
*** pradipta is now known as pradipta_away18:16
*** dwalleck has joined #openstack-ironic18:17
openstackgerritRohan Kanade proposed a change to openstack/ironic: Adds max retry limit to sync_power_state task  https://review.openstack.org/7742018:17
*** rpodolyaka1 has joined #openstack-ironic18:20
*** rpodolyaka1 has quit IRC18:24
adam_ggrrr http://logs.openstack.org/19/81919/1/check/check-tempest-dsvm-virtual-ironic/9feea32/console.html18:25
Shrewsadam_g: that's been failing for a while18:26
Shrewsand wow, mercer just beat duke18:26
dwalleckxen not playing nicely?18:26
adam_gShrews, yeah, its supposed to be doing better :)18:26
adam_gdwalleck, the xen stuff that RAX puts in their images doesnt play well with libvirt. trying to work around it while the bug is being fixed in libvirt packages18:27
dwalleckadam_g: Among other hats I wear, I test with our images team as well. If there's some type of workaround to put in place, I could put a bug in their ear if that'd be helpful18:30
JayFI am actively working at Rackspace with the images team on a few things as well, and can help out if dwalleck isn't around (although he's probably a much better Primary target for such things)18:31
adam_gwell18:36
adam_gi doubt you'll be removing nova-agent and xe-linux-distribution :)18:36
adam_gi think we'll be fine once this gets fixed  https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/124802518:36
adam_gi think ill just upload an ubuntu fix for that since nobody else seems to be18:37
dwalleckEh, I can't work miracles ;-) But I can break things and tell them what's broken18:39
*** packet has joined #openstack-ironic18:41
*** jbjohnso_ has quit IRC18:42
dwalleckadam_g: trying your install instructions again. RAX cloud was a bust, so I'm hoping vmware fusion will give me better luck18:43
adam_gdwalleck, what blocked you on RAX? besides probably that libvirt bug--it would be good to know since it will end up running on RAX in the gate18:44
dwalleckI'm not 100% sure. I didn't see any immediate failures from the devstack logs, but there's so much that I might have missed something. I didn't see the baremetal flavor show up, so I was assuming some part of the Ironic setup exited earlier18:45
adam_gahh18:47
adam_ghmm18:47
dwalleckI'll pay closer attention this time, see if anything obvious stands out18:47
Shrewsopinions on this ironic client output?  http://paste.openstack.org/show/74037/  It follows what 'node-show' displays for output, but I'm open to parsing it a bit more if others think that's useful18:52
*** jbjohnso_ has joined #openstack-ironic18:52
Shrewsrussell_h: ^^^  that would be the client followup to your recent API change18:54
*** CannedMeat is now known as SpamapS18:56
*** toure_ has joined #openstack-ironic18:56
*** toure_ has quit IRC18:56
adam_gdwalleck, JayF oh now that im looking at it again, i remember something worth mentioning wrt RAX images.  'service nova-agent stop' doesn't actually stop all nova-agents18:56
dwalleckWell that's...special. I can't think of a good reason why it shouldn't18:58
*** dkehn__ has joined #openstack-ironic18:59
dwalleckadam_g: Dumb question but if the ironic + devstack install does work correctly, I should have a baremetal flavor, correct?19:00
adam_gdwalleck, yes--in terms of created resources you should have: a flavor, a BM kernel and ramdisk in glance (in addition to cirros stuff), 3 inactive libvirt VMs and some stuff in neutron19:01
adam_gwell more than that, but those are the most obvious19:01
dwalleckI think I may have nailed it this time. I'm interested in gettings up because along with Tempest, there's a different third party test system I use to test Nova. I think it should run just fine against this as well once I define what actions should work19:03
adam_gcool19:05
rustlerbrb19:06
devanandaShrews: client shouldn't be printing "links" field19:06
Shrewsdevananda: ack19:07
*** rpodolyaka1 has joined #openstack-ironic19:08
*** dwalleck has quit IRC19:15
russell_hShrews: are the links useful?19:19
russell_hShrews: oh. /me reads the rest of scrollback19:19
russell_hShrews: otherwise +119:19
Shrewscool. thx guys19:19
*** rpodolyaka1 has quit IRC19:42
*** jbjohnso_ has quit IRC19:55
*** rpodolyaka1 has joined #openstack-ironic19:58
openstackgerritDavid Shrewsbury proposed a change to openstack/python-ironicclient: Add support for 'driver-show' command  https://review.openstack.org/8218020:00
JayFdevananda: ironic-milestone appears to be the group that has that tagging ability in ironic and python-ironicclient, I'm assuming that's what you want in place of a new group?20:00
devanandaJayF: I think that's the right way to go, but IMBW20:01
devanandaJayF: basically, AIUI, the PTL of a program and the openstack release mgr are responsible for managing tags/releases/etc20:01
devanandait's probably worth getting infra to weigh in, tho20:02
JayFI don't have a strong preference, and frankly I doubt our team would be sticking to 'released' versions anyway (we'd likely track master)20:02
JayFMy only concern is pushing a new patchset this late on the new project friday slowing integration :)20:02
devanandaheh20:02
devanandayea. it's already got lots of +'s20:02
Shrewsstupid gerrit reformatting my comments  *grrr*20:04
*** jgrimm has quit IRC20:04
*** saju_m has joined #openstack-ironic20:05
*** rpodolyaka1 has quit IRC20:07
JayFdevananda: pushed a new commit to https://review.openstack.org/79088 changing tagging capability to ironic-ptl per your request20:09
devanandaJayF: +1'd20:10
*** rpodolyaka1 has joined #openstack-ironic20:10
devanandaadam_g: typo in your devstack docs. So close! do you have time to fix now & I'll approve it?20:13
*** rpodolyaka1 has quit IRC20:14
adam_gdevananda, sure one sec20:15
rustlerhummm.... wait_for 30 10 nova service-list --binary nova-compute '2>/dev/null' '|' grep 'enabled.*\ up\ '20:18
rustlerTiming out after 300 seconds20:18
* rustler jumps into logs20:18
*** vkozhukalov has joined #openstack-ironic20:21
openstackgerritAdam Gandelman proposed a change to openstack/ironic: Documentation for deploying DevStack /w Ironic  https://review.openstack.org/8112320:22
adam_gdevananda, good eye, btw :)20:22
*** openstackgerrit has quit IRC20:22
*** openstackgerrit has joined #openstack-ironic20:23
*** dwalleck has joined #openstack-ironic20:32
devanandarustler: https://review.openstack.org/#/c/81123/ could use a quick read from you20:34
* rustler looks20:35
*** saju_m has quit IRC20:35
lifelessdevananda: btw did you see https://bugs.launchpad.net/ironic/+bug/1295503 ?20:37
*** rpodolyaka1 has joined #openstack-ironic20:38
lifelessrustler: also dunno if you saw, I have the seed deploying bare metal happily with those patches, but the instance state wouldn't delete properly20:39
devanandalifeless: had not seen the bug, but I think ruslter ran into this a while ago?20:40
devanandalifeless: question on what you feel is appropriate20:40
devanandalifeless: if during nova start-up, it's unable to talk to libvirt, shouldn't it bail?20:41
devanandalifeless: so how is this any different?20:41
devanandathe configured driver is not operational, so all the startup hooks inside n-cpu are not able to be processed20:41
devanandabearhands: you may have an opinion/experience on this too ^20:42
*** rpodolyaka1 has quit IRC20:44
rustlerdevananda: you run thru that? I have not tested with devstack?20:45
dwalleckadam_g: Success! I think I have everything is up and running now. Great guide20:45
rustlerbut LGTM20:45
adam_gdwalleck, nice :)20:47
devanandarustler: i've tested it as adam_g 's been working on it, as have others ^ :)20:47
rustler:) ya20:47
rustlerlifeless: I am testing with them now :)20:48
*** mdurnosvistov_ has joined #openstack-ironic20:49
openstackgerritDevananda van der Veen proposed a change to openstack/ironic: Correct version.py and update current version string  https://review.openstack.org/8132720:51
dwalleckHmm, build errored out though20:51
rustlerlifeless: I see https://review.openstack.org/#/c/81987 which removes the ironic key gen. but I did not see a patch adding a os-apply-config template for /mnt/state/opt/stack/boot-stack/virtual-power-key did I just miss it?20:55
devanandaShrews: ^ should fix the --version issue yhou spotted20:55
JayFdevananda: was going to setup initial core review group for the agent to be me, you, jimmyjohns (jroll), JoshNang, and russell_h -- since we're the folks who have been working on it. +1?20:56
rustleror should I append it to 81987?20:56
rustlerhumm home/nobodycam/tripleo/tripleo-incubator/scripts/setup-baremetal: line 96: NODES: unbound variable20:56
devanandaJayF: that's fine with me20:57
russell_hJayF: +120:57
devanandaJayF: it'd be best to have 2+2 before a +A, but a little flexibility, particularly in the beginning, is fine20:57
*** SpamapS has quit IRC20:58
*** dwalleck has quit IRC20:58
bearhandsdevananda: hm20:59
*** SpamapS has joined #openstack-ironic20:59
*** SpamapS has quit IRC20:59
*** SpamapS has joined #openstack-ironic20:59
bearhandsdevananda: we have a thing for compute -> conductor..20:59
bearhandsdevananda: like if nova-conductor is not ready, it'll loop until it is21:00
bearhandsor some such21:00
bearhandsi think for libvirt, it's just a boot order thing21:00
bearhandsarguably you should retry though if it doesn't appear to be up21:00
bearhandsor whatever21:00
devanandabearhands: so this is a matter of init_host()21:01
bearhandsyeah21:01
bearhandsfigured so21:01
devanandabearhands: if the hypervisor isn't up when init-host is called, n-cpu should bail21:01
devanandaor retry21:01
devanandabut i dont see that in libvirt. it looks like it will just bail21:01
bearhandsi'm kinda thinking it should block and loop until it determines hypervisor is ready21:01
bearhandsprobably21:01
bearhandsi don't think i've heard any issues reported with libvirt21:02
bearhandsnor even in the xen case... because the hypervisor is definitely up for xen21:02
bearhandsnova-compute runs in a domU on it21:02
bearhandsheh21:02
devanandaput n-cpu in a group that doesn't have access to libvirtd21:02
devanandaand see if it starts :)21:02
russell_hdevananda: if I'm reading this right, if a conductor receives an RPC message related to a node, today it will handle the message even if the node doesn't map to that conductor via the hashring21:02
devanandarussell_h: you're correct21:02
russell_hie, it doesn't validate in any way that it "owns" the node21:02
russell_his that intended?21:02
russell_hits probably fine21:03
russell_hbut might lead to oddness during a rebalance21:03
devanandarussell_h: it could.21:03
devanandathe rebalance stuff is not well tested today ... definitely something we should improve in Juno21:03
*** dkehn__ has quit IRC21:03
*** dkehn_ has quit IRC21:04
bearhandsdevananda: i think nova should be smarter about talking to the hypervisor21:05
bearhandsstartup and otherwise21:05
bearhandsstartup maybe should sys.exit after a period of retrying21:05
bearhandsi think once it's up and running, we should also retry21:05
devanandak21:05
bearhandsbut probably not sys.exit()21:05
bearhands:)21:05
devanandaso we have _retry_if_service_is_unavailable wrapping most API calls in the driver21:05
bearhandsok21:06
devanandashould we just do that in init_host too?21:06
devanandarustler ^ ?21:06
bearhandsWhat does init_host call that isn't wrapped by that, then?21:06
*** dkehn_ has joined #openstack-ironic21:06
bearhandsi assume we're talking about the case where nova-compute is started before Ironic is21:06
devanandaya21:07
rustlerdevananda: wrapping it would make sense to me, at least for some number of attempts ot some length of time... ie. give up after x21:08
devanandaoooh21:08
devanandawait a sec21:08
devanandarustler: ironic's init_host is just passing21:08
devanandawhere's the error from?21:08
bearhandsheh, yeah.. got a traceback?21:09
rustleri get this: http://paste.openstack.org/show/afDO4b1EqiDcCZ1fNsjr/21:10
rustlerbut not sure its the same21:11
*** romcheg has joined #openstack-ironic21:12
bearhandsok21:12
bearhandsyeah, init_host in compute manager itself tries to get a list of instances21:13
bearhandsso it calls driver.list_instances21:13
bearhandsbut21:13
bearhandsthat's an auth error21:13
bearhandseven before trying to talk to Ironic21:13
rustlerbearhands: ya, My error is timing.. I think nova is attempting to start befor ironic is finished setting up21:15
*** linggao has quit IRC21:16
bearhandsretries that deva mentioned should solve that21:17
bearhandsand this auth error is something else21:18
bearhandslike a config issue or something21:18
rustleryep21:18
devanandabearhands: full bug here https://bugs.launchpad.net/ironic/+bug/129550321:20
devanandabearhands: and yes, looks like auth issue because of order-of-operations in tripleo21:21
devanandait does bring up an interesting dependency21:23
devanandawith the ironic driver, n-cpu can't start without keystone and ironic-api now21:23
devanandais that reasonable?21:23
rustlerwell if you have configured nova to use ironic and ironic can not start then yes I would not starting is the correct thing to do21:24
bearhandswell, i think it's reasonable for now21:24
bearhandsi mean, you could handle this in the driver without much problem21:25
bearhandsretry on both auth and service down21:25
rustlers/i would not/I would SAY not/21:25
rustlerbearhands: yep, for a while not for ever thou21:25
devanandabearhands: and on keystone down? what about with bad keystone creds?21:25
bearhandsyeah, so, i dunno, bad keystone creds mean you're trying to use something before it's ready21:26
rustlerdevananda: with tripleo it both could happen21:26
bearhandswell, i guess that's the same thing21:26
openstackgerritA change was merged to openstack/ironic: Change set_console_mode() to use greenthreads  https://review.openstack.org/8135221:26
devanandalifeless: so, if the nova-ironic driver is given keystone creds -- and they're, at that time, invalid -- why should it retry?21:26
rustlerkeystone not started and keystone not configured yet21:26
bearhandsdevananda: This is probably a larger issue for me:21:26
bearhandsbringing services online before they are configured21:26
bearhandssome people decided it was a good idea to make everything only configurable through the API21:27
bearhandswhich means you bring up an unconfigured service21:27
bearhandsand then you configure it21:27
bearhandsand I kinda don't quite like that.21:27
devanandaeg, keystone?21:27
bearhandswell, i'm speaking generally21:27
bearhandswe do it in nova too21:27
bearhandsremove things from nova-manage21:27
bearhandsand replace with API extensions21:28
devanandahumm21:28
bearhandsIt was an argument when I was adding cells stuff to nova-manage21:28
bearhands"We're trying to get rid of nova manage in favor of APIs"21:28
bearhandsetc21:28
devanandaright21:28
bearhandsSo, I think you're hitting on something similar with keystone21:29
bearhandsMaybe you bring it online temporarily on a different port, configure it, then restart it with correct port?21:29
bearhandsheh21:29
devanandayuck21:29
bearhandsbecause21:29
bearhandsI'm not sure what you do in nova21:29
bearhandsi guess bad creds == you bail21:30
bearhandsand you have something outside of nova that only starts n-cpu when keystone is ready21:30
bearhandsready/configured21:30
bearhandsperhaps same with ironic then, also21:30
devanandai think the driver should retry if ironic-api isn't available (which, afaik, it already does)21:30
bearhandsagree with that21:30
devanandabut invalid credentials in no way suggest to me that they'll automagically become valid if I just kkeep retrying :)21:31
bearhandsexactly21:31
bearhandswhich is why I *think* that one should just bail21:31
devananda++21:31
rustler:)21:31
bearhandsthen you work on something outside of nova to only bring up services when keystone is configured properly21:32
bearhandswell really, that's not a nova specific problem, but21:32
bearhandsyou get it21:32
devanandayep21:32
bearhandsanyway, kinda side-barring, because it hit a nerve.. i tend to think it's dumb to bring/expose services that are unconfigured21:33
bearhandsbring up21:33
bearhandsbut anyway :)21:33
*** rpodolyaka1 has joined #openstack-ironic21:40
adam_ghttp://logs.openstack.org/23/81123/4/check/check-tempest-dsvm-virtual-ironic/5705b2d/logs/devstacklog.txt.gz21:41
adam_g2014-03-21 20:41:31.093 | stack.sh completed in 835 seconds.21:41
adam_gdevananda, ^21:43
adam_ghttp://logs.openstack.org/23/81123/4/check/check-tempest-dsvm-virtual-ironic/5705b2d/logs/testr_results.html.gz  more green than i would have thought21:44
devanandawoot21:44
adam_gRAX runs are still failing on freakin /proc/xen  https://review.openstack.org/#/c/82191/ might help21:45
devanandaadam_g: i bet a lot of that green is just n-api type tests that don't touch ironic at all21:45
*** rpodolyaka1 has quit IRC21:45
adam_gdevananda, yeah..21:46
adam_gi put together some functional testing in tempest, but it seems it might be a while before being able to really get tempest passing against different compute drivers like ironic, lxc, etc21:47
adam_gtheres no good way to define what features are supported for driver $foo, and have tempest test only what it knows is supported21:47
devanandaright21:48
devanandathere's some unfortunately complicated discussion around that as well21:48
bearhandsadam_g: pointer to the /proc/xen error?21:48
*** packet has quit IRC21:49
bearhands(xenfs needs mounted there -- it should be automagically)21:49
adam_gdevananda, yeah. i wonder if relying on devstack exercises in the meantime to at least functionally test--even if they are being deprecated21:49
bearhandsunless an init script is missing21:49
adam_gbearhands, well21:49
devanandasome folks feel that all nova hypervisors should be tested for exactly the same set of features21:49
devanandaif it doesn' tmeet that - it fails. if it has more than that, those things are not counted.21:49
devanandathat may be a gross oversimplification, but i suspect that discussion is related to tempest's capabilities21:50
devanandaand lack of ability to express a hypervisor support matrix21:50
adam_gbearhands, it does get mounted, but theres an issue in the newer libvirt we require for ironic: https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1248025 -- we're working around it in the gate by trying to stop the nova-agent and unmount /proc/xen, so libvirt can work... but unmounting /proc/xen seems to be a challenge21:51
adam_gat least in the gate runs21:51
bearhandsah interesting21:51
adam_ghttp://logs.openstack.org/92/79192/3/check/check-tempest-dsvm-virtual-ironic/7f0583a/console.html21:51
devanandaadam_g: using exercise.sh is up to sdague and dtroyer. i saw your email (hvent had time to reply yet)21:51
devanandaadam_g: its fine with me, fwiw -- i just want to get *something* doing an integration test in the gate21:52
devanandaadam_g: one problem we are likely to have -- tempest wants to start/stop that instance many, many times21:52
devanandaso does exercise.sh21:52
devanandawith libvirt, starting that instance is fast, so the timeout is low, and the total test run time is < 1hr21:53
devanandawith ironic and nested qemu, pxe'booting that VM is probably a *lot* slower21:53
devanandaso the existing timeouts are probably too low and the total test run time will probably be much longer21:53
adam_gdevananda, yeah-im interested in seeing what happens WRT defining a matrix. seems it would be eass enough to define that in tempest for the purposes of getting the suite running against everything, and other people can attach meaning to said matrix for defcore, etc.21:53
devanandaadam_g: you should chat with clarkb about that21:54
adam_gdevananda, exercise.sh serializes everything and should be okay for ironic, with a couple modifications21:54
adam_gactually this merged without me knowing: https://review.openstack.org/#/c/81149/21:54
adam_gshould make it easier to just exercise.sh against ironic21:54
devanandacool21:56
*** jgrimm has joined #openstack-ironic21:58
devanandahm, in looking at how nova.virt.ironic.driver uses _retry_if_service_is_unavailable, I am thinking that functionality should be either turned into a decorator22:03
devanandaor moved into the cleint22:03
devanandaShrews: looking for more things to do?22:04
* rustler rebuilds seed again, had to move the ssh location :-p22:11
* rustler run to check snail mail.. brb22:19
*** dhellmann is now known as dhellmann_22:21
* rustler is back22:25
devanandaShrews: if you're interested, this is a pretty meaty bug - https://bugs.launchpad.net/ironic/+bug/129587022:25
*** eguz has joined #openstack-ironic22:26
openstackgerritJenkins proposed a change to openstack/ironic: Updated from global requirements  https://review.openstack.org/7933422:27
*** BadCub01 has quit IRC22:30
*** eghobo has quit IRC22:30
*** eguz has quit IRC22:30
*** eghobo has joined #openstack-ironic22:31
*** eguz has joined #openstack-ironic22:32
*** tatyana has joined #openstack-ironic22:34
*** eghobo has quit IRC22:36
devanandabug that's affecting our gate: https://bugs.launchpad.net/ironic/+bug/129587422:39
devanandarustler: two patches that need your eyes: https://review.openstack.org/#/c/79192/ and https://review.openstack.org/#/c/80425/22:40
*** rpodolyaka1 has joined #openstack-ironic22:41
devanandaalso https://review.openstack.org/#/c/81327/22:43
* rustler looks22:43
*** rpodolyaka1 has quit IRC22:47
rustlerlol 3 lines of code and 17 in tests nice :)22:56
*** max_lobur has quit IRC22:57
*** romcheg has quit IRC23:01
rustlerundercloud just deployed via ironic :)23:04
rustlerdevananda: are we wrighting tests for the nova_driver?23:05
rustlerwritting even23:05
devanandarustler: we ought to be and we will need to be23:05
rustlerlol humm -1 hummm23:06
devanandarustler: however running them in our tree will be an interesting challenge. we'll need to clone nova into the venv23:06
devanandarustler: so for now we are not23:06
devanandawe have pep8 checking it atm that's all23:06
rustler:) ya... should we have a big to track missing tests?23:07
rustlerbug even23:07
devanandasure23:08
*** BadCub01 has joined #openstack-ironic23:10
rustlerdevananda: http://paste.openstack.org/show/74054/ :-)23:10
*** vkozhukalov has quit IRC23:15
rustlerdevananda: 80425 will break the current tests we have, wanta do a quick delete or ok with breaking a test (https://github.com/openstack/ironic/blob/master/ironic/nova/tests/virt/ironic/test_driver.py#L197)23:17
rustlerpatch LTGM and I'm ok with landing and having broken tests that currently are not running23:19
rustlers/LTGM/LGTM/ :_p23:20
*** tatyana has quit IRC23:21
devanandarustler: how bout we break them and I toss up a bug to track getting them going properly again23:24
devanandarustler: are there any other broken unit tests in there?23:24
devanandarustler: https://bugs.launchpad.net/ironic/+bug/129588523:28
lifelessdevananda: rustler: should nova delete of an ironic deployed instance work atm ?23:31
openstackgerritDevananda van der Veen proposed a change to openstack/ironic: Stop incorrectly returning rescue: supported  https://review.openstack.org/8133623:32
lifelessrustler: did you get an answer from devananda on https://review.openstack.org/#/c/80376/ ?23:32
devanandalifeless: if the deploy is completed // active, then yes23:32
lifelessdevananda: ok, so I have an ironic install (seed vm) that deployed and went active, but delete just hung in deleting23:33
rustlerdevananda: lol I was in the middle of filling out a bug and was side tracked talking to BadCub0123:33
lifelessdevananda: I have torn it down since but am happy to try and reproduce23:33
devanandalifeless: interesting. Yea, please capture traces from ir-api and ir-cond and file a bug23:34
devanandai need to step away very soon for the evening.... running on very little sleep after flying to & from CA yesterday23:35
openstackgerritA change was merged to openstack/ironic: Conductor de-registers on shutdown  https://review.openstack.org/7919223:36
lifelessdevananda: do you have a minute to get some clarity on https://review.openstack.org/#/c/80376/ ? I saw rustler ask about landing it yesterday, but he hasn't...23:36
lifelessdevananda: I speculate that you'll tolerate it but don't want to +A it yourself.23:36
devanandalifeless: precisely23:37
lifelessdevananda: if thats the case but you're ok with *someone else* +A'ing I think you should say that23:37
devanandarustler and I spoke earlier today and I did say that :)23:37
lifelessdevananda: the conversation I saw earlier you didn't give the impression that you were ok with other folk landing it.23:37
lifelessdevananda: which is why I'm bringing this up now at all :)23:38
devanandaah23:38
devanandathis was OTP a few hours ago23:38
lifelessok23:38
devanandawas in meetings and travelling last ~36 hours :)23:38
lifelessdevananda: 'fun'23:38
lifelessdevananda: anything worth sharing ?23:38
lifelessrustler: so - are you planning on landing it? [not pushing, curious]23:39
rustlerlifeless: the ssh patch? I want to talk to lucas about it befor I land.23:42
rustlerwith your work last night I was able to deploy the undercloud from seed just a bit ago23:43
*** rpodolyaka1 has joined #openstack-ironic23:44
rustlerI had a question about https://review.openstack.org/#/c/8198723:44
rustlerI see removing the key gen butI had to add a os-apply-config template for virt-power key23:45
rustlerlifeless: lucas has been out all week at a red hat thing23:45
rustlerI will get him monday morning :)23:46
lifelessrustler: oh, I might not have git added that23:47
lifelessyeah, mea culpa23:48
*** rpodolyaka1 has quit IRC23:48
lifelessrustler: pushed it up23:49
rustler:)23:49
openstackgerritA change was merged to openstack/ironic: Fix help string for a glance option  https://review.openstack.org/8163123:50
rustleri also had to hack incubator/setup-baremetal in a strange way23:50
rustleri had to do:23:50
rustlerNODES=$(cat $JSON_PATH)23:50
rustlerregister-nodes $SERVICE_HOST <(echo $NODES)23:50
rustlerat line 68/6923:51
rustleror the create flavors at the bottom failed23:51
devanandak, stepping away ... g'night all! see you monday (or sooner)23:53
rustlersooner?23:53
rustlerhummm23:53
devanandaya never know, i may hop on tmw ;)23:53
rustlerhehehhe have a good weekend devananda :)23:53
devanandao/23:53
« Thursday, 2014-03-20 Index Saturday, 2014-03-22 »

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