Friday, 2014-02-21

*** rbrady-afk has quit IRC00:15
*** cd-undercloud has joined #tripleo00:18
cd-undercloud************** overcloud complete status=1 ************00:18
*** cd-undercloud has quit IRC00:18
*** e0ne has joined #tripleo00:21
*** CaptTofu has joined #tripleo00:23
*** StevenK_ is now known as StevenK00:25
*** StevenK is now known as Guest9969500:26
*** e0ne has quit IRC00:26
*** Guest99695 is now known as StevenK00:26
StevenKlifeless: I wasn't touching power yet00:30
*** matsuhashi has joined #tripleo00:31
lifelessStevenK: all the same thing :)00:31
*** lazy_prince has quit IRC00:31
*** cadenzajon has quit IRC00:32
*** killer_prince has joined #tripleo00:32
lifelessSpamapS: still up ?00:33
*** hashar has quit IRC00:35
*** edmund has quit IRC00:38
*** dkehn__ has joined #tripleo00:46
*** howleyt_ has joined #tripleo00:48
*** matsuhas_ has joined #tripleo00:48
*** tteggel_ has joined #tripleo00:48
*** nijaba_ has joined #tripleo00:48
*** dkehn___ has joined #tripleo00:49
*** csd has quit IRC00:49
*** AaronGreen has joined #tripleo00:49
*** csd has joined #tripleo00:50
*** matty_du1s|gone has joined #tripleo00:51
openstackgerritA change was merged to openstack/tripleo-image-elements: Restart libvirt-bin when reconfiguring.
openstackgerritA change was merged to openstack/tripleo-image-elements: Stop using deprecated libvirt_type setting.
openstackgerritA change was merged to openstack/tripleo-image-elements: Enable live block migration for libvirt.
*** matsuhashi has quit IRC00:55
*** nijaba has quit IRC00:55
*** tteggel has quit IRC00:55
*** dkehn_ has quit IRC00:55
*** dkehn has quit IRC00:55
*** lifeless has quit IRC00:55
*** howleyt has quit IRC00:55
*** AaronGr has quit IRC00:55
*** Daviey has quit IRC00:55
*** matty_dubs|gone has quit IRC00:55
*** tteggel_ is now known as tteggel00:56
*** cd-undercloud has joined #tripleo00:57
cd-undercloud************** overcloud complete status=1 ************00:57
*** cd-undercloud has quit IRC00:57
*** AaronGreen is now known as AaronGr00:57
*** lifeless has joined #tripleo00:57
*** sets mode: +v lifeless00:57
openstackgerritA change was merged to openstack/tripleo-heat-templates: Add a Merge::Map feature.
*** Daviey has joined #tripleo00:58
lifelessgo you beauty00:58
openstackgerritA change was merged to openstack/tripleo-heat-templates: Gather all hypervisor names into /etc/hosts.
*** lifeless has quit IRC01:02
*** lifeless has joined #tripleo01:02
*** sets mode: +v lifeless01:02
greghaynesanyone made an overcloud recently? mine seems to not be starting amqp01:10
lifelessgreghaynes: you're missing rabbitmq-server01:11
lifelessgreghaynes: pull incubator01:11
greghaynesthat would probably do it :p01:11
lifelessand look at the log01:11
*** weshay has joined #tripleo01:15
*** openstackgerrit has quit IRC01:18
*** openstackgerrit has joined #tripleo01:19
*** cd-undercloud has joined #tripleo01:21
cd-undercloud************** overcloud complete status=1 ************01:21
*** cd-undercloud has quit IRC01:21
*** e0ne has joined #tripleo01:21
*** e0ne has quit IRC01:26
openstackgerritAlex Holden proposed a change to openstack/tripleo-image-elements: Fixing error in readme
*** edmund has joined #tripleo01:32
*** nosnos has joined #tripleo01:34
greghaynesseems somewhat odd that setting DIB_EXTRA_ARGS actually removed rabitmq-server01:37
greghaynesnot that I have a better solution in mind01:37
jog0lifeless: ping01:38
*** kfox1111 has quit IRC01:43
*** cd-undercloud has joined #tripleo01:45
cd-undercloud************** overcloud complete status=1 ************01:45
*** cd-undercloud has quit IRC01:45
*** cd-undercloud has joined #tripleo01:47
cd-undercloud************** overcloud complete status=128 ************01:47
*** cd-undercloud has quit IRC01:47
lifelessjog0: pong01:48
jog0lifeless: bug 1282842 is ugly01:48
jog0wanted to move over here because infra is a little overwhealmed01:48
jog0with other work01:48
lifelessyeah, thats cool01:48
lifelessI know01:48
lifelessits a shotgun bug01:48
*** newell has quit IRC01:48
jog0heh yeah01:49
jog0so I am not surprised01:49
jog0only surprising thing is how come you didn't notice earlier?01:49
lifelessof 40 instances nodepool spawned 22 ended up in error.01:49
lifelesswell, I've whinged to you about this with nova baremetal before01:49
jog0so I does tripleo cloud preserve instances accross updates now?01:49
lifelessthis time I caught it in the act on kvm01:50
lifelessjog0: yes, and we have live migration too01:50
jog0thats what I thought nice01:50
lifelessthough see
lifelessnova default is not live01:50
jog0oh yeah saw that01:50
jog0anyway winding down for the day01:51
lifelessthe ci-overcloud isn't being updated though01:51
lifelessso there are no updates to worry about here01:51
jog0wasted 2 hours fighting HP's neutron or rather quantum01:51
jog0it didn't set up a default network for me01:51
lifelessin the public cloud ?01:52
jog0anyway I am happy to help dig on your shotgun bug01:52
jog0it was because I am special flower01:52
lifelessso its 'nothing by default' ?01:52
lifelessor you has some special setting ?01:52
lifelessanyhow yeah it it might be e.g. keystone failing to be useful01:53
jog0its because my account was created differently01:53
*** shakayumi has quit IRC01:53
lifelessonly 400K tokens01:53
jog0yeah this bug sounds like keystone derping01:53
lifelessso shouldn't be the usual fail01:54
lifelessSpamapS: did we automate that tuning we did ?01:54
* jog0 goes AFK01:54
lifelesshmm, i need another +2er01:54
lifelessmarios_: around ?01:55
*** cd-undercloud has joined #tripleo01:57
cd-undercloud************** overcloud complete status=6 ************01:57
*** cd-undercloud has quit IRC01:57
StevenKHm, 6?01:57
*** dkehn___ is now known as dkehn02:07
*** cwolferh has joined #tripleo02:07
lifelessFetching Base Image02:08
lifelesscurl: (6) Couldn't resolve host ''02:08
lifelessFetching Base Image02:08
lifelesscurl: (6) Couldn't resolve host ''02:08
*** cd-undercloud has joined #tripleo02:08
cd-undercloud************** overcloud complete status=128 ************02:08
*** cd-undercloud has quit IRC02:08
lifelesserror: Unable to find 6f8f611cbeb83ea38563ca09abcde925e05e32fd under
lifelessCannot obtain needed tree 6f8f611cbeb83ea38563ca09abcde925e05e32fd02:09
lifelesswhile processing commit a94be9443d639dc0ac8498d8a35fed2832379ee7.02:09
lifelesserror: Fetch failed.02:09
StevenKlifeless: AKA "Hm, where did the Internet go?"02:10
*** dkehn__ is now known as dkehn_02:10
*** e0ne has joined #tripleo02:21
openstackgerritlifeless proposed a change to openstack/tripleo-image-elements: Don't use /cgit/ urls for git.
*** e0ne has quit IRC02:26
*** weshay has quit IRC02:28
*** tzumainn has quit IRC02:28
*** saurabhs has quit IRC02:32
*** matsuhas_ has quit IRC02:34
*** matsuhas_ has joined #tripleo02:36
*** marun has quit IRC02:37
*** tzumainn has joined #tripleo02:42
SpamapSlifeless: remove the /cgit02:43
SpamapSlifeless: it doesn't really work02:43
lifelessSpamapS: which one did I leave?02:44
lifelessSpamapS: also good...morning ?02:44
lifelessSpamapS: or did you just get back from irish hospitality ?02:44
openstackgerritlifeless proposed a change to openstack/tripleo-heat-templates: Fix the hosts setting - nova uses novalocal.
lifelessSpamapS: one small erraa ^02:45
SpamapSlifeless: I'm still in .ie .. flight leaves in ~ 7 hours02:45
SpamapSofficially in the clutches of the lag monster02:46
lifelessSpamapS: can you +2 75287 before you flap flap flap away02:46
SpamapSlifeless: isn't that just the example?02:47
openstackgerritlifeless proposed a change to openstack/tripleo-heat-templates: Fix the hosts setting - nova uses novalocal.
openstackgerritlifeless proposed a change to openstack/diskimage-builder: Better apt-sources docs.
*** cd-undercloud has joined #tripleo03:02
cd-undercloud************** overcloud complete status=1 ************03:02
*** cd-undercloud has quit IRC03:02
*** matsuhas_ has quit IRC03:03
*** cd-undercloud has joined #tripleo03:03
cd-undercloud************** overcloud complete status=7 ************03:03
*** cd-undercloud has quit IRC03:03
*** matsuhas_ has joined #tripleo03:06
lifelessSpamapS: dunno, looking03:10
lifelessSpamapS: two reviews up there plox ;)03:10
SpamapS\o/ testing03:12
*** cd-undercloud has joined #tripleo03:12
cd-undercloud************** overcloud complete status=6 ************03:12
*** cd-undercloud has quit IRC03:12
lifelessdats a norty word03:13
*** lifeless changes topic to "tripleo-cd running preserve-ephemeral WIP patches and | Using OpenStack to deploy OpenStack;meetings Tuesday 1900 | Using OpenStack to deploy OpenStack;meetings Tuesday 1900 UTC in #openstack-meeting-alt"03:15
*** jcooley_ has joined #tripleo03:17
openstackgerritA change was merged to openstack-infra/tripleo-ci: Get logs from undercloud
*** e0ne has joined #tripleo03:21
*** cd-undercloud has joined #tripleo03:24
cd-undercloud************** overcloud complete status=128 ************03:24
*** cd-undercloud has quit IRC03:24
*** e0ne has quit IRC03:26
lifelessSpamapS: you might want to +A the cgit thing to unbreak tripleo-cd03:29
*** matsuhas_ has quit IRC03:29
lifelessseed didn't come up03:32
lifelesswe're broken atm03:32
openstackgerritA change was merged to openstack/tripleo-image-elements: Don't use /cgit/ urls for git.
*** jcooley_ has quit IRC03:34
*** jcooley_ has joined #tripleo03:35
*** ramishra has joined #tripleo03:37
*** killer_prince2 has joined #tripleo03:39
*** jcooley_ has quit IRC03:39
*** killer_prince2 is now known as lazy_prince03:39
*** lazy_prince has quit IRC03:50
StevenKlifeless: So, yes, DIB_DISTRIBUTION_MIRROR is a subset, but I think you could mention both, now that the security.u.c change has landed. "If you want to point to a closer Ubuntu mirror, or an internal mirror, use DIB_DISTRIBUTION_MIRROR. If you'd like fine-grained control over the contents of the deployed sources.list, perhaps by adding other repositories, use the apt-sources element."04:01
*** CaptTofu has quit IRC04:02
SpamapSlifeless: btw, regarding mellanox, I have it on good authority that our mellanox networking troubles will go away if we install the latest driver and firmware.04:04
*** cd-undercloud has joined #tripleo04:09
cd-undercloud************** overcloud complete status=1 ************04:09
*** cd-undercloud has quit IRC04:09
SpamapSah back to failing the right way ... ;)04:09
lifelessStevenK: I could; tell you what - if you have prose in mind, push an iteration with that.04:11
*** csd has quit IRC04:11
lifelessStevenK: I wanted to capture pointers for folk, not embed an entire dib manual :) - but I'm not against more.04:11
lifelessStevenK: I just have little motivation for more.04:12
StevenKlifeless: I was thinking of a few sentences04:17
*** ramishra has quit IRC04:19
lifelessStevenK: thats fine, I'm just saying - I won't push back against it, but I don't really feel the need to add it.04:19
*** ramishra has joined #tripleo04:19
lifelessStevenK: so I'd be delighted if you want to add it in04:20
*** e0ne has joined #tripleo04:21
lifelessStevenK: for context, I didn't set out to improve the docs, I just had a seniors moment when I went to use my local proxy, so captured a pointer in the relevant script04:23
openstackgerritA change was merged to openstack/tripleo-incubator: Remove duplicate configure-vm call in create-nodes
StevenKMore yak shaving from yesterday landing04:25
*** matsuhashi has joined #tripleo04:25
StevenKlifeless: So I reach for DIB_DISTRIBUTION_MIRROR first, but that's probably because I wrote the support for it.04:25
*** e0ne has quit IRC04:26
lifelessthe reason I like apt-sources is I can point it at my actual apt-sources04:26
lifelessand it DTRT thing to build as long as my laptop is configured for wherever I am04:26
*** epim has joined #tripleo04:33
SpamapSlifeless: shouldn't it be DIB_PYPI_MIRROR_URL ? Or is PYPI_MIRROR_URL something special ?04:35
*** jcooley_ has joined #tripleo04:35
lifelessSpamapS: its not special, so I can change it if you like04:37
SpamapSlifeless: we have been pretty consistent with that prefix04:38
SpamapSIt is very late/early here depending on how you look at it04:38
lifelessso, I don't really care either way, rebase is easy.04:38
SpamapSand I may have just done way too many reviews on an empty stomach, and may be a bit grumpy now04:38
lifelessbut I want to get away from variables anyway04:38
lifelessyour call04:39
SpamapSso let's do it and we'll trash the variables later04:39
*** jcooley_ has quit IRC04:40
openstackgerritA change was merged to openstack/diskimage-builder: Permit using arbitrary PyPI urls.
SpamapSlifeless: the recent status=1 fail was 500's from nova-api04:44
SpamapSlifeless: we have a bit of convincing to do in #heat soon... there's hesitance to just ignore/retry on 500... :-/04:44
SpamapSbut anyway.. I need a nap... bbl04:44
lifelesshave they actually used heat?04:44
StevenKlifeless: I think I'm ready to test my nodes change, I suspect kicking off a hpcloud node to use it for devtest would be a terrible idea?04:44
lifelessStevenK: you don't have 16GB yet?04:46
StevenKlifeless: Nope04:46
StevenKI can probably just run it locally like my existing setup04:48
lifelesswhats your existing setup ?04:48
lifelesscloud nodes are terrrrribly slow due to nested virt04:48
StevenKlifeless: The devtest I have currently on my local desktop04:49
lifelessStevenK: you could just tear that down and run in its place, no?04:49
StevenKRight, hence "I can probably just run it locally like my existing setup"04:50
lifelessit would be nice to have a test job that deploys a CI node and runs regular gate checks against it04:51
lifelessbut we either end up with nested virt, or we need undercloud user keys for it04:52
lifelessso it would need to be secured and I don't think zuul is quite up to that04:52
*** cd-undercloud has joined #tripleo04:53
cd-undercloud************** overcloud complete status=1 ************04:53
*** cd-undercloud has quit IRC04:53
*** epim has quit IRC04:53
*** lazy_prince has joined #tripleo04:55
*** noslzzp has quit IRC04:55
StevenKlifeless: I shall have to troll Craige mercilessly on your behalf when I see him tonight04:55
*** jcooley_ has joined #tripleo05:03
*** akuznets_ has quit IRC05:07
*** jcooley_ has quit IRC05:11
*** michchap has quit IRC05:14
*** cwolferh has quit IRC05:14
lifelesshmm overcloud failing05:15
*** akuznetsov has joined #tripleo05:20
*** e0ne has joined #tripleo05:21
*** matsuhashi has quit IRC05:23
*** e0ne has quit IRC05:26
*** matsuhashi has joined #tripleo05:33
*** jcooley_ has joined #tripleo05:47
*** killer_prince has quit IRC06:01
*** killer_p- has joined #tripleo06:02
*** killer_p- is now known as killer_prince06:02
*** CaptTofu has joined #tripleo06:03
*** akuznetsov has quit IRC06:04
*** CaptTofu has quit IRC06:08
*** cd-undercloud has joined #tripleo06:08
cd-undercloud************** overcloud complete status=255 ************06:08
*** cd-undercloud has quit IRC06:08
*** akuznetsov has joined #tripleo06:09
*** coolsvap has joined #tripleo06:13
lifeless255 now06:17
lifelesswtf is this lotto06:17
*** mrunge has joined #tripleo06:17
*** e0ne has joined #tripleo06:21
*** saurabhs has joined #tripleo06:22
*** e0ne has quit IRC06:26
*** vkozhukalov has joined #tripleo06:33
*** saju_m has joined #tripleo06:36
SpamapSlifeless: fail to ssh into the instance.. that's a fairly encouraging failure. :-P06:38
*** lsmola has joined #tripleo06:42
lifelessSpamapS: its certainly a new one06:42
*** lsmola has quit IRC06:42
*** lsmola has joined #tripleo06:43
lsmolalifeless: hello, can i have a quick question?06:43
lsmolalifeless: we are trying to solve initialization of overcloud in tuskar06:43
lsmolalifeless: in short, these API calls
lsmolalifeless: does it make sense for us to pack this all to TuskarAPI, and then use tuskar client in the devtest?06:44
lsmolalifeless: or that initialization will be soon moved into template? So we should rather use the devtest scripts06:45
lsmolaSpamapS: ^06:45
lsmolalifeless: btw. using the devtest scripts would mean, we will just invoke it from tuskarAPI, which will be very ugly...06:47
*** jprovazn has joined #tripleo06:49
lifelessthere are two things there06:53
lifelessone is sshing in06:53
lifelesswhich is ugly06:53
lifelesswe should find a good reliable way to avoid that06:53
lifelessthe second is a very small number (2 I think) of API calls to make keystone remotely usable06:54
lsmolalifeless: yeah, the first one, right?06:54
lifelessthen the third thing (sorry, off by one error:P) is a bunch of normal API calls that wouldin a non-scripted setup be done via the overcloud horizon06:54
lsmolalifeless: right, I am skipping that, that is the creating of demo user, uploading image, twc. yes that will be done in overcloud horizon06:55
lifelesswe can actually trim init-keystone down06:55
lifelessonly need admin + admin user + endpoint registration IIRC06:55
lsmolalifeless: ok cool06:56
lsmolalifeless: so does it make sense for us to do this in TuskarAPI?06:56
lsmolalifeless: when it will be ready, we would put TuskarAPI CLI inside devtest?06:56
lifelessjust thinking, be a few:)06:57
lsmolahehe, ok :-)06:57
lifelessok so the problem is this with keystone06:58
lifelesswe need the same certs on all the boxes06:58
lifelessso we need something syncing them06:58
lifelessprobably we need to generate them and inject into the template in the absence of barbican06:58
lsmolalifeless: right06:59
lifelessand secondly we don't have tuskarAPI to bring up the seed06:59
lsmolalifeless: yes06:59
lsmolalifeless: so that was my second question06:59
lsmolalifeless: he have been thinking about tuskar only in Overcloud ocntext07:00
lifelessso I think we want some code that can be called from tuskarAPI and from shell07:00
lifelessI don't mind if its python with a CLI wrapper or more shell or $whatever07:00
lifelessbut lets make sure that tuskar can wrap it nicely07:01
lsmolalifeless: ok, so Undercloud would use shell, and overcloud tuskarAPI07:01
lifelesse.g. no parsing freetext output or anything07:01
lifelessseed will use shell07:01
lsmolalifeless: oh yeah, sry07:01
lifelessundercloud shell or tuskar CLI (pref tuskar CLI), overcloud shell or tuskar CLI (pref tuskar CLI)07:01
lsmolalifeless: well if we leave out the certificates, it should be just API calls, so we can do that clean Openstack way07:02
lsmolalifeless: ok, so I'll start with admin + admin user + endpoint07:03
lsmolalifeless: should I create a bug for the certificate?07:03
lsmolalifeless: is it easy to pack it to template?07:03
lifelessso taking a step back07:04
lifelesshere is what I think07:04
lifelesswe need to aim for:07:04
lifeless - shared code for the bring-up07:04
lifeless - tuskar has all the complex scaling code07:05
lifeless - seed deployment is a static configuration so doesn't need any of that07:05
lifelessI think calling shell from within tuskarAPI is ok, its no worse than calling e.g. C07:06
lifelessI'll walk through that ether pad and annotate line by line in a little bit for you07:06
lsmolalifeless: hm07:06
lsmolalifeless: so we will start with wrapping those scripts in tuskarAPI?07:07
lsmolalifeless: but if we will be wrapping just API calls, we might rewrite that to using pythonclients07:08
lsmolalifeless: and just wrap the one shell with certificate, until we get rid of that07:09
lifelesslsmola: sure, but then you'll need to expose that rewritten python with a CLI07:09
lifelesslsmola: for 'shared code for the bring-up'07:09
lsmolalifeless: yeah07:09
lsmolalifeless: we already have cli07:09
lifelesslsmola: we don't want to maintain two copies of it, whatever it is.07:09
lsmolalifeless: so that should not be hard07:09
lifelesslsmola: not tuskar-cli07:09
lifelesslsmola: the seed case.07:09
lsmolalifeless: right07:09
lsmolalifeless: darn07:10
lifelessso whats unique between the seed and (undercloud, overcloud)07:10
lifelessthey both have varying networks to create, which are user modifiable07:11
lifelessso a little bit of data processing07:11
lsmolalifeless: well, it should be the same07:11
lsmolalifeless: except templates07:11
lifelessyeah broadly07:11
lifelesstemplate forseed is static07:11
lifelesstemplate for undercloud/overcloud is tuskar owned07:11
*** jcooley_ has quit IRC07:12
lsmolalifeless: ok07:12
lsmolalifeless: so, if we will expose the rewritten code in TuskarCLI, the general part?07:13
lsmolalifeless: and we make also CLI code specially for Overcloud?07:13
lsmolalifeless: seed could use tuskar too right? it would just use some general function with few API calls in it07:14
lifelessseed can't use tuskar API07:14
lifelesswe can install the CLI and run it yes07:14
lifelessbut It would be ugly IMO to have a circular dep tuskar API -> tuskar CLI -> tuskar API07:15
*** jcooley_ has joined #tripleo07:15
lifelesslets have a specific thing - library or whatever - for this one-time-setup stuff07:15
lsmolalifeless: hm the problem is that CLi should just call API07:15
lsmolalifeless: well not sure if general library will help07:16
lifelesslook at it like a graph07:16
lifelesswe have a thing07:16
lsmolalifeless: cause the right way, when using UI, we should skip the CLI, and call pythonclients07:16
lifelesscall it 'cloud-setup'07:16
lifelessits some code we need to call07:16
lifelessbut nice and ugly07:16
lifeless-api uses it07:16
lifelessand we setup the seed with it too07:17
*** jcooley_ has quit IRC07:18
lifelessso it needs a python api (the most basic of which is and a CLI for the seed setup07:18
lifelessbut some folk may want a GUI to setup the seed07:18
lifelesslocal html5 or something07:18
lifelessI dunno - but that has to run privileged calls etc etc, so - lets not engineer it just yet07:18
lsmolalifeless: just a little step back, what is the problem with seed having tuskar installed?07:19
*** e0ne has joined #tripleo07:21
lsmolalifeless: if we have API call cloud_setup, that will be used by CLi and just take params like passwords?07:21
lsmolalifeless: so for seed we would use tuskar setup-cloud IP PARAMS=...07:22
lsmolalifeless: the same we would use for overcloud07:22
lsmolalifeless: and we would just use puthonclinet from UI, that would call the same07:23
lifelesslsmola: what API does tuskar stuff depend on?07:23
lsmolalifeless: should not be different from any other service07:24
lifelessI meant HTTP API07:24
lifelesstuskarAPI uses heat right07:24
lifelesswhat heat would you use to deploy the seed07:24
lsmolalifeless: yeah it call heat now07:24
lsmolalifeless: and it would use also keystone07:25
lsmolalifeless: that should be it07:25
lsmolalifeless: so we would just have to setup it correctly07:25
lifelessbut there is no heat until we *have configured the seed*07:26
lsmolalifeless: though, I am not sure how far we should go with this, should all of this init be part of the template in J?07:26
*** e0ne has quit IRC07:26
lifelessthe template can't orchestrate things in the cloud above07:26
lifelessmaybe in future we'll teach it how to do that, but as it stands, it can't07:26
lsmolalifeless: hm, the heat should not be a problem, unless you call the heat related action07:26
lifelesslsmola: and there's no nova07:26
lsmolalifeless: this action in tuskar API would just create some stuff in keystone DB07:27
lsmolalifeless: so for this call, it wouldn't need the other stuff07:27
lifelesslsmola: it seems very odd07:28
lsmolalifeless: if you would call overcloud_create, it would fail with no heat service available07:28
lifelesslsmola: that you'd have an API server which you can't use except for some-end stuff07:28
tzumainnlsmola, just to check - are you talking about having seed creation depend upon tuskar api?07:28
lsmolalifeless: or07:28
lsmolalifeless: the third option is07:29
lsmolalifeless: we will call the keystone api call directly from UI07:29
lsmolalifeless: which is not bad07:29
lifelesslsmola: then we have two copies of the setup code07:29
lsmolalifeless: the problem here is that it should not take very long07:29
lsmolalifeless: well, the CLI way of install and Horizon way are always different i think07:30
lsmolalifeless: because Horizon doesn't call CLI07:30
lsmolalifeless: that should be alright i suppose07:30
lifelessso, what do you dislike about my suggestion?07:30
*** vkozhukalov has quit IRC07:31
lsmolalifeless: to pack it in a general lib that will be used by everything?07:31
lifelessyes, in a common place07:31
lsmolatzumainn: yea something like that07:31
tzumainnlsmola, I think that expands the scope of tuskar in a dangerous way07:31
lsmolalifeless: hm ok07:32
lsmolalifeless: and inside of this lib, we would call the pythonclients?07:32
lsmolalifeless: and we would expose API and CLI for it?07:33
lifelessits an abstraction, we can do whatever we want07:33
lifelessit would have a Python API for TuskarAPI and a CLI for seed setup07:33
lsmolalifeless: and this would mean to start another service for it?07:34
lsmolalifeless: like TripleoCloudSetup ?07:34
lifelessno, no service07:34
SpamapSlifeless: I have another reason to have Barbican as raised by a nova person... we need to do cert setups for live migration as well07:34
lifelessSpamapS: the reference docs dont do that ;)07:35
lifelessSpamapS: but yes07:35
lsmolalifeless: so where would it live, just as some egg?07:36
lifelesslsmola: yes07:36
lsmolalifeless: and we would just sent it AdminCredentials, so it can talk to anything?07:37
lifelessI don't know all that we'd pass it. Quite a bit I imagine.07:37
lsmolalifeless: hm, I can't think about anything better right now07:38
lsmolalifeless: this will keep the coed in one place07:39
lsmolalifeless: we will need another repo, under tripleo project07:39
lifelessyes, don't want tuskar-api depending on incubator :)07:40
lsmolalifeless: yeah, this is probably the best we can do right now07:40
lifelessthis stuff has been fairly stable for a bit now07:40
lifelessso I think its time07:40
lifelessbrainstorming for names...07:40
lsmolalifeless: ok, now we need to find some volunteers to extract it07:41
lsmolalifeless: tripleo-cloud-setup sounds nice07:41
lsmolalifeless: though we can pack there the init and also the "create a demo horizon for me with some vms"07:42
lifelessI still need one more +2 on this
lifelessand we can close the live migration card07:42
lifelesslsmola: lets not conflate the demo stuff07:42
lsmolalifeless: ok, just an idea :-)07:43
*** rdopieralski has joined #tripleo07:43
lsmolalol Purple07:43
lifelesslsmola: so - can has review please  ^ :)- its not related to the failures (affects overcloud only and the seed failed)07:44
lsmolalifeless: so do you have anybody up there who can extract it like in 2 weeks? :-D so we can use it in Icehouse..07:44
lsmolalifeless: on it07:45
*** jang1 has joined #tripleo07:45
lsmolalifeless: ok, the failures looks scary, but approving07:46
openstackgerritA change was merged to openstack/tripleo-heat-templates: Fix the hosts setting - nova uses novalocal.
SpamapSregarding the initialize keystone stuff..07:48
SpamapSkind of feels like that should be built into keystone-manage07:48
SpamapSbut... finally... time to go to the airport07:48
*** jistr has joined #tripleo07:48
*** jcooley_ has joined #tripleo07:49
lifelessSpamapS: yeah, we need to talk to uhm $name about that07:49
lsmolaSpamapS: yeah, that might be the long term placement07:49
lsmolalifeless: lol07:49
lsmolalifeless: do you see this landing in Icehouse timeframe?07:50
saurabhslifeless: I am trying to understand how is metadata, cloudinit and userdata is related.07:50
lifelesslsmola: we can do the common lib pretty quickly07:50
saurabhsWhen I am trying to add metadata to instance using heat template07:50
saurabhs    "Metadata" : {07:50
saurabhs       "admin-password": "Test123",07:50
saurabhs       "AWS::CloudFormation::Init" : {07:50
saurabhs         "config" : {07:50
saurabhs           "packages" : {07:50
saurabhs             "yum" : {07:50
saurabhs               "mysql"        : [],07:50
saurabhs                "mysql-server" : [],07:50
saurabhs                "httpd"        : [],07:50
saurabhs                "wordpress"    : []07:50
saurabhs              }07:50
saurabhs            }07:50
saurabhs       }07:50
lifelesslsmola: it may be heinous, but still07:50
lsmolalifeless: :-) ok I've never done it so I don't know07:50
lsmolalifeless: though I 've created plenty of gems, that was almost easy :-)07:51
lsmolalifeless: will you have time to take this one?07:51
lifelesslsmola: sure07:53
lsmolalifeless: excellent, thank you07:53
lsmolalifeless: then it will probably take some more time to rewrite it into API, cli...07:54
*** jcooley_ has quit IRC07:54
lsmolalifeless: but for now we are good with something that we can import to python and call, shellscript or whatever :-)07:54
lsmolalifeless: without rewriting it to other place07:55
lifelesssaurabhs: go on07:55
saurabhslifeless: When I am trying to add metadata to instance using heat template07:56
lifelesslsmola: I've added a card for it07:56
saurabhsit is showing up in07:56
saurabhsand not in07:56
saurabhscult curl
lsmolalifeless: excellent, thank you07:56
lifelesssaurabhs: ok, so thats the ec2 metadata server, which (today) only has things nova knows about07:57
lifelesssaurabhs: there is a separate CFN metadata url we use which is random and delivered to us by the user-data from nova07:57
lifelesssaurabhs: but also note that for golden imagse we don't choose software packages in heat templates- we do that in the image build07:58
lifelesssaurabhs: we only deliver configuration07:58
saurabhswhen I query : curl
openstackgerritandrea-frittoli proposed a change to openstack/tripleo-image-elements: Fixes links in
saurabhsit gives me
saurabhswhat is all that data the actual user_data script in the template was very small08:01
*** e0ne has joined #tripleo08:01
lifelessthats heats bootstrap code08:01
lifelessyou should basically ignore all of this :)08:01
saurabhslifeless: how to get hold of CFN metadata url08:02
lifelessthat bootstrap code writes it to disk08:03
lifeless/var/lib/cloud/ I think08:03
*** CaptTofu has joined #tripleo08:04
*** e0ne has quit IRC08:05
saurabhsits giving me : how do I get the metadata from here now08:06
lifelessuse os-collect-config or cfn-hup08:07
saurabhslifeless: are you saying that os-collect-config internally uses this url to get the data ?08:07
lifelessI haven't looked at the plumbing layer for ages08:08
*** CaptTofu has quit IRC08:09
lifelessI think its great you're interested, but you'll need to deep dive through the code - these days its just an abstraction we depend on08:09
lifelessnot something we interact with directly08:09
saurabhslifeless: going back to the os-config-refresh and os-config-apply.08:11
saurabhsif I have to use these for building trove images, do I have to write these elements seperately for trove image elements or I will have to extend these elements there.08:11
saurabhstrove will have different configuration keys and multiple configuration files, how do we plug the metadata into those files correctly08:11
lifelesssaurabhs: os-refresh-config and os-apply-config are both generic, you never change them. You use them.08:12
lifelesssaurabhs: have you looked into them yet?08:13
*** cd-undercloud has joined #tripleo08:15
cd-undercloud************** overcloud complete status=1 ************08:15
*** cd-undercloud has quit IRC08:15
rpodolyakamorning tripleo08:15
saurabhslifeless: I was going though os-apply-config earlier. I think the use of this is to read the parameters from metadata and write that into the configs right?08:16
saurabhsdid I interpret that incorrectly?08:17
*** bauzas has joined #tripleo08:18
lifelessit does two things08:18
lifelessa) it gives you a language (moustache) to capture metadata into config files08:18
lifelessso you got that08:18
lifelessand b) it gives a command line query language for selecting specific values out of the metadata08:19
saurabhsso now what to read and where to write, how do we control that?08:19
*** shardy_afk is now known as shardy08:20
saurabhsok I guess I should aks how to I use os-apply-config to control what to read and where to write? can you point me to a sample?08:20
jprovaznlifeless, ping08:22
lifelessjprovazn: pong08:23
lifelesssaurabhs: in general, you could do worse than looking around in that git repo and the tripleo-heat-templates repo08:24
jprovaznlifeless, Hi, myaybe you might help me with one issue - when defining a list of nodes for rabbit mq cluster, I also define "master" node - to distinguish first node in cluster which doesn't try to connect to others and which other nodes will try to connect to. Disadvantage of this is that it's "extra" setting for one specific node in heat template08:26
jprovaznlifeless, which makes scaling conf worse08:26
lifelessjprovazn: is that a rabbit requirement ?08:26
jprovaznrabbit requirement is to start first node in a different way than others08:27
jprovaznlifeless, ^08:27
lifelessjprovazn: what happens when that node crashes, or if we have to reboot it?08:27
lifelessor is this a one-time setup thing?08:28
jprovaznlifeless, one-time08:28
lifelessso after that, when we restart it, it has the same config file as the others08:29
jprovaznlifeless, when we restart the "master" node? yes - then it remembers that it was joined in a cluster before08:30
lifelessok, so what greghaynes is proposing for galera is to use the $0 node to do that init.08:30
lifelessjprovazn: how is this extra setting expressed to rabbit ?08:31
lifelesslike, is it command line, or a config file option, or a client command you run ?08:31
*** saju_m has quit IRC08:31
jprovaznlifeless, well, both is possible but I use command line08:32
lifelessok, so whats the extra setting ?08:33
lifelessin the heat template?08:33
jprovaznlifeless, the "master" node08:33
greghaynesIn the xtradb cluster its a config file setting - you start the first node without specifying rest of the cluster and then once others join you change config file to match rest of cluster incase of restart08:34
jprovazngreghaynes, you can use an extra param (mysqld pxc) to start the first node - to avoid editing config file twice08:35
lifelessjprovazn: so, is the question 'how does a node know it is/is not the initial master' ?08:35
lifelessjprovazn: could you just see if its hostname ends in [^\d]0 ?08:35
jprovaznlifeless, no, question is: is there a better way how to recognize that some node is "first" to avoid defining the master node08:35
greghaynesI do like the idea of being explicit about which is master rather than implicit, the 0th node being a more implicit system08:36
lifelessI'd like to actually run paxos to select a node08:36
lifelessbut maybe I'm a little nuts08:36
greghaynesYou might be the only one of us that actually read a paxos whitepaper :)08:36
greghaynesalthough maybe thats not a bad idea08:37
*** al has quit IRC08:37
lifelessit would mean that we don't try to start the cluster we have a quorum at least transiently available08:37
lifelesswhich I think is actually a nice property08:37
saurabhslifeless: I understood os-apply-config. also as per my understanding os-collect-config only triggeres os-refresh is that right? or it does something more than that?08:37
*** al has joined #tripleo08:37
jprovaznlifeless, greghaynes: well, I could check for "[^\d]0" if it would be preferred way08:37
lifelessjprovazn: do you have a sane answer in the templates today ?08:38
greghaynesoh, I was arguing against that, it was just an idea I had for now08:38
lifelesssaurabhs: it queries all the metadata sources (ec2, cfn, local file, possibly chef and or puppet) and provides all of their metadata for the node in a cache to o-r-c08:38
saurabhsand then os-refresh-config provides that a way of running predefined scripts. is that right?08:40
lifelesssaurabhs: yes08:40
jprovaznlifeless, I don't know about any existing param in our templates which could help, for this reason I explicitly add ""master" node, I was just wondering if there is a better way08:40
lifelessjprovazn: ok so heres a thought08:41
lifelesslets make an abstract interface08:41
lifelessa 5 line shell script for now08:41
*** jcooley_ has joined #tripleo08:41
lifelessos-am-i-master <servicename>08:41
lifelessit should look at the hostname for now08:41
lifelessdo the 'ends in 0 but not [0-9]0 :)08:42
*** pblaho has joined #tripleo08:42
lifelesswe can loop back and make it fancy without having to rejigger a bunch of elements08:42
jprovaznlifeless, ok08:42
lifelessbecause having a special node in scaling groups will be uhm, hard.08:42
lifelessand I don't think we should add a dependency on unique attributes for that reason08:43
saurabhslifeless: so os-collect-config triggers os-refresh-config but then who triggeres os-apply-config? I mean if there is an change in metadata, and it refers to one of the keys in the configs the config will be update by os-apply-config right? if so somebody has to trigger it08:43
*** panda has joined #tripleo08:43
lifelessbut I don't like the hostname trick, so I want to know precisely where it lives so we can take it out back and sacrifice it when the time is right08:43
lifelesssaurabhs: look at elements/os-apply-config/os-refresh-config/configure.d/50-os-config-applier in the tripleo-image-elements tree08:44
*** jang1 has quit IRC08:46
saurabhslifeless: thanks. I will trouble you again tomorrow with further set of questions08:47
lifelesssaurabhs: tomorrow I'll be spending time with family - saturday :)08:47
lifelesssaurabhs: but there are plenty of folk here that live in the past08:48
lifelesssaurabhs: so just ask here, I'm reasonably sure someone will chime in08:48
*** e0ne has joined #tripleo08:48
saurabhssure thanks08:48
*** saurabhs has quit IRC08:48
*** saurabhs has joined #tripleo08:49
*** jcooley_ has quit IRC08:51
*** saurabhs has left #tripleo08:52
lifelessthinking out loud about the init problem08:52
lifelesswhat we really want is some sort of distributed while true: try:init and break; except:pass08:52
lifelesse.g. paxos to pick someone to do init; wait while they do it and if they go away go back to paxos to find a new leader, keep doing that until08:53
lifelessthe init reports completed, then get additional nodes replicated, and if that process fails due to the master going away, go back to paxos again, and loop...08:54
lifelesswipe all trace of dbs whenever someone gets elected08:54
*** jtomasek has joined #tripleo08:56
*** pblaho has quit IRC08:56
*** akrivoka has joined #tripleo08:57
lifelessoh cool08:59
*** saju_m has joined #tripleo09:01
*** saju_m has quit IRC09:02
lifelessSpamapS: ^ when you get off your plane; new shiny09:05
lifelessLicense: 3-Clause BSD09:05
*** vkozhukalov has joined #tripleo09:07
rdopieralskiI didn't know someone still uses 3-clause one09:07
lifelessrdopieralski: thats the stock BSD :)09:08
*** derekh has joined #tripleo09:10
lifelessderekh: hai09:17
derekhlifeless: hi09:18
lifelessderekh: we're back with CI09:19
lifelessderekh: overcloud jobs have the timeout increased but failingm not sure why09:19
lifelessderekh: I merged your log change but suggested an alternate implementation in the review09:19
derekhlifeless: looking at it at the moment, it looks like its a problem getting the log info09:19
lifelessderekh: I'm thinking the next stage is to run tempest against the overcloud job09:19
lifelessderekh: from the jenkins slave09:20
*** saju_m has joined #tripleo09:20
derekhlifeless: ok, we could add that tempest element I created to the overcloud and run it09:21
derekhlifeless: Although it hasn't been used in a long time so may need some updates09:22
derekhlifeless: I see a fedora template09:23
derekhlifeless: is that just left over after some tests?09:23
derekhmust be : | OS-SRV-USG:launched_at               | 2014-02-14T19:44:09.000000                               |09:23
lifelessderekh: no element needed09:23
lifelessderekh: run it like d-g does from the slave09:23
derekhlifeless: k09:24
*** martyntaylor has joined #tripleo09:24
openstackgerritA change was merged to openstack/tuskar: Fix default path in tools/
lifelessderekh: we should be able to reuse a bunch of d-g code09:26
lifelessthat way09:26
derekhlifeless: ok09:26
openstackgerritA change was merged to openstack/tuskar: Removed unnecessary nova client
jistrlsmola: a nit - please give also +2 when giving Approved. Patches should have at least two +2 reviews before going approved. E.g. this patch has only one +2 review then in "Notes":
*** pbelanyi has joined #tripleo09:30
lsmolajistr: yeah I do that randomly09:30
lsmolajistr: cause giving just approved is safer, in case I have misread it and there is no +209:30
*** cd-undercloud has joined #tripleo09:30
cd-undercloud************** overcloud complete status=1 ************09:30
*** cd-undercloud has quit IRC09:30
lsmolajistr: if I give both, I can approve it all by myself09:31
lsmolajistr: :-)09:31
lsmolajistr: I did that by mistake once09:31
jistrlsmola: i think best is not to misread and make sure there are two +2 as the OpenStack rules say :)D09:31
lsmolajistr: yeah, prerequisite for the is having a good sleep :-D09:32
jistryeah :D09:32
*** lucasagomes has joined #tripleo09:33
*** dtantsur has joined #tripleo09:33
openstackgerritDerek Higgins proposed a change to openstack-infra/tripleo-ci: Write host_info.txt with "sudo dd"
derekhlifeless: lets see if that sorts out the log issue .... in 2 hours :-(09:39
*** max_lobur_afk is now known as max_lobur09:43
*** dtantsur has quit IRC09:46
*** jcooley_ has joined #tripleo09:47
lifelessderekh: heh09:48
*** dtantsur has joined #tripleo09:48
derekhlifeless: adding idempotentcy to test env scripts,09:48
lifelessderekh: I've read your etherpad on test optimisation, have some thoughts but haven't serialised them yet09:48
lifelessderekh: awesome!09:48
lifelessderekh: while you're there, you might like to start calling to make the environment09:49
lifelessderekh: its intended for reuse09:49
lifelessI need to halt() - gnight!09:49
derekhlifeless: will do, night09:50
*** jcooley_ has quit IRC09:52
*** akrivoka has quit IRC09:56
*** CaptTofu has joined #tripleo10:05
*** CaptTofu has quit IRC10:09
*** derekh has quit IRC10:27
*** derekh has joined #tripleo10:31
*** jp_at_hp has joined #tripleo10:34
openstackgerritJon-Paul Sullivan proposed a change to openstack/diskimage-builder: Set noclobber around element consolidation
*** jcooley_ has joined #tripleo10:41
*** akrivoka has joined #tripleo10:44
*** athomas has joined #tripleo10:45
*** coolsvap has quit IRC10:45
*** jcooley_ has quit IRC10:46
openstackgerritRadomir Dopieralski proposed a change to openstack/tuskar-ui: Actually get node profiles from the API
*** jcooley_ has joined #tripleo10:47
* jistr relocates10:48
*** jistr has quit IRC10:48
*** jcooley_ has quit IRC10:52
*** coolsvap has joined #tripleo10:52
*** bauzas has quit IRC10:53
*** athomas has quit IRC10:54
*** cd-undercloud has joined #tripleo10:56
cd-undercloud************** overcloud complete status=1 ************10:56
*** cd-undercloud has quit IRC10:56
*** CaptTofu has joined #tripleo10:57
*** nosnos has quit IRC10:58
*** geerdest has joined #tripleo11:01
*** matsuhashi has quit IRC11:02
*** athomas has joined #tripleo11:03
*** jistr has joined #tripleo11:05
*** vkozhukalov has quit IRC11:08
*** bauzas has joined #tripleo11:19
*** vkozhukalov has joined #tripleo11:20
*** coolsvap has quit IRC11:28
jp_at_hpSpamapS: Can you re-review please?11:31
*** jcooley_ has joined #tripleo11:48
openstackgerritJon-Paul Sullivan proposed a change to openstack/diskimage-builder: Enable custom apt.conf in apt-conf element
*** jcooley_ has quit IRC11:53
*** jcoufal has joined #tripleo11:54
*** cd-undercloud has joined #tripleo12:00
cd-undercloud************** overcloud complete status=1 ************12:00
*** cd-undercloud has quit IRC12:00
*** dshurtle has joined #tripleo12:32
*** vkozhukalov has quit IRC12:39
*** cd-undercloud has joined #tripleo12:48
cd-undercloud************** overcloud complete status=1 ************12:48
*** cd-undercloud has quit IRC12:48
*** e0ne_ has joined #tripleo12:49
*** cd-undercloud has joined #tripleo12:49
cd-undercloud************** overcloud complete status=6 ************12:49
*** cd-undercloud has quit IRC12:49
*** weshay has joined #tripleo12:51
*** slagle has joined #tripleo12:51
*** e0ne has quit IRC12:53
*** vkozhukalov has joined #tripleo12:55
*** cd-undercloud has joined #tripleo13:03
cd-undercloud************** overcloud complete status=128 ************13:03
*** cd-undercloud has quit IRC13:03
*** bauzas has quit IRC13:03
openstackgerritDerek Higgins proposed a change to openstack-infra/tripleo-ci: Gracefully handle commands that outlive test env
openstackgerritDerek Higgins proposed a change to openstack-infra/tripleo-ci: Increase the testenv-worker timeout
openstackgerritDerek Higgins proposed a change to openstack-infra/tripleo-ci: Don't truncate the logged json
derekhAnybody looking at ci jobs for testenv overcloud will see it complete but jenkins thinks it failed should fix it13:04
*** mrunge has quit IRC13:06
slaglederekh: small nit there in the comment for 7540213:10
*** cd-undercloud has joined #tripleo13:10
cd-undercloud************** overcloud complete status=128 ************13:10
*** cd-undercloud has quit IRC13:10
*** julim has joined #tripleo13:13
*** julim has quit IRC13:14
openstackgerritRadomir Dopieralski proposed a change to openstack/tuskar-ui: Actually get node profiles from the API
openstackgerritRadomir Dopieralski proposed a change to openstack/tuskar-ui: Update the deployment design page
*** CaptTofu has quit IRC13:19
openstackgerritRadomir Dopieralski proposed a change to openstack/tuskar-ui: Actually get node profiles from the API
*** jdob has joined #tripleo13:24
*** bauzas has joined #tripleo13:25
*** max_lobur is now known as max_lobur_afk13:41
openstackgerritJames Slagle proposed a change to openstack-infra/tripleo-ci: Gracefully handle commands that outlive test env
slaglederekh: do you know how we trigger a new check job for things like: ?13:43
*** cd-undercloud has joined #tripleo13:44
cd-undercloud************** overcloud complete status=100 ************13:44
*** cd-undercloud has quit IRC13:44
*** jcooley_ has joined #tripleo13:47
derekhslagle: try "recheck no bug", although I'm not 100% sure with the new queues that we added last night13:47
derekhslagle: thanks for the fixup13:47
derekhslagle: actually13:48
derekhslagle: maybe unapprove and approve again, I'll do it since you were author13:49
openstackgerritA change was merged to openstack/tripleo-image-elements: Fix neutron package install
slagleoh, there it goes :)13:49
slagleah, you +A'd again. thx13:50
derekhslagle: yup, np13:50
*** lblanchard has joined #tripleo13:52
*** jcooley_ has quit IRC13:52
SpamapSand one more airport13:53
* SpamapS is at LHR13:53
openstackgerritDerek Higgins proposed a change to openstack-infra/tripleo-ci: Don't truncate the logged json
openstackgerritDerek Higgins proposed a change to openstack-infra/tripleo-ci: Increase the testenv-worker timeout
*** dprince has joined #tripleo13:55
openstackgerritLadislav Smola proposed a change to openstack/tuskar-ui: Showing configuration tab of deployed overcloud
*** saju_m has quit IRC14:10
dprincederekh: re: Increase the testenv-worker timeout (review 75403). Does that mean the tests take 3 hours?14:10
derekhdprince: the overcloud test takes just over 2 hours, so the jenkins timeout is 2hrs 25 minutes
*** julim has joined #tripleo14:12
derekhdprince: the timeout on the testenv-work MUST be longer then the jenkins timout if its not we get false negative results and risk14:12
derekhdprince: having 2 jobs using the same test env at the same time14:13
derekhdprince: this handles the false negatives or at least we'll know once its finished running14:13
dprincederekh: yep. that is a super long test cycle... but we'll take what we can get I guess.14:14
derekhdprince: and thios ensures we don't doubley allocate the test env
dprincederekh: be interesting to see if things run faster on our rack once it is up and running.14:14
derekhdprince: but a new test env will need to be built for the TE worker change to take effect14:14
derekhdprince: yes it is super long and will be interesting14:15
derekhdprince: (and anybody else who has been following), we have plans in place to speed this up a bit,
derekhdprince: we have to decide on the best option, I've jotted down my ideas there so we can assess the options properly14:17
dprincederekh: So each image gets 4MB of RAM right now? Could we tinker with increasing that... and decreasing the number of workers per server?14:25
openstackgerritLadislav Smola proposed a change to openstack/tuskar-ui: Showing configuration tab of deployed overcloud
derekhdprince: I think most of the time is spent building the images on the jenkins node so I don't think tinkering with the test env will help much14:27
dprincederekh: okay14:28
* dprince looks some more14:28
derekhdprince: looks like the test for that testenv-clinet change failed
derekhlooks unrelated14:28
*** noslzzp has joined #tripleo14:29
derekhdprince: actually it restarted it on another jenkins
dprincederekh: magical14:30
openstackgerritA change was merged to openstack/diskimage-builder: Add 00-fedora-fixup-vim
dprincederekh: I think our biggest bang might be to go for shared venvs then14:32
dprincederekh: which is what devstack does14:32
*** akrivoka has quit IRC14:33
dprincederekh: seems like we are spending a bunch of time compiling the same old things in our DIB images...14:33
dprincederekh: so if we shared venvs... or better yet pre-created a beefed up venv which would get used by all-the-apps then we'd save a good bit of time I think on the image building...14:34
openstackgerritRadomir Dopieralski proposed a change to openstack/tuskar-ui: Actually get node profiles from the API
openstackgerritRadomir Dopieralski proposed a change to openstack/tuskar-ui: Update the deployment design page
dprincederekh: or we could just use real packages which solve this problem nicely as well :)14:34
*** dmitry_tantsur has joined #tripleo14:35
derekhdprince: yup, it would, I was about to suggest a little mad, we could find common lines in all the requirements files, build a venv with common entries and then copy that as a base for the others, so we still avoid conflicting requirments14:35
*** akrivoka has joined #tripleo14:36
derekhdprince: I might give that a go and see if it works14:36
derekhdprince: yup, packages would do it too :-)14:36
*** jprovazn has quit IRC14:37
*** dtantsur has quit IRC14:38
dprincederekh: I would say we can do whatever devstack does here and get by w/ it. If devstack uses a shared venv then we should be able to as well.14:38
derekhdprince: so what happens when one project has a conflicting requirement with another ?14:39
dprincederekh: there is a global requirements file now as well so that should help to slow down the adoption of random new versions of things for each project14:39
*** jtomasek has quit IRC14:39
derekhdprince: looks like it could still have problems at a glance in various projects we have14:41
derekhand lots or other example of differences14:41
dprincederekh: with WebOb, perhaps. The others should be retaining backwards compat.14:41
dprincederekh: plus it isn't like distros do that w/ packages.14:41
dprincederekh: stable clients should be deprecating things so there is an upgrade path...14:42
dprincederekh: so in general we'd use the newest version I think and it should work14:42
derekhdprince: agreed they should be14:42
derekhdprince: this handles the false negatives or at least we'll know once its finished running14:42
*** hewbrocc` is now known as hewbrocca14:43
derekhdprince: ^^ pasted from history ignore14:43
*** ramishra_ has joined #tripleo14:43
*** ramishra has quit IRC14:43
*** markmc has joined #tripleo14:43
*** matty_du1s|gone is now known as matty_dubs14:48
*** afrittoli has joined #tripleo14:48
*** jcooley_ has joined #tripleo14:48
*** max_lobur_afk is now known as max_lobur14:49
*** jprovazn has joined #tripleo14:50
*** CaptTofu has joined #tripleo14:50
*** jcooley_ has quit IRC14:54
*** jtomasek has joined #tripleo14:55
*** jprovazn has quit IRC14:58
*** slagle has quit IRC15:05
openstackgerritandrea-frittoli proposed a change to openstack/tripleo-image-elements: Fixes links in
*** jcooley_ has joined #tripleo15:09
*** coolsvap has joined #tripleo15:10
*** akuznetsov has quit IRC15:19
*** akrivoka has quit IRC15:41
*** spzala has joined #tripleo15:41
*** lazy_prince has quit IRC15:44
*** max_lobur is now known as max_lobur_afk15:45
openstackgerritMichael Kerrin proposed a change to openstack/tripleo-image-elements: Make the pypi mirror configurable via a DIB_PYPI_INDEX_URL
openstackgerritMichael Kerrin proposed a change to openstack/tripleo-image-elements: Install latest version of pip and virtualenv
*** tserong has quit IRC15:56
*** jcoufal has quit IRC15:57
*** jcoufal has joined #tripleo15:57
*** tserong has joined #tripleo15:59
*** jcooley_ has quit IRC16:00
*** jcooley_ has joined #tripleo16:01
*** spzala has quit IRC16:02
derekhci jobs failing now because nova isn't ready (not keystone as I originally thought), its almost as if the "nova-manage db sync" in os-refresh-config is taking longer then usual16:03
*** ramishra_ has quit IRC16:03
*** bauzas has quit IRC16:05
*** tserong has quit IRC16:05
*** jcooley_ has quit IRC16:06
*** coolsvap has quit IRC16:08
openstackgerritRalf Haferkamp proposed a change to openstack/tripleo-incubator: Use baremetal element to extract kernel and initrd for undercloud
*** coolsvap1 has joined #tripleo16:08
*** coolsvap1 has quit IRC16:08
*** slagle has joined #tripleo16:09
*** jcooley_ has joined #tripleo16:10
*** coolsvap1 has joined #tripleo16:14
openstackgerritRalf Haferkamp proposed a change to openstack/diskimage-builder: Update openSUSE package mapping for libffi-dev
*** CaptTofu has quit IRC16:16
*** jcoufal has quit IRC16:16
*** jtomasek has quit IRC16:17
*** Lotus907efi_ has quit IRC16:19
*** dmitry_tantsur has quit IRC16:26
*** CaptTofu has joined #tripleo16:30
*** tserong has joined #tripleo16:34
*** tserong has joined #tripleo16:34
*** rdopieralski has quit IRC16:34
*** jistr has quit IRC16:37
*** rha has quit IRC16:43
*** jcooley_ has quit IRC16:44
*** jcooley_ has joined #tripleo16:44
*** al has quit IRC16:45
*** al has joined #tripleo16:45
*** max_lobur_afk is now known as max_lobur16:54
*** jcooley_ has quit IRC16:57
*** jcooley_ has joined #tripleo16:58
*** matty_dubs is now known as matty_dubs|lunch16:58
*** ccrouch has quit IRC16:59
*** saurabhs has joined #tripleo17:01
*** jcooley_ has quit IRC17:02
*** derekh has quit IRC17:04
*** akuznetsov has joined #tripleo17:04
*** jp_at_hp has quit IRC17:07
*** cwolferh has joined #tripleo17:14
*** epim has joined #tripleo17:16
*** jcooley_ has joined #tripleo17:17
*** epim_ has joined #tripleo17:19
*** e0ne_ has quit IRC17:19
*** ccrouch has joined #tripleo17:19
*** e0ne has joined #tripleo17:19
*** lsmola has quit IRC17:20
*** epim has quit IRC17:20
*** epim_ is now known as epim17:20
*** jcooley_ has quit IRC17:21
*** newell has joined #tripleo17:21
*** spzala has joined #tripleo17:22
*** jcooley_ has joined #tripleo17:22
*** e0ne has quit IRC17:24
*** jcooley_ has quit IRC17:24
*** markmc has quit IRC17:25
*** geerdest has quit IRC17:29
*** jog0 is now known as flashgordon17:30
*** jcooley_ has joined #tripleo17:31
*** rha has joined #tripleo17:32
*** jcooley_ has quit IRC17:32
*** akuznetsov has quit IRC17:34
*** cadenzajon has joined #tripleo17:39
*** vkozhukalov has quit IRC17:40
*** epim has quit IRC17:48
*** akuznetsov has joined #tripleo17:49
*** matty_dubs|lunch is now known as matty_dubs17:51
*** max_lobur is now known as max_lobur_afk17:55
*** wendar has quit IRC17:57
*** pbelanyi has quit IRC17:59
openstackgerritTzu-Mainn Chen proposed a change to openstack/tuskar: Remove object storage from initial data
*** kfox1111 has joined #tripleo18:02
openstackgerritBen Nemec proposed a change to openstack/diskimage-builder: Add RHEL 7 image element
openstackgerritBen Nemec proposed a change to openstack/diskimage-builder: Generalize install-packages for yum
openstackgerritBen Nemec proposed a change to openstack/diskimage-builder: Add redhat-common element
openstackgerritBen Nemec proposed a change to openstack/diskimage-builder: Use redhat-common in fedora element
kfox1111I'm following: just built a seed vm. launched it. need to get in to configure it. don't know the pw. do I need to root it or is there a default?18:04
kfox1111the shadow file has no pw entires. guess I did have to root it.18:11
*** marun has joined #tripleo18:12
*** akuznetsov has quit IRC18:14
*** akuznetsov has joined #tripleo18:17
*** cadenzajon has quit IRC18:21
NobodyCamStevenK: happen to be around?18:22
*** lucasagomes has quit IRC18:24
NobodyCamkfox1111: did you include the stack user element18:24
*** UtahDave has joined #tripleo18:24
kfox1111no. the doc didn't say to.18:25
kfox1111that element lets you configure the user?18:25
NobodyCamthat element add a user "stack" with same pass18:26
kfox1111do you have the ability to amend the doc?18:27
NobodyCamkfox1111: yes and so do you... would need a review put up for
*** cadenzajon has joined #tripleo18:28
NobodyCamthou I use: this Doc:
NobodyCamwhich now seem to have broken links18:29
kfox1111how do I do the review?18:29
NobodyCamhave you used git?18:30
*** cody-somerville has joined #tripleo18:30
*** cody-somerville has joined #tripleo18:30
kfox1111I was afraid you were going to say that. :/ My company has been fighting through the CLA issues since September. :/18:31
kfox1111still an indeterminate amount of time left before I can contribute. :/18:31
NobodyCamit is requied18:31
NobodyCambut that is the work flow18:31
kfox1111yeah. I'm going to avoid creating any more patches until the layers can get their acts together.18:31
kfox1111I already have three patches waiting since September.18:32
NobodyCam:) :-p18:32
kfox1111they are all public domain, and prior art up the wazoo, but everyone's scared of being sued.18:32
kfox1111one of the patches is a sort comparater function. :/ Sort this way instead of that. its depressing.18:33
NobodyCamlol yea, but the laywers got get paid too :-p18:34
kfox1111I dont mind them getting paid. I just mind them taking more then 4 months to let me contribute a patch that took litterally 1 hour to write.18:34
kfox1111so, I rooted the seed. Its up and running.18:35
kfox1111no nova services running.18:35
NobodyCamare then installed in /opt/stack?18:35
kfox1111yeah. lots of stuff.18:36
kfox1111does the vm need an eth0 and eth1? I only have an eth0 at the moment and looks like upstart may not be starting things cause of it?18:37
NobodyCamyour logs should be in /var/log/upstart18:37
lifelessthe seed needs eth1 yes18:38
lifelessbut its a vm, should have been created that way automatically18:38
kfox1111network is showing its trying to dhcp off of eth0. I don't have it setup that way. do I make that a nat interface?18:39
kfox1111I manually created the libvirt xml file. didn't see an example in the doc.18:39
NobodyCamkfox1111: have you seen the script?18:40
kfox1111no. was trying to hack things into a sort of working environment. wasn't sure things would work if I didn't manually configure some stuff.18:41
kfox1111thats in incubation repository?18:41
NobodyCamyea -
kfox1111I have an older box hosting the seed. I had to qemu-img convert the image just to get it to boot. I built it on a newer ubuntu box.18:42
NobodyCamI recamend starting with a clean slate! the devtest script should get you up and running from a clean start18:43
NobodyCammorning lifeless18:44
NobodyCamhey quick question ... what time zone is StevenK in?18:44
kfox1111yeah. I was trying to get someting up and working quickly to kick some tires.18:45
lifelessNobodyCam: UTC+10[11]18:45
lifelessNobodyCam: 11 right now18:46
kfox1111I've got access to a bit of old hpc hardware for just a little bit.18:46
NobodyCamahh :) :)18:46
kfox1111maybe I should switch gears and try:
*** dprince has quit IRC18:46
NobodyCamkfox1111: I have not tested that at all18:46
lifelesskfox1111: the script setup-seed-vm will setup a seed vm correctly for you and you can then copy your disk image over the top18:47
lifelesskfox1111: but, there have been lots of changes and bugfixes to nova etc; you may find yourself debugging issues that were solved weeks ago18:47
kfox1111so... more complicated then that. My laptop where i can build images is on the internet and far away from the rack of nodes I can do testing with.18:47
kfox1111the rack of hardware in  question is on a private only network so I can't build things there since all the tripleo scripts assume they can pull lots of random things in.18:48
lifelesskfox1111: thats ok, the USE_CACHE=1 flag will tell the scripts not to build images18:48
kfox1111we usually use rpm mirroring to work around that.18:48
kfox1111ah. ok.18:48
lifelesswe don't yet have a good 'and these are the files you need to copy across' thing though18:48
lifelesswe do have some build here, run-there support which we use for tripleo-ci18:49
lifelessso perhaps you would have success taking that and improving it for baremetal18:49
openstackgerritDustin J. Mitchell proposed a change to openstack/diskimage-builder: order phase subdirs by execution order, add subheadings, fix formatting
kfox1111I think I'm going to give the rdo prebuilt images a shot real quick. would help me ensure I have a working hardware/vm setup without having to worry about ensuring all the vm images are built properly too.18:51
kfox1111hmm.. maybe not.18:52
kfox1111the instructions look like everything may be virtual.18:53
matty_dubskfox1111: Let me see if I can find the instructions -- that's the default, but it doesn't have to be all virtual18:53
kfox1111if just the overcloud was physical that would work.18:53
matty_dubsOh wow, the devtest.html page's formatting is all screwed up18:55
NobodyCammatty_dubs: ya18:55
kfox1111matty_dubs: thanks.18:56
matty_dubsSo I can't find my notes. I've gone through this guide (but without broken formatting) and just used physical machines instead of VMs.18:57
matty_dubsI thought I had a bookmark with exactly what I did, but no such luck :(18:58
kfox1111no worries. thanks for checking.19:01
kfox1111no major problems though?19:01
matty_dubsNo. The whole thing does take a long time, but it all worked last time I ran through it.19:02
kfox1111cool. thanks. :)19:02
*** athomas has quit IRC19:02
NobodyCamkfox1111: when i was testing on real hardware i recall having to replace the DIB vm element with baremetal element... i think i recall that any way.19:04
NobodyCamthere may be other elements like that too.. its been awhile sense i've had real hardware19:05
kfox1111hmm.. so I might have to hack an image slightly to get it to boot.19:05
kfox1111was kind of expecting that anyway. booting on real hardware is always touchy... :)19:06
NobodyCamthings changing the19:06
kfox1111this hardware has those weird cciss raid stuff.19:06
NobodyCamVirtualPowerManager to IPMI and such19:06
kfox1111ah. yeah.19:07
*** athomas has joined #tripleo19:10
*** wendar has joined #tripleo19:11
kfox1111hmm.. that doc says to look at instead.19:13
kfox1111looks good but its again pointing to go git pull stuff. :/19:13
kfox1111I guess I can pull on one box and copy the whole tree over.19:13
NobodyCamrsync is your friend19:14
kfox1111If I were going to do it alot, I could probably set it up so I could git clone it, and push from it to the isolated box.19:16
kfox1111gits very nice that way. I'm so glad to be mostly off of svn now.19:16
matty_dubsA previous employer tried to switch from svn to git, and I pushed back because it seemed really complicated.19:17
matty_dubsI was _sooo_ wrong.19:17
matty_dubsWell, totally right about it being complicated, but totally wrong about it not being worth it.19:18
kfox1111ah. so this set is putting a generic set of binaries in the image, and doing all the config with configdrive. clever.19:18
kfox1111yeah. the few extra steps seems like it would be slowing you down all the time. but way worth it. :)19:18
kfox1111I was extatic the first time I was able to do a partial single file commit.19:19
NobodyCami liked bazzar19:19
kfox1111didn't play too much with bazar but it seemed nice.19:19
NobodyCamwe should all just go mack to M$ sourceSafe19:20
NobodyCamand never git anything done again19:20
kfox1111you say that in jest, but there still are folks around that... ;)19:21
* NobodyCam was old school foxpro... and has spent many many hours fighting with / fixing / cursing M$ SS19:22
cody-somervilleNobodyCam: I liked bzr as well. However, I used it the other day and was like "wow. I forgot how slow this really was".19:22
kfox1111I'm sorry. :)19:22
* kfox1111 wonders what will replace git some day19:23
kfox1111cvs was so much better then rcs. svn was so much better then cvs. git is so much better then svn...19:24
*** vkozhukalov has joined #tripleo19:25
NobodyCamcody-somerville: ya.. its been quite awhile sense i've use bzr19:27
*** UtahDave has quit IRC19:29
*** panda___ has joined #tripleo19:31
*** kfox1111 has quit IRC19:34
*** panda has quit IRC19:34
*** kfox1111 has joined #tripleo19:35
*** hashar has joined #tripleo20:12
kfox1111ok. got the rdo undercloud vm image booted. :)20:14
NobodyCamkfox1111: :) awesome ness20:16
NobodyCamhumm devtest currently has not idea of what the virt_type is for the nodes its is managing20:18
*** afrittoli has quit IRC20:18
kfox1111just gota get the nodes enrolled now...20:23
saurabhsI have a doubt around building tripleO image. As per the instruction from
saurabhsTo build a tripleO image I can use20:24
saurabhsdiskimage-builder/bin/disk-image-create -u base vm bootstrap local-config stackuser heat-cfntools -a i386 -o bootstrap20:24
saurabhswant to understand where does this hooks up os-config-apply, os-config-refresh, os-collect-config?20:25
saurabhsdo I have to explicitly mention every element I want to include their or these 3 always get included whether we specify it or not20:25
greghaynesIs there a fix for up for review somewhere?20:27
*** cadenzajon has quit IRC20:29
NobodyCamsaurabhs: is bootstrap still a valid element?20:30
NobodyCamthat may be a old create line20:30
saurabhsNobodyCam: looks like it give me error20:32
saurabhsERROR: Element 'bootstrap' not found in 'tripleo-image-elements/elements:diskimage-builder/bin/../elements'20:32
saurabhsI will replace it with ubuntu20:32
saurabhsbut I am not able to figure out if I have specify os-confg-apply explicitly or they get included by default20:32
bnemecsaurabhs: os-refresh-config is listed in element-deps for the elements that use it.  Or should be anyway.20:33
NobodyCamsaurabhs: you may want to give this script a read thru:
NobodyCamnote the elements it is including20:34
*** ccrouch1 has joined #tripleo20:40
slagleanyone know what's up with the devtest formatting:
slaglei generated locally, and they look ok20:41
greghaynesslagle: how do you gen docs localy?20:42
slagleactually, i should probably blow away my tox venvs and try again20:42
slagletox -e venv python build_sphinx20:42
slaglegreghaynes: ^20:42
slaglethen they are under doc/build20:42
*** ccrouch has quit IRC20:43
slagleyea, still look ok locally20:43
*** ekarlso has quit IRC20:50
*** ekarlso has joined #tripleo20:50
*** martyntaylor has quit IRC20:58
*** e0ne has joined #tripleo20:59
*** cadenzajon has joined #tripleo21:13
*** e0ne has quit IRC21:16
*** e0ne has joined #tripleo21:18
*** noslzzp has quit IRC21:20
*** e0ne has quit IRC21:22
*** noslzzp has joined #tripleo21:24
greghaynesI first noticed the doc thing last night, and I dont think anything has merged to t-i since then....21:25
tchaypoI noticed that about a week ago, been meaning to look at it on monday21:33
tchaypoit looks like there's a bit of awk that parses a script ( and pulls out the comments, and it's not taking out the ## at the start of the comments21:34
greghaynesodd because it seems to work fine locally21:35
tchaypo is the script that does the extracting21:36
tchaypoIt's broken for me locally :)21:36
greghaynesyou have a useful setup :)21:36
tchaypolet me check that it hasn't mysteriously fixed itself..21:39
tchaypostill broken.21:40
tchaypobut I am being called away. I'll dig into it some more later on. I can't see any recent changes to the extract-docs, extract-docs.awk,and the changes to seem innocuous21:44
tchayposo I'm suspect awk or sphinx version changes21:44
*** jcooley_ has joined #tripleo21:50
kfox1111ok. finally got the bare metal nodes enroled. just gota upload an image and give it a try. :)21:57
*** jdob has quit IRC22:00
*** lblanchard has quit IRC22:03
*** matty_dubs is now known as matty_dubs|gone22:05
*** hashar has quit IRC22:09
*** CaptTofu has quit IRC22:10
*** CaptTofu has joined #tripleo22:12
kfox1111does the console work with nova bare metal?22:18
kfox1111Can an admin get to the console without booting an instance?22:18
NobodyCamie console before there anything to connect it to?22:20
*** cwolferh_ has joined #tripleo22:22
*** cwolferh has quit IRC22:25
kfox1111I know some of these nodes probably don't have hardware virtualization turned on in the bios.22:27
kfox1111I guess I could just boot an instance and console attach right away if that works.22:27
NobodyCamkfox1111: I believe console does work ( i have not tested it ) but have heard many a tail of others using it22:29
kfox1111k. thx.22:29
kfox1111do you need to use a kernel/initrd seperate or does a regular grub/linux kerel boot work?22:30
NobodyCamits pxe booting each time not just for the deploy22:32
kfox1111ah. ok. so the deploy kernel/image are for getting the image onto the node, not for booting the node.22:35
kfox1111or both? Does it reboot after it puts the image on the node, booting local?22:36
*** julim has quit IRC22:36
lifelesskfox1111: ot netboots every time22:38
lifelesskfox1111: deploy is one boot, run is a second boot with the users provided kernel + ramdisk22:38
kfox1111can the user not provide a kernel/ramdisk but just use whats in the image?22:40
lifelessnot today22:40
lifelessbut the user typically pulls their kernel/ramdisk out of the image22:40
lifelessso same same22:40
kfox1111k. so I'll have to extract it then. Thanks.22:41
lifelesskfox1111: load-image will do that for you (and the baremetal element does it at image build time for you)22:42
NobodyCamha lifeless is faster then i I was going to reffer you to
kfox1111ah. ok.22:43
*** edmund has quit IRC22:44
kfox1111ah. I see load-image calling out to disk image builder's script to pull the kernel bits. yeah.22:46
*** CaptTofu has quit IRC22:54
*** ccrouch has joined #tripleo22:58
*** jcooley_ has quit IRC22:59
*** ccrouch1 has quit IRC23:00
kfox1111where does tripleo put log files?23:09
lifelessrunning ubuntu? /var/log/upstart/23:10
kfox1111it a fedora build.23:12
kfox1111ah. so systemd.23:12
kfox1111its just throwing things in /var/log/messages.23:18
kfox1111ok. trying to boot a vm gets no valid host found...23:19
kfox1111ah. looks like maybe x86_64 vs amd64 difference somewhere...23:20
lifelessnova.conf has an arch value for nova baremetal23:20
lifelessthe extra specs on the flavor have to match23:20
lifelessand the servicehost on the node has to match the nova-compute agent name23:21
lifelessunless both match it won't advertise the node to the scheduler23:21
lifelessnova hypervisor-stats lets you see whats advertised without poking around log files23:21
kfox1111ah. cool. thanks.23:21
NobodyCamnova hypervisor-stats nice /me way way to use to check the log23:22
NobodyCamTY lifeless23:23
kfox1111ok. changing nova.conf to match x86_64 got farter.23:25
kfox1111looks like it wont make any further progress until I extract the kernel/initrd from the image. was hoping I could atleast get to the bios first. oh well.23:26
lifelessthere's also nova host-list and nova host-descriobe23:26
lifelesskfox1111: we use amd64 everywhere FWIW23:26
lifelesskfox1111: consistency.23:26
kfox1111 theres a step configure-vm '--arch x86_64'23:27
kfox1111I thought that was what ARCH was set to. I guessed wrong I guess.23:28
lifelessall our tooling uses amd6423:34
lifelessslagle has documented a different key there23:34
lifelessI presume its slagle anyhow, he did a bunch of nice tooling support for RHer's getting started with this23:34
lifelessI'd prefer to see that also say amd64, just to avoid confusion like you have just experienced.23:35
kfox1111going to be confusion both ways I'm afraid. the debian/ubuntu guys use amd64. the redhatish folks use x86_64 all over the place. its really odd to me to see amd64, though it makes sense.23:39
kfox1111odd... Error: Image 7ea-d86e-4174-b696-19c56a191c87 could not be found.23:44
kfox1111there is no 7ea as far as I can tell, so its right...23:45
*** sdake_ has joined #tripleo23:45
*** sdake_ has quit IRC23:45
*** sdake_ has joined #tripleo23:45
kfox1111do I have to restart something after I upload the deploy images to glance?23:46
*** athomas has quit IRC23:46
NobodyCamjust deploy with that image23:48
openstackgerritGregory Haynes proposed a change to openstack/tripleo-image-elements: Use xtradb-cluster in place of mysql
kfox1111does it make up a uuid for the customized image? is it complaining about that, or is it saying the deploy image it cant find or something?23:52
NobodyCamoh you did remove the other images23:53
NobodyCami'll bet23:53
NobodyCamglance image-list23:53
NobodyCamglance image-delete UUID23:53
kfox1111deleting and booting the vm clames it cant find 7ea-d86e-4174-b696-19c56a191c8723:54
kfox1111so, same image...23:54
kfox1111hmm.. the kerenl image I uploaded is "639197ea-d86e-4174-b696-19c56a191c87"23:56
kfox1111bare metal deploy kernel...23:56
kfox1111flavor metadata maybe?23:56
kfox1111yeah. the string is truncated there.23:57
NobodyCamkfox1111: ya extra specs23:58
NobodyCamon the flavor23:58
kfox1111there we go....23:59

Generated by 2.14.0 by Marius Gedminas - find it at!