*** manous has quit IRC | 00:00 | |
*** sdake_ has joined #rdo | 00:06 | |
*** stanchan has quit IRC | 00:07 | |
*** sdake has quit IRC | 00:09 | |
*** nixpenguin has quit IRC | 00:14 | |
*** nixpenguin has joined #rdo | 00:14 | |
*** aortega has joined #rdo | 00:17 | |
*** sarath has quit IRC | 00:19 | |
*** limao has joined #rdo | 00:25 | |
*** chlong has quit IRC | 00:32 | |
*** stanchan has joined #rdo | 00:33 | |
*** mburned is now known as mburned_out | 00:34 | |
*** dmacpher has joined #rdo | 00:43 | |
*** rhallisey has joined #rdo | 00:43 | |
*** aortega has quit IRC | 00:46 | |
*** maeca1 has quit IRC | 00:47 | |
*** stanchan has quit IRC | 00:54 | |
*** gildub has joined #rdo | 01:05 | |
*** ayoung has joined #rdo | 01:07 | |
*** jcoufal has joined #rdo | 01:13 | |
ggillies | pabelanger: ping | 01:50 |
---|---|---|
zodbot | ggillies: Ping with data, please: https://fedoraproject.org/wiki/No_naked_pings | 01:50 |
rdobot | [sensu] NEW: master.monitoring.rdoproject.org - check-delorean-newton-head-current @ http://tinyurl.com/hc82n3z |#| Build failure on centos7-master-head/current: nova: http://trunk.rdoproject.org/centos7-master-head/report.html | 01:50 |
ggillies | pabelanger: is there any plans for openstackid.net to expand beyond offering oauth2 | 01:51 |
ggillies | such as openidc connect or saml or something | 01:51 |
rdobot | [sensu] RESOLVED: master.monitoring.rdoproject.org - check-delorean-newton-head-current @ http://tinyurl.com/hc82n3z |#| No build failures detected: http://trunk.rdoproject.org/centos7-master-head/report.html | 01:53 |
*** fzdarsky_ has joined #rdo | 01:56 | |
*** fzdarsky has quit IRC | 01:59 | |
dmsimard | ggillies: I've seen ipsilon (same thing as fedora account system) mentioned somewhere | 01:59 |
dmsimard | https://fedorahosted.org/ipsilon/ | 02:00 |
dmsimard | https://id.fedoraproject.org/ | 02:00 |
dmsimard | i.e, to help replace launchpad and integrate openstackid | 02:01 |
*** sdake has joined #rdo | 02:03 | |
*** sdake_ has quit IRC | 02:04 | |
*** rhallisey has quit IRC | 02:05 | |
ggillies | dmsimard: interesting | 02:05 |
ggillies | just looking at federation options for RDO Cloud because I really don't want to be stuck doing password management :V | 02:05 |
ggillies | hmmm not sure if I am a fan of ipsilons "lets just be this intermediary wrapped in a black box installer for you" | 02:06 |
*** myoung is now known as myoung|afk | 02:06 | |
*** mburned_out is now known as mburned | 02:08 | |
*** myoung|afk is now known as myoung | 02:10 | |
*** fragatina has joined #rdo | 02:12 | |
*** stanchan has joined #rdo | 02:13 | |
*** sdake_ has joined #rdo | 02:14 | |
*** danielbruno has quit IRC | 02:16 | |
*** sdake has quit IRC | 02:17 | |
*** jhershbe has joined #rdo | 02:17 | |
*** ashw has joined #rdo | 02:21 | |
*** maeca1 has joined #rdo | 02:22 | |
*** stanchan has quit IRC | 02:24 | |
*** fragatina has quit IRC | 02:27 | |
*** julim has joined #rdo | 02:28 | |
dmsimard | ggillies: chat with fungi about it | 02:32 |
dmsimard | ggillies: in #openstack-infra | 02:32 |
dmsimard | ggillies: anything would be better than trystack's facebook auth :D | 02:33 |
dmsimard | and looks like our friends ayoung and richm have already discussed keystone/ipsilon http://lists.openstack.org/pipermail/openstack-dev/2015-June/066856.html | 02:35 |
ayoung | dmsimard, RHSSO is going to be based on Keycloak. I am actively testing it right now | 02:35 |
pabelanger | ggillies: IIRC, we are planning on implementing ipsilon upstream. As is supports more backends | 02:36 |
*** crossbuilder has joined #rdo | 02:36 | |
*** crossbuilder_ has quit IRC | 02:36 | |
*** jhershbe has quit IRC | 02:41 | |
ggillies | dmsimard: where are the links to DLRN repos that have passed CI? I'm not seeing them on http://trunk.rdoproject.org/ | 02:43 |
ayoung | Ipsilon is the Fedora AS replacement, and it is pretty nice and light weight. DOes not do OpenIDC yet last I checked. Keycloak is a hell of a lot more full featured, which is a mixed blessing. It does both SAML2 and OpenIDC, with OIDC being the first and better supported. | 02:44 |
ayoung | And it is the basis for RHSSO, so we kindof have to make sure it works | 02:44 |
ggillies | ayoung: from a very quick glance | 02:45 |
ggillies | it looks like it's based off you having to run these "magic scripts" (ipsilon-client-install) on hosts to set it up | 02:45 |
ayoung | ggillies, no magic. Just python | 02:45 |
ggillies | it is magic | 02:45 |
ggillies | that pattern in software is cancer | 02:45 |
ggillies | I don't need magic scripts or buttons | 02:45 |
ayoung | ggillies, most of it is configuring HTTPD | 02:45 |
*** nyechiel has joined #rdo | 02:45 | |
ggillies | yes that | 02:46 |
ggillies | that is wrong | 02:46 |
ayoung | it is the complexity of SAML that calls for it | 02:46 |
ggillies | *I* want to configure httpd based off how I need | 02:46 |
ggillies | through my configuration management engine | 02:46 |
ggillies | no | 02:46 |
ggillies | people *think* it's too complicated, but it's not, the doco just sucks | 02:46 |
ayoung | ggillies, I think you are reading it backwards | 02:46 |
ayoung | you can configure SAML or OIDC however you want | 02:46 |
ayoung | these scripts are designed to do it the least surpsing way | 02:46 |
ggillies | right | 02:47 |
ggillies | so is ipsilon just providing a saml endpoint I can integrate to | 02:47 |
ggillies | or openidc, or oauth, or whatever the flavor of the month is | 02:47 |
ayoung | ggillies, I'd watch the word "just" there | 02:47 |
ayoung | SAML is like short term X509 | 02:48 |
ayoung | with all the key signing and 2 way auth | 02:48 |
ayoung | done in the HTTP layer. I'd be much happier with X509 client Auth for the most part, or Kerberos, but those suck for other reasons. | 02:48 |
ggillies | yes | 02:49 |
ggillies | but all of those are possible for me to configure and integrate with using my configuration management system of choice | 02:49 |
ayoung | But what the scripts do is not magic. They make a call to the IdP and registers the service provider, fetch the files from the IdP, and generate the local files to get them signed | 02:49 |
ggillies | that is configuration management | 02:50 |
ayoung | ggillies, I'm actually doing somethig like that right now | 02:50 |
ggillies | I have an ecosystem to solve that problem | 02:50 |
ggillies | but developers keep thinking they have to solve this problem over and over and keep writing their own systems to do it for their software | 02:50 |
ggillies | often poorly | 02:50 |
*** rasca has quit IRC | 02:50 | |
ayoung | ggillies, so the way I see this working is that for Tripleo, we would run this on the undercloud, cache the fiels in a tarball, and delivber them to the keystone nodes | 02:50 |
*** rasca has joined #rdo | 02:50 | |
ggillies | no thankyou | 02:51 |
ayoung | ggillies, with HA, you need to make sure the three (or more) nodes all respond as the same SP | 02:51 |
ggillies | I can deliver it to the nodes through heat/puppet and the appropriate hooks | 02:51 |
ggillies | right | 02:51 |
ayoung | ggillies, yes, that is what I mean | 02:51 |
ayoung | Heat will do the delivery | 02:52 |
ayoung | but the files will be pre-created | 02:52 |
ggillies | that's fine, but I want it all done through the configuration management and remote execution frameworks I have | 02:52 |
ayoung | Heat has the ability to deliver the tarball...however, I think it might deliver it too early. We need to make sure, at a minimum, that mod_auth_mellon or mod_auth_oidc is installed in the webserver before we try to run | 02:53 |
ayoung | ggillies, that is a give | 02:53 |
ayoung | given | 02:53 |
ayoung | ggillies, part of my problem right now is that Heat does not provide an easy way to run pre-execution scripts to generate things like this prior to an openestack server install. Is there a way to do that, or do I need to wait for Mistral? | 02:54 |
ayoung | I'd rather not have to do it manually | 02:54 |
ggillies | there are hooks when the server is booted before any further configuration is done | 02:54 |
ayoung | Nope too late | 02:55 |
ayoung | I need it on the Heat server prior to running on the booted node | 02:55 |
ggillies | why should it be done on the undercloud | 02:55 |
ggillies | also anything like x509 certs and related | 02:55 |
ggillies | I want to generate and store in my config management system so I can track them | 02:55 |
ggillies | I don't want them generated on the fly at deploy time | 02:56 |
ggillies | that's a horrible idea | 02:56 |
ayoung | THe registration of Identity needs to be done by a trusted server | 02:56 |
*** steveg_afk has quit IRC | 02:57 | |
ayoung | ggillies, of course they get stored in the config management system | 02:57 |
*** Tenhi has quit IRC | 02:57 | |
ayoung | but I don't want to generate them on one node, them pull them into the CMS to be stuck on another | 02:57 |
ggillies | right I'm saying I would pregenerate and then inject into my deploy | 02:59 |
ggillies | which I think is what you were saying? | 02:59 |
ayoung | ggillies, yep | 02:59 |
*** Tenhi has joined #rdo | 03:00 | |
ayoung | ggillies, I've just started getting my head around how Heat and the os-*-config programs work to drive puppet. | 03:00 |
ayoung | I was doing Ansible all last year, and the mapping is kindof hard from one to the other. | 03:00 |
ayoung | But I think I get it now. I was able to work through the SSL deploy steps with FQDN earlier today | 03:01 |
ggillies | I'm actually working on an example now I hope to post publically (groundwork for RDO Cloud) that shows using tripleo for full static ip assignment for all interfaces on all networks, static assignment of vips, and 1:1 node placement | 03:02 |
ggillies | the step after that is federation integration | 03:02 |
ggillies | hence my question to begin with | 03:02 |
ayoung | ggillies, I should have something for you with Federation integration shortly | 03:03 |
ayoung | Help getting it production ready would be greatly appreciated | 03:03 |
ayoung | We had a proof of concept of Ipsilon with Packstack back in Tokyo. The trick is mapping those steps to Tripleo, and cleaning up some of the demo-isms | 03:03 |
ggillies | I've done most the work I've needed already in puppet-keystone | 03:04 |
ggillies | it's just a bunch of stuff passed as an environment template | 03:04 |
ggillies | then some extra steps to upload the mapping etc | 03:04 |
ayoung | ggillies, have you submitted it to puppet-keystone, at least as a WIP? I'd like to look at it. | 03:04 |
ggillies | ayoung: yeah it's been merged | 03:05 |
ggillies | the first bit | 03:05 |
ggillies | https://github.com/openstack/puppet-keystone/blob/master/manifests/federation/openidc.pp | 03:05 |
ggillies | it isn't the complete set, you still need to run some keystone commands | 03:06 |
ayoung | # The allowed values are: 'external', 'password', 'token', 'oauth1', 'saml2' | 03:06 |
ayoung | Heh | 03:06 |
ggillies | but I'm pondering the best way to automate that with puppet-keystone at the moment | 03:06 |
ayoung | I think we allow oidc now too | 03:06 |
ggillies | woops | 03:06 |
ggillies | you can put in anything there | 03:06 |
ggillies | you can put in asdf1234 | 03:06 |
ggillies | as long as you have a section for it to map to | 03:07 |
ayoung | The way we configure auth mechanisms in Keystone is too centralized. It should be on the protocol URL, not one for the whole server | 03:07 |
ayoung | yeah, they all use the mapped Keystone auth plugin | 03:07 |
ggillies | yeah fair enough | 03:07 |
ggillies | yeah | 03:08 |
ggillies | it seems like a lot of this federation stuff is in flux | 03:08 |
ggillies | like slowly getting reworked | 03:08 |
ggillies | but I dunno | 03:08 |
ayoung | ggillies, well, I should be working on it, but Learning Tripleo has eaten my time for this release | 03:08 |
*** steveg_afk has joined #rdo | 03:09 | |
ayoung | I don't think any radical changes are queued up for Federation, though. I just think a lot of people have not tried to really make it work yet. We are hitting some of the snarls | 03:09 |
ggillies | :/ | 03:10 |
ayoung | The puppet code was obviously written with only K2K federation in mind. There was some work done on untangling that | 03:10 |
ggillies | It seems keystone to keystone is a lot stronger | 03:10 |
ggillies | yeah | 03:10 |
*** Son_Goku has quit IRC | 03:11 | |
ayoung | Anyway, with Keycloak/RHSSO, we should have a single IdP that can handle either Saml2 or OIDC | 03:11 |
ayoung | I've not put a lot of time into OIDC, as the demand from our side has been on SAML. | 03:12 |
ayoung | Good to know that you;'ve been beating on it. | 03:12 |
*** Son_Goku has joined #rdo | 03:12 | |
ayoung | There was far more interest in OIDC at the last summit than SAML | 03:12 |
ayoung | ggillies, https://github.com/openstack/puppet-keystone/blob/master/manifests/federation/openidc.pp#L23 is it necessary to have secrets stored in the Hiera data? That really is poor security hygene. | 03:18 |
ayoung | We are starting to get beat up on compliance issues...rightfully so, too. | 03:19 |
*** ashw has quit IRC | 03:25 | |
*** ayoung is now known as ayoung_ZZZzzz | 03:26 | |
*** Tenhi has quit IRC | 03:27 | |
ggillies | unfortunately we haven't got much of a choice | 03:29 |
ggillies | and what are the compliance issues your talking about? | 03:29 |
ayoung_ZZZzzz | Key management for the most part. PCI for passwords, etc. | 03:30 |
ayoung_ZZZzzz | ggillies, Gotta crash. Thanks for leetting me know what you are working on, I'll catch up and sync with you once I have something functional. | 03:31 |
ggillies | so is the problem unencrypted at rest? | 03:31 |
*** julim has quit IRC | 03:31 | |
*** Tenhi has joined #rdo | 03:35 | |
*** vimal has joined #rdo | 03:36 | |
*** stanchan has joined #rdo | 03:41 | |
*** sdake_ has quit IRC | 03:42 | |
*** ohochman has left #rdo | 03:46 | |
*** CustosL1men has quit IRC | 03:47 | |
*** dmacpher has quit IRC | 03:48 | |
*** CustosL1men has joined #rdo | 03:48 | |
*** chlong has joined #rdo | 03:49 | |
*** JonSnowcl has joined #rdo | 03:49 | |
*** dmacpher has joined #rdo | 03:50 | |
*** JonSnowcl has quit IRC | 03:50 | |
*** richm has quit IRC | 03:51 | |
*** stanchan has quit IRC | 03:52 | |
*** sdake has joined #rdo | 03:57 | |
*** jdob has quit IRC | 04:01 | |
*** jcoufal has quit IRC | 04:02 | |
*** rlandy has quit IRC | 04:08 | |
*** nyechiel has quit IRC | 04:10 | |
*** vimal has quit IRC | 04:12 | |
*** sdake has quit IRC | 04:14 | |
*** numans has joined #rdo | 04:16 | |
*** chandankumar has joined #rdo | 04:22 | |
*** sdake has joined #rdo | 04:24 | |
*** fragatina has joined #rdo | 04:27 | |
*** fragatina has quit IRC | 04:31 | |
*** limao has quit IRC | 04:31 | |
*** limao has joined #rdo | 04:32 | |
*** anilvenkata has joined #rdo | 04:32 | |
*** vimal has joined #rdo | 04:32 | |
*** dhill_ has quit IRC | 04:38 | |
*** jaosorior has joined #rdo | 04:45 | |
*** jaosorior has quit IRC | 04:51 | |
*** nehar has joined #rdo | 04:52 | |
*** jaosorior has joined #rdo | 04:52 | |
*** Poornima has joined #rdo | 04:53 | |
*** jaosorior has quit IRC | 04:54 | |
*** jaosorior has joined #rdo | 04:55 | |
*** hynekm has joined #rdo | 04:55 | |
*** eliska has joined #rdo | 05:09 | |
*** jaosorior has quit IRC | 05:14 | |
*** jaosorior has joined #rdo | 05:15 | |
*** dmacpher has quit IRC | 05:15 | |
*** stanchan has joined #rdo | 05:16 | |
*** maeca1 has quit IRC | 05:17 | |
*** jaosorior has quit IRC | 05:17 | |
*** dmacpher has joined #rdo | 05:18 | |
*** jaosorior has joined #rdo | 05:18 | |
*** rcernin has joined #rdo | 05:31 | |
*** livelace has joined #rdo | 05:34 | |
*** links has joined #rdo | 05:36 | |
*** limao_ has joined #rdo | 05:42 | |
*** nmagnezi has joined #rdo | 05:44 | |
*** limao has quit IRC | 05:45 | |
*** mcornea has joined #rdo | 05:47 | |
*** pgadiya has joined #rdo | 05:57 | |
*** ganesh has joined #rdo | 06:01 | |
*** ganesh is now known as Guest65384 | 06:02 | |
*** Guest65384 is now known as gkadam | 06:02 | |
*** chlong has quit IRC | 06:08 | |
*** aufi has joined #rdo | 06:09 | |
*** coolsvap has joined #rdo | 06:14 | |
*** jhershbe has joined #rdo | 06:23 | |
*** pnavarro has joined #rdo | 06:35 | |
*** stanchan has quit IRC | 06:36 | |
*** chlong has joined #rdo | 06:40 | |
*** danpawlik has joined #rdo | 06:44 | |
*** jtomasek has joined #rdo | 06:44 | |
*** chlong has quit IRC | 06:52 | |
*** kaminohana has quit IRC | 06:55 | |
*** Poornima has quit IRC | 06:56 | |
*** nyechiel has joined #rdo | 06:56 | |
*** jhershbe has quit IRC | 06:59 | |
*** pnavarro has quit IRC | 06:59 | |
*** nyechiel has quit IRC | 07:03 | |
*** belmoreira has joined #rdo | 07:03 | |
*** satya4ever has joined #rdo | 07:04 | |
*** dgurtner has joined #rdo | 07:05 | |
*** mosulica has joined #rdo | 07:06 | |
*** gildub has quit IRC | 07:07 | |
*** Poornima has joined #rdo | 07:12 | |
*** saneax_AFK is now known as saneax | 07:12 | |
*** stanchan has joined #rdo | 07:12 | |
*** nyechiel has joined #rdo | 07:13 | |
*** paramite has joined #rdo | 07:14 | |
*** nyechiel has quit IRC | 07:15 | |
*** nyechiel has joined #rdo | 07:17 | |
*** nyechiel has quit IRC | 07:23 | |
*** panda|zz is now known as panda | 07:24 | |
*** rdas has joined #rdo | 07:24 | |
*** jpena|off is now known as jpena | 07:26 | |
jpena | good morning! | 07:27 |
jaosorior | jpena: Hey; yesterday a pin was done to mistral due to an error in it? Is it still in place? The fix just landed | 07:28 |
jpena | jaosorior: we ended up not doing it, apevec wanted to give the fix some time to land (https://review.rdoproject.org/r/1792) | 07:28 |
flepied | good morning! | 07:30 |
*** stanchan has quit IRC | 07:31 | |
flepied | jpena: do you know where is the jjb repo for the RDO promotion CI? | 07:31 |
*** gfidente has joined #rdo | 07:32 | |
jpena | flepied: I remember some talks yesterday about it using gerrithub, let me check | 07:32 |
jpena | flepied: https://github.com/rdo-infra/ci-config | 07:33 |
flepied | thx jpena | 07:33 |
*** flaper87 has joined #rdo | 07:36 | |
*** flaper87 is now known as Guest73548 | 07:36 | |
*** jpich has joined #rdo | 07:39 | |
*** dmacpher has quit IRC | 07:43 | |
*** eliska has quit IRC | 07:44 | |
*** eliska has joined #rdo | 07:51 | |
*** eliska has quit IRC | 07:55 | |
*** shardy has joined #rdo | 07:55 | |
*** eliska has joined #rdo | 07:55 | |
*** dtantsur|afk is now known as dtantsur | 07:56 | |
*** ohamada has joined #rdo | 07:58 | |
*** dgurtner has quit IRC | 07:58 | |
chandankumar | jpena, flepied number80 \o/ | 08:00 |
jpena | o/ | 08:00 |
*** migi_ is now known as migi | 08:02 | |
*** openstackgerrit has quit IRC | 08:03 | |
*** mosulica has quit IRC | 08:03 | |
*** openstackgerrit has joined #rdo | 08:04 | |
*** ChanServ sets mode: +v openstackgerrit | 08:04 | |
*** stanchan has joined #rdo | 08:04 | |
*** jhershbe has joined #rdo | 08:05 | |
*** Amita has joined #rdo | 08:08 | |
*** pkovar has joined #rdo | 08:09 | |
*** leanderthal|afk is now known as leanderthal | 08:09 | |
*** jaosorior has quit IRC | 08:13 | |
*** shardy has quit IRC | 08:14 | |
*** jaosorior has joined #rdo | 08:14 | |
*** toanju has joined #rdo | 08:14 | |
*** shardy has joined #rdo | 08:14 | |
amoralej | good morning | 08:14 |
*** nmagnezi has quit IRC | 08:15 | |
*** tesseract- has joined #rdo | 08:16 | |
*** Goneri has joined #rdo | 08:16 | |
jpena | welcome back amoralej! | 08:18 |
amoralej | thanks jpena!!! | 08:18 |
number80 | amoralej: hey, welcome back :) | 08:19 |
amoralej | o/ | 08:19 |
chandankumar | amoralej, \o/ | 08:23 |
amoralej | hi chandankumar!! | 08:23 |
*** jhershbe has quit IRC | 08:25 | |
*** pnavarro has joined #rdo | 08:27 | |
*** lucas|dinner is now known as lucasagomes | 08:28 | |
*** stanchan has quit IRC | 08:31 | |
*** sasha2 has quit IRC | 08:31 | |
*** d0ugal has quit IRC | 08:32 | |
*** d0ugal has joined #rdo | 08:33 | |
*** sasha2 has joined #rdo | 08:33 | |
*** d0ugal has quit IRC | 08:33 | |
*** d0ugal has joined #rdo | 08:33 | |
*** sdake has quit IRC | 08:37 | |
*** Guest73548 is now known as flaper87 | 08:39 | |
*** flaper87 is now known as Guest78641 | 08:39 | |
*** jlibosva has joined #rdo | 08:39 | |
*** jbadiapa has joined #rdo | 08:39 | |
*** Guest78641 has quit IRC | 08:44 | |
rdogerrit | Alan Pevec proposed rdoinfo: Add openstack-tripleo-ui http://review.rdoproject.org/r/1733 | 08:45 |
rdogerrit | Alan Pevec proposed rdoinfo: Add openstack-tripleo-validations http://review.rdoproject.org/r/1763 | 08:46 |
*** mdavidson has joined #rdo | 08:52 | |
*** pkovar has quit IRC | 08:59 | |
*** iranzo has joined #rdo | 09:06 | |
*** iranzo has quit IRC | 09:06 | |
*** iranzo has joined #rdo | 09:06 | |
*** dgurtner has joined #rdo | 09:08 | |
*** dgurtner has quit IRC | 09:08 | |
*** dgurtner has joined #rdo | 09:08 | |
rdogerrit | Merged openstack/oslo-vmware-distgit: Update dependencies http://review.rdoproject.org/r/1789 | 09:11 |
*** Goneri has quit IRC | 09:15 | |
*** mdavidson has quit IRC | 09:16 | |
*** Ryjedo_ has joined #rdo | 09:17 | |
*** Ryjedo has quit IRC | 09:17 | |
*** Ryjedo_ is now known as Ryjedo | 09:17 | |
*** Goneri has joined #rdo | 09:18 | |
social | hmm mistral db-sync is broken? | 09:27 |
* social ;.; https://paste.fedoraproject.org/401612/89254147 | 09:28 | |
*** imcleod has joined #rdo | 09:29 | |
*** abehl has joined #rdo | 09:30 | |
*** jbadiapa has quit IRC | 09:31 | |
*** degorenko has joined #rdo | 09:31 | |
*** puzzled has joined #rdo | 09:31 | |
d0ugal | social: Yeah, I think it is | 09:31 |
*** iberezovskiy has joined #rdo | 09:32 | |
*** dgurtner has quit IRC | 09:33 | |
d0ugal | social: From #openstack-mistral earlier today. http://paste.openstack.org/show/550203/ | 09:34 |
d0ugal | I think that is related anyway... | 09:34 |
*** iranzo has quit IRC | 09:35 | |
*** paragan has quit IRC | 09:36 | |
*** iranzo has joined #rdo | 09:37 | |
social | d0ugal: yes it seems to be the issue | 09:38 |
flepied | adarazs: I started a pad to keep track of the best practices around debugging: https://review.rdoproject.org/etherpad/p/rdo-ci-debug-howto | 09:39 |
*** saneax is now known as saneax_AFK | 09:39 | |
adarazs | flepied: ack, nice. | 09:39 |
*** flaper87 has joined #rdo | 09:40 | |
*** flaper87 has quit IRC | 09:40 | |
*** flaper87 has joined #rdo | 09:40 | |
*** iranzo has quit IRC | 09:40 | |
*** iranzo has joined #rdo | 09:41 | |
*** iranzo has quit IRC | 09:41 | |
*** iranzo has joined #rdo | 09:41 | |
*** iranzo has quit IRC | 09:41 | |
*** limao_ has quit IRC | 09:41 | |
rdogerrit | Fabien Boucher created config: Activate experimental CBS non-scratch build in the gate pipeline http://review.rdoproject.org/r/1801 | 09:44 |
lucasagomes | (not sure if I should ask here but...) I have a BZ to include a new package for ironic-staging-drivers hanging for a while. Any chance someone can take a look at https://bugzilla.redhat.com/show_bug.cgi?id=1312328 when you find the right time? Thanks | 09:45 |
openstack | bugzilla.redhat.com bug 1312328 in Package Review "New Package: openstack-ironic-staging-drivers" [Unspecified,New] - Assigned to hguemar | 09:45 |
*** jbadiapa has joined #rdo | 09:45 | |
*** amoralej is now known as amoralej|brb | 09:46 | |
chandankumar | lucasagomes, I am doing a initial review of the package | 09:46 |
lucasagomes | chandankumar, cool, thank you :-) | 09:46 |
*** saneax_AFK is now known as saneax | 09:48 | |
*** iranzo has joined #rdo | 09:49 | |
*** paragan has joined #rdo | 09:51 | |
*** paragan has joined #rdo | 09:51 | |
*** nyechiel has joined #rdo | 09:51 | |
*** Amita has quit IRC | 09:53 | |
*** chandankumar_ has joined #rdo | 09:54 | |
*** chandankumar has quit IRC | 09:54 | |
*** jbadiapa has quit IRC | 09:54 | |
*** dgurtner has joined #rdo | 09:59 | |
*** chandankumar_ is now known as chandankumar | 09:59 | |
*** jcoufal has joined #rdo | 10:00 | |
jpena | apevec: would it be a good time now to purge the old centos-master instance in the dlrn server? It'll probably take a long time, not sure if leaving it running all over the weekend would be best | 10:02 |
*** amoralej|brb is now known as amoralej | 10:06 | |
*** nyechiel has quit IRC | 10:08 | |
rdogerrit | Javier Peña created rdo-infra/puppet-dlrn: Make cron schedule for workers configurable http://review.rdoproject.org/r/1802 | 10:10 |
rdogerrit | Javier Peña created rdo-infra/puppet-dlrn: Use virtualenv --system-site-packages http://review.rdoproject.org/r/1803 | 10:16 |
rdogerrit | Javier Peña proposed rdo-infra/puppet-dlrn: Use virtualenv --system-site-packages http://review.rdoproject.org/r/1803 | 10:16 |
*** alexismonville has joined #rdo | 10:21 | |
*** akrivoka has joined #rdo | 10:23 | |
*** Goneri has quit IRC | 10:29 | |
*** Goneri has joined #rdo | 10:31 | |
*** imcleod has quit IRC | 10:40 | |
*** tosky has joined #rdo | 10:43 | |
*** chandankumar has quit IRC | 10:48 | |
*** chandankumar has joined #rdo | 10:51 | |
*** dgurtner has quit IRC | 10:56 | |
*** alexismonville has quit IRC | 10:56 | |
gfidente | hey guys, looks like we're unable to build python-tripleoclient DLRN | 10:58 |
gfidente | as apparently some of the deps added in https://review.rdoproject.org/r/#/c/1756/5 are missing | 10:58 |
gfidente | anybody willing to help? I am unsure why some of those should be missing and also unsure about in which repo | 10:59 |
jpena | gfidente: link? | 10:59 |
gfidente | https://review.rdoproject.org/r/#/c/1756/5 | 10:59 |
gfidente | ah sorry, https://review.openstack.org/#/c/348412/8 | 10:59 |
gfidente | jpena, thanks | 10:59 |
dmsimard | amoralej: woohoo, welcome back | 11:01 |
dmsimard | jpena: not sure you saw this before you left yesterday | 11:01 |
*** jcoufal has quit IRC | 11:01 | |
dmsimard | jpena: but why is -head so much faster than s/-head// ? | 11:01 |
jpena | gfidente: the error happened yesterday, while we were switching repos. Let me check if the missing packages are there now, just in case, before rechecking | 11:02 |
amoralej | thanks dmsimard | 11:04 |
jpena | dmsimard: that's what I was wondering yesterday | 11:05 |
*** rhallisey has joined #rdo | 11:07 | |
dmsimard | bahahaha.. troubleshooting an issue with a browser dev and the guy assumed I was on Windows | 11:09 |
jpena | gfidente: I think the issue is in the repos the CI job is using for dependencies while building python-tripleoclient. I see mock is using the following repo configuration: https://paste.fedoraproject.org/401650/95392147/ | 11:09 |
dmsimard | BLASPHEMY | 11:10 |
jpena | it's finding some packages (such as python-cliff) because they are available from the openstack-newton repo at buildlogs, but the failing ones are not, because they're at the DLRN repos | 11:10 |
dmsimard | flepied: did you get the answers you needed for the rdo ci promotion ? | 11:11 |
jpena | gfidente: where is the job expecting the dlrn packages to be present? | 11:11 |
*** lucasagomes is now known as lucas-hungry | 11:11 | |
gfidente | jpena, sec I think I can get the .repo files from the tripleo-ci | 11:12 |
flepied | dmsimard: I'm trying to collect the best practices here: https://review.rdoproject.org/etherpad/p/rdo-ci-debug-howto . if you can add some content for the weirdo jobs, it could be cool. | 11:12 |
dmsimard | flepied: I have something basic for weirdo jobs, I'll put it there, sec | 11:13 |
*** pnavarro is now known as pnavarro|lunch | 11:13 | |
*** spr1 has quit IRC | 11:19 | |
*** fultonj has joined #rdo | 11:19 | |
*** alexismonville has joined #rdo | 11:24 | |
*** nyechiel has joined #rdo | 11:26 | |
*** florianf has joined #rdo | 11:27 | |
*** limao has joined #rdo | 11:29 | |
*** nehar has quit IRC | 11:32 | |
Jehane | hello | 11:32 |
*** nyechiel has quit IRC | 11:32 | |
*** limao has quit IRC | 11:32 | |
Jehane | I'm still fighting with keystone. | 11:32 |
Jehane | When trying to install a qualif platform (I have a free server who wasn't used) we are hitting https://bugzilla.redhat.com/show_bug.cgi?id=1330289 | 11:33 |
openstack | bugzilla.redhat.com bug 1330289 in openstack-packstack "Failure to install Controller/Network&&Compute Cluster on RDO Mitaka with keystone API V3" [High,Post] - Assigned to jpena | 11:33 |
*** imcleod has joined #rdo | 11:34 | |
Jehane | is the up-to-date package not (yet) pushed ? | 11:34 |
*** gszasz has joined #rdo | 11:36 | |
jpena | Jehane: I see the latest openstack-puppet-modules build on http://buildlogs.centos.org/centos/7/cloud/x86_64/openstack-mitaka/ dates back to early May, so the official packages are not updated | 11:36 |
jpena | you can use the updated packages from RDO Trunk stable/mitaka in the meantime | 11:36 |
jpena | those would be https://trunk.rdoproject.org/centos7-mitaka/current/openstack-packstack-8.0.0-0.20160708211308.af375b0.el7.centos.noarch.rpm, https://trunk.rdoproject.org/centos7-mitaka/current/openstack-packstack-puppet-8.0.0-0.20160708211308.af375b0.el7.centos.noarch.rpm and https://trunk.rdoproject.org/centos7-mitaka/current/openstack-puppet-modules-8.1.7-0.20160728171754.56f1812.el7.centos.noarch.rpm | 11:37 |
Jehane | thanks, I was searching the link | 11:37 |
jpena | (openstack-packstack, openstack-packstack-puppet, and openstack-puppet-modules) | 11:37 |
Jehane | I force it manually but the version number of the trunk package is lower than the one on the centos repos | 11:43 |
Jehane | be careful with that if you're the one pushing the update | 11:43 |
jpena | mmm... that means we have something wrong with tagging, /me thinks | 11:45 |
*** paramite has quit IRC | 11:46 | |
*** dgurtner has joined #rdo | 11:47 | |
*** dgurtner has joined #rdo | 11:47 | |
Jehane | there is 0 who shouldn't be here or something close | 11:47 |
*** morazi has joined #rdo | 11:49 | |
rdogerrit | David Moreau Simard created rdo-infra/weirdo: When building docs for readthedocs, pre-fetch the roles http://review.rdoproject.org/r/1804 | 11:53 |
rdogerrit | David Moreau Simard created rdo-infra/weirdo: Remove kolla from ansible-role-requirements.yml http://review.rdoproject.org/r/1805 | 11:54 |
*** weshay has joined #rdo | 11:59 | |
Jehane | ok, new error (yet another), packstack wan't to install the non-available "python-aodhclient" rpm | 12:00 |
Jehane | I will open a bug | 12:00 |
*** Poornima has quit IRC | 12:01 | |
rdogerrit | David Moreau Simard proposed rdo-infra/weirdo: Always make sure the WeIRDO roles are there before building the docs http://review.rdoproject.org/r/1804 | 12:02 |
gfidente | jpena, so I think it's using the following | 12:02 |
gfidente | http://trunk.rdoproject.org/centos7/delorean-deps.repo | 12:02 |
gfidente | http://trunk.rdoproject.org/centos7/current/delorean.repo (as delorean-current.repo locally) | 12:02 |
rdogerrit | David Moreau Simard proposed rdo-infra/weirdo: Remove kolla from ansible-role-requirements.yml http://review.rdoproject.org/r/1805 | 12:02 |
*** toanju has quit IRC | 12:02 | |
gfidente | http://trunk.rdoproject.org/centos7/current-tripleo/delorean.repo | 12:02 |
*** Goneri has quit IRC | 12:03 | |
gfidente | jpena, is it missing something or ? | 12:03 |
jpena | gfidente: no, but... it's like it isn't using the current-tripleo repo when building. Let me check why | 12:04 |
gfidente | jpena, delorean-current you mean? | 12:06 |
gfidente | not sure what is current-tripleo? | 12:06 |
jpena | gfidente: I mean http://trunk.rdoproject.org/centos7/current-tripleo/delorean.repo | 12:06 |
gfidente | ah right it's not using that | 12:07 |
jpena | that repo contains the missing deps | 12:07 |
jpena | which repo contains tripleo.sh? | 12:07 |
*** rodrigods has quit IRC | 12:07 | |
*** rodrigods has joined #rdo | 12:08 | |
*** Son_Goku has quit IRC | 12:08 | |
*** saneax is now known as saneax_AFK | 12:08 | |
*** Son_Goku has joined #rdo | 12:09 | |
gfidente | jpena, those three | 12:09 |
jpena | ok, I think I know what it is | 12:09 |
jpena | https://github.com/openstack-infra/tripleo-ci/blob/master/scripts/tripleo.sh#L389 | 12:09 |
jpena | I thought this was fixed :? | 12:09 |
*** amoralej is now known as amoralej|lunch | 12:09 | |
jpena | the command line uses --build-env DELOREAN_DEV=1 , which relies on dlrn internals (which changed some time ago) | 12:10 |
jpena | it should be --dev instead | 12:10 |
gfidente | jpena, ack can you push a change for this? | 12:10 |
jpena | gotta run (lunch time), but I can propose a patch to tripleo-ci, then recheck your patch. Sounds like a plan? | 12:10 |
gfidente | sure, thanks | 12:10 |
*** julim has joined #rdo | 12:12 | |
*** dprince has joined #rdo | 12:13 | |
jpena | gfidente: https://review.openstack.org/351682. I'm not sure if it could break some other stuff (maybe that's why it used the env var to begin with), will check after lunch | 12:13 |
*** jpena is now known as jpena|lunch | 12:13 | |
*** Goneri has joined #rdo | 12:14 | |
*** saneax_AFK is now known as saneax | 12:16 | |
*** lmiccini_ has joined #rdo | 12:17 | |
*** lmiccini has quit IRC | 12:17 | |
*** pradk has joined #rdo | 12:18 | |
*** jayg|g0n3 is now known as jayg | 12:20 | |
*** nstrug has quit IRC | 12:24 | |
*** lucas-hungry is now known as lucasagomes | 12:24 | |
*** nstrug has joined #rdo | 12:25 | |
*** nstrug_ has joined #rdo | 12:26 | |
*** nstrug_ has quit IRC | 12:27 | |
*** ashw has joined #rdo | 12:30 | |
*** nyechiel has joined #rdo | 12:31 | |
*** mlammon has joined #rdo | 12:31 | |
social | d0ugal: I'm quite unhappy from that, isn't there rule that master never breaks? | 12:33 |
d0ugal | social: Not that I know of, you could try asking in #openstack-mistral | 12:35 |
d0ugal | social: I mean, I don't make the rules but I don't know of them being written down anywhere for the project | 12:35 |
*** chandankumar has quit IRC | 12:36 | |
*** jdob has joined #rdo | 12:36 | |
*** jhershbe has joined #rdo | 12:37 | |
*** chandankumar has joined #rdo | 12:37 | |
*** rlandy has joined #rdo | 12:43 | |
*** aortega has joined #rdo | 12:45 | |
*** jhershbe has quit IRC | 12:48 | |
*** Goneri has quit IRC | 12:48 | |
*** jmelvin has joined #rdo | 12:49 | |
*** nyechiel has quit IRC | 12:51 | |
*** jcoufal has joined #rdo | 12:52 | |
EmilienM | dmsimard: barbican is now part of scenario002 in our gate | 12:53 |
*** ccamacho is now known as ccamacho|lunch | 12:53 | |
*** jaosorior has quit IRC | 12:54 | |
dmsimard | EmilienM++ | 12:54 |
*** nyechiel has joined #rdo | 12:55 | |
EmilienM | trying to re-enable mistral again on scenario003 | 12:57 |
EmilienM | but tempest tests are super flappy | 12:57 |
*** paramite has joined #rdo | 12:59 | |
*** rbowen has joined #rdo | 13:00 | |
*** ChanServ sets mode: +o rbowen | 13:00 | |
*** stanchan has joined #rdo | 13:00 | |
EmilienM | dmsimard: what happens with designate? | 13:01 |
*** imcleod has quit IRC | 13:01 | |
*** pnavarro|lunch has quit IRC | 13:01 | |
*** stanchan has quit IRC | 13:02 | |
*** Goneri has joined #rdo | 13:03 | |
*** saneax is now known as saneax_AFK | 13:03 | |
*** nyechiel has quit IRC | 13:04 | |
weshay | adarazs, we tracking https://ci.centos.org/artifacts/rdo/jenkins-tripleo-quickstart-promote-master-delorean-minimal-458/undercloud/home/stack/undercloud_post_install.log.gz | 13:05 |
weshay | ? | 13:05 |
adarazs | weshay: I was just about to add it to the issue list :/ | 13:08 |
weshay | adarazs, rockin | 13:08 |
weshay | adarazs, hey panda and I are chatting about the instack-virt-setup blueprint | 13:08 |
weshay | do you want to join? | 13:08 |
dmsimard | EmilienM: nothing much I'm focused on other things right now | 13:10 |
dmsimard | EmilienM: but I had insomnia last night so I rebased all my p-o-i things :p | 13:10 |
EmilienM | oh ok | 13:10 |
EmilienM | but why is it failing? | 13:10 |
dmsimard | EmilienM: a bunch of random stuff, remember how it makes tempest not being able to find the aodh endpoint and mistral failures and things like that | 13:10 |
dmsimard | EmilienM: http://logs.openstack.org/60/337860/19/check/gate-puppet-openstack-integration-3-scenario003-tempest-centos-7/86a6cb4/console.html#_2016-08-05_04_55_55_969775 | 13:11 |
*** sdake has joined #rdo | 13:11 | |
adarazs | weshay: oh sorry I didn't see the notification. okay. | 13:12 |
adarazs | weshay: you're bluejeans? | 13:12 |
EmilienM | that is super wweird | 13:12 |
weshay | ya | 13:12 |
*** saneax_AFK is now known as saneax | 13:12 | |
dmsimard | EmilienM: it is | 13:13 |
dmsimard | EmilienM: it's something from designate interfering for sure, just gotta find what | 13:13 |
dmsimard | maybe something from the designate tempest plugin or I don't know | 13:13 |
dmsimard | EmilienM: it wouldn't be the first time a tempest plugin broke tempest (see: neutron monkey patching) | 13:15 |
*** nyechiel has joined #rdo | 13:15 | |
*** lmiccini has joined #rdo | 13:15 | |
hewbrocca | Yeah I think that's pretty much the expected behavior for tempest plugins, isn't it? | 13:17 |
EmilienM | dmsimard: i'll look into it | 13:17 |
*** lmiccini_ has quit IRC | 13:18 | |
*** dustins has joined #rdo | 13:18 | |
dmsimard | hewbrocca: I see what you did there :D | 13:20 |
*** amoralej|lunch is now known as amoralej | 13:20 | |
dmsimard | EmilienM: Appreciate the help. I'm fairly certain the designate tempest failures are (to some extent) legit and we'll need to troubleshoot them eventually but if we could at least not have random aodh/mistral/etc errors .. | 13:20 |
dmsimard | EmilienM: an interesting experiment would be to see if Ubuntu Cloud Archive packages the plugin and install it (to see if it would be something weird with RDO packaging) | 13:22 |
EmilienM | you're asking me to test ubuntu | 13:22 |
EmilienM | dude it's Friday, it's supposed to be a good day. | 13:22 |
dmsimard | EmilienM: Does UCA package the tempest plugins at all ? | 13:23 |
dmsimard | zigo: what about Debian ? ^ | 13:23 |
EmilienM | dmsimard: they include it in the services | 13:23 |
*** saneax is now known as saneax_AFK | 13:24 | |
dmsimard | EmilienM: oh, so you apt-get install python-tempest (or something) and the plugin is (maybe?) installed in-tree ? | 13:24 |
dmsimard | or er, python-nova | 13:24 |
dmsimard | or python-designate, whatever | 13:24 |
* dmsimard tired | 13:24 | |
EmilienM | dmsimard: yes | 13:24 |
zigo | dmsimard: I'm not testing designate in my Tempest CI yet, because it's very complicated to install. | 13:24 |
*** stanchan has joined #rdo | 13:25 | |
EmilienM | zigo: use puppet | 13:25 |
*** jpena|lunch is now known as jpena | 13:25 | |
dmsimard | zigo: it was mostly a general question around how (if at all) tempest plugins in general are packaged | 13:25 |
EmilienM | I think I told you that 92384 times | 13:25 |
zigo | dmsimard: I do package them all. | 13:25 |
*** gkadam has quit IRC | 13:25 | |
zigo | dmsimard: Though Designate is Canonical maintained, they don't sync it from Debian, so I can't tell what they do. | 13:26 |
zigo | (you'd have to check by yourself) | 13:26 |
zigo | EmilienM: I always wanted to do *both* tempest *and* packaged install. | 13:26 |
*** rdas has quit IRC | 13:26 | |
zigo | EmilienM: The packaging-deb project is finally getting into shape, with the few last infra patches going through, so we'll get there ! :) | 13:27 |
*** egafford has joined #rdo | 13:27 | |
*** ohochman has joined #rdo | 13:32 | |
*** ccamacho|lunch is now known as ccamacho | 13:33 | |
panda | pabelanger: Hi, can you send a brief email with an outline of all the changes you want to make in quickstart, and what is the long-term goal ? I'm a bit lost .. | 13:34 |
jpena | gfidente: I've updated https://review.openstack.org/351682 with what I think is the right fix. | 13:34 |
*** jeckersb_gone is now known as jeckersb | 13:34 | |
*** ohochman has quit IRC | 13:34 | |
*** ohochman has joined #rdo | 13:35 | |
*** eliska has quit IRC | 13:36 | |
*** richm has joined #rdo | 13:38 | |
Jehane | jpena: there is a small issue with the packstack rpm you show me a few hours ago | 13:38 |
*** stanchan has quit IRC | 13:38 | |
* jpena is afraid of "small issues" | 13:38 | |
*** Son_Goku has quit IRC | 13:38 | |
Jehane | for aodh, it's trying to install the rpm python-aodhclient but the package is called python2-aodhclient | 13:38 |
Jehane | jpena: easy fix but I don't know where to look in the code :) | 13:39 |
*** dcain has joined #rdo | 13:39 | |
jpena | oh damn virtul packages | 13:39 |
*** dgurtner has quit IRC | 13:39 | |
jpena | virtual, I mean | 13:39 |
Jehane | yep, that was my supposition | 13:39 |
*** stanchan has joined #rdo | 13:39 | |
pabelanger | panda: Sure, for the most part I am experimenting with some things in quickstart. End game would be to run tripleo-quickstart upstream in OpenStack, which we just landed an experimental job to help test that. | 13:40 |
*** maeca1 has joined #rdo | 13:41 | |
jpena | Jehane: the fastest fix is to install it manually, then re-run packstack with the same answer file. | 13:41 |
pabelanger | panda: but, yes, I plan to send out an email explaining some of the changes once I have examples of it working | 13:41 |
jpena | Jehane: I proposed https://review.openstack.org/336089 for a more permanent fix some time ago | 13:41 |
Jehane | jpena: already done ;) but I wanted to report it can be fixed for everyone | 13:41 |
panda | pabelanger: ok, I'll wait for rewiewing more then | 13:42 |
*** danpawlik has quit IRC | 13:42 | |
*** chandankumar has quit IRC | 13:45 | |
*** kgiusti has joined #rdo | 13:47 | |
pabelanger | panda: sure, if you want to see progress on the experimental job, you can use 'check experimental' in a quickstart review which will start the experimental job. | 13:48 |
pabelanger | panda: it won't work right now, but its the first step to running it in the gate | 13:48 |
*** jhershbe has joined #rdo | 13:49 | |
*** dcain1 has joined #rdo | 13:58 | |
*** dcain has quit IRC | 14:00 | |
*** jhershbe has quit IRC | 14:02 | |
pabelanger | is http://mirror.centos.org/centos/7/cloud/x86_64/openstack-mitaka/ == http://trunk.rdoproject.org/centos7-mitaka/current/ ? | 14:02 |
*** openstackgerrit has quit IRC | 14:03 | |
*** satya4ever has quit IRC | 14:03 | |
jpena | pabelanger: no. The first one is the official RDO repo, built from CBS. The second one is current stable/mitaka status, packaged by DLRN | 14:04 |
*** openstackgerrit has joined #rdo | 14:04 | |
*** ChanServ sets mode: +v openstackgerrit | 14:04 | |
pabelanger | jpena: okay, does dlrn eventually flow into mirror.centos.org? | 14:05 |
*** stanchan has quit IRC | 14:07 | |
jpena | pabelanger: it should, but it's a manual process. dlrn follows the git repos, and CBS uses tagged versions, but eventually any fix added to dlrn should make its way into CBS | 14:08 |
pabelanger | jpena: great, that helps. Thanks | 14:08 |
*** stanchan has joined #rdo | 14:08 | |
*** lmiccini has quit IRC | 14:09 | |
*** dgurtner has joined #rdo | 14:11 | |
*** coolsvap has quit IRC | 14:11 | |
*** dgurtner has quit IRC | 14:13 | |
*** lmiccini has joined #rdo | 14:19 | |
*** vimal has quit IRC | 14:21 | |
*** mlammon1 has joined #rdo | 14:23 | |
*** pgadiya has quit IRC | 14:25 | |
*** jeckersb is now known as jeckersb_gone | 14:26 | |
*** mlammon has quit IRC | 14:26 | |
*** stanchan has quit IRC | 14:28 | |
*** jeckersb_gone is now known as jeckersb | 14:29 | |
*** shaunm has joined #rdo | 14:29 | |
*** stanchan has joined #rdo | 14:30 | |
*** stanchan has quit IRC | 14:31 | |
*** aortega has quit IRC | 14:32 | |
*** stanchan has joined #rdo | 14:34 | |
rdogerrit | Merged rdo-infra/weirdo: Always make sure the WeIRDO roles are there before building the docs http://review.rdoproject.org/r/1804 | 14:35 |
rdogerrit | Merged rdo-infra/weirdo: Remove kolla from ansible-role-requirements.yml http://review.rdoproject.org/r/1805 | 14:35 |
*** aortega has joined #rdo | 14:36 | |
*** Goneri has quit IRC | 14:36 | |
myoung | weshay, dmsimard: it looks like some of the image gates were failing last night because the overcloud validate is pushing job times > 180 mins. Via web UI I increased the max job time --> 240 to get things unstuck & passing,but that's a temporary stop-gap measure. IMHO 3 hours is already too long. also, who pushes JJB for those jobs (trown|outtypewww)? | 14:38 |
*** pgadiya has joined #rdo | 14:39 | |
dmsimard | myoung: 3 hours is pretty damn long, yeah. What is that job doing anyway ? Just installing packages inside an image ? | 14:39 |
dmsimard | or is that the whole undercloud image thing | 14:39 |
myoung | dmsimard: no, for the image build role for tripleo-quickstart, we build an image, then verify that a basic tripleo-quickstart deploy works (basic tempest smoke, and some ping tests to make sure overcloud is actually up and operational) | 14:40 |
myoung | this is when "images-ci-check" appears in a comment (gerrit hook) | 14:40 |
dmsimard | myoung: so basically it's almost a full pipeline job | 14:40 |
dmsimard | i.e, build image, then test that image | 14:40 |
myoung | as policy for ansible-role-tripleo-image-build we like to see green gate jobs for things that affect images since breaking that is bad news :) | 14:41 |
dmsimard | that does take pretty long in the promote pielines .. | 14:41 |
dmsimard | fair, there needs to be a way to make ooo-q in pipeline faster in general | 14:41 |
myoung | the jobs used to hover around 2:40 total, but it looks like a little while ago we added some additional checks in the overcloud validate (this is a good thing btw) and it just pushes things over by a bit | 14:41 |
dmsimard | timeouts have been an issue for many jobs, increasing them just hides the issue under the carpet | 14:42 |
*** chandankumar has joined #rdo | 14:42 | |
dmsimard | we can blame the (slower) CPUs all we want but >1:30hr for a job is pretty darn long | 14:42 |
*** ayoung_ZZZzzz is now known as ayoung | 14:42 | |
myoung | dmsimard: aye...I've done some initial analysis. we already do some good things (like building IPA and overcloud DIB passes in parallel), but there is lots more that can be done. I have a card from months ago enumerating ways to speed things up I'll revive and port over to RDO board. | 14:42 |
dmsimard | myoung: we could work with a cached base image or something | 14:43 |
dmsimard | myoung: like iirc we download the centos cloud image and then install stuff in it | 14:43 |
dmsimard | myoung: we could install everything that is not openstack-related in that image, and then download that "pre-installed" image instead | 14:43 |
dmsimard | myoung: for example for weirdo jobs, there are >300 packages that aren't openstack related that are installed throughout the jobs | 14:45 |
dmsimard | myoung: https://review.rdoproject.org/r/gitweb?p=config.git;a=blob;f=nodepool/rdo-weirdo-get-packages.sh;h=c4169c016a890484c518b2d62e1c302fd65cb7c4;hb=HEAD | 14:45 |
myoung | dmsimard: {nod} there's a variety of things we can do. Perhaps we should just schedule a bit of time to chat, I've been looking into this and have some ideas. I'll move/create the card, and here's how it works today: https://github.com/redhat-openstack/ansible-role-tripleo-image-build#how-does-it-work | 14:45 |
dmsimard | myoung: https://review.rdoproject.org/r/gitweb?p=config.git;a=blob;f=nodepool/weirdo-packages.txt;h=e37195fbf5b440b1ca91fb78f82c065dd5fb168a;hb=HEAD | 14:45 |
* myoung looks | 14:46 | |
*** fragatina has joined #rdo | 14:46 | |
myoung | ooooo. i like. I'll spin a patch for artib to leverage this. How much time in practice do you see saved? | 14:47 |
*** iranzo has quit IRC | 14:47 | |
dmsimard | myoung: whatever time it took to install 365 packages | 14:47 |
dmsimard | :p | 14:48 |
*** fragatin_ has joined #rdo | 14:48 | |
myoung | btw the actual image building is only one component of the 3+ hour job times, but anything we can do to speed up image building (pile of network + i/o on slow machines...) makes sense to do | 14:48 |
*** eharney has joined #rdo | 14:49 | |
*** nyechiel has quit IRC | 14:49 | |
*** anilvenkata has quit IRC | 14:49 | |
dmsimard | myoung: so what you need to do to leverage something like that is to 1) run that script on a oooq journalctl log 2) recover list of pacakges 3) download base image 4) install those packages in base image 5) upload base image to artifacts ... 6) download that base image in relevant jobs (instead of original cloud image) 7) profit | 14:50 |
*** lmiccini_ has joined #rdo | 14:51 | |
*** fragatina has quit IRC | 14:51 | |
dmsimard | there's an assumption in that script that expects a gzipped journalctl (that's how it ends up in weirdo..) not sure about ooq | 14:53 |
*** jeckersb is now known as jeckersb_gone | 14:54 | |
*** lmiccini has quit IRC | 14:54 | |
myoung | dmsimard: hrm...i kind of like the idea of a periodic job that does all of that (with minimal tire kicking). presently the image building role starts with a user image (starting point) that can be just passed as a parameter: https://github.com/redhat-openstack/ansible-role-tripleo-image-build#role-variables. It currently defaults to cloud.centos.org/.../GenericCloud.qcow2 | 14:57 |
myoung | dmsimard ^ s/user// | 14:57 |
dmsimard | yeah that default to cloud.centos.org could be our new pre-built image | 14:58 |
dmsimard | That would be generated by a periodic job.. could be once per 24hr or something | 14:58 |
*** aortega has quit IRC | 14:59 | |
*** aortega has joined #rdo | 15:01 | |
*** jeckersb_gone is now known as jeckersb | 15:02 | |
*** dgurtner has joined #rdo | 15:02 | |
*** d0ugal has quit IRC | 15:05 | |
flepied | à'é(çé | 15:05 |
*** aufi has quit IRC | 15:07 | |
*** shardy is now known as shardy_afk | 15:07 | |
*** dgurtner has quit IRC | 15:07 | |
*** hynekm has quit IRC | 15:07 | |
hewbrocca | flepied: řéřččá | 15:08 |
flepied | hewbrocca: you speak yubikey natively? :-) | 15:08 |
*** dhill_ has joined #rdo | 15:09 | |
*** dhill_ has quit IRC | 15:10 | |
*** dhill_ has joined #rdo | 15:10 | |
*** fragatin_ has quit IRC | 15:10 | |
*** gszasz has quit IRC | 15:13 | |
*** jistr is now known as jistr|mtg | 15:14 | |
*** stanchan has quit IRC | 15:14 | |
*** gszasz has joined #rdo | 15:16 | |
*** stanchan has joined #rdo | 15:18 | |
*** d0ugal has joined #rdo | 15:18 | |
*** tesseract- has quit IRC | 15:21 | |
*** hewbrocca is now known as hewbrocca-afk | 15:24 | |
*** rcernin has quit IRC | 15:26 | |
*** stanchan has quit IRC | 15:28 | |
*** stanchan has joined #rdo | 15:29 | |
*** links has quit IRC | 15:33 | |
*** belmoreira has quit IRC | 15:40 | |
*** stanchan has quit IRC | 15:42 | |
*** beagles is now known as beagles_food | 15:42 | |
*** stanchan has joined #rdo | 15:42 | |
*** mdnadeem has quit IRC | 15:43 | |
*** stanchan has quit IRC | 15:43 | |
*** dgurtner has joined #rdo | 15:44 | |
*** dgurtner has joined #rdo | 15:44 | |
*** pgadiya has quit IRC | 15:48 | |
*** florianf has quit IRC | 15:48 | |
*** florianf has joined #rdo | 15:49 | |
*** stanchan has joined #rdo | 15:50 | |
*** florianf has quit IRC | 15:53 | |
*** abehl has quit IRC | 15:54 | |
*** iberezovskiy is now known as iberezovskiy|off | 15:55 | |
*** sdake has quit IRC | 15:56 | |
*** sdake has joined #rdo | 15:56 | |
*** stanchan has quit IRC | 16:01 | |
*** paragan has quit IRC | 16:03 | |
*** dgurtner has quit IRC | 16:04 | |
*** stanchan has joined #rdo | 16:05 | |
*** pcaruana has quit IRC | 16:05 | |
*** paragan has joined #rdo | 16:06 | |
*** beagles_food is now known as beagles | 16:07 | |
*** dtantsur is now known as dtantsur|afk | 16:08 | |
*** fragatina has joined #rdo | 16:08 | |
*** fragatina has quit IRC | 16:09 | |
*** jeckersb is now known as jeckersb_gone | 16:12 | |
*** shardy_afk is now known as shardy | 16:12 | |
*** jeckersb_gone is now known as jeckersb | 16:12 | |
*** amoralej is now known as amoralej|off | 16:13 | |
*** READ10 has joined #rdo | 16:15 | |
*** stanchan has quit IRC | 16:16 | |
*** stanchan has joined #rdo | 16:17 | |
*** mcornea has quit IRC | 16:17 | |
rdogerrit | Javier Peña created DLRN: [WIP] Parallel mock builds in DLRN http://review.rdoproject.org/r/1806 | 16:19 |
*** tumble has quit IRC | 16:20 | |
*** degorenko has quit IRC | 16:21 | |
*** hrw has quit IRC | 16:21 | |
*** hrw has joined #rdo | 16:23 | |
*** fragatina has joined #rdo | 16:24 | |
*** rasca has quit IRC | 16:27 | |
*** jlibosva has quit IRC | 16:27 | |
*** Tenhi has quit IRC | 16:28 | |
*** rwsu has quit IRC | 16:28 | |
*** coolsvap has joined #rdo | 16:29 | |
*** myoung has quit IRC | 16:29 | |
*** Tenhi has joined #rdo | 16:30 | |
*** myoung has joined #rdo | 16:30 | |
*** ohamada has quit IRC | 16:30 | |
rdogerrit | Javier Peña proposed DLRN: [WIP] Parallel mock builds in DLRN http://review.rdoproject.org/r/1806 | 16:31 |
*** snecklifter has joined #rdo | 16:31 | |
*** rwsu has joined #rdo | 16:31 | |
*** paramite has quit IRC | 16:33 | |
*** lucasagomes is now known as lucas|dinner | 16:36 | |
*** jpich has quit IRC | 16:36 | |
*** gszasz has quit IRC | 16:39 | |
*** stanchan has quit IRC | 16:41 | |
*** d0ugal has quit IRC | 16:44 | |
*** paragan has quit IRC | 16:47 | |
imcsk8 | jpena: just merged: https://review.openstack.org/336089 | 16:53 |
*** fragatin_ has joined #rdo | 16:53 | |
*** fragatina has quit IRC | 16:55 | |
jpena | imcsk8: thanks! if you have time, pls have a look at https://review.openstack.org/349885, we need it to unblock packstack CI | 16:55 |
imcsk8 | jpena: i will, i'm checking reviews | 16:56 |
jpena | cool, thx! | 16:56 |
*** paragan has joined #rdo | 16:56 | |
*** paragan has joined #rdo | 16:56 | |
*** mlammon1 has quit IRC | 16:56 | |
jpena | time to go, have a nice weekend everyone! | 16:57 |
*** jpena is now known as jpena|off | 16:58 | |
imcsk8 | thanks! | 16:58 |
*** numans has quit IRC | 17:00 | |
*** stanchan has joined #rdo | 17:01 | |
*** bandini has quit IRC | 17:03 | |
rdogerrit | rdo-trunk created openstack/horizon-distgit: python-django-horizon: failed to build 216e673 http://review.rdoproject.org/r/1807 | 17:04 |
*** d0ugal has joined #rdo | 17:12 | |
rdobot | [sensu] NEW: master.monitoring.rdoproject.org - check-delorean-newton-current @ http://tinyurl.com/gud2vup |#| Build failure on centos7-master/current: horizon: http://trunk.rdoproject.org/centos7-master/report.html | 17:13 |
*** itlinux has quit IRC | 17:15 | |
*** Tenhi has quit IRC | 17:17 | |
*** dustins has quit IRC | 17:17 | |
*** Tenhi has joined #rdo | 17:25 | |
*** puzzled has quit IRC | 17:26 | |
*** tosky has quit IRC | 17:27 | |
*** pradk has quit IRC | 17:30 | |
*** akrivoka has quit IRC | 17:32 | |
*** pradk has joined #rdo | 17:33 | |
*** stanchan has quit IRC | 17:34 | |
*** paragan has quit IRC | 17:36 | |
*** chandankumar has quit IRC | 17:40 | |
rdogerrit | rdo-trunk created openstack/cliff-distgit: python-cliff: failed to build 9f42088 http://review.rdoproject.org/r/1808 | 17:42 |
*** dgurtner has joined #rdo | 17:43 | |
openstackgerrit | Merged openstack/packstack: Fix pep8 and py27 tests in Ubuntu Xenial https://review.openstack.org/349885 | 17:44 |
openstackgerrit | Merged openstack/packstack: Create Gnocchi legacy resources https://review.openstack.org/330478 | 17:44 |
*** abehl has joined #rdo | 17:45 | |
*** Tenhi has quit IRC | 17:47 | |
*** leanderthal is now known as leanderthal|afk | 17:47 | |
*** ddmitriev has quit IRC | 17:48 | |
*** paragan has joined #rdo | 17:48 | |
*** eharney has quit IRC | 17:49 | |
*** Tenhi has joined #rdo | 17:50 | |
*** dustins has joined #rdo | 17:51 | |
*** fragatin_ has quit IRC | 17:52 | |
*** fragatina has joined #rdo | 17:53 | |
rdobot | [sensu] NEW: master.monitoring.rdoproject.org - check-delorean-newton-head-current @ http://tinyurl.com/hc82n3z |#| Build failure on centos7-master-head/current: cliff: http://trunk.rdoproject.org/centos7-master-head/report.html | 17:53 |
*** weshay has quit IRC | 17:58 | |
*** sarath has joined #rdo | 17:59 | |
*** weshay has joined #rdo | 18:01 | |
*** shardy is now known as shardy_afk | 18:03 | |
*** eharney has joined #rdo | 18:03 | |
*** stanchan has joined #rdo | 18:10 | |
*** JuanDRay has joined #rdo | 18:10 | |
*** eharney_ has joined #rdo | 18:13 | |
*** JuanDRay has quit IRC | 18:14 | |
*** eharney has quit IRC | 18:15 | |
*** eharney_ has quit IRC | 18:19 | |
*** stanchan has quit IRC | 18:26 | |
*** eharney_ has joined #rdo | 18:33 | |
*** gfidente has quit IRC | 18:36 | |
*** nstrug has quit IRC | 18:40 | |
*** iranzo has joined #rdo | 18:44 | |
*** READ10 has quit IRC | 18:44 | |
*** dcain1 has quit IRC | 18:46 | |
*** iranzo has quit IRC | 18:52 | |
*** livelace has quit IRC | 18:52 | |
*** stanchan has joined #rdo | 18:53 | |
*** dhill_ has quit IRC | 18:54 | |
*** dyasny has quit IRC | 18:58 | |
*** fcoelho has quit IRC | 18:58 | |
*** dhill_ has joined #rdo | 18:58 | |
*** dgurtner has quit IRC | 19:02 | |
*** stanchan has quit IRC | 19:05 | |
*** dyasny has joined #rdo | 19:07 | |
*** shaunm has quit IRC | 19:08 | |
rdobot | [sensu] NEW: master.monitoring.rdoproject.org - check-delorean-newton-current @ http://tinyurl.com/gud2vup |#| Build failure on centos7-master/current: horizon: http://trunk.rdoproject.org/centos7-master/report.html | 19:10 |
dmsimard | myoung: protip http://eavesdrop.openstack.org/irclogs/%23rdo/%23rdo.2016-08-05.log.html#t2016-08-05T14:38:55 | 19:10 |
myoung | dmsimard: oooooo. neat. | 19:10 |
* myoung thanks dmsimard | 19:10 | |
*** eharney_ is now known as eharney | 19:16 | |
dmsimard | I added a tl;dr description to the card | 19:16 |
*** Son_Goku has joined #rdo | 19:17 | |
*** dneary has quit IRC | 19:21 | |
*** RoozbehShafiee has joined #rdo | 19:24 | |
*** sdake has quit IRC | 19:26 | |
*** sdake has joined #rdo | 19:28 | |
pabelanger | myoung: dmsimard: so, would be interesting to talk to both of you about some jobs upstream I am wanting to try. Specifically building the overcloud-full.qcow2 image. While it might not do everything that ansible-role-tripleo-image-build will, I think there is an option to make better pipelines then have a single job build for 3+ hours | 19:31 |
pabelanger | myoung: dmsimard: review in question is https://review.openstack.org/#/c/351278/ and adds an experimental job into python-tripleoclient | 19:32 |
dmsimard | pabelanger: I'm not very knowledgeable about triple-o (and ooo-q) but I can try to help | 19:33 |
myoung | pabelanger: ack, rope me in. note that the bulk of that time isn't spent building an image but testing it... | 19:33 |
myoung | actual image build times are more like 35ish mins | 19:33 |
myoung | and i'm working on ways to make that even faster | 19:34 |
myoung | also these gates are only run when we need to make a change to the image building role *itself* | 19:34 |
pabelanger | myoung: dmsimard: how I see it working, is we'll launch parallel jobs to build all the images, and only start validation once all image builds are green. Then maybe at a later time or in another pipeline, do that validation of the images | 19:34 |
dmsimard | myoung: it takes around 1hr on average to build the image, not negligible https://ci.centos.org/job/tripleo-quickstart-promote-master-delorean-build-images/buildTimeTrend | 19:35 |
pabelanger | myoung: right, this is POC to see how things look. controling when we run the jobs can be updated to be when ever we want | 19:35 |
pabelanger | dmsimard: is that all images? | 19:35 |
dmsimard | pabelanger: hum, that's just the job that builds the undercloud image and uploads it, I think. | 19:35 |
dmsimard | for one release (i.e, in this case delorean newton) | 19:36 |
pabelanger | myoung: another thing we have upstream, is a pretty good caching system in place. So we can help reduce the time downloading bits, but adding things to the cache | 19:36 |
pabelanger | dmsimard: I can see that, it takes us about 1 hour to build a new DIB for nodepool per label | 19:36 |
pabelanger | have our images are 5G, not sure what they are for delorean | 19:37 |
dmsimard | ~3.5GB | 19:37 |
dmsimard | pabelanger: not built with dib however | 19:37 |
pabelanger | dmsimard: oh? | 19:37 |
dmsimard | pabelanger: it downloads base centos image and builds on top of that | 19:37 |
dmsimard | https://ci.centos.org/artifacts/rdo/images/master/delorean/stable/ | 19:38 |
myoung | pabelanger, dmsimard: we're talking about a few different things here. build role gates themselves actually build am image, then use it in place for the validation, so there are no copies | 19:38 |
pabelanger | dmsimard: Ya, I want to change that. and use centos-minimal, since your basically using a binary as your base. If that image was to be deleted, you have zero way of reproducing it | 19:38 |
myoung | there is also a gating role (oooq) that can be used to modify existing images quickly in response to a patch...quick virt-customize --> rpm changes --> test | 19:38 |
myoung | ^ https://github.com/redhat-openstack/ansible-role-tripleo-gate | 19:39 |
dmsimard | pabelanger: well to be fair, dib images aren't exactly guaranteed to be reproducible either | 19:39 |
pabelanger | myoung: right, I'm looking at just building images right now and only building them in the gate. Validation will only happen after I get that process finished | 19:39 |
dmsimard | pabelanger: repositories can change and all that | 19:39 |
*** alexismonville has quit IRC | 19:40 | |
myoung | dmsimard: re: 38 vs 1 hr, correct, I had wires crossed. RDO machines seem to be a little slower :( | 19:40 |
pabelanger | dmsimard: Right, but using -minimal images, gets you closer to that. We had to switch to it after canonical started deleting ubuntu cloud qcow2 images from upstream | 19:40 |
*** alexismonville has joined #rdo | 19:42 | |
*** alexismonville has quit IRC | 19:43 | |
*** rbowen has quit IRC | 19:48 | |
*** gfidente has joined #rdo | 19:50 | |
rdobot | [sensu] NEW: master.monitoring.rdoproject.org - check-delorean-newton-head-current @ http://tinyurl.com/hc82n3z |#| Build failure on centos7-master-head/current: cliff: http://trunk.rdoproject.org/centos7-master-head/report.html | 19:50 |
*** RoozbehShafiee has quit IRC | 19:51 | |
*** aortega has quit IRC | 19:53 | |
bnemec | It looks like we're using master gnocchiclient in Mitaka, and they just merged a patch that broke us: https://github.com/openstack/python-gnocchiclient/commit/dcb53e7040655d48fa96beaf39dd6be22c4f684d | 19:54 |
bnemec | osc-lib doesn't exist in Mitaka so that can't work. | 19:54 |
rdogerrit | David Moreau Simard proposed openstack/cliff-distgit: Remove announce.rst from packaging http://review.rdoproject.org/r/1808 | 19:55 |
dmsimard | bnemec: we're probably not using master gnocchiclient, no | 19:56 |
bnemec | Example failure caused by this: http://logs.openstack.org/62/348862/5/check-tripleo/gate-tripleo-ci-centos-7-ovb-ha-mitaka/c0bebf4/logs/undercloud/var/log/undercloud_install.txt.gz#_2016-08-05_17_00_36_000 | 19:56 |
dmsimard | bnemec: swhich repo, delorean or stable ? | 19:56 |
bnemec | dmsimard: It's a mitaka CI job. | 19:56 |
dmsimard | bnemec: mitaka delorean or mitaka stable ? :) | 19:56 |
*** ppowell has joined #rdo | 19:57 | |
bnemec | I believe we use all dlrn repos. | 19:58 |
dmsimard | bnemec: hrm | 19:59 |
dmsimard | pradk: o/ | 19:59 |
dmsimard | pradk: should gnocchiclient have source branches set ? https://github.com/redhat-openstack/rdoinfo/blob/master/rdo.yml | 19:59 |
dmsimard | er, https://github.com/redhat-openstack/rdoinfo/blob/master/rdo.yml#L962 | 20:00 |
pradk | dmsimard, we could, but generally client doesnt go by releases. gnocchiclient is expected to be compatible with multiple releases.. unless otherwise specified.. so i guess we just set it to latest source version? | 20:02 |
dmsimard | pradk: bnemec mentioned that a recent commit might've broken mitaka | 20:03 |
pradk | oh | 20:03 |
* pradk reads back | 20:03 | |
bnemec | Yeah, the osc-lib commit breaks older releases. | 20:03 |
bnemec | Because osc-lib is a shiny new thing. :-) | 20:03 |
dmsimard | pradk: https://github.com/openstack/python-gnocchiclient/commit/36af0d94f98ec7f4ade790ba7382a76b3ac1f70f | 20:03 |
dmsimard | bnemec: telemetry likes shiny new things :P | 20:04 |
pradk | hmm i see.. we could set mitaka to 2.1 if that alleviate this issue | 20:04 |
dmsimard | pradk: gnocchiclient is somehow at 2.4.1 in dlrn, not sure how that can happen | 20:05 |
pradk | dmsimard, thats building off master? | 20:06 |
dmsimard | oh, nevermind | 20:06 |
dmsimard | yeah I guess it takes the latest (2.4.0) and builds off of that | 20:06 |
dmsimard | hence 2.4.1 | 20:06 |
pradk | yea we just grab setup.py --version | 20:06 |
dmsimard | probably 2.3.1 would be fine ? /me shrugs | 20:06 |
pradk | dmsimard, i would think it so .. we need atleast 2.1 for sure, we can just set it to 2.2 | 20:07 |
dmsimard | you're the pro | 20:07 |
pradk | dmsimard, so if we set the source to 2.2 it will only build from that tag? | 20:08 |
pradk | source here is tag or branch? | 20:08 |
pradk | dmsimard, reason i ask is.. in gnocchiclient we have https://github.com/openstack/python-gnocchiclient/tree/stable/2.2 | 20:09 |
pradk | thats the latest stable branch | 20:09 |
pradk | if we can point mitaka to stable/2.2 we should be safe | 20:10 |
*** ohochman has left #rdo | 20:12 | |
dmsimard | ah, yes, 2.2... I was looking at the releases | 20:12 |
dmsimard | which goes up to 2.4 | 20:12 |
*** pnavarro|lunch has joined #rdo | 20:13 | |
*** sarath has quit IRC | 20:13 | |
rdogerrit | Merged openstack/cliff-distgit: Remove announce.rst from packaging http://review.rdoproject.org/r/1808 | 20:15 |
pradk | dmsimard, ok i can set that in rdoinfo.yml and submit a patch? | 20:16 |
dmsimard | sounds sane to me, I think we would need to do something similar for liberty -- or was gnocchiclient not there yet for liberty ? | 20:16 |
pradk | so want to clarify whats source-branch here referring to? | 20:18 |
pradk | the tag or the branch? | 20:18 |
dmsimard | it's a ref, it can be either | 20:19 |
dmsimard | except if you put a tag, dlrn is effectively not going to be building new commits | 20:20 |
dmsimard | a branch, however, will have new commits in it | 20:20 |
pabelanger | dmsimard: 36af0d94f98ec7f4ade790ba7382a76b3ac1f70f would have been a perfect commit to test the 3rdparty CI jobs were talking about. I would expect the %check macro from python-gnocchiclient to fail | 20:20 |
dmsimard | pradk: we just, for example, created the newton-uc concept where libs and clients are pinned to the tags defined in upstream uppper-constraints | 20:20 |
pabelanger | dmsimard: also, I'm free if you are to work on zuul.conf.ergb | 20:21 |
pabelanger | erb* | 20:21 |
dmsimard | pabelanger: more or less, gnocchiclient would have built successfully | 20:21 |
dmsimard | pabelanger: it's a runtime dependency, not a build dependency | 20:21 |
pradk | dmsimard, k so i'll set source-branch: stable/2.0 then | 20:21 |
dmsimard | pabelanger: or else we would've got a failure to build from source long before it would've landed in bnemec's broken CI :) | 20:21 |
pabelanger | dmsimard: right, but hopefully unit tests would have caught it | 20:21 |
dmsimard | pabelanger: I'm not sure if we typically run unit test systematically in the builds | 20:22 |
pabelanger | dmsimard: since we need to install Requires into spec file | 20:22 |
*** jayg is now known as jayg|g0n3 | 20:22 | |
*** Son_Goku has quit IRC | 20:22 | |
dmsimard | pabelanger: right, but that's a runtime dependency | 20:22 |
pabelanger | dmsimard: right, I think the number is 24% atm of spec files run testr | 20:22 |
*** beagles is now known as beagles_bl8r | 20:22 | |
pabelanger | agreed | 20:22 |
pabelanger | going to look at spec file now | 20:22 |
rdogerrit | Pradeep Kilambi created rdoinfo: Set gnocchiclient mitaka to stable/2.2 http://review.rdoproject.org/r/1809 | 20:23 |
pradk | dmsimard, ^^ | 20:23 |
dmsimard | pabelanger: i.e, gnocchiclient could have for example sqlalchemy as a runtime dependency.. but since it's not used/imported at build time, dlrn will build it successfully and it will crash rather horrible once the package lands in the repos | 20:23 |
dmsimard | pradk: can you add a note that this is to pin before the introduction of osc_lib or something which does not exist in <newton ? | 20:24 |
pradk | sure | 20:24 |
dmsimard | pradk: I can't quite +2 rdoinfo reviews but I'll +1 and add number80 and apevec :) | 20:24 |
pradk | dmsimard, note in the yml or commi message? | 20:24 |
dmsimard | pradk: commit | 20:24 |
pradk | will do | 20:24 |
dmsimard | pradk: put the review too: https://review.openstack.org/#/c/343877/ | 20:25 |
pabelanger | dmsimard: I would expect it to look like https://review.rdoproject.org/r/#/c/1673/2/python-troveclient.spec which number80 recently did | 20:25 |
dmsimard | pradk: and btw just noticed aodhclient is doing the same thing in https://review.openstack.org/#/c/343868/ | 20:25 |
pabelanger | dmsimard: also, I cannot find the python-gnocchiclient spec file. Are you able to help with that? | 20:26 |
dmsimard | ahhh, jschlueter :) | 20:26 |
dmsimard | pabelanger: https://github.com/rdo-packages/?utf8=✓&query=gnocchi | 20:26 |
pradk | he's doing it for everything https://review.openstack.org/#/q/owner:jschluet%2540redhat.com+status:open :) | 20:26 |
pabelanger | dmsimard: ah, thanks. Looking in the wrong repo | 20:26 |
dmsimard | pradk: so he's going to break everything | 20:26 |
pradk | lol | 20:27 |
dmsimard | and then jschlueter comes to us, all sad, when we break things ! | 20:27 |
dmsimard | /friday | 20:27 |
pabelanger | dmsimard: yes, we don't run testr for python-gnocchiclient. Let me see if I can fix that | 20:27 |
dmsimard | pabelanger: you're probably going to need quite a lot more build dependencies for that | 20:28 |
rdogerrit | Pradeep Kilambi proposed rdoinfo: Set gnocchiclient mitaka to stable/2.2 http://review.rdoproject.org/r/1809 | 20:28 |
pabelanger | dmsimard: yup | 20:28 |
rdogerrit | David Moreau Simard proposed rdoinfo: Set gnocchiclient mitaka to stable/2.2 http://review.rdoproject.org/r/1809 | 20:29 |
jschlueter | dmsimard, pradk: sorry if I mixed things up but that deprecation warning that is in master of openstackclient really does mess up installs, if it's too soon we can put a -2 on the reviews or -1 workflow | 20:30 |
* jschlueter has to run but will check back later | 20:30 | |
dmsimard | jschlueter: http://logs.openstack.org/62/348862/5/check-tripleo/gate-tripleo-ci-centos-7-ovb-ha-mitaka/c0bebf4/logs/undercloud/var/log/undercloud_install.txt.gz#_2016-08-05_17_00_36_000 | 20:30 |
jschlueter | yep | 20:31 |
*** null_ref has quit IRC | 20:32 | |
jschlueter | there are 4 or 5 instances across the packages that may be hit ... and a revert of that patch would then be in order ... why was that not caught at gate? | 20:32 |
dmsimard | jschlueter: which gate ? | 20:33 |
*** aortega has joined #rdo | 20:33 | |
dmsimard | that's a tripleo failure, there's no tripleo CI on gnocchiclient :) | 20:34 |
*** ohochman has joined #rdo | 20:34 | |
*** ohochman has left #rdo | 20:35 | |
pabelanger | jschlueter: were working on a POC to run rpmbuild from zuul.rdoproject.org against patches to review.openstack.org, 36af0d94f98ec7f4ade790ba7382a76b3ac1f70f would have been a perfect commit to potential be exposed if the POC was working | 20:36 |
pabelanger | jschlueter: and alert people from 18days ago to 13hour when it was merged | 20:37 |
*** shardy_afk is now known as shardy | 20:38 | |
*** zaneb has quit IRC | 20:39 | |
jschlueter | dmsimard: see https://review.openstack.org/#/q/topic:osc-lib for related patches | 20:40 |
jschlueter | pabelanger: yep it would have been | 20:40 |
pabelanger | jschlueter: it won't take long to get the POC working, working with dmsimard on that. But should give downstream plenty of heads on of breaking changes hit review.o.o | 20:42 |
*** ppowell has quit IRC | 20:42 | |
dmsimard | pabelanger: finishing up on a failure to build and I'm all yours | 20:42 |
pabelanger | dmsimard: \o/ | 20:43 |
rdogerrit | Honza Pokorny proposed rdoinfo: Add openstack-tripleo-ui http://review.rdoproject.org/r/1733 | 20:43 |
dmsimard | jruzicka: you're probably gone by now, yes ? :) | 20:44 |
dmsimard | pabelanger: crap, this might take a while :/ | 20:45 |
dmsimard | pabelanger: and your CI would've also caught this one :) | 20:45 |
rdobot | [sensu] RESOLVED: master.monitoring.rdoproject.org - check-delorean-newton-head-current @ http://tinyurl.com/hc82n3z |#| No build failures detected: http://trunk.rdoproject.org/centos7-master-head/report.html | 20:45 |
*** jeckersb is now known as jeckersb_gone | 20:46 | |
rdogerrit | Honza Pokorny proposed rdoinfo: Add openstack-tripleo-ui http://review.rdoproject.org/r/1733 | 20:47 |
pabelanger | dmsimard: ya, for catching more things | 20:47 |
dmsimard | pabelanger: horizon just introduced 3 new deps that are not packaged anywhere | 20:47 |
dmsimard | this is the worst kind of failures :( | 20:47 |
rdogerrit | pabelanger created openstack/gnocchiclient-distgit: Run unit tests during package build http://review.rdoproject.org/r/1810 | 20:47 |
pabelanger | dmsimard: which commit? | 20:48 |
dmsimard | pabelanger: https://review.openstack.org/#/c/332745/25/requirements.txt | 20:48 |
rdogerrit | Gabriele Cerami created openstack/horizon-distgit: Increase systemd start timeout http://review.rdoproject.org/r/1811 | 20:49 |
dmsimard | pabelanger: also TIL gerrit can render looks of disapproval in reviews https://review.openstack.org/#/c/332745/20/horizon/static/framework/widgets/form/fields/array.html | 20:49 |
dmsimard | pabelanger: I shall make use of that. | 20:49 |
pabelanger | dmsimard: making an etherpad to track some of this | 20:50 |
rdogerrit | Honza Pokorny proposed rdoinfo: Add openstack-tripleo-ui http://review.rdoproject.org/r/1733 | 20:50 |
*** Son_Goku has joined #rdo | 20:51 | |
*** zaneb has joined #rdo | 20:52 | |
*** beagles_bl8r is now known as beagles | 20:54 | |
dmsimard | pabelanger: I'm going to... try to pick up the three new deps | 20:55 |
dmsimard | pabelanger: sorry, I have to get the ball rolling so we're not too far behind | 20:55 |
dmsimard | pabelanger: if you want I can paste you the zuul config so we have a starting point or sometihng | 20:56 |
dmsimard | number80: three new xstatic dependencies to package.. I'll give it a shot | 20:56 |
pabelanger | dmsimard: sure, let me get the syntax to you that needs to be added | 20:57 |
*** jdob has quit IRC | 20:58 | |
*** Son_Goku has quit IRC | 20:58 | |
*** shardy has quit IRC | 20:58 | |
*** kgiusti has left #rdo | 21:00 | |
*** marrusl has quit IRC | 21:02 | |
*** pradk has quit IRC | 21:03 | |
*** rbrady is now known as rbrady-afk | 21:07 | |
pabelanger | dmsimard: lets use this as a metric to track failures we think 3rdparty CI would catch: https://etherpad.openstack.org/p/rdo-3rdparty | 21:08 |
rdobot | [sensu] NEW: master.monitoring.rdoproject.org - check-delorean-newton-current @ http://tinyurl.com/gud2vup |#| Build failure on centos7-master/current: horizon: http://trunk.rdoproject.org/centos7-master/report.html | 21:10 |
*** sdake has quit IRC | 21:12 | |
*** dprince has quit IRC | 21:12 | |
*** pnavarro|lunch has quit IRC | 21:12 | |
dmsimard | pabelanger: probably a deal of those too: https://etherpad.openstack.org/p/master-ci-trunk-issues | 21:15 |
pabelanger | dmsimard: yes, maybe I'll add the info their | 21:16 |
pabelanger | there* | 21:16 |
dmsimard | pabelanger: some of those aren't packaging-specific issues (some are) | 21:16 |
* pabelanger nods | 21:16 | |
dmsimard | a lot of them stem from using master against master | 21:16 |
rdogerrit | rdo-trunk created openstack/horizon-distgit: python-django-horizon: failed to build 87818bc http://review.rdoproject.org/r/1812 | 21:22 |
pabelanger | dmsimard: this is basically the syntax we need to add to zuul.conf.erb: http://paste.openstack.org/show/550678/ | 21:22 |
pabelanger | hardcoded values obviously | 21:23 |
pabelanger | bobsmith is our user in review.openstack.org | 21:23 |
pabelanger | dmsimard: once that is in place, and zuul has been restarted, we should see events from gerrit but no jobs will build, since we haven't updated layout.yaml | 21:23 |
pabelanger | lets do that first, then setup the pipelines | 21:24 |
*** Son_Goku has joined #rdo | 21:24 | |
*** stanchan has joined #rdo | 21:24 | |
*** fzdarsky_ has quit IRC | 21:25 | |
rdobot | [sensu] NEW: master.monitoring.rdoproject.org - check-delorean-newton-head-current @ http://tinyurl.com/hc82n3z |#| Build failure on centos7-master-head/current: horizon: http://trunk.rdoproject.org/centos7-master-head/report.html | 21:30 |
*** ashw has quit IRC | 21:32 | |
*** bnemec is now known as beekneemech | 21:36 | |
rdogerrit | Merged rdoinfo: Set gnocchiclient mitaka to stable/2.2 http://review.rdoproject.org/r/1809 | 21:39 |
dmsimard | pabelanger: you minimum familiar with koji ? | 21:44 |
dmsimard | pabelanger: I'm a super duper koji noob and all the pros are already out for the weekend :) | 21:45 |
dmsimard | pabelanger: tried a scratch build and it failed, says to read build.log but where is that ? http://koji.fedoraproject.org/koji/taskinfo?taskID=15147078 | 21:45 |
*** stanchan has quit IRC | 21:46 | |
dmsimard | pabelanger: ok, found it | 21:51 |
*** fultonj has quit IRC | 21:53 | |
*** sdake has joined #rdo | 21:54 | |
*** dustins has quit IRC | 21:54 | |
*** morazi has quit IRC | 21:58 | |
*** tdasilva has quit IRC | 21:59 | |
*** rlandy has quit IRC | 22:00 | |
*** jcoufal_ has joined #rdo | 22:00 | |
*** weshay has quit IRC | 22:02 | |
*** jcoufal has quit IRC | 22:04 | |
*** paragan has quit IRC | 22:04 | |
*** paragan has joined #rdo | 22:05 | |
*** paragan has joined #rdo | 22:05 | |
*** panda is now known as panda|zz | 22:06 | |
*** stanchan has joined #rdo | 22:11 | |
*** milan has quit IRC | 22:12 | |
*** jcoufal_ has quit IRC | 22:15 | |
*** thrash is now known as thrash|wknd | 22:16 | |
*** jcoufal has joined #rdo | 22:20 | |
*** maeca1 has quit IRC | 22:21 | |
*** gfidente has quit IRC | 22:22 | |
*** stanchan has quit IRC | 22:25 | |
*** tdasilva has joined #rdo | 22:27 | |
*** stanchan has joined #rdo | 22:32 | |
*** jcoufal has quit IRC | 22:35 | |
*** jcoufal has joined #rdo | 22:40 | |
*** imcsk8 has quit IRC | 22:50 | |
*** imcsk8 has joined #rdo | 22:50 | |
*** scorcoran has joined #rdo | 23:01 | |
*** stanchan has quit IRC | 23:08 | |
*** stanchan has joined #rdo | 23:09 | |
rdobot | [sensu] NEW: master.monitoring.rdoproject.org - check-delorean-newton-current @ http://tinyurl.com/gud2vup |#| Build failure on centos7-master/current: horizon: http://trunk.rdoproject.org/centos7-master/report.html | 23:10 |
*** rhallisey has quit IRC | 23:13 | |
*** stanchan has quit IRC | 23:23 | |
rdobot | [sensu] NEW: master.monitoring.rdoproject.org - check-delorean-newton-head-current @ http://tinyurl.com/hc82n3z |#| Build failure on centos7-master-head/current: horizon: http://trunk.rdoproject.org/centos7-master-head/report.html | 23:27 |
*** jcoufal has quit IRC | 23:30 | |
*** sdake has quit IRC | 23:31 | |
*** aortega has quit IRC | 23:31 | |
EmilienM | dmsimard: I guess you're aware about https://bugs.launchpad.net/tripleo/+bug/1610411 | 23:32 |
openstack | Launchpad bug 1610411 in tripleo "Mitaka jobs broken by gnocchiclient move to osc-lib" [Critical,Triaged] | 23:32 |
EmilienM | I haven't followed all the bits | 23:32 |
EmilienM | but is there anything I can do? | 23:32 |
dmsimard | EmilienM: pradk submitted an update to rdoinfo and it was merged http://review.rdoproject.org/r/1809 | 23:33 |
dmsimard | tl;dr gnocchiclient should build 2.2 instead of 2.4.1 for mitaka | 23:33 |
EmilienM | k | 23:33 |
dmsimard | but I don't know if/how/when it would be rebuilt 2.2 | 23:33 |
EmilienM | so fixed now? | 23:33 |
*** manous has joined #rdo | 23:40 | |
*** jcoufal has joined #rdo | 23:41 | |
*** stanchan has joined #rdo | 23:41 | |
*** dtrainor has quit IRC | 23:41 | |
*** dtrainor has joined #rdo | 23:41 | |
*** myoung is now known as myoung|afk | 23:46 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!