Monday, 2014-10-27

« Sunday, 2014-10-26 Index Tuesday, 2014-10-28 »
*** ramishra has joined #tripleo00:16
*** ramishra has quit IRC00:21
*** pcrews has quit IRC00:39
*** fayablazer has joined #tripleo01:15
*** ramishra has joined #tripleo01:17
*** Marga_ has joined #tripleo01:20
*** ramishra has quit IRC01:21
*** fayablazer has quit IRC01:39
*** yamahata has joined #tripleo01:39
*** dsneddon has quit IRC01:40
*** dsneddon has joined #tripleo01:42
*** noslzzp has joined #tripleo01:59
*** dsneddon has quit IRC02:01
*** dsneddon has joined #tripleo02:03
*** nosnos has joined #tripleo02:07
*** ramishra has joined #tripleo02:17
*** ramishra has quit IRC02:22
*** jerryz has joined #tripleo02:29
*** ramishra has joined #tripleo02:31
*** lazy_prince has quit IRC02:39
*** julim has joined #tripleo02:40
*** julim has quit IRC02:45
*** nijaba has quit IRC02:53
*** nijaba has joined #tripleo02:54
*** nijaba has quit IRC03:01
*** Marga_ has quit IRC03:12
*** killer_prince has joined #tripleo03:12
*** killer_prince is now known as lazy_prince03:12
*** Marga_ has joined #tripleo03:16
*** nosnos has quit IRC03:22
*** nosnos has joined #tripleo03:23
*** nosnos has quit IRC03:27
*** noslzzp has quit IRC03:33
openstackgerritVenkat Sundaram proposed a change to stackforge/kolla: Containerizing barbican service  https://review.openstack.org/13101903:35
openstackgerritSteve Kowalik proposed a change to openstack/os-cloud-config: Permit configuration of DHCP when creating subnets  https://review.openstack.org/13007503:51
openstackgerritSteve Kowalik proposed a change to openstack/os-cloud-config: Support adding extra routes when creating subnets  https://review.openstack.org/13010803:52
*** untriaged-bot has joined #tripleo04:00
untriaged-botUntriaged bugs so far:04:00
untriaged-bothttps://bugs.launchpad.net/python-tuskarclient/+bug/138246804:00
untriaged-bothttps://bugs.launchpad.net/tripleo/+bug/138523404:00
uvirtbotLaunchpad bug 1382468 in tuskar "Bad parameter exception" [Undecided,Triaged]04:00
untriaged-bothttps://bugs.launchpad.net/tuskar/+bug/138246804:00
uvirtbotLaunchpad bug 1385234 in tripleo "OVS tunneling between multiple neutron nodes misconfigured if amqp is restarted" [High,New]04:00
*** untriaged-bot has quit IRC04:00
*** nosnos has joined #tripleo04:04
*** pradeep has joined #tripleo04:04
*** Marga_ has quit IRC04:07
*** Marga_ has joined #tripleo04:11
*** pcrews has joined #tripleo04:13
*** ramishra has quit IRC04:18
*** ramishra_ has joined #tripleo04:22
*** jerryz has quit IRC04:40
*** pcrews has quit IRC05:00
*** bvivek has joined #tripleo05:14
*** pradeep has quit IRC05:32
*** pradeep has joined #tripleo05:33
*** Marga_ has quit IRC05:58
*** nosnos has quit IRC05:59
*** nosnos_ has joined #tripleo05:59
*** loki184 has joined #tripleo06:00
openstackgerritA change was merged to openstack/diskimage-builder: Updated from global requirements  https://review.openstack.org/13031806:03
*** avozza is now known as zz_avozza06:08
*** xuhaiwei_ has joined #tripleo06:10
*** pradeep has quit IRC06:15
*** pradeep has joined #tripleo06:17
*** k4n0 has joined #tripleo06:18
*** loki184 has quit IRC06:30
xuhaiwei_Filed a bug report. https://bugs.launchpad.net/tripleo/+bug/138605606:34
uvirtbotLaunchpad bug 1386056 in tripleo "undercloud ironic dbsync failed" [Undecided,New]06:34
*** lazy_prince has quit IRC06:35
*** ifarkas has joined #tripleo06:35
*** lazy_prince has joined #tripleo06:36
*** ramishra_ has quit IRC06:45
*** loki184 has joined #tripleo06:47
*** ramishra has joined #tripleo06:51
loki184please review https://review.openstack.org/#/c/106727/06:51
*** bvivek has quit IRC07:05
*** ramishra has quit IRC07:13
*** nosnos_ has quit IRC07:13
*** nosnos has joined #tripleo07:13
*** ramishra has joined #tripleo07:20
*** mrunge has joined #tripleo07:24
*** jprovazn has joined #tripleo07:29
*** bvivek has joined #tripleo07:34
GheRiveromornign all07:35
marios_o/07:40
*** jang1 has joined #tripleo07:47
*** mrunge has quit IRC07:48
*** rdopieralski has joined #tripleo07:52
*** rdopieralski has quit IRC07:53
*** rdopieralski has joined #tripleo07:53
xuhaiwei_I found there is no undercloud-bm-ironic.yaml, does it mean I can not build baremetal environment with ironic? The undercloud-bm.yaml seems not supporting ironic baremetal deployment07:58
*** mrunge has joined #tripleo08:03
openstackgerritA change was merged to openstack/tripleo-heat-templates: Simplify the layout of haproxy net binds config  https://review.openstack.org/12925808:07
*** bvivek has quit IRC08:08
openstackgerritA change was merged to openstack/tripleo-image-elements: The Nova dhcp_options_enabled flag doesn't exist  https://review.openstack.org/13020808:10
*** jprovazn has left #tripleo08:13
*** jprovazn has quit IRC08:14
*** jprovazn has joined #tripleo08:14
*** jtomasek has joined #tripleo08:15
*** yamahata has quit IRC08:17
*** bvivek has joined #tripleo08:23
openstackgerritA change was merged to openstack/diskimage-builder: Move busybox binary-dep to ramdisk element  https://review.openstack.org/13003408:30
*** marun_ has joined #tripleo08:32
*** Marga_ has joined #tripleo08:46
*** zz_avozza is now known as avozza08:48
*** jistr has joined #tripleo08:49
*** dtantsur|afk is now known as dtantsur08:59
*** jang1 has quit IRC09:26
*** andreaf has joined #tripleo09:44
*** xuhaiwei_ has quit IRC09:47
*** athomas has joined #tripleo09:50
*** Marga_ has quit IRC09:51
*** loki184 has quit IRC09:55
*** nijaba has joined #tripleo09:56
*** nijaba has quit IRC09:56
*** nijaba has joined #tripleo09:56
*** fayablazer has joined #tripleo09:58
*** akrivoka has joined #tripleo10:00
*** untriaged-bot has joined #tripleo10:00
untriaged-botUntriaged bugs so far:10:00
untriaged-bothttps://bugs.launchpad.net/python-tuskarclient/+bug/138246810:00
untriaged-bothttps://bugs.launchpad.net/tripleo/+bug/138523410:00
uvirtbotLaunchpad bug 1382468 in tuskar "Bad parameter exception" [Undecided,Triaged]10:00
untriaged-bothttps://bugs.launchpad.net/tuskar/+bug/138246810:00
uvirtbotLaunchpad bug 1385234 in tripleo "OVS tunneling between multiple neutron nodes misconfigured if amqp is restarted" [High,New]10:00
*** untriaged-bot has quit IRC10:00
*** bvivek has quit IRC10:27
*** rlandy has joined #tripleo10:31
*** pradeep has quit IRC10:47
*** bvivek has joined #tripleo11:07
*** gfidente has joined #tripleo11:14
*** julim has joined #tripleo11:20
*** lazy_prince has quit IRC11:30
*** ramishra has quit IRC11:33
*** dtantsur is now known as dtantsur|brb11:36
*** jerryz has joined #tripleo11:43
*** pblaho has joined #tripleo11:54
*** bcrochet has quit IRC12:03
*** bcrochet has joined #tripleo12:05
*** killer_prince has joined #tripleo12:07
*** killer_prince is now known as lazy_prince12:07
*** lazy_prince has quit IRC12:07
*** jerryz has quit IRC12:08
*** lazy_prince has joined #tripleo12:08
*** yamahata has joined #tripleo12:11
*** rhallisey has joined #tripleo12:11
*** akrivoka has quit IRC12:14
*** jprovazn has quit IRC12:15
*** openstackgerrit has quit IRC12:19
*** openstackgerrit has joined #tripleo12:20
*** yamahata has quit IRC12:23
*** akrivoka has joined #tripleo12:27
*** weshay has joined #tripleo12:28
*** yamahata has joined #tripleo12:28
*** jdob has joined #tripleo12:33
*** openstackgerrit has quit IRC12:34
*** openstackgerrit has joined #tripleo12:34
*** dtantsur|brb is now known as dtantsur12:36
*** ramishra has joined #tripleo12:36
*** jdob has quit IRC12:41
*** jdob has joined #tripleo12:41
*** pradeep has joined #tripleo12:48
*** bvivek has quit IRC12:49
*** ccrouch has joined #tripleo12:52
*** radez_g0n3 is now known as radez12:55
*** dprince has joined #tripleo12:58
*** tzumainn has joined #tripleo13:02
*** spzala has joined #tripleo13:11
*** jehb has joined #tripleo13:12
*** pradeep has quit IRC13:14
*** morazi has joined #tripleo13:15
*** ramishra has quit IRC13:17
*** rlandy_ has joined #tripleo13:27
*** noslzzp has joined #tripleo13:28
*** rlandy has quit IRC13:29
*** pblaho_ has joined #tripleo13:34
*** pblaho has quit IRC13:34
*** fayablazer has quit IRC13:38
*** ramishra has joined #tripleo13:46
*** k4n0 has quit IRC13:51
lxslihttps://review.openstack.org/#/c/115524/ is failing CI, I think because it's using oac 0.1.22 instead of 0.1.2413:53
lxsliAnyone know how I can bump that please?13:53
*** fayablazer has joined #tripleo13:53
openstackgerritAlexis Lee proposed a change to openstack/tripleo-image-elements: Fix RabbitMQ element clustering start and stop  https://review.openstack.org/11552414:00
*** pradeep has joined #tripleo14:01
*** radez is now known as radez_g0n314:01
*** beekneemech is now known as bnemec14:03
*** avozza is now known as zz_avozza14:04
*** nosnos has quit IRC14:04
*** nosnos has joined #tripleo14:05
*** pblaho_ has quit IRC14:06
*** pblaho has joined #tripleo14:08
*** Marga_ has joined #tripleo14:08
*** nosnos has quit IRC14:09
*** jinxly has joined #tripleo14:10
openstackgerritAlexis Lee proposed a change to openstack/tripleo-image-elements: Use oac to create .erlang.cookie  https://review.openstack.org/13115814:16
*** radez_g0n3 is now known as radez14:17
openstackgerritA change was merged to openstack/tripleo-image-elements: Add package install support for tuskar-ui  https://review.openstack.org/12931114:19
openstackgerritA change was merged to openstack/tripleo-image-elements: Add package install support for tuskar  https://review.openstack.org/12931014:20
*** jprovazn has joined #tripleo14:28
*** sdake has joined #tripleo14:31
*** sdake has quit IRC14:31
*** sdake has joined #tripleo14:31
*** sdake has quit IRC14:32
*** sdake has joined #tripleo14:32
*** sdake has quit IRC14:32
*** sdake has joined #tripleo14:32
sdakemorning14:32
*** ramishra has quit IRC14:32
larskshowdy.14:36
*** spzala has quit IRC14:40
*** zz_avozza is now known as avozza14:46
*** Marga_ has quit IRC14:51
*** radez is now known as radez_g0n314:53
*** pradeep1 has joined #tripleo14:53
*** pradeep has quit IRC14:55
*** mrunge has quit IRC14:56
openstackgerritRadomir Dopieralski proposed a change to openstack/tuskar-ui: Improving the looks of the plan design page, part I  https://review.openstack.org/13080615:02
openstackgerritRyan Brady proposed a change to openstack/diskimage-builder: Enable RHEL Registration  https://review.openstack.org/12314515:03
openstackgerritTzu-Mainn Chen proposed a change to openstack/tuskar-ui: Re-factor nodes index tables  https://review.openstack.org/13117215:11
openstackgerritPhil Neal proposed a change to openstack/tripleo-image-elements: Add elements for Ceilometer alarm services  https://review.openstack.org/11364715:11
*** pcrews has joined #tripleo15:11
*** asparks has joined #tripleo15:11
*** ramishra has joined #tripleo15:13
*** asparks has quit IRC15:15
*** pradeep1 has quit IRC15:20
*** pradeep has joined #tripleo15:21
*** yamahata has quit IRC15:27
openstackgerritRadomir Dopieralski proposed a change to openstack/tuskar-ui: Improving the looks of the plan design page, part I  https://review.openstack.org/13080615:27
*** andreaf has quit IRC15:30
*** avozza is now known as zz_avozza15:30
*** ramishra has quit IRC15:38
*** bnemec has quit IRC15:41
sdakeSpamapS is it too late to submit sessions for design sessions re kolla?15:41
*** bnemec has joined #tripleo15:42
SpamapSsdake: no15:43
sdakeSpamapS do you ahve anything you would like to see re kolla?15:43
sdakeI wont be there, but larsks will be15:43
SpamapSsdake: was just actually looking at the etherpad trying to decide how to divide up our time between the 2x 40 min sessions and the day-long contributor meetup.15:43
sdakealso do you know which day?15:43
SpamapSsdake: http://kilodesignsummit.sched.org/15:43
SpamapSsdake: http://kilodesignsummit.sched.org/overview/type/tripleo#.VE5oQIWdTcs15:44
sdakehow many slots do you hvae?15:44
SpamapSperhaps more useful :)15:44
sdakeiirc larsks wont be there friday15:44
sdakelarsks are you alive :)15:44
larsksYeah, traveling that day.15:44
SpamapSsdake: so we have 2 scheduled sessions that are more for general-public stuff, and then a day which is more for us to put our heads together and design stuff.15:44
SpamapSlarsks: you will miss basically all of the TripleO part of the summit. :(15:44
SpamapSthe two 40 minute sessions will be interesting, but it sounds like we need to make sure kolla squeezes in there15:45
larsksAlas for scheduling on the last day! :)15:45
larsksMy return flight actually leaves at 4:10pm local time, so I might be around in the morning.15:46
SpamapSlarsks: oh yeah that's fine then.15:46
SpamapS4:10 is an eternity from 9:30. :)15:46
sdakeSpamapS we do have a 45 min design session scheduled15:46
*** flashgordon is now known as jogo15:46
sdakeI think monday or tuesday15:46
sdakeso maybe that will work15:46
larsksTuesday.15:46
larskshttp://kilodesignsummit.sched.org/?s=kolla15:47
larsksAt dinnertime.15:47
SpamapSsdake: you have a what?15:47
sdakeI dont, larsks does15:47
sdakeI actually wont be at summit15:47
SpamapSOhhhhhh ok I see, you went in the other barrel.15:48
sdakeother barrel?15:48
SpamapS"Other" is the category.15:48
sdakeoh right15:48
sdakeit was offered, so I took :)15:48
SpamapSWhich is fine. We likely won't have time to fit everything into TripleO's 2 sessions.15:48
sdakeok so then we are all set spamaps larsks?15:48
larsksI think so.15:49
sdakejust trying to facilitate communication since I don't think you two know each other too well15:49
SpamapSNo but we should.15:49
*** panda has quit IRC15:49
sdakethere will be lots o drinking in paris!15:49
sdakeor if you don't drink lots of water!15:49
SpamapSMost of our communication has been me beating larsks with the os-cloud-config stick, so something less aggressive is in order. :)15:49
sdakeSpamapS I think we totally want to go there15:50
sdakepretty much i harrassed the entire kolla team to hit our deadline come hell or high water15:50
sdakeneither came :)15:50
SpamapSYeah I realized that and have stepped aside a bit. :)15:50
* larsks is heading out for some lunch. Back in a bit!15:50
SpamapSsdake: way I see it, Kolla has alternative implementations for software delivery, orchestration, and configuration... what I'm interested in is which of those can be replaced without replacing all of them.15:51
SpamapSOh and bootstrapping.. but I don't think that is desired. :)15:51
*** zz_avozza is now known as avozza15:52
*** rdopieralski has quit IRC15:54
*** mrunge has joined #tripleo15:57
*** Marga_ has joined #tripleo15:57
*** untriaged-bot has joined #tripleo16:00
untriaged-botUntriaged bugs so far:16:00
untriaged-bothttps://bugs.launchpad.net/python-tuskarclient/+bug/138246816:00
uvirtbotLaunchpad bug 1382468 in tuskar "Bad parameter exception" [Undecided,Triaged]16:00
untriaged-bothttps://bugs.launchpad.net/tripleo/+bug/138523416:00
uvirtbotLaunchpad bug 1385234 in tripleo "OVS tunneling between multiple neutron nodes misconfigured if amqp is restarted" [High,New]16:00
untriaged-bothttps://bugs.launchpad.net/tuskar/+bug/138246816:00
*** untriaged-bot has quit IRC16:00
*** jprovazn is now known as jprovazn_biab16:04
sdakeSpamapS sounds good - wish I could be there16:04
sdakebut larsks is sharp, he can handle your qs, hope you folks come up with solutions at the design sessions16:05
sdakeand take notes :)16:05
*** panda has joined #tripleo16:09
openstackgerritAna Krivokapic proposed a change to openstack/tuskar-ui: Improve flavor index view  https://review.openstack.org/13118816:11
*** radez_g0n3 is now known as radez16:11
*** Hefeweizen has quit IRC16:11
greghaynesStevenK: Any idea why we didnt use cliff for os-cloud-config/cmd/*?16:14
greghaynesTrying to figure out my summit schedule and im learning about all sorts of new projects ;)16:15
greghaynesI guess were not doing a bunch of output16:16
greghayneshrm16:17
*** dkehn has quit IRC16:21
*** dkehn has joined #tripleo16:23
greghaynesslagle: hrmm http://logs.openstack.org/93/120293/5/check-tripleo/check-tripleo-ironic-undercloud-precise-nonha/bfac532/console.html.gz#_2014-10-24_20_21_26_60116:25
greghaynesthats on the switch a bunch of stuff to pkg-map16:25
greghayneser, package-installs16:26
*** Hefeweizen has joined #tripleo16:30
openstackgerritTzu-Mainn Chen proposed a change to openstack/tripleo-image-elements: Make 101-tuskar-ui re-runnable  https://review.openstack.org/13066116:31
greghaynesah16:35
*** rbrady has joined #tripleo16:35
*** mrunge has quit IRC16:43
sdakelarsks nice change for the maintainers16:43
openstackgerritgreghaynes proposed a change to openstack/diskimage-builder: package-installs assumes packages have a pkg-map  https://review.openstack.org/13120416:43
greghaynesslagle: ^16:43
sdakeI don't like how docker requires a maintainer line16:43
openstackgerritazher ullah khan proposed a change to openstack/diskimage-builder: Optimize speed of deletion in find command  https://review.openstack.org/13120516:44
openstackgerritAna Krivokapic proposed a change to openstack/tuskar-ui: Improve flavor index view  https://review.openstack.org/13118816:47
ccrouchsdake: quick question. Are you guys receptive to having kolla/docker based containers running outside of K8s? e.g. in plain docker16:47
ccrouchi'm just talking about having blah_SERVICE_HOST default to one of the standing docker "link" env vars if its not set already16:48
ccrouchthat sort of thing16:48
sdakeccrouch absolutely16:48
sdakeso I don't know about the second point16:49
sdakeI am not hung up on k8s16:49
*** ramishra has joined #tripleo16:49
sdakek8s = fast prototype16:49
sdakewe got rocking in 2 weeks with a full implementation16:49
sdaketo do that natively would have taken months16:49
ccrouchthats pretty fast, no arguments there :-)16:49
*** fayablazer has quit IRC16:49
sdakeif openstack had a container scheduler we would be using that and we would be all set16:50
sdakethen we could tackle the bootstrappiing issue easily as well SpamapS :)16:50
*** cody-somerville has quit IRC16:51
*** ccrouch has quit IRC16:52
*** ccrouch has joined #tripleo16:52
*** ramishra has quit IRC16:53
ccrouchsdake: re: the second point, I just meant something like this in start.sh16:54
ccrouch$MARIADB_SERVICE_HOST=$MARIADB_PORT_3306_TCP_ADDR16:54
ccrouchso if you are not using k8s, therefore no _SERVICE_HOST, you could use container linking and it should all continue to work16:54
*** jistr has quit IRC16:55
openstackgerritDan Prince proposed a change to openstack/diskimage-builder: Support installing packages by default  https://review.openstack.org/13120616:57
sdakewhat is container linking?16:57
sdakeccrouch we could handle that in kolla-common pretty easily I htink (the mappings)16:58
sdakemariadb_port_3306 comes from k8s as well as mariadb_service_host16:58
sdakethey are both envs set by k8s16:58
SpamapSsdake: so I wonder if Kubernetes could be made multi-tenant the same way nova makes libvirt multi-tenant... with a thin (haha) openstack shim in front of it?16:58
sdakeSpamapS yes could be possible16:58
ccrouchsdake: https://docs.docker.com/userguide/dockerlinks/16:58
sdakeSpamapS possibly better would be a native scheduler + endpoint manager16:59
SpamapSor, ideally, Kubernetes would just support Keystone auth and one could simply use the Kubernetes API16:59
sdakek8s is not multitentant or secure for end users imo16:59
sdakeand it also only has one namespace16:59
sdakeso even ify ou wanted to pretend its multitenant you cant :(16:59
*** cwolferh has joined #tripleo16:59
SpamapSsdake: indeed, ok, so yeah needs a shim in front16:59
ccrouch(11:58:23 AM) sdake: they are both envs set by k8s16:59
ccrouchack, but mariadb_service_host *only* comes from k8s, plain "docker run --link mariadb:mariadb..." will provide mariadb_port_330616:59
SpamapSbut I like the idea of _not_ using Nova for that.17:00
SpamapSbecause containers just aren't VMs17:00
SpamapSand separating it means being able to possibly let users allocate vms separate from containers17:00
sdakeccrouch I see, I was not aware of that17:00
SpamapSThen again could we also just make it easier to spin up your own kubernetes on your own vms.17:00
sdakeSpamapS there is "magnum"17:01
SpamapSbecause that would also make it easier to deploy kolla on a regular old "undercloud"17:01
sdakedoes the undercloud virtualize the overcloud?17:01
sdakeor does it run basically on bare metal17:01
sdakeit being the overcloud17:01
sdakelet me restate17:01
ccrouchsdake: this is the only confirmation i found of the k8s behavior17:02
ccrouch https://github.com/GoogleCloudPlatform/kubernetes/blob/master/DESIGN.md#kubernetes-proxy17:02
sdakedoes the overcloud end up using the virt hardware extensions to do its basic operations17:02
sdakeccrouch got it, so *TCP_3306 comes from docker, SERVICE_HOST comes from k8s17:03
SpamapSsdake: the current arch has the undercloud only doing baremetal for the overcloud17:03
sdakethat seems easy enough to adjust17:03
daneyon_larsks: sdake: I have a model for multiple container network interfaces working besides one issue.17:03
SpamapSsdake: but it is entirely _possible_ to do vms too if you give the undercloud some compute nodes.17:03
sdakedaneyon_ nice!17:03
sdakeSpamapS I think we want to run on bare metal for perf reasons no?17:03
*** lazy_prince is now known as killer_prince17:03
SpamapSsdake: I believe one must use host aggregates to achieve this.17:03
sdakeso running k8s on openstack wouldn't fix it17:03
SpamapSsdake: we _must_ run baremetal for vm hosts.17:04
larsksdaneyon_: sounds spiffy.  got a link or something?17:04
sdakeright17:04
daneyon_sdake: larsks: do u have a few minutes to look at some go code?17:04
larsksdaneyon_: sure.17:04
sdakeI can look at go code17:04
sdakedaneyon_ did you write it yourself or something your using?17:04
SpamapSsdake: I'm quite certain we'd _LOVE_ to migrate non virt loads to vms.17:05
*** cody-somerville has joined #tripleo17:05
*** cody-somerville has joined #tripleo17:05
sdakeSpamapS wouldn't that make the performance suck?17:05
sdakevirt on virt is terribly slow on my system17:05
ccrouchsdake: *TCP_3306 comes from docker *and* k8s, SERVICE_HOST comes from k8s only17:05
sdakeeven running something in virt is slow because of all the memcpy and extra work the cpu is doing17:05
SpamapSsdake: I've failed to communicate what I mean.17:05
sdakeccrouch i see17:05
larsksccrouch: that's correct.  The *_service_host variables are only from k8s.17:06
SpamapSsdake: virt on virt is what we're avoiding by using baremetal.17:06
sdakeI failed to listen lets try again :)17:06
sdakeright17:06
SpamapSsdake: but things like keystone and nova-api don't need baremetal.17:06
sdakeoh and you want to virt those workloads?17:06
SpamapSCurrently we give that to them because it's simpler to not use host aggregates and mix and match.17:06
daneyon_sdake: larsks: So this is what I'm doing. I'm having docker spotter connect to the docker daemon and looking for events (start/restart/stop) related to a container with a name nova-network. When docker-spotter sees the event, it will issue a pipework command to create a veth on a kbr-ex bridge on the minion. https://github.com/discordianfish/docker-spotter/blob/master/spotter.go17:06
sdakekeystone needs all the cpu bandwidth it  can get imho :)_17:06
SpamapSsdake: sure, but it's easier to give it more CPU bandwidth if you can spread it out over the whole cloud. :)17:07
sdakeright17:07
sdakeand you don't want  to figure out how to horizontally scale bare metal I suppose17:07
daneyon_sdake: Larsks: the problem I have is the container name. By looking at the docker-spotter code, can I use a regex to specify a container name that includes nova-network since k8s appends a hash to the end of the k8s_nova-network container name?17:08
daneyon_sdake: larsks: or can i modify the spotter code to look for a hostname instead of a container name?17:09
*** Marga_ has quit IRC17:09
larsksdaneyon_: This smells a little hacky. I still think that if we really want to solve this problem it means patching k8s to perform the necessary network setup in the network container before starting up other containers in the pod.17:09
daneyon_sdake: larsks: as we discussed before, this approach gives us runway until k8s natively supports multiple interfaces.17:10
openstackgerritHaneef Ali proposed a change to openstack/tripleo-image-elements: Install openstackclient  https://review.openstack.org/13060817:10
sdakeok with hacky for a few weeks17:10
sdakeneutron is scheduled for friday (milestone #2)17:11
larsksdaneyon_: w/r/t to identifying containers correctly, can we work with the kubernetes api?  Either getting events from there, or using that to authoritatively identify that a container belongs to an appropriate pod?  I don't know.17:11
sdakedoubt we can sort out upstream in 1 week's time17:11
daneyon_larsks: agreed, this just gives us runway. In my testing (as long as the container is named nova-network), it works well. It is using tools such as pipework that are well known to the community.17:11
larskssdake: I think friday for having neutron ready is not realistic.17:11
sdakelarsks I added a dnote to the blueprint for comments from daneyon last week17:11
sdakeor this weekend I don't recall17:12
sdakeif he thinks its possible we will go with it17:12
openstackgerritSteven Hardy proposed a change to openstack/python-tuskarclient: Add --version option  https://review.openstack.org/13121117:12
larsksI have strong reservations.  But it's not my decision, so I will just eat popcorn and watch, I guess.17:12
sdakehaha its a team discussion17:12
sdakelets agenda for today's meeting17:12
sdakewe can vote if there is contention17:12
sdakewhich ther eappares to be :)17:13
*** gfidente has quit IRC17:13
sdakehow should I word the agenda item larsks?17:13
sdake"Integrate daneyon's slightly hacky solution for neutron into milestone #2?"17:13
daneyon_sdake: haha!17:14
larsks"Review goals regarding neutron in milestone 2"?17:14
larsksOr something?17:14
*** dshulyak has joined #tripleo17:18
SpamapSsdake, larsks btw.. if you haven't seen https://etherpad.openstack.org/p/kilo-tripleo-summit-topics .. I suggest maybe adding Kolla to this just so we have it on any agendas.17:20
larsksSpamapS: aye aye17:20
openstackgerritA change was merged to openstack/tripleo-image-elements: Neutron is not Ironic  https://review.openstack.org/13069117:21
*** Marga_ has joined #tripleo17:21
greghaynes^ FACT17:26
sdakelarsks sounds good17:27
sdakei'll do after this short meeting17:27
*** Marga_ has quit IRC17:35
*** Marga_ has joined #tripleo17:35
*** pradeep has quit IRC17:35
*** wfoster has quit IRC17:44
*** dshulyak has quit IRC17:46
*** killer_prince is now known as lazy_prince17:47
*** daneyon_ has quit IRC17:50
*** daneyon has joined #tripleo17:51
*** dshulyak has joined #tripleo17:51
greghaynesclarkb: lets talk here17:58
greghaynesclarkb: since this is somewhat relevant ;)17:58
greghaynesor maybe ill just sumarise17:59
greghaynespypi.o.o is supposedly dead17:59
greghaynesand we need to use a region specific one in our CI17:59
greghaynesI am going to change to pypi.p.o for now17:59
openstackgerritgreghaynes proposed a change to openstack-infra/tripleo-ci: pypi.openstack.org/simple is dead  https://review.openstack.org/13122918:01
greghaynesclarkb: ^ correct?18:01
clarkboui18:02
greghaynesSpamapS: lifeless dprince other cores, CI is all red due to ^18:03
openstackgerritSteven Hardy proposed a change to openstack/python-tuskarclient: Sync oslo-incubator  https://review.openstack.org/13123018:03
openstackgerritSteven Hardy proposed a change to openstack/python-tuskarclient: Move utils arg/env to oslo cliutils  https://review.openstack.org/13123118:03
openstackgerritSteven Hardy proposed a change to openstack/python-tuskarclient: Add a --debug option  https://review.openstack.org/13123218:03
bnemecgreghaynes: But we'll move to a region-specific one later, right?18:07
greghaynesbnemec: we should18:07
greghaynesthis is just the fix everything being on fire18:07
bnemecgreghaynes: Yeah, sounds good.  Do we have a bug open on this to track switching to a proper mirror?18:08
clarkbyou have a region specific one though right18:08
clarkbthe first index is local18:08
greghaynesclarkb: thats not an actualy mirror18:08
greghaynesclarkb: just a couple packages (maybe)18:08
greghaynesbnemec: negative, ill file one18:08
bnemecgreghaynes: Cool, +2.18:09
*** bnemec changes topic to "CI: Broken due to pypi.o.o death. Fix proposed here: https://review.openstack.org/131229 | Using OpenStack to deploy OpenStack; meetings Tuesday 1900/0800 UTC in #openstack-meeting-alt"18:10
SpamapSso what are these regional mirrors?18:12
greghaynesclarkb: ^18:12
SpamapSalso should we just maintain mirrors ourselves in each CI region and use a DNS search domain or something?18:12
greghaynesSpamapS: I think we basically need to spin one up ourselves18:12
greghaynesSpamapS: I think so, we should ask -infra how they do it18:13
*** ifarkas has quit IRC18:13
*** pradeep has joined #tripleo18:13
SpamapSwould be great if we could just fetch the tarball of the old one or something. :_P18:14
*** lazy_prince is now known as killer_prince18:15
SpamapSgreghaynes: why was the old one so rapidly removed?18:16
greghaynesThese are all good questions18:16
greghaynesI just noticed the red and fixed it18:16
sdakelarsks is there a dnf package manager for centos packaged you are aawre of (jpeeler also?)18:17
clarkbSpamapS: currently pypi.region.openstack.org18:18
clarkbso ord, iad, dfw, and geo-1.region-b18:18
clarkbif tripleo* clouds have disk and cpu to spare we can spin up there too likely18:19
*** Marga_ has quit IRC18:19
*** Marga_ has joined #tripleo18:19
larskssdake: i don't know the status of dnf outside Fedora. You can just use curl to fetch the repo config18:20
clarkbbut I think infra has been wary of doing long lived service nodes in those regions beacuse 1) those regions seem to disappear at a regular rate and 2) they are resource constrained18:20
*** pblaho has quit IRC18:21
jpeelersdake: looks like dnf is packaged for epel 7 - not sure if that's helpful18:22
sdakejpeeler link?18:25
sdakelarsks I just used ADD18:25
jpeelersdake: i take that back, they have a very old spec file in dist-git. but no builds: https://apps.fedoraproject.org/packages/dnf18:26
openstackgerritPhil Neal proposed a change to openstack/tripleo-heat-templates: Add overcloud options for SSL RabbitMQ connections  https://review.openstack.org/11783018:26
*** dshulyak has quit IRC18:28
*** wfoster has joined #tripleo18:29
*** dshulyak has joined #tripleo18:29
SpamapSclarkb: they don't disappear that often anymore. CPU I'm surprised by. How does that figure?18:30
*** jerryz has joined #tripleo18:30
clarkbSpamapS: you need a VM to update the mirror18:31
clarkband iirc you guys have sized everything out without considering services like that18:31
clarkbso all resources are accounted for18:31
SpamapSclarkb: I think our vms are not super packed.. they're limited by our testenvs which live outside the cloud vms.18:32
SpamapSclarkb: if the vm was just 2CPU and 4G .. I bet we could find room for something like that.18:32
*** pradeep1 has joined #tripleo18:32
SpamapSgreghaynes: ^ concur?18:33
clarkbthat appears to be about what we are using now18:33
clarkbthat and 200GB of disk18:33
greghaynesYep, wed need to see about the rh region also18:33
*** pradeep has quit IRC18:33
SpamapS2014-10-27 18:05:18.512 | --2014-10-27 18:05:18--  http://dl.fedoraproject.org/pub/fedora/linux/updates/20/Images/i386/18:34
SpamapS2014-10-27 18:05:18.513 | Connecting to 192.168.1.100:3128... failed: Connection refused.18:34
SpamapSit's sad mirror day18:34
clarkbdl.fedoraproject doesn't use their mirrors iirc18:34
clarkbwe recently switching devstack to using the mirrors instead of the dl address18:34
clarkb(because it was taking hours to get one image in rax land018:34
SpamapSoh wait18:34
SpamapSthat's a failure on the squid18:34
greghaynes192.168.1?18:35
greghaynesyea18:35
SpamapS2014-10-27 18:01:49.792 | Building remotely on tripleo-f20-tripleo-test-cloud-rh1-2797831 (tripleo-f20) in workspace /home/jenkins/workspace/check-tripleo-ironic-overcloud-f20-nonha18:35
SpamapSgreghaynes: can you show me (for the millionth time) where info on how to log in to rh1 is?18:35
greghaynesThere is a spreadsheet that I am not a part of currently18:36
greghaynes(I think)18:36
greghaynesalso, can we log into rh1?18:36
SpamapSI thought we could18:37
SpamapSdprince: ^ halp?18:37
greghaynesdprince: ^18:37
greghayneshehe18:37
* SpamapS mutters about derekh and bouncers18:37
greghaynesmaybe we should put that on the sumit etherpad18:37
SpamapSwhat? "be better at ops" ?18:38
greghaynes"get derekh a persistent irc"18:38
dprincegreghaynes: you should have access.18:38
greghaynesoh!18:38
jerryzSpamapS: sorry to interrupt, but i need some help on rabbitmq behind haproxy and keepalived. i am talking about the service check-in failure issue i met the other day. later i point all nova-compute and neutron ovs agents directly to the active controller, everything went to normal.18:38
greghaynesdprince: to the spreadsheet?18:39
*** jprovazn_biab is now known as jprovazn18:39
jerryzSpamapS: it seemed there were many timeouts and socket closed when connecting rabbitmq through haproxy.18:40
SpamapSjerryz: I'm actually not really very excited about running rabbitmq behind haproxy, but people tell me it should work. :-/18:41
dprincegreghaynes: if you want spreadsheet access I'll add you. I need your gmail ID18:41
dprincegreghaynes: it is a google doc18:42
greghaynesdprince: greg@greghaynes.net18:42
greghaynesshould be google-ey18:42
* SpamapS grumbles something about git18:42
dprincegreghaynes: added you. FWIW there are 9 people now who have access to this. Not just derek :)18:43
*** pradeep1 has quit IRC18:43
greghaynessweet18:43
*** dshulyak has quit IRC18:43
SpamapSjerryz: I would expect that you would want longer timeouts for rabbitmq since there may not be activity for a little while. If the defaults were not working, could you file a bug at https://bugs.launchpad.net/tripleo/+filebug  ?18:43
greghaynesjerryz: SpamapS There was a patch for just that a while ago18:43
SpamapSgreghaynes: AFAIK jerryz is deploying from master18:44
SpamapSwait no18:44
jerryzSpamapS: how could i disable haproxy when building the disk image? even though i reduced controller scale to 1, haproxy is still configured18:44
SpamapSjerryz: are you doing juno stable branches?18:44
*** Marga_ has quit IRC18:44
greghaynesIIRC we set rmq to have infinite timeouts at the app level in haproxy and then set the tcp_keepalive to have a value18:44
jerryzSpamapS: no, i am building the overcloud image off the trunk18:44
jerryzthe trunk 2 weeks ago18:45
SpamapSah that may be it18:45
SpamapSgreghaynes: could you track that commit down?18:45
greghaynesIt was way before then /me finds anyways18:45
lifelessmorning18:46
greghaynesohai18:47
greghaynesjerryz: SpamapS https://bugs.launchpad.net/nova/+bug/85676418:48
uvirtbotLaunchpad bug 856764 in oslo.messaging "RabbitMQ connections lack heartbeat or TCP keepalives" [Critical,In progress]18:48
greghaynesthats the original bug18:48
greghayneshttps://review.openstack.org/#/c/108647/18:49
greghaynesSo that sets timeout to 0 which should be infinite, but we do have tcpka set so it doesnt mean dead connections would live on forever18:50
openstackgerritPhil Neal proposed a change to openstack/tripleo-image-elements: Add ssl options for ceilometer rabbitmq connection  https://review.openstack.org/11782118:53
openstackgerritSteven Hardy proposed a change to openstack/python-tuskarclient: Move utils arg/env to oslo cliutils  https://review.openstack.org/13123118:54
openstackgerritSteven Hardy proposed a change to openstack/python-tuskarclient: Add a --debug option  https://review.openstack.org/13123218:54
jerryzgreghaynes: the patch was already merged when i built the overcloud image.18:55
greghaynesjerryz: I wonder if your keepelived is spinning18:55
jerryzi checked the overcloud.yaml, it has the timeout 0 set18:55
greghaynesjerryz: have you looked in your syslog to see if its spamming stuff in there about failing over18:56
jerryzgreghaynes: now i only have one controller18:56
greghayneshrm, we might still use a VIP in the one node setup case18:56
jerryzgreghaynes: VIP is always reachable18:57
greghaynes*shrug*18:57
jerryzgreghaynes: i can call the api through keepalived+haproxy18:58
jerryzgreghayes: just that rabbitmq has many errors in haproxy status18:59
jerryzstatistics18:59
jerryzgreghayes: http://picpaste.com/pics/Capture-bgpbqoqC.1414436673.JPG  i am not sure if it is related19:05
openstackgerritSteven Dake proposed a change to stackforge/kolla: Prepare for multi-base image support in Kolla  https://review.openstack.org/13124519:05
openstackgerritSteven Dake proposed a change to stackforge/kolla: Move the links down one level for all of the build scripts  https://review.openstack.org/13124619:05
openstackgerritSteven Dake proposed a change to stackforge/kolla: Add a centos-rdo-base container  https://review.openstack.org/13124719:05
sdakewhy is fedora-rdo-base 1.4gb, and centos-rdo-base 256mb?19:05
sdakeguess fedora needs trimming19:06
sdakeusing centos should speed up load times on caching significantly since the images will be smaller19:07
*** dshulyak has joined #tripleo19:10
openstackgerritDan Prince proposed a change to openstack/diskimage-builder: Support installing packages by default  https://review.openstack.org/13120619:13
*** sdake has quit IRC19:26
*** dshulyak has quit IRC19:26
*** spzala has joined #tripleo19:27
*** sdake has joined #tripleo19:28
jerryzSpamapS: is there a way to run a stand alone rabbitmq?19:32
openstackgerritPhil Neal proposed a change to openstack/tripleo-heat-templates: Add overcloud options for SSL RabbitMQ connections  https://review.openstack.org/11783019:36
SpamapSjerryz: meaning without the haproxy? I think it might require a slight modification to the Heat template.19:40
*** sdake has quit IRC19:41
*** victor_lowther__ has quit IRC19:43
*** victor_lowther__ has joined #tripleo19:44
*** sdake has joined #tripleo19:45
*** sdake has quit IRC19:45
*** sdake has joined #tripleo19:45
*** dprince has quit IRC19:46
jerryzSpamapS: it is weird that after i changed rabbit host back to VIP , everything is still fine now.19:47
*** lucasagomes has joined #tripleo19:48
SpamapSjerryz: maybe keepalived was having issues before?19:50
jerryzSpamapS: maybe. after heat stack-update19:50
*** sdake_ has joined #tripleo19:51
*** sdake_ has quit IRC19:51
*** sdake_ has joined #tripleo19:51
jerryzSpamapS: let see if another heat stack-update will reproduce it.19:51
*** marun_ has quit IRC19:52
*** jpeeler has quit IRC19:53
jerryzSpamapS: btw, why almost everytime i do the heat stack-update, first time would fail because of ip conflict telling me "IpAddressInUseClient: Unable to complete operation for network 1ec0c6a1-dc70-426e-a1b2-d36147e3a302. The IP address 162.3.121.200 is in use."19:54
jerryzSpamapS: i need retry.19:55
*** jpeeler has joined #tripleo19:55
jerryzSpamapS: it did reproduce the issue20:04
*** avozza is now known as zz_avozza20:04
*** zz_avozza is now known as avozza20:06
*** Marga_ has joined #tripleo20:11
lifelessbnemec: / slagle - os-cloud-config. Do you care about py26 support? https://review.openstack.org/#/c/129435/5/zuul/layout.yaml20:20
bnemeclifeless: I don't believe so.  We're not supporting tripleo on rhel6.20:21
*** dtantsur is now known as dtantsur|afk20:21
lifelessthanks20:21
jerryzSpamapS, greghaynes: https://bugs.launchpad.net/tripleo/+bug/138640620:22
uvirtbotLaunchpad bug 1386406 in tripleo "Nova compute services check-in timeouts when register to rabbitmq behind haproxy+keepalived" [Undecided,New]20:22
*** Marga_ has quit IRC20:24
openstackgerritSteven Hardy proposed a change to openstack/python-tuskarclient: Add a --debug option  https://review.openstack.org/13123220:25
*** jehb has quit IRC20:28
*** andreaf has joined #tripleo20:31
*** jtomasek has quit IRC20:33
*** lucasagomes has quit IRC20:38
*** jerryz has quit IRC20:39
*** rlandy_ has quit IRC20:43
*** jdob has quit IRC20:53
*** radez is now known as radez_g0n320:55
*** spzala has quit IRC20:55
jpeelerlarsks: do you think something like fig could be used to bring up kolla? "fig up" and go would be pretty neat20:57
larsksjpeeler: maaaaaybe. Fig will not handle network configuration the same way as k8s...20:58
*** yamahata has joined #tripleo20:59
larsksjpeeler: I've only looked at it briefly, and somehow it never quite does what I want :)20:59
jpeelerjust wondering - i didn't know enough to really judge21:00
*** jang1 has joined #tripleo21:19
greghayneshrm, so do we want to wait on http://logs.openstack.org/29/131229/1/check-tripleo/check-tripleo-ironic-overcloud-f20-nonha/2c53a02/console.html to merge https://review.openstack.org/#/c/131229/21:19
greghaynesIMO it should be fine to just ram it through...21:19
*** akrivoka has quit IRC21:23
openstackgerritA change was merged to openstack-infra/tripleo-ci: pypi.openstack.org/simple is dead  https://review.openstack.org/13122921:27
bnemecgreghaynes: Agreed.21:27
greghaynesw00t21:27
*** jprovazn has quit IRC21:28
*** andreaf has quit IRC21:43
*** andreaf has joined #tripleo21:43
*** andreaf has quit IRC21:55
*** weshay has quit IRC21:55
*** andreaf has joined #tripleo21:56
openstackgerritgreghaynes proposed a change to openstack/tripleo-image-elements: Migrate to package-installs scripts  https://review.openstack.org/12029321:58
*** untriaged-bot has joined #tripleo22:00
untriaged-botUntriaged bugs so far:22:00
untriaged-bothttps://bugs.launchpad.net/python-tuskarclient/+bug/138246822:00
untriaged-bothttps://bugs.launchpad.net/tripleo/+bug/138640622:00
uvirtbotLaunchpad bug 1382468 in tuskar "Bad parameter exception" [Undecided,Triaged]22:00
uvirtbotLaunchpad bug 1386406 in tripleo "Nova compute services check-in timeouts when register to rabbitmq behind haproxy+keepalived" [Undecided,New]22:00
untriaged-bothttps://bugs.launchpad.net/diskimage-builder/+bug/138635322:00
uvirtbotLaunchpad bug 1386353 in diskimage-builder "Image builds on Ubuntu 14.04 fails with errors" [Undecided,New]22:00
untriaged-bothttps://bugs.launchpad.net/tripleo/+bug/138523422:00
uvirtbotLaunchpad bug 1385234 in tripleo "OVS tunneling between multiple neutron nodes misconfigured if amqp is restarted" [High,New]22:00
untriaged-bothttps://bugs.launchpad.net/tuskar/+bug/138246822:00
*** untriaged-bot has quit IRC22:00
*** rhallisey has quit IRC22:05
sdakedaneyon about?22:06
daneyonsdake: hey22:06
*** yamahata has quit IRC22:10
*** morazi has quit IRC22:11
sdakelarsks we will need to move to fc21 (I think we are on 20 atm)22:14
sdakemany of the newer openstack services are only available there22:14
sdakewe may not have to move, i'll find out22:14
daneyonsdake: what does fc mean?22:16
sdakefedora core22:16
ccrouchsdake: F21 isn't due for more than another month right?22:17
ccrouchhttps://fedoraproject.org/wiki/Releases/21/Schedule22:17
sdakeccrouch yes, I'm not sure why its in my repo list on my dev box22:18
sdakemaybe its already built in RDO and tagged fc2122:18
sdakewe use fedora + RDO22:18
ccrouchhttps://repos.fedorapeople.org/repos/openstack/openstack-juno/fedora-20/?C=M;O=D22:18
sdakewe may be ok with just fc2022:18
ccrouchright, i think thats it22:18
ccrouchwe ran fc20 + RDO icehouse tagged with fc21 fine22:19
ccrouchand have been testing with22:19
sdakecool22:19
ccrouchfc20 + RDO juno tagged with fc2222:19
ccrouchso i think you should be fine22:19
*** andreaf has quit IRC22:20
*** andreaf has joined #tripleo22:21
*** spzala has joined #tripleo22:24
*** sseago has quit IRC22:30
*** sseago has joined #tripleo22:30
sdakedaneyon did you catch my PMs?22:39
daneyonsdake: I must have missed it.22:40
daneyonsdake: looking at them now.22:40
sdakeoops I just c&P :)22:40
ccrouchlarsks: fyi: http://kilodesignsummit.sched.org/event/4f22d2febeaca88fe9f380c9d1dfdf8a22:44
*** spzala has quit IRC22:47
openstackgerritSrinivas Sakhamuri proposed a change to openstack/tripleo-heat-templates: HA proxy to limit the ceilometer-api maxconns  https://review.openstack.org/11237022:51
*** andreaf has quit IRC23:03
*** jang1 has quit IRC23:20
openstackgerritHaneef Ali proposed a change to openstack/tripleo-image-elements: Install openstackclient  https://review.openstack.org/13060823:20
*** sdake_ has quit IRC23:31
*** avozza is now known as zz_avozza23:34
*** rhallisey has joined #tripleo23:36
StevenKgreghaynes: I didn't know about cliff23:55
greghaynesyea, I was just poking around and saw it23:57
greghaynesmight be a neat thing to do at some point23:57
« Sunday, 2014-10-26 Index Tuesday, 2014-10-28 »

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