Tuesday, 2016-11-15

*** dave-mccowan has joined #openstack-containers00:17
*** selena_091 has joined #openstack-containers00:33
*** yuanying_ has joined #openstack-containers00:34
*** yuanying_ has quit IRC00:38
*** catintheroof has quit IRC00:47
*** catintheroof has joined #openstack-containers00:48
*** catintheroof has quit IRC00:52
*** hongbin has joined #openstack-containers00:58
*** chris_hultin is now known as chris_hultin|AWA00:58
*** tovin07_ has joined #openstack-containers01:05
*** selena_091 has quit IRC01:07
*** tkarke has quit IRC01:09
*** EricGonczer_ has joined #openstack-containers01:09
*** dave-mccowan has quit IRC01:45
*** fragatin_ has joined #openstack-containers01:46
*** fragatin_ has quit IRC01:47
*** fragatin_ has joined #openstack-containers01:48
*** fragatina has quit IRC01:49
*** fragatin_ has quit IRC01:50
*** chetnakh has quit IRC02:04
*** syed__ has quit IRC02:25
*** wangbo has joined #openstack-containers02:34
*** yamamoto_ has joined #openstack-containers02:39
*** Drago has quit IRC02:39
*** wangbo has quit IRC02:39
*** Drago has joined #openstack-containers02:40
*** kevinz has joined #openstack-containers02:46
*** yuanying has quit IRC02:51
yatinDrago: Thanks02:53
*** wangbo has joined #openstack-containers02:53
yatinchetnakh, tonanhngo: The reason i heard/discussed in some patch for two security groups was: One security group for general purpose(like 22,80,443) and other secgroup coe specific(kubernetes, swarm, etc)02:55
*** kevin__ has joined #openstack-containers02:59
*** hongbin has quit IRC03:00
*** hongbin has joined #openstack-containers03:01
*** jwcroppe has quit IRC03:05
*** jwcroppe has joined #openstack-containers03:05
*** EricGonc_ has joined #openstack-containers03:05
*** EricGonczer_ has quit IRC03:06
*** jwcroppe has quit IRC03:06
*** jwcroppe has joined #openstack-containers03:08
*** GB21 has joined #openstack-containers03:08
openstackgerrityatin proposed openstack/magnum: Set config param [DEFAULT]/host to hostname  https://review.openstack.org/38223503:11
openstackgerritTony Breeds proposed openstack/python-magnumclient: Revert "Fix: Creation of cluster without floating ip"  https://review.openstack.org/39750303:14
*** vijendar has joined #openstack-containers03:16
*** vijendar1 has joined #openstack-containers03:16
*** GB21 has quit IRC03:19
*** dimtruck is now known as zz_dimtruck03:27
*** fragatina has joined #openstack-containers03:29
*** sdake has quit IRC03:36
*** yuanying has joined #openstack-containers03:39
*** vijendar has quit IRC03:40
*** sdake has joined #openstack-containers03:44
*** yuanying has quit IRC03:45
*** vijendar has joined #openstack-containers03:47
*** yuanying has joined #openstack-containers03:49
*** sdake has quit IRC04:03
*** Drago has quit IRC04:03
*** vijendar has quit IRC04:04
*** vijendar has joined #openstack-containers04:12
*** zz_dimtruck is now known as dimtruck04:16
*** vijendar has quit IRC04:19
*** vijendar has joined #openstack-containers04:20
*** sdake has joined #openstack-containers04:22
*** hongbin has quit IRC04:38
*** hongbin has joined #openstack-containers04:39
*** EricGonc_ has quit IRC04:42
*** EricGonczer_ has joined #openstack-containers04:42
*** vijendar has quit IRC04:50
*** vijendar has joined #openstack-containers04:55
*** chetnakh has joined #openstack-containers04:57
*** dimtruck is now known as zz_dimtruck04:57
*** chetnakh has quit IRC04:58
*** khamtamtun has joined #openstack-containers05:04
*** vijendar has quit IRC05:09
*** vijendar has joined #openstack-containers05:19
*** vmud213 has joined #openstack-containers05:20
*** khamtamtun has quit IRC05:21
*** chetnakh has joined #openstack-containers05:28
*** chetnakh has quit IRC05:29
*** fragatina has quit IRC05:30
*** fragatina has joined #openstack-containers05:31
*** fragatina has quit IRC05:35
*** hongbin has quit IRC05:36
*** vijendar has quit IRC05:39
*** syed__ has joined #openstack-containers05:43
*** chetnakh has joined #openstack-containers05:44
*** vimal has joined #openstack-containers05:45
*** chetnakh has quit IRC05:46
*** manikanta_tadi has joined #openstack-containers05:46
*** vijendar has joined #openstack-containers05:49
*** Jeffrey4l has quit IRC06:03
*** kevinz has quit IRC06:07
*** vijendar has quit IRC06:08
*** vijendar has joined #openstack-containers06:16
*** khamtamtun has joined #openstack-containers06:17
tonanhngoHi yatin,  that would makes sense if we share templates between the COE's.06:18
*** chetnakh has joined #openstack-containers06:19
tonanhngoIn this case, the Kubernetes set of templates is used by itself06:20
*** tkarke has joined #openstack-containers06:20
*** chetnakh has quit IRC06:20
*** bks has joined #openstack-containers06:21
*** tkarke has quit IRC06:22
*** yamamoto_ has quit IRC06:27
yatintonanhngo: I think idea was to create two secgroup(coe specific/ general). coe specific is fixed(cannot be changed at cluster creation/updation time) The other secgroup can be changed during cluster-create and cluster-update06:33
*** khamtamtun has quit IRC06:35
*** tkarke has joined #openstack-containers06:36
*** tkarke has quit IRC06:37
*** Jeffrey4l has joined #openstack-containers06:41
*** mjura has joined #openstack-containers06:43
*** vijendar has quit IRC06:43
*** khamtamtun has joined #openstack-containers06:48
*** chhavi has joined #openstack-containers06:51
*** vijendar has joined #openstack-containers06:52
*** nadeem_ has joined #openstack-containers07:01
*** vijendar has quit IRC07:03
*** vijendar has joined #openstack-containers07:04
*** vijendar has quit IRC07:06
*** nadeem_ has quit IRC07:06
*** masayukig_ has joined #openstack-containers07:11
*** masayukig has quit IRC07:12
*** masayukig_ is now known as masayukig07:13
*** Chris_BKK has joined #openstack-containers07:14
Chris_BKKHello guys quick question to magnum07:14
Chris_BKKdoe sit require keystone v3 or is v2.0 also Ok ?07:14
*** vijendar has joined #openstack-containers07:15
Chris_BKKok found it: https://blueprints.launchpad.net/magnum/+spec/pluggable-keystone-model07:21
Chris_BKKwhere does i need to do the changes?07:21
yatinChris_BKK: It should work with v2.0 as well,07:21
*** rcernin has quit IRC07:23
*** manikanta_tadi has quit IRC07:27
*** manikanta_tadi has joined #openstack-containers07:32
*** nadeem_ has joined #openstack-containers07:32
*** belmoreira has joined #openstack-containers07:33
*** chetnakh has joined #openstack-containers07:34
*** chetnakh has quit IRC07:36
*** vijendar has quit IRC07:40
*** tonanhngo has quit IRC07:42
*** yamamot__ has joined #openstack-containers07:46
*** nadeem_ has quit IRC07:47
*** vijendar has joined #openstack-containers07:47
*** sdake_ has joined #openstack-containers07:49
*** sdake has quit IRC07:53
*** syed__ has quit IRC07:55
*** vijendar has quit IRC08:00
*** tonanhngo has joined #openstack-containers08:01
*** tonanhngo has quit IRC08:02
*** nadeem_ has joined #openstack-containers08:04
*** chhavi has quit IRC08:05
*** vijendar has joined #openstack-containers08:07
*** tkarke has joined #openstack-containers08:11
*** tkarke has quit IRC08:12
*** xiangxinyong has quit IRC08:13
*** chetnakh has joined #openstack-containers08:19
*** pcaruana has joined #openstack-containers08:22
*** tovin07_ has quit IRC08:24
*** chhavi has joined #openstack-containers08:24
*** tkarke has joined #openstack-containers08:27
*** shu-mutou is now known as shu-mutou-AWAY08:27
*** tkarke has quit IRC08:28
*** zul has quit IRC08:33
*** vijendar has quit IRC08:37
*** zul has joined #openstack-containers08:44
*** vijendar has joined #openstack-containers08:47
openstackgerritOpenStack Proposal Bot proposed openstack/magnum-ui: Imported Translations from Zanata  https://review.openstack.org/39758408:51
*** Chris_BKK has quit IRC09:01
*** vijendar has quit IRC09:10
*** openstackgerrit has quit IRC09:18
*** openstackgerrit has joined #openstack-containers09:19
*** vijendar has joined #openstack-containers09:20
*** jperry has joined #openstack-containers09:21
*** sdake_ has quit IRC09:23
*** chhavi has quit IRC09:25
*** vijendar has quit IRC09:50
*** vijendar has joined #openstack-containers09:58
*** chetnakh has quit IRC09:59
*** sdake has joined #openstack-containers10:02
*** chetnakh has joined #openstack-containers10:05
*** chetnakh has quit IRC10:06
*** sdake_ has joined #openstack-containers10:08
*** sdake has quit IRC10:11
*** vijendar has quit IRC10:13
*** tonanhngo has joined #openstack-containers10:18
*** tonanhngo has quit IRC10:18
*** vijendar has joined #openstack-containers10:22
*** fragatina has joined #openstack-containers10:31
*** fragatina has quit IRC10:36
*** tkarke has joined #openstack-containers10:38
*** khamtamtun has quit IRC10:39
*** tkarke has quit IRC10:40
*** sdake_ is now known as sdake10:40
openstackgerritMathieu Velten proposed openstack/magnum: Use Kubernetes config defined in /etc/kubernetes/* files to launch the scheduler and controller-manager.  https://review.openstack.org/39766810:43
openstackgerritMathieu Velten proposed openstack/magnum: Use Kubernetes config to launch services pods.  https://review.openstack.org/39766810:44
openstackgerritMathieu Velten proposed openstack/magnum: Move kube-apiserver to a pod container.  https://review.openstack.org/39767010:48
openstackgerritMerged openstack/magnum: Set config param [DEFAULT]/host to hostname  https://review.openstack.org/38223510:51
*** tkarke has joined #openstack-containers10:54
*** tkarke has quit IRC10:55
*** vijendar has quit IRC10:55
*** chhavi has joined #openstack-containers10:58
*** khamtamtun has joined #openstack-containers11:01
*** vijendar has joined #openstack-containers11:03
*** tkarke has joined #openstack-containers11:07
*** tkarke has quit IRC11:08
*** tkarke has joined #openstack-containers11:09
*** chetnakh has joined #openstack-containers11:09
*** tkarke has quit IRC11:10
*** chetnakh has quit IRC11:11
*** vijendar has quit IRC11:18
*** tkarke has joined #openstack-containers11:24
*** tkarke has quit IRC11:25
*** vijendar has joined #openstack-containers11:29
*** zenoway has joined #openstack-containers11:33
*** houming has joined #openstack-containers11:40
*** sdake has quit IRC11:45
*** chetnakh has joined #openstack-containers11:48
*** houming has quit IRC11:49
*** chetnakh has quit IRC11:50
*** chetnakh has joined #openstack-containers11:52
*** vijendar has quit IRC11:53
*** chetnakh has quit IRC11:54
*** bks has quit IRC11:57
*** vijendar has joined #openstack-containers12:01
*** zenoway has quit IRC12:04
*** zenoway has joined #openstack-containers12:04
*** zenoway has quit IRC12:09
*** tonanhngo has joined #openstack-containers12:13
*** tonanhngo has quit IRC12:15
*** zenoway has joined #openstack-containers12:18
*** vijendar has quit IRC12:20
*** zenoway has quit IRC12:22
*** yamamot__ has quit IRC12:25
*** zenoway has joined #openstack-containers12:27
*** catintheroof has joined #openstack-containers12:28
*** vijendar has joined #openstack-containers12:29
*** zenoway has quit IRC12:36
*** zenoway has joined #openstack-containers12:36
*** vmud213 has quit IRC12:43
*** chetnakh has joined #openstack-containers12:44
*** chetnakh has quit IRC12:45
*** nadeem_ has quit IRC12:48
*** vijendar has quit IRC12:53
*** chhavi has quit IRC12:59
*** vijendar has joined #openstack-containers13:00
*** dave-mccowan has joined #openstack-containers13:03
*** khamtamtun has quit IRC13:16
*** sdake has joined #openstack-containers13:16
*** vijendar has quit IRC13:16
*** chetnakh has joined #openstack-containers13:19
*** chetnakh has quit IRC13:21
*** vijendar has joined #openstack-containers13:27
*** chhavi has joined #openstack-containers13:32
openstackgerritMichal Jura proposed openstack/magnum: [suse] Build openSUSE Leap 42.1 OpenStack Magnum image  https://review.openstack.org/39774313:37
*** yolanda has quit IRC13:39
*** jwcroppe has quit IRC13:47
*** chetnakh has joined #openstack-containers13:50
*** chetnakh has quit IRC13:52
*** jperry has quit IRC13:53
*** vijendar has quit IRC13:54
*** yolanda has joined #openstack-containers13:55
*** khamtamtun has joined #openstack-containers13:58
*** vijendar has joined #openstack-containers14:03
*** yolanda has quit IRC14:07
*** vijendar has quit IRC14:10
*** jperry has joined #openstack-containers14:10
*** tonanhngo has joined #openstack-containers14:12
*** tonanhngo has quit IRC14:13
*** pcaruana has quit IRC14:13
*** khamtamtun has quit IRC14:16
*** randallburt has joined #openstack-containers14:17
*** chris_hultin|AWA is now known as chris_hultin14:18
*** vijendar has joined #openstack-containers14:19
*** randallburt has quit IRC14:22
*** yolanda has joined #openstack-containers14:23
*** sdake has quit IRC14:23
*** chris_hultin is now known as chris_hultin|AWA14:24
*** sdake has joined #openstack-containers14:24
*** vimal has quit IRC14:25
*** zz_dimtruck is now known as dimtruck14:26
*** chris_hultin|AWA is now known as chris_hultin14:26
*** randallburt has joined #openstack-containers14:29
*** vijendar has quit IRC14:36
*** yamamoto has joined #openstack-containers14:36
*** chetnakh has joined #openstack-containers14:37
*** yamamoto_ has joined #openstack-containers14:37
*** chetnakh has quit IRC14:38
*** yamamoto has quit IRC14:41
*** dimtruck is now known as zz_dimtruck14:42
*** vijendar has joined #openstack-containers14:44
*** khamtamtun has joined #openstack-containers14:45
*** khamtamtun has quit IRC14:50
*** vimal has joined #openstack-containers14:51
*** Drago1 has joined #openstack-containers14:51
*** Drago1 has quit IRC14:55
*** vimal has quit IRC14:55
*** jwcroppe has joined #openstack-containers14:56
*** vijendar has quit IRC14:58
*** yamamoto_ has quit IRC14:58
*** yamamoto has joined #openstack-containers15:00
*** rpothier has joined #openstack-containers15:00
*** jasond has joined #openstack-containers15:01
*** Drago1 has joined #openstack-containers15:01
*** jmckind has joined #openstack-containers15:01
*** zz_dimtruck is now known as dimtruck15:01
*** EricGonczer_ has quit IRC15:02
*** tkarke has joined #openstack-containers15:04
*** EricGonczer_ has joined #openstack-containers15:05
*** vijendar has joined #openstack-containers15:06
*** tkarke has quit IRC15:07
*** yamamoto has quit IRC15:07
*** tonanhngo has joined #openstack-containers15:08
*** tonanhngo has quit IRC15:09
*** syed__ has joined #openstack-containers15:10
*** yamamoto has joined #openstack-containers15:11
*** yamamoto has quit IRC15:12
*** hongbin has joined #openstack-containers15:14
*** yamamoto has joined #openstack-containers15:14
*** yamamoto has quit IRC15:15
*** vijendar has quit IRC15:18
*** chhavi has quit IRC15:22
*** yamamoto has joined #openstack-containers15:23
*** yamamoto has quit IRC15:23
*** chetnakh has joined #openstack-containers15:25
*** chetnakh has quit IRC15:26
*** vijendar has joined #openstack-containers15:27
*** EricGonc_ has joined #openstack-containers15:30
*** yamamoto has joined #openstack-containers15:31
*** yamamoto has quit IRC15:32
*** tonanhngo has joined #openstack-containers15:32
*** EricGonczer_ has quit IRC15:33
*** chetnakh has joined #openstack-containers15:35
*** chetnakh has quit IRC15:36
*** vijendar has quit IRC15:42
*** vijendar has joined #openstack-containers15:42
*** adrian_otto has joined #openstack-containers15:43
*** jvgrant__ has quit IRC15:45
*** vijendar2 has joined #openstack-containers15:45
adrian_ottoOur team meeting will begin in 15 minutes in #openstack-meeting-alt at 1600 UTC.15:45
*** jvgrant has joined #openstack-containers15:45
*** dimtruck is now known as zz_dimtruck15:45
*** vijendar1 has quit IRC15:46
*** openstackgerrit has quit IRC15:48
*** openstackgerrit has joined #openstack-containers15:48
*** mvelten has joined #openstack-containers15:49
*** absubram has joined #openstack-containers15:50
openstackgerritSpyros Trigazis proposed openstack/magnum: [WIP] Add cluster_attributes table  https://review.openstack.org/39501215:53
*** yamamoto has joined #openstack-containers15:57
*** dave-mccowan has quit IRC15:59
*** hieulq_ has joined #openstack-containers15:59
*** yamamoto has quit IRC16:01
*** yamamoto has joined #openstack-containers16:01
*** pcaruana has joined #openstack-containers16:02
*** zz_dimtruck is now known as dimtruck16:02
*** Drago1 is now known as Drago16:02
*** vijendar has quit IRC16:04
*** vijendar2 has quit IRC16:04
*** dimtruck is now known as zz_dimtruck16:05
*** yamamoto has quit IRC16:06
*** hieulq__ has joined #openstack-containers16:08
*** hieulq__ has quit IRC16:09
openstackgerritMerged openstack/magnum-specs: Initial commit  https://review.openstack.org/39567316:10
*** cooldharma06 has joined #openstack-containers16:13
*** zenoway has quit IRC16:17
*** dave-mccowan has joined #openstack-containers16:17
*** zenoway has joined #openstack-containers16:17
*** ramishra has quit IRC16:19
*** ramishra has joined #openstack-containers16:20
*** cooldharma06 has quit IRC16:21
*** zenoway has quit IRC16:22
*** belmoreira has quit IRC16:22
*** chetnakh has joined #openstack-containers16:23
*** chetnakh has quit IRC16:24
*** yamamoto has joined #openstack-containers16:27
*** yamamoto has quit IRC16:35
*** zz_dimtruck is now known as dimtruck16:40
*** swatson_ has joined #openstack-containers16:42
swatson_adrian_otto: Ping16:51
adrian_ottohi16:51
adrian_ottowhat's up?16:51
swatson_https://review.openstack.org/#/c/388215/16:51
swatson_Can you check out the replies on that patch?16:51
adrian_otto"I have one implementation patch that's out now. Do you think I should move these strings to that patch?" answer: yes, please16:52
adrian_ottonot super important, but just would be a little bit better form from a purity perpsective.16:52
adrian_ottoand if there is any counter argument, I'm happy to adjust my vote to proceed16:53
adrian_ottomy current vote is "0  - No Score"16:54
swatson_My only counter is just that the change is pretty basic anyway16:54
adrian_ottoawaiting your input16:54
*** fragatina has joined #openstack-containers16:54
adrian_ottoyeah, I totally agree16:54
swatson_But if you're more comfortable with it being split to the implementation I'll go ahead and move it16:54
adrian_ottook, cool16:54
*** fragatina has quit IRC16:54
adrian_ottogood work on that patch by the way. I really like it.16:55
swatson_adrian_otto: Thanks :)16:55
openstackgerritStephen Watson proposed openstack/python-magnumclient: Implement Parameter Deprecation  https://review.openstack.org/38821516:57
*** chhavi has joined #openstack-containers16:59
openstackgerritStephen Watson proposed openstack/python-magnumclient: Renames *-id parameters  https://review.openstack.org/38824017:01
yatinadrian_otto: Also need to look at https://review.openstack.org/#/c/397503/17:03
*** mjura has quit IRC17:03
adrian_ottoyatin: ok17:04
jrollrandom question, does magnum have some sort of hook for adding additional user-data to the instances it uses?17:05
swatson_adrian_otto yatin: https://review.openstack.org/#/c/388215/ and https://review.openstack.org/#/c/388240/ should be good to go17:06
rochaportoadrian_otto: this one https://blueprints.launchpad.net/magnum/+spec/run-kube-as-container17:07
rochaportowould it be ok to reassign to mvelten? it's been there for a while now, he has two patches in17:08
rochaportoyatin: do you need more input for this https://blueprints.launchpad.net/magnum/+spec/secure-etcd-cluster-coe?17:08
randallburtjroll:  not in the default drivers I don't think.17:08
rochaportothe only thing from us is that we have flat/provider networks, so we can't rely on blocking access to the magnum cluster priv network (we run with a local patch for this)17:09
yatinrochaporto: i want to know about kubernetes version in the current fedora image and it's support for connecting to secure etcd cluster17:09
jrollrandallburt: thanks17:09
randallburtnp17:09
rochaportojroll: there's this https://blueprints.launchpad.net/magnum/+spec/allow-user-softwareconfig17:09
rochaportobut we moved to having our own custom drivers instead (with the required changes)17:10
rochaportodo it probably doesn't make sense any longer17:10
jrollah17:10
mveltenyatin good luck with that.... I tore down the rpm a while ago and it matches nothing. To be precise it matches a commit on github that is not available on any public branch17:11
rochaportoyatin: this is the latest one (docker 1.10, latest kube 1.2.x): http://tarballs.openstack.org/magnum/images/fedora-atomic-23-dib.qcow217:12
*** chhavi has quit IRC17:12
yatinmvelten: I also saw some different commits on kubernetes for adding and removing those etcd parameters17:12
yatinrochaporto: i used the image that devstack adds automatically with magnum17:13
yatinrochaporto: Ok i will check this image as well.17:13
*** tonanhngo has quit IRC17:14
mveltenrochaporto it is not the last 1.2.x for sure, it is some custom baked version17:15
mveltenlast comment in bhodi :17:15
mveltenhttps://bodhi.fedoraproject.org/updates/FEDORA-2016-1911d6ed9717:15
mveltenUpdate to origin v1.2.117:15
*** adrian_otto has quit IRC17:15
yatinI also think mvelten is right, but let's confirm17:15
yatinrochaporto: I think there is a plan to introduce kuberentes 1.3 in this release, may be spyros is working on that.17:16
rochaportoactually locally we're aiming at 1.4 :)17:16
mveltenyes strigazi have been playing with the kubernetes official rpm packages that comes with 1.417:17
Dragostrigazi: Has it been fun17:17
rochaportoDrago: he looooooves it!17:17
DragoCoreOS's tutorials are actually pretty good17:17
*** tkarke has joined #openstack-containers17:18
openstackgerritJaycen Grant proposed openstack/magnum: Spec for adding template versions  https://review.openstack.org/39232717:20
*** yamamoto has joined #openstack-containers17:21
*** jwcroppe has quit IRC17:21
*** jwcroppe has joined #openstack-containers17:22
*** yamamoto has quit IRC17:22
*** yamamoto has joined #openstack-containers17:22
yatinHas someone seen this message : <wangbo> yatin, we tried to integrate harbor into magnum. there are some synchronize calls between magnum-conductor and harbor server and these api calls take long time.17:24
yatin<wangbo> these calls occupied one magnum-conductor fully17:24
yatin<wangbo> such as image-upload17:24
yatin<wangbo> so I want to find way to run multiple processes.17:24
yatin<wangbo> yatin: another problem is that all magnum-api rpc to magnum-conductor use synchronize calls. so we need multiple magnum-api processes.17:25
*** jwcroppe has quit IRC17:26
*** jwcroppe has joined #openstack-containers17:27
*** adrian_otto has joined #openstack-containers17:27
*** yamamoto has quit IRC17:34
*** Jeffrey4l has quit IRC17:34
*** fragatina has joined #openstack-containers17:38
*** fragatina has quit IRC17:38
*** fragatina has joined #openstack-containers17:39
*** fragatin_ has joined #openstack-containers17:40
*** fragatin_ has quit IRC17:42
*** zenoway has joined #openstack-containers17:43
*** fragatina has quit IRC17:44
*** fragatina has joined #openstack-containers17:44
*** zenoway has quit IRC17:47
openstackgerritRandall Burt proposed openstack/magnum-specs: Add cluster driver encapsulation spec  https://review.openstack.org/39785817:48
openstackgerritJaycen Grant proposed openstack/magnum: Spec for adding template versions  https://review.openstack.org/39232717:51
Dragothanks jvgrant17:52
jvgrantno problem, it does make it clearer17:52
rochaportoyatin: we'll try f25 with kube 1.4 from f26 builds (a quick look and the deps look ok)17:54
rochaportohttp://koji.fedoraproject.org/koji/packageinfo?packageID=1920217:54
rochaportothe binaries and rpms look compatible with the existing magnum templates17:55
rochaportoif you have something to try for the etcd tls we can give it a go17:55
randallburtwe've managed to build an f24 with k8s1.4 and latest docker from their repos but its not atomic17:58
rochaportotheir repos == fedora repos?18:00
Dragorochaporto: I think randallburt means the k8s yum repo18:01
rochaportook. but that has changes in the binaries right?18:01
rochaportoideally we move to containers (if it works), using the f26 builds would be a way to get kube 1.4 with no magnum changes (worth trying)18:01
Dragorochaporto: Ha ha, no magnum changes18:02
Dragorochaporto: You mean the driver, right?18:02
*** jwcroppe has quit IRC18:02
rochaportoyes18:02
Dragorochaporto: Well, I guess if you can get the image just right, yes18:02
*** jwcroppe has joined #openstack-containers18:03
rochaportowe'll try that, some people are asking for 1.4 around here18:03
Dragorochaporto: Do you guys have a good kube-proxy and kube-dns manifest?18:03
Dragorochaporto: I've been working on 1.4 and relying on kubeadm to bootstrap it, but it doesn't drop a manifest for those two18:04
DragoAnd kube-discovery too18:04
DragoI'm not sure that kube discovery is necessary though18:05
rochaportokube-proxy and kube-dns, yes18:05
rochaportobut the dns is a patch on the magnum templates18:05
*** jwcroppe has quit IRC18:07
Dragorochaporto: Why?18:08
rochaportoi'm fetching the manifest give me a sec :)18:08
Dragorochaporto: Oh, is it because it doesn't exist in the current driver?18:08
rochaportoyup18:08
*** jperry has quit IRC18:09
rochaportohttps://stikked.web.cern.ch/stikked/view/e1673bbf18:10
rochaportocan you access this?18:10
rochaportowe've had it for a while now18:10
Dragorochaporto: What is your CERN username and password? :D18:10
Dragorochaporto: Just kidding18:11
DragoI can't18:11
randallburtsorry, yes rochaporto I meant the k8s and docker yum repos18:11
randallburtI built using diskimage builder and a small custom element18:11
*** hieulq_ has quit IRC18:11
rochaportohttps://justpaste.it/10fqg18:12
rochaportorandallburt: ok thanks, we've done something similar to add a couple packages we need18:13
randallburtrochaporto:  same18:13
Dragorochaporto: Thanks! Is there one for kube-proxy?18:13
rochaportoDrago: isn't that one in upstream?18:14
Dragorochaporto: And is there any reason 10.254.10.10 was picked?18:14
*** sdake_ has joined #openstack-containers18:14
Dragorochaporto: As I understand it, kube-proxy needs to be a Daemonset18:14
Dragoso that there is exactly one on each node18:15
*** sdake has quit IRC18:17
rochaportothe ip i don't remember, maybe mvelten does18:20
rochaportofor kube-proxy, this does it? magnum/drivers/common/templates/kubernetes/fragments/enable-kube-proxy-minion.sh?18:20
rochaportosimilar one for master18:20
rochaportotime to go here, talk tomorrow :)18:21
*** tonanhngo has joined #openstack-containers18:21
DragoLater rochaporto18:22
*** jwcroppe has joined #openstack-containers18:24
*** tonanhngo has quit IRC18:26
*** tonanhngo has joined #openstack-containers18:26
*** fragatina has quit IRC18:28
*** fragatina has joined #openstack-containers18:29
*** flaviodsr has quit IRC18:34
*** harlowja has quit IRC18:38
*** fragatina has quit IRC18:44
*** fragatina has joined #openstack-containers18:45
hongbinstrigazi: there?18:51
*** rpothier has quit IRC19:00
*** rcernin has joined #openstack-containers19:03
*** harlowja has joined #openstack-containers19:06
openstackgerritMerged openstack/magnum-ui: Imported Translations from Zanata  https://review.openstack.org/39758419:19
*** sdake_ has quit IRC19:25
*** sdake has joined #openstack-containers19:27
openstackgerritMerged openstack/python-magnumclient: Implement Parameter Deprecation  https://review.openstack.org/38821519:28
*** EricGonc_ has quit IRC19:30
*** EricGonczer_ has joined #openstack-containers19:31
Dragojvgrant: Enjoying your +2 powers?19:32
Drago:)19:32
openstackgerritMerged openstack/python-magnumclient: Renames *-id parameters  https://review.openstack.org/38824019:35
jvgrantDrago: yep :)19:35
swatson_Drago jvgrant: Will https://review.openstack.org/#/c/396729/ get merged automatically now?19:36
swatson_I marked it as Depends-on the one in client that just got merged19:36
swatson_But the doc changes were +2'd and workflowed already19:36
jvgranti'd imagine it would automatically merge now, but i can check19:37
swatson_I would have thought so too, but yatin's comment on the doc patch made me think otherwise19:37
openstackgerritJason Dunsmore proposed openstack/magnum: Don't parse k8s node capacity if node is out of disk  https://review.openstack.org/39790219:39
jvgrantswatson_: i gave it a third workflow +1. I think it should work this time :)19:39
swatson_jvgrant: Fine by me19:40
Dragojvgrant: nice19:40
jvgrantswatson_: you might have record for most workflow +1's on a review ever19:40
swatson_That'll look good on stackalytics I'm sure B)19:41
swatson_3 +1s, 3 +2s, 3 WF+1. People LOVE my doc changes19:42
DragoPatches so unmergeable that 3 workflow +1s are needed19:42
swatson_Drago: :P19:42
openstackgerritMerged openstack/magnum: Updates Documentation for non-ID Params  https://review.openstack.org/39672919:42
swatson_The parameter deprecation changes are done! Feels good19:43
jvgrantDrago: going to take one more pass at the nodegroup spec for the remaining TODOs. Should I put some generic heat stack info in or do we just want to reference a separate spec?19:43
Dragojvgrant: I would just reference a separate spec. I think all you could really put that would be safe is "template things will change"19:44
Dragoswatson: Good job!19:44
jvgrantDrago: ok, will do. I'll say that the implementation details of the heat stack changes will be covered in a separate spec. I think that should be enough to remove the WIP and get it moving forward19:45
Dragojvgrant: Put up a blueprint for it and reference that in the NodeGroup spec19:46
jvgrantDrago: ok19:46
jvgrantDrago: heat-templates-for-nodegroups should like a good name for it?19:47
Dragojvgrant: no19:47
Dragojvgrant: nodegroups-heat-templates?19:48
Dragojvgrant: nodegroups-heat-changes?19:48
Dragojvgrant: nodegroups-heat?19:48
Dragojvgrant: nodegroups-stacks?19:48
DragoI like nodegroups-heat19:48
jvgrantDrago: nodegroups-stacks is probably the most generic, but nodegroups-heat is probably better19:49
DragoI'd rather have nodegroups as the prefix regardless19:49
jvgrantok19:49
*** EricGonc_ has joined #openstack-containers20:03
*** EricGonczer_ has quit IRC20:04
*** jperry has joined #openstack-containers20:06
*** chris_hultin is now known as chris_hultin|AWA20:06
*** Drago has quit IRC20:17
*** Drago has joined #openstack-containers20:18
*** chris_hultin|AWA is now known as chris_hultin20:19
*** Drago has quit IRC20:22
*** Drago has joined #openstack-containers20:31
*** sdake_ has joined #openstack-containers21:00
*** sdake has quit IRC21:04
*** vijendar has joined #openstack-containers21:06
*** adrian_otto has quit IRC21:13
*** adrian_otto has joined #openstack-containers21:14
*** vijendar1 has joined #openstack-containers21:17
*** pcastets has quit IRC21:17
*** pcastets has joined #openstack-containers21:18
hongbinjvgrant: Drago : have time to discuss the cluster upgrade spec?21:27
Dragohongbin: Yes21:27
hongbinDrago: see my latest comment?21:27
Dragohongbin: Yes21:28
hongbinDrago: I just think it is undesirable if we lose containers after a upgrade21:28
*** sdake_ has quit IRC21:28
jvgrantreading it now21:29
hongbinIf during a upgrade, we lose container. This make Magnum immediate out of production requirement21:30
Dragohongbin: How can users avoid losing data if they switch to an entirely new cluster?21:31
hongbinDrago: I am not proposing to switch the cluster21:31
Dragohongbin: Switching the cluster is what we have right now21:31
hongbinDrago: I am proposing to reinstall COE in the node21:31
jvgrantthe rolling upgrades is an improvement to what we have right now, but hongbin is requesting something more than this middle ground21:31
*** pcaruana has quit IRC21:32
hongbini think my requirement is common: after a upgrade, all the containers are still there21:32
DragoLet's back up a bit21:32
DragoIf you're deploying an application on a cluster, and one of the nodes dies, and you lose data because you're not replicating your data across multiple containers/nodes/hosts, that means you designed your application wrong21:33
jvgrantit is probably possible in some upgrades and not others. I could see that might be different between COE's and what is being upgraded21:33
DragoOr your application is not production-ready anyway21:34
hongbinDrago: then, that means if there is an accident, you blame your customers :)21:34
Drago"If during a upgrade, we lose container. This make Magnum immediate out of production requirement"21:35
jvgrantfor HA Openstack setups the rolling upgrades using image replacement will work well, but all upgrades might not need that21:35
DragoSimilarly, If a node goes down and you lose data, your application is obviously not production-ready21:35
*** vijendar1 has quit IRC21:36
hongbinlet's assume the node is not gone21:36
*** vijendar has quit IRC21:36
jvgrantthe spec all ready indicates 2 types of upgrades: rolling and in-place. I'd suggest that the spec allows each driver to implement what makes most sense for each COE driver21:37
hongbinif during a upgrade, we don't take down any node, we can keep all the containers21:37
DragoNow, if a solution was proposed where the client warned the user about what happened in the upgrade, I think it would make sense, because sure, the data in the containers is gone21:37
DragoAnd/or if it's documented properly21:37
openstackgerritRandall Burt proposed openstack/magnum: [WiP] Move cluster sync into the Heat driver  https://review.openstack.org/39796121:37
*** filler has quit IRC21:38
hongbinwell, if it is possible to keep the data, why we do it in a distructive way21:38
hongbini would argue that magnum should do the upgrade in the less destructive way if it is possible21:38
Dragohongbin: Possible is not the same as easy21:39
jvgranthongbin: agreed, which is why I think the image replacement is just one way to implement the upgrade21:39
Dragohongbin: It's also possible to do an upgrade with zero impact, but it's not easy21:39
hongbinDrago: jvgrant : why not just reinstall the COE in the nodes21:39
*** filler has joined #openstack-containers21:40
DragoI think less destructive is good too, but I think it would be easier and generally-applicable (to multiple COEs) by doing a node-replace rolling upgrade21:40
DragoFurthermore, if you do an in-node upgrade, you can't have immutable infrastructure21:40
hongbincoreos?21:41
DragoIf it would make everyone happy, we could remove any implementation details about how the upgrades will be implemented21:41
Drago*upgrades will be done21:41
jvgrantin my opinion the spec should leave that up to each driver implementation. We should document upgrade types: replace(destructive), rolling, and in-image. the driver can implement 1 or all of the options and we can document what level of "interference" each type has21:42
hongbinthat will work21:42
DragoThe k8s fedora atomic driver will not have in-image upgrades until kubernetes doesn't need to be installed on the image21:43
hongbinbut i guess we need to choose a default upgrade type? (among replace, in-image?)21:43
hongbinDrago: how about os-tree upgrade?21:43
jvgranthongbin: the reason for not just the reinstall could be that there might be other changes in the image version that is now used as latest that might not be upgrade with just a coe upgrade. So it wouldn't be 100% the latest version of the template21:43
Dragohongbin: requires a reboot, doesn't it?21:43
*** sdake has joined #openstack-containers21:44
openstackgerritRandall Burt proposed openstack/magnum: [WiP] Move cluster sync into the Heat driver  https://review.openstack.org/39796121:45
hongbinjvgrant: well, agree. this could be problems. but there should be ways to solve it21:45
jvgranthongbin: agreed. that is why i'd like to remove the "image replace" implementation from spec and just define types of upgrades. Then on a driver by driver basis we can implement the best option for each and improve them to make them less invasive as often as we can21:46
hongbinDrago: maybe, not sure. however, a reboot doesn't cause a lose of containers and its data. right?21:46
hongbinjvgrant: ok, sounds good to me21:47
jvgranthongbin: would be downtime though, unless in HA setup21:47
Dragojvgrant, hongbin: Replacing the whole node covers lots of cases and works basically the same way for multiple COEs and is relatively easy to implement. I think it makes sense to start with that, *warn users through docs and CLI warnings* and continue to make the upgrade story better as we go along, rather than block ourselves on things that can be in the second phase.21:47
jvgranthongin: and if in an HA setup then the rolling upgrades would work well21:47
hongbinjvgrant: yes..., but i am not sure how to do HA in container level21:48
hongbinDrago: ok, if image replace is easier to implement, then it is reasonable to start with that21:48
jvgrantDrago: i agree that the image replacement is the cleanest/easiest and will likely be the first upgrade implementation we see, but we should make sure we leave it open differences and improvement21:49
Dragohongbin: We are in agreement that other upgrades may be better, for reasons such as data loss. I'm not saying we should avoid those things, just not do them first :)21:49
jvgranti added a note to the spec that we needed to figure out how to communicate what will happen if the user goes forward with a upgrade21:50
hongbinok, sounds good21:50
Dragohongbin: This is just like certificate revocation21:50
Dragohongbin: We have the "nuclear" option, which revokes all users at once. It's a lot worse than per-user revocation, but still better than not being able to do it at all.21:51
DragoAnd easier to implement than the per-user option21:52
hongbinyes, i remembered21:54
*** harlowja has quit IRC21:55
*** jwcroppe has quit IRC22:00
*** jwcroppe has joined #openstack-containers22:02
*** harlowja has joined #openstack-containers22:13
*** jwcroppe has quit IRC22:14
*** randallburt has quit IRC22:17
*** catinthe_ has joined #openstack-containers22:18
*** catintheroof has quit IRC22:21
*** vijendar has joined #openstack-containers22:26
*** adrian_otto has quit IRC22:28
*** khamtamtun has joined #openstack-containers22:29
*** adrian_otto has joined #openstack-containers22:31
*** catintheroof has joined #openstack-containers22:35
*** catinthe_ has quit IRC22:36
*** vijendar1 has joined #openstack-containers22:38
*** rcernin has quit IRC22:42
*** dimtruck is now known as zz_dimtruck22:44
*** adrian_otto has quit IRC22:45
*** vijendar has quit IRC22:57
*** vijendar1 has quit IRC22:58
*** khamtamtun has quit IRC22:59
*** jperry has quit IRC23:00
*** vijendar has joined #openstack-containers23:00
*** vijendar1 has joined #openstack-containers23:06
*** vijendar has quit IRC23:06
*** vijendar1 has quit IRC23:08
*** jasond has quit IRC23:09
*** vijendar has joined #openstack-containers23:10
*** chris_hultin is now known as chris_hultin|AWA23:11
*** dave-mccowan has quit IRC23:12
*** jwcroppe has joined #openstack-containers23:15
*** vijendar has quit IRC23:16
*** vijendar has joined #openstack-containers23:18
*** vijendar1 has joined #openstack-containers23:18
*** sdake has quit IRC23:22
*** vijendar1 has quit IRC23:23
*** vijendar has quit IRC23:23
*** vijendar has joined #openstack-containers23:24
*** chetna has joined #openstack-containers23:26
*** vijendar has quit IRC23:28
*** zz_dimtruck is now known as dimtruck23:39
*** swatson_ has quit IRC23:43
*** jvgrant_ has joined #openstack-containers23:49
*** jvgrant has quit IRC23:52
*** catintheroof has quit IRC23:52
*** catintheroof has joined #openstack-containers23:54
tkarkeHi, can someone tell me what is the current version of Magnum API?23:56
*** catintheroof has quit IRC23:58
*** jmckind has quit IRC23:58

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