Wednesday, 2019-04-03

*** sthussey has quit IRC00:07
tao_liuportdirect: The ks-endpoints.sh uses the region name set in OS_REGION_NAME, to create the endpoint.  There is no way to create an endpoint in a different region, other than the region set in OS_REGION_NAME. Since, in a multiple regions environment, there is a need to configure the endpoints in a different region, is there a plan to support multi-region configuration?  If the answer is no, would it be acceptable to make a simple change to get my00:16
tao_liuenvironment up and running?00:16
*** Liang__ has joined #openstack-helm00:46
*** mfuller_ has joined #openstack-helm01:10
*** bodgix_ has joined #openstack-helm01:11
*** jamesgu__ has quit IRC01:12
*** cfriesen has quit IRC01:12
*** mcfuller has quit IRC01:12
*** bodgix has quit IRC01:12
openstackgerritGage Hugo proposed openstack/openstack-helm master: [DNM] Testing resource type  https://review.openstack.org/64931801:12
*** jamesgu__ has joined #openstack-helm01:13
*** Liang__ is now known as LiangFang01:20
portdirecttao_liu: at this point we have not yet considered multi region support01:35
portdirectBut it would be a great thing to have01:36
portdirectI dont think the changes required would be to great, if you pushed a poc we could probably iterate into a spec from there01:36
portdirectfdegir: we currently default to ocata for deployments01:37
*** nguyenhai has joined #openstack-helm02:09
*** spsurya has joined #openstack-helm02:12
*** michael-beaver has quit IRC02:29
*** michaelbeaver has joined #openstack-helm02:29
*** michaelbeaver has quit IRC02:31
*** gkadam has joined #openstack-helm02:37
*** cheng1 has quit IRC02:55
*** cheng1 has joined #openstack-helm02:58
openstackgerritGage Hugo proposed openstack/openstack-helm master: [DNM] Testing resource type  https://review.openstack.org/64931803:02
*** roman_g has quit IRC03:22
*** cheng1 has quit IRC03:30
*** cheng1 has joined #openstack-helm03:33
openstackgerritAngie Wang proposed openstack/openstack-helm master: Add Aodh Chart  https://review.openstack.org/64151803:48
openstackgerritAngie Wang proposed openstack/openstack-helm master: Ceilometer chart: add the ability to publish events to panko  https://review.openstack.org/64151903:48
openstackgerritAngie Wang proposed openstack/openstack-helm master: Add ubuntu multinode job as experimental check  https://review.openstack.org/64778703:48
*** itlinux has joined #openstack-helm04:41
itlinuxhello all.. does openstack-helm work on centos?04:42
itlinuxthanks04:42
jayahnitlinux: yes, i does. you mean centos as host os, right?04:42
itlinuxyes04:42
itlinuxI have 4 machines and I want to test it out..04:43
itlinuxdo you have any docs on how to make it work?04:43
itlinuxthanks04:43
itlinuxreading this but ..https://docs.openstack.org/openstack-helm/latest/install/multinode.html04:43
itlinuxuhm!04:43
itlinuxmaybe you can give me a rundown.. since I have done lots on OOO..04:44
itlinuxthanks jayahn:04:44
itlinuxthis shows ubunto thanks again https://docs.openstack.org/openstack-helm/latest/install/kubernetes-gate.html04:45
jayahntrue, basically gate runs on ubuntu server. once you have kubernetes installed, deploying openstack-helm charts will not have much difference between host os.04:55
jayahnhm.. probably we need to write a document running basic osh (including kubernetes) on centos and/or suse.04:56
jayahnitlinux: sorry not being much of help for now. I will see what we can provide shortly. or when pordirect: wakes up, he might have more hints.05:00
itlinuxok05:00
itlinuxis pordirect in the US or EMEA?05:02
itlinuxjayahn:05:02
jayahnsince our organization deploy kubernetes using kubespary, it is a bit different from how osh gating does.05:02
itlinuxcan you tell me if the primary can use any of the services? or is that just the deployment node?05:03
itlinuxdoes kubespray work on centos?05:03
itlinuxjayahn: this looks like https://itnext.io/install-kubernetes-on-bare-metal-centos7-fba40e9bb3de05:04
itlinuxwhat's your thought on this...05:05
jayahnkubespray works on centos, and your link looks good.05:10
itlinuxok will install that and see :) thanks05:10
itlinuxthen deploy helm since looks like the basic is kube05:11
itlinuxas a common denominator..05:11
*** lemko has joined #openstack-helm05:42
*** jamesgu has quit IRC05:51
evrardjpportdirect: a few days ago I asked you a question about the images in your docker namespace, and the "assets" images for novnc/spice. Could you answer it?06:33
*** pgaxatte has joined #openstack-helm06:44
*** dpawlik has joined #openstack-helm06:47
*** cheng1 has quit IRC07:02
*** cheng1 has joined #openstack-helm07:04
*** jistr is now known as jistr|afk07:22
*** witek has joined #openstack-helm07:37
*** roman_g has joined #openstack-helm07:42
*** lemko has quit IRC08:31
openstackgerritItxaka Serrano Garcia proposed openstack/openstack-helm master: Allow more generic overrides for ceilometer-api  https://review.openstack.org/64180208:42
*** LiangFang has quit IRC08:59
happyhemanthello folks, I am facing some strange error building loci images.  evrardjp, portdirect  do you familier with this? https://www.irccloud.com/pastebin/60wOtwF2/10:08
evrardjphappyhemant: I would need more context: Which branch are you using, can you show me the registry you are using and the arguments you are passing ?10:09
happyhemantevrardjp: here you go https://www.irccloud.com/pastebin/sLtrISHl/10:12
happyhemantevrardjp: i am using master branch and my own docker account10:13
evrardjpso out of curiosity the requirements repo is properly built then?10:14
evrardjpand the failure comes from the fetching of the requirments wheels?10:14
evrardjpbuild and uploaded into the registry10:14
evrardjpI am glad someone is using this code, I thought I was the only one :)10:14
dimitris_Hello all I'm installing  OSH on Ubuntu 18.04 but I'm facing problem with deploy-k8s (which I didn't had on ubuntu 16.04) https://hastebin.com/gevegunise.http. Is this have to do with Ubuntu release?10:22
evrardjpdimitris_: I think we are not testing on bionic yet10:24
dimitris_evrardjp, thank you evrardjp, have check the logs, because from these logs I cannot tell that the problem is the bionic release10:29
evrardjpI am not sure -- just saying I haven't tested this10:29
dimitris_evrardjp, ok10:30
happyhemantevrardjp: yes you are right10:31
evrardjphappyhemant: this shell script should build requirement image, publish it, and then the others should use what was built10:31
happyhemanthttps://www.irccloud.com/pastebin/nGLgVnJ9/evrardjp%3A%20the%20equirements%20repo%20is%20properly%20built%2010:32
evrardjpqueens-master-ubuntu_xenial looks very weird :)10:32
evrardjpcould you simply run the queens wrapper, see how it goes?10:33
evrardjphttps://github.com/openstack/openstack-helm-images/blob/4aa86c67cb5d2b938a8614641b0a0e55ea2aa229/openstack/loci/build-queens.sh10:33
happyhemantevrardjp: i didnt change anything in the script these are tags. I am newbee in building loci images10:34
evrardjpI have to refactor those scripts, do you want to help me there?10:34
evrardjpI want to simplify them10:34
evrardjphappyhemant: anyway, the current code should work10:35
evrardjpyou have shown me it's built, but not if it's published10:35
happyhemantevrardjp: where ?10:36
evrardjpI guess it depends on your registry10:36
evrardjpI suppose the requirements should be built on your dockerhub?10:36
evrardjpthat's what you said?10:36
evrardjphttps://github.com/openstack/openstack-helm-images/blob/master/openstack/loci/build.sh#L27010:36
evrardjpit looks like it exists: https://hub.docker.com/r/hemsonawane/requirements/tags10:37
evrardjpso I guess could you paste the docker build arguments that are used for building the project images?10:37
happyhemantevrardjp: yeah it is being push to my docker hub10:37
evrardjpas you can see parallel processes that file: https://github.com/openstack/openstack-helm-images/blob/master/openstack/loci/build.sh#L28010:38
evrardjpcheck if everything sounds right there10:38
happyhemantevrardjp: it seems everything is good there10:39
evrardjpcould you paste it somewhere?10:40
happyhemantevrardjp: I didnt get you ? :( sorry10:40
evrardjpcould you paste that actions file somewhere, so I can help you?10:42
evrardjp(maybe manually run the command that failed, and put the whole log somewhere)10:42
happyhemantevrardjp:  yeah there is one log file in /tmp/ directory10:43
evrardjpthe actions file tells you what to build in parallel, you should see the commands that are run there. Could you copy / paste it to paste.openstack.org please?10:44
evrardjpthat's what I meant :D10:44
happyhemantevrardjp:  sorry i will do it right away :(10:45
evrardjpno worries, don't be sad :)10:45
evrardjpI just can't help you without a log10:45
evrardjpwithout 2 logs sorry10:45
evrardjp1) the commands you are running (the actions)10:46
happyhemanthttps://www.irccloud.com/pastebin/7DChzPcr/10:46
happyhemanthere it is thanks alot :)10:46
evrardjp2) what those commands do when they fail10:46
evrardjpso that's log 110:46
happyhemantyeah10:46
evrardjpso it seems there is something wrong10:47
happyhemantwhat it could be ?10:47
evrardjpdo you want to build master or queens?10:47
evrardjphttps://hub.docker.com/r/hemsonawane/requirements/tags10:48
happyhemantqueens i build it previously10:48
evrardjpok10:48
evrardjpso there is something wrong out there10:48
happyhemanti pulled it from there to my new account10:48
happyhemantso what could be the solution to get rid of this issue10:49
evrardjpcould you run the build-queens.sh instead, and just override the REGISTRY_URI var?10:49
happyhemantsure10:49
happyhemanti will keep the resigesty REGISTRY_URI  as it is10:50
evrardjpfor example export REGISTRY_URI="docker.io/hemsonawane" && ./openstack/loci/build-queens.sh10:50
evrardjpsee how it goes10:50
happyhemantexecuted the command let see how it goes10:52
happyhemantjust hoping it should work10:52
happyhemantevrardjp:  it doesn't work :(10:56
happyhemanthttps://www.irccloud.com/pastebin/KkJdDC4V/10:57
happyhemantit has the same issue10:57
evrardjpso here the names are correct10:59
evrardjpI tested all those things with a local requirements registry, so I am using it in a different way than yourself11:00
happyhemantok11:01
evrardjpbut I have the impression this might be the cause11:01
evrardjpbecause http://docker.io/v2/ doesn't sound like it's the right url11:01
*** jistr|afk is now known as jistr11:02
happyhemantthat i was looking from since from yesterday\11:02
happyhemanti dont know whats wrong with that11:02
happyhemanti have given right registery URI11:02
evrardjphappyhemant: agreed with you --build-arg WHEELS=docker.io/hemsonawane/requirements:<tag> sounds okay to me11:04
evrardjpoh wait11:04
evrardjpcould you try setting11:04
happyhemantyes sure11:05
evrardjpexport REGISTRY_URI="hemsonawane" simply?11:05
happyhemantsounds good :) let me try11:05
evrardjpsorry to make you test and retry11:05
evrardjpI haven't tested that case11:05
happyhemantohh come on dont say that i am newbee here and you are the only hope for me to make it work11:06
evrardjphaha :)11:09
evrardjpI am like you, just a little different background :)11:09
*** schwicht has joined #openstack-helm11:10
happyhemanthaha :) really i dont think so ;)11:10
happyhemantevrardjp:  and finally it works :) thank you very much11:12
happyhemantevrardjp:  just one silly mistake of mine that i was using docker.io/hemsonawane all the time :(11:15
*** schwicht has quit IRC11:29
*** schwicht has joined #openstack-helm11:29
happyhemantevrardjp:  really thank you very much I tested my horizon image and it works really fine :)11:31
evrardjphappyhemant: I think the comment is very confusing, could you remove it from build.sh ?11:32
evrardjpor edit it11:32
happyhemantevrardjp:  you mean docker.io/hemsonawane right?11:33
*** schwicht has quit IRC11:34
evrardjphttps://github.com/openstack/openstack-helm-images/blob/master/openstack/loci/build.sh#L165-L16611:34
evrardjpthere is docker.io/ there , which is confusing11:34
*** schwicht has joined #openstack-helm11:34
happyhemantevrardjp:  yeah i already did that :) thanks again11:35
happyhemantevrardjp: can i ask you one more question if you dont mind ?11:36
evrardjpyeah sure11:38
evrardjpI am never forced to answer, am I? :D11:38
evrardjphahahahah11:38
evrardjplet's hope I can help you11:38
happyhemantahaha :)11:39
happyhemanti have mine customized dashboard and i want to horizon image with my git horizon repository and i am confused how am i suppose to do it using loci11:40
happyhemanti know how to build it with kolla11:40
happyhemantbut loci is bit diffrent11:40
happyhemantevrardjp: i have mine customized dashboard and i want to build horizon image with my git horizon repository and i am confused how am i suppose to do it using loci *11:41
evrardjpok so11:42
evrardjphttps://github.com/openstack/loci main page is a readme11:43
evrardjpas you can see PROJECT_REPO is the git repo11:43
happyhemantyeah11:43
happyhemanthttps://github.com/hemantsonawane95/horizon/tree/pisek11:43
evrardjpif you override PROJECT_REPO and PROJECT_REF (for the branch) you would be building with this11:43
happyhemantthis is my horizon repo11:43
evrardjpit's public so it's good11:43
happyhemantthanks :11:44
evrardjpyou need to point PROJECT_REPO to the git url of https://github.com/hemantsonawane95/horizon11:44
evrardjphttps://github.com/hemantsonawane95/horizon.git11:44
evrardjpand PROJECT_REF to pisek11:44
happyhemantyeah yeah sure11:44
happyhemantThanks for explaining me in such details :D11:44
evrardjpplease keep in mind that you can do that in the wrapper script (by exporting something which will override these 2 vars)11:45
evrardjpbut I would start with doing so manually11:45
evrardjpplease make sure the requirements are matching too11:45
evrardjp(so requirements built today, means code from your branch built today)11:45
happyhemantit is not build today :(11:46
evrardjpyou can try11:46
happyhemantok11:46
evrardjpmaybe the requirements won't be so different11:46
happyhemantcould be possible :)11:46
evrardjpbut the openstack/requirements are shifting frequently11:46
evrardjpkeep at least the branch the same :)11:46
happyhemantsure11:47
evrardjpbasically you have the power to do what you want, just use it carefully :D11:47
happyhemantindeed :D11:47
*** schwicht has quit IRC11:54
happyhemantevrardjp: I think i should chnage here right? https://www.irccloud.com/pastebin/KN5MZ44P/11:55
happyhemantevrardjp: or something like this12:01
happyhemanthttps://www.irccloud.com/pastebin/5wgZvteT/12:01
evrardjpyeah so you don't need to change the code (first link), you can just export horizon_project_repo for that12:07
evrardjpthe second link is incorrect because there are extra spaces12:07
evrardjpI think12:07
happyhemantyeah yeah sure12:09
*** odyssey4me has quit IRC12:12
*** schwicht has joined #openstack-helm12:16
happyhemantevrardjp: how can we add (orchestration) heat dashboard there ?12:18
evrardjpthat, I have no clue.12:20
evrardjpI think it would require modification of the helm charts12:20
evrardjpmaybe itxaka can help there?12:20
happyhemantok cool i will ask him12:20
happyhemantthanks12:20
*** schwicht has quit IRC12:22
happyhemantitxaka: hey hi I am building loci horizon image and I would like to know how we can add (orchestration) heat dashboard there ? help will be appreciated :)12:23
*** odyssey4me has joined #openstack-helm12:26
*** schwicht has joined #openstack-helm12:29
*** nguyenhai has quit IRC12:33
*** schwicht has quit IRC12:42
happyhemantevrardjp: now i am stuck here. what do you think its because of requirements ? https://www.irccloud.com/pastebin/CJSBfkwR/12:46
evrardjpwow that's weird12:47
evrardjpwhat's the project you are building?12:48
evrardjpit seems unset12:48
evrardjp(again, a bigger log would help)12:48
happyhemant i am building my repo customize horizon image  https://www.irccloud.com/pastebin/G79DodYa/12:50
happyhemantdocker build https://github.com/openstack/loci.git --build-arg FROM=gcr.io/google_containers/ubuntu-slim:0.14 --build-arg PROJECT_REPO=https://github.com/hemantsonawane95/horizon.git --build-arg PROJECT_REF=pisek --build-arg WHEELS=hemsonawane/requirements:latest-queens-ubuntu_xenial --build-arg PROFILES='fluent apache' --tag new12:50
happyhemantevrardjp: this is command i am using to buld it  https://www.irccloud.com/pastebin/uOZ2cUe8/12:51
evrardjpyeah you need to pass PROJECT horizon too IIRC12:53
evrardjpyup12:54
evrardjphttps://github.com/openstack/loci12:54
happyhemantevrardjp: i didnt get you :(12:55
evrardjpyou are missing a build-arg12:55
evrardjpPROJECT12:55
evrardjpyou can keep PROJECT_REPO and all, but PROJECT still has to be set12:56
happyhemant--build arg PROJECT=horizon ?12:56
evrardjpcorrect12:56
happyhemantohhh stupid me :D Thanks12:57
evrardjphappyhemant: build-arg12:59
happyhemantevrardjp: yeah done that too :D12:59
happyhemantevrardjp: thanks alot for your help i manage to deploy my customize horizon now just looking for heat stack13:07
itxakasorry was on lunch happyhemant, did you get to add the heat dashboard to the horizon chart?13:07
itxakayou should be able to use https://review.openstack.org/#/c/645543/ if it gets merged or you can just pull that patch locally for your horizon chart :)13:08
happyhemantitxaka: no prob  yeah thats right i want to add heat13:08
*** schwicht has joined #openstack-helm13:09
happyhemantitxaka: I build image wiith loci will it work there ?13:09
itxakaonce you get that patch, heat dashboard should be available but you would first need to install it as part of your image building13:10
*** lemko has joined #openstack-helm13:10
itxakanope if you dont add the heat-dashboard to the pydep.txt and rebuild the image13:10
itxakathis is how you can do it, but it may not work as it has no versioning added: https://review.openstack.org/#/c/645545/13:11
itxakaso more work is needed to create something that works for horizon and its plugins (mountVolumes maybe?)13:12
*** Nishant_ has joined #openstack-helm13:13
lemkoNice. I think ti13:13
lemkoI think it was really needed for horizon chart.13:14
*** howell has joined #openstack-helm13:21
evrardjphello lemko :)13:22
evrardjphow are you today?13:22
lemkoHello evrardjp, it's a  nice day in Prague :) thanks for asking.  That project is really getting good. Octavia was even working out of the box. What about you?13:31
*** kranthikirang has joined #openstack-helm13:31
evrardjpthings are busy, the pressure is perceptible :)13:39
evrardjpand it's sunny in Belgium :)13:39
portdirectGot a copy of some the slides from susecon yesterday evrardjp - looks great :)13:40
evrardjpglad to hear that13:41
openstackgerritMerged openstack/openstack-helm master: Fix globbing  https://review.openstack.org/64830813:42
tao_liuportdirect: cool, thanks.13:43
*** Liang__ has joined #openstack-helm13:45
*** cheng1__ has joined #openstack-helm13:48
happyhemantitxaka, well thanks alot heat works for me now  :)13:49
evrardjpgreat to hear that13:50
evrardjpand look at the activity in the channel, it's great to hear!13:50
evrardjpWe should do that everyday!13:51
itxakaawesome, nice to hear happyhemant!13:57
itxakado we have "office hours" in the channel? Like some hours where we are all open for questions and direct help?13:58
evrardjpnot yet, but we are working on it, see this tuesday's meeting14:01
*** Liang__ has quit IRC14:03
openstackgerritSteve Wilkerson proposed openstack/openstack-helm-infra master: Ceph charts: Add missing templates, update dependency values  https://review.openstack.org/64485914:09
openstackgerritOleh Hryhorov proposed openstack/openstack-helm-infra master: [WIP] Add dnsmasq chart  https://review.openstack.org/64489714:17
openstackgerritSteve Wilkerson proposed openstack/openstack-helm master: Armada job: Add upgrade options to rabbitmq chart  https://review.openstack.org/64963114:21
*** ohryhorov has left #openstack-helm14:22
*** ohryhorov has joined #openstack-helm14:23
openstackgerritSteve Wilkerson proposed openstack/openstack-helm master: WIP/DNM: Test consolidation of openstack armada chart groups  https://review.openstack.org/64963214:25
*** jamesgu has joined #openstack-helm14:27
*** cheng1__ has quit IRC14:27
openstackgerritSteve Wilkerson proposed openstack/openstack-helm-infra master: WIP/DNM: Test osh-infra armada chart group consolidation  https://review.openstack.org/64963314:30
openstackgerritSteve Wilkerson proposed openstack/openstack-helm master: WIP/DNM: Test consolidation of openstack armada chart groups  https://review.openstack.org/64963214:31
*** cheng1 has quit IRC14:32
*** cheng1 has joined #openstack-helm14:35
openstackgerritGage Hugo proposed openstack/openstack-helm master: [DNM] Testing resource type  https://review.openstack.org/64931814:42
*** dpawlik has quit IRC14:44
*** aaronsheffield has joined #openstack-helm14:52
itlinuxcan someone suggest how to fix this error thanks http://pastebin.mattei.co/index.php/view/28d91d0214:55
itlinuxtrying to build kube base to deploy openstack-helm thanks14:55
*** sc68cal has joined #openstack-helm15:00
openstackgerritPete Birley proposed openstack/openstack-helm master: WIP: enable compute kit to be deployed from outwidth the cluster  https://review.openstack.org/64762915:08
*** sthussey has joined #openstack-helm15:19
sc68calhey folks, just poking at openstack-helm for the first time - my first question is the chart server - it'd be nice to be able to configure that URL to have one central location, if you have multiple regions that you are looking to deploy15:20
sc68calany support for that? since it appears all the charts are hardcoded to localhost15:20
itlinuxyou and I sc68cal:15:21
evrardjpmmm15:21
sc68cal:)15:21
evrardjpcould you show me that?15:21
itlinuxtrying to install kube with kubespray ;)15:21
evrardjpitlinux: you might want to speak with jayahn15:21
sc68calhttps://github.com/openstack/openstack-helm/blob/master/neutron/requirements.yaml#L1715:21
itlinuxok..I talked to him last night :) I am getting this strange error..15:22
*** witek has quit IRC15:45
portdirectsc68cal: thats a byproduct of our make process15:49
portdirectthe built tarballs can be consumed from anywhere15:49
*** pgaxatte has quit IRC15:50
openstackgerritPRATEEK REDDY DODDA proposed openstack/openstack-helm master: Neutron: Add pod/container security context  https://review.openstack.org/64587515:54
sc68calportdirect: ok - I'm just following https://docs.openstack.org/openstack-helm/latest/install/multinode.html and the `./tools/deployment/multinode/010-setup-client.sh` step, which calls `make all`15:58
sc68calwhich requires you to have the chart server running, etc15:58
sc68calso I was assuming that for every region, I'd have to do that step, which means I'd have multiple chart servers running, when really I'd like to just point to a central location16:00
*** dpawlik has joined #openstack-helm16:02
portdirectsc68cal: once the charts are built into tarballs, you can serve these via any webserver16:14
portdirectwhich allows you to deploy from a known artifact on each site, rather than build JIT each time16:15
portdirecteg `helm install http://foo.bar/hey.tgz`16:16
openstackgerritGage Hugo proposed openstack/openstack-helm master: Add wait.resources.type to armada manifest  https://review.openstack.org/64850316:16
sc68calportdirect: ack. So, if I build the tarballs and place them somewhere, for the next region I build, what step do I start at, since the 010-setup-client.sh script builds the charts, which will have already been completed previously. 020-ingress.sh ?16:16
portdirectpretty much, though itw worth bearing in mind, that at present we dont have a `production` installer for osh in the repo16:17
*** michael-beaver has joined #openstack-helm16:17
portdirectand it is expected that either people will construct their own scripts/playbooks/armada-manifests16:17
portdirectthis is somthing we need to work on - but for now the focus has been getting the raw charts solid, without adding opinion as to how they are deployed.16:18
*** howell has quit IRC16:18
sc68calAck. Is the `production` installer script/playbook/etc just orchestrating the invocation of `helm install` of the openstack components and configuration?16:19
sc68cal(the piece I would have to write)16:19
portdirectin essence yes - there are numerious efforts underway for this at the moment16:20
*** howell has joined #openstack-helm16:20
portdirectthe simplest would just be to use the scripts, slightly modified for your env16:20
portdirectbut there is also thinks like the airship project that look to also cater for the lifecycel of the underling infrastructure as well as the management of helm charts16:21
portdirectand taco: https://github.com/sktelecom-oslab/taco-scripts from skt16:21
openstackgerritMeghan Heisler proposed openstack/openstack-helm-infra master: Add wait.resource for LMA services to armada manifest  https://review.openstack.org/64914116:22
sc68calgotcha. I built this cluster using `kubeadm-ansible` - and some internal ansible scripts for Arista EOS - so bare metal is already handled16:22
portdirectas well as the early stages of an effort from SuSE: https://github.com/SUSE-Cloud/socok8s16:22
portdirectin that case this may help: https://review.openstack.org/#/c/648723/16:23
portdirectwhich is what i have been using to deploy osh on azures' aks for development16:23
portdirectthough it relies upon this ps: https://review.openstack.org/#/c/647629/ that im also in the process of cleaning up16:23
sc68calah ok - so probably lines 74 thru 119-ish are what I'm looking for?16:25
sc68calin 64872316:25
portdirectthis should get you running (i hope)16:26
portdirectprovided you have this: https://review.openstack.org/#/c/647629/3/tools/deployment/multinode/140-compute-kit.sh16:26
sc68calack16:26
portdirect(please forgive the uglyness there, but i give myself a pass as they are WIP)16:26
openstackgerritMeghan Heisler proposed openstack/openstack-helm-infra master: Add wait.resource for LMA services to armada manifest  https://review.openstack.org/64914116:27
sc68calI'm totally fine with WIP16:27
sc68calgetting openstack-helm working for our needs is my full time job now, and i have physical hardware, so I'll start contributiing16:29
sc68calmy env is a little different though - since i am using Calico as the CNI plugin, which was installed via kubeadm16:29
portdirectThat would rock!16:29
sc68calI haven't gotten to the part of openstack-helm that deals with neutron - but i did see a top level var for configuring calico16:29
portdirectThat env sounds similar to my own lab, and what att use with airship16:30
portdirectThe calico integration with neutron will have bit rotted I suspect16:31
sc68calyeah. Right now we've been using `openstack-ansible` and honestly it's a real dead end. It has zero idempotence16:31
sc68calso as part of a refresh we're looking to move to something that's able to handle day 2 operations16:31
sc68calso yeah, we reached for k8s and containers for the openstack control plane, hence, here i am16:32
portdirectNice, we still have some rough edges, but things are starting to shape up well16:33
portdirectTheres a few production deployments out there now, where they heavily rely upon the idempotent nature of the charts16:33
portdirectAs they blanket helm upgrade the sites at least once a month16:34
sc68calyeah. The thing I'm scratching my head on, is the fact that the nodes that will be running neutron and nova are already running `calico/node` daemonsets (because they're k8s nodes already), so I don't know how the calico piece that openstack-helm will want to orchestrate is going to work16:34
portdirectAh ok16:34
portdirectSo here we dont use the cni at all16:35
portdirectAnd default to ovs16:35
portdirectThough you can also use linux bridge out of the box too16:35
portdirectTheres work underway to add tungsten fabric support, and I think skt are using the sdn from cord16:36
*** cfriesen has joined #openstack-helm16:36
portdirectAt some point I'd like to add ovn, but just have lacked the time :(16:36
sc68calif there was just a neutron plugin that just plumbed the VM and just added a route to the hypervisor's routing table, everything would work16:36
sc68calsince I am already running calico, peered with the ToRs, etc etc16:37
portdirectThat would be really nice16:38
sc68calkinda sucks - it's been a while but it sounds like since I did all the legwork outside of openstack for networking, I now have to write a thin ml2 driver :(16:42
*** gkadam has quit IRC16:42
sc68caleither that or I have to have the calico that openstack-helm runs, peer with a switch higer up in the fabric that none of the nodes are already peered with, so it can push routes in16:43
sc68calsort of like what metallb has to do, to work around conflicting with calico16:44
sc68calhttps://metallb.universe.tf/configuration/calico/16:44
*** spsurya has quit IRC16:46
openstackgerritMeghan Heisler proposed openstack/openstack-helm-infra master: Add wait.resource for LMA services to armada manifest  https://review.openstack.org/64914117:12
openstackgerritMeghan Heisler proposed openstack/openstack-helm-infra master: Add wait.resource for LMA services to armada manifest  https://review.openstack.org/64914117:16
openstackgerritdiwakar thyagaraj proposed openstack/openstack-helm-infra master: Add default AppArmor profile to Fluentbit  https://review.openstack.org/64788117:18
*** schwicht has quit IRC17:26
*** spsurya has joined #openstack-helm17:27
*** tao_liu has quit IRC17:30
*** tao_liu has joined #openstack-helm17:30
openstackgerritdiwakar thyagaraj proposed openstack/openstack-helm-infra master: Add default AppArmor profile to Fluentbit  https://review.openstack.org/64788117:32
openstackgerritdiwakar thyagaraj proposed openstack/openstack-helm-infra master: Add default AppArmor profile to Fluentbit  https://review.openstack.org/64788117:33
*** dpawlik has quit IRC17:45
openstackgerritdiwakar thyagaraj proposed openstack/openstack-helm-infra master: Add default AppArmor profile to Fluentbit  https://review.openstack.org/64788117:46
openstackgerritIan Howell proposed openstack/openstack-helm master: WIP/DNM - Add the upgrade lifecycle step  https://review.openstack.org/64918917:46
openstackgerritdiwakar thyagaraj proposed openstack/openstack-helm-infra master: Add default AppArmor profile to Fluentbit  https://review.openstack.org/64788117:49
*** tao_liu has quit IRC17:51
*** tao_liu has joined #openstack-helm17:51
itlinuxhello jayahn: are you around?17:53
*** schwicht has joined #openstack-helm17:54
openstackgerritPete Birley proposed openstack/openstack-helm-infra master: MariaDB: update password cleanly  https://review.openstack.org/64968518:05
openstackgerritPete Birley proposed openstack/openstack-helm-infra master: MariaDB: update password cleanly  https://review.openstack.org/64968518:07
openstackgerritKoffi Nogbe proposed openstack/openstack-helm-infra master: Fix mysqldump default option  https://review.openstack.org/64969018:21
*** kranthikirang has quit IRC18:52
openstackgerritGhanshyam Mann proposed openstack/openstack-helm-infra master: Migrate migrate  Ubuntu jobs from xenial to Bionic  https://review.openstack.org/64969818:58
openstackgerritGage Hugo proposed openstack/openstack-helm master: Add wait.resources.type to armada manifest  https://review.openstack.org/64850318:58
*** anticw has joined #openstack-helm19:00
openstackgerritGhanshyam Mann proposed openstack/openstack-helm master: DNM: testing bionic  https://review.openstack.org/64969919:00
openstackgerritKoffi Nogbe proposed openstack/openstack-helm-infra master: Fix mysqldump default option  https://review.openstack.org/64969019:42
openstackgerritPete Birley proposed openstack/openstack-helm-infra master: WIP: MariaDB: Only expose service when min number of active nodes  https://review.openstack.org/64971119:47
*** howell has quit IRC19:48
itlinuxis there anyone else other than jayahn: that can answer some questions on kubespray so I can get openstack helm going? Thanks19:49
*** kranthikirang has joined #openstack-helm19:49
portdirecti cant help with kubespray - but can with deploying on a running k8s cluster19:49
portdirect(not set up via our dev-env scripts)19:50
*** michael-beaver has quit IRC20:02
*** michael-beaver has joined #openstack-helm20:03
*** sgrasley has quit IRC20:35
*** michaelbeaver has joined #openstack-helm20:39
itlinuxok portdirect. let me see if jayahn: and then I can try another option.. do you have a link to look at the setup (centos 7 bm)20:40
*** michael-beaver has quit IRC20:42
openstackgerritMerged openstack/openstack-helm-infra master: MariaDB: update password cleanly  https://review.openstack.org/64968520:49
*** tao_liu has quit IRC20:50
*** spsurya has quit IRC20:59
openstackgerritPete Birley proposed openstack/openstack-helm-infra master: WIP: Mariadb: Add security policy functions  https://review.openstack.org/64973621:13
openstackgerritDeepak proposed openstack/openstack-helm-images master: Makefile and Docker file for calicoctl-utility containers.  https://review.openstack.org/64974221:16
openstackgerritdiwakar thyagaraj proposed openstack/openstack-helm-infra master: Add default AppArmor profile to Fluentbit  https://review.openstack.org/64788121:19
openstackgerritDeepak proposed openstack/openstack-helm-images master: Makefile and Docker file for calicoctl-utility containers.  https://review.openstack.org/64974221:20
*** schwicht has quit IRC21:29
*** sgrasley has joined #openstack-helm21:36
openstackgerritPete Birley proposed openstack/openstack-helm-infra master: WIP: Mariadb: Add security policy functions  https://review.openstack.org/64973621:41
*** sgrasley_ has joined #openstack-helm21:43
*** sgrasley has quit IRC21:47
itlinuxportdirect: do you have any guidance for the full install on BM?21:48
portdirectso once you have k8s running21:48
itlinuxok21:48
portdirectthe convo i had this morning should help21:48
itlinuxtrying to find the best way to get the K8 on21:48
portdirecthttps://review.openstack.org/#/c/648723/21:48
itlinuxwhich one? I do not remember we talked ;) but it has been a long day21:49
itlinuxthe installation I have with kubespray is having an issue with 3 ethd21:51
itlinuxshould I just keep this as 1 master?21:51
itlinuxI have setup this as 3 master ;)21:51
itlinuxor at least trying since it fails on the cert now21:52
itlinuxhttp://pastebin.mattei.co/index.php/view/8b14626521:52
itlinuxthanks portdirect:21:53
itlinuxdoes it matter which way k8 is deployed?21:53
itlinuxsince jayahn: last night was telling me to talk to you!21:53
itlinuxthanks21:53
openstackgerritDeepak proposed openstack/openstack-helm-images master: Makefile and Docker file for calicoctl-utility containers.  https://review.openstack.org/64974221:59
openstackgerritPete Birley proposed openstack/openstack-helm-infra master: Mariadb: Add security policy functions  https://review.openstack.org/64973622:05
*** michaelbeaver has quit IRC22:10
jayahnitlinux: hi, the reason I might want to talk portdirect, was to get guide to do what gate script does with centos host os. :) I thought following community provided way at the first time would be a good way. Anyway, I am just waking up and won't have time today. However, I will certainly try to find if one of my team can help you today with kubespray. (fyi, I am in Korea, so we are just starting a day, 7am now)22:14
openstackgerritChris Wedgwood proposed openstack/openstack-helm-infra master: [Calico] Update to v3.4.4  https://review.openstack.org/64447122:55
openstackgerritChris Wedgwood proposed openstack/openstack-helm-infra master: [DNM] Calico: Update to v3.6.1  https://review.openstack.org/64448022:58
openstackgerritHemachandra Reddy proposed openstack/openstack-helm-infra master: Make network policy template a little more generic  https://review.openstack.org/64540022:59
itlinuxjayahn: can you give me a tip on this error http://pastebin.mattei.co/index.php/view/03ff2cd923:17
itlinuxtried to clean up and redeploy after doing a reset.yml23:17
itlinuxthank you!23:17
itlinuxjayahn: in Seol? where in Korea?23:37
itlinuxa very good friend in the lang team is there.. he is pretty cool. I am doing the italian section.. but really behind on some of the docs23:39

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!