*** Pavo has joined #openstack-kolla | 00:00 | |
kfox1111 | http://logs.openstack.org/59/403959/10/experimental/gate-kolla-kubernetes-deploy-centos-binary-external-ovs-nv/a357ae7/logs/pods/ shows no openvswitch stuff running, which is very odd. | 00:00 |
---|---|---|
kfox1111 | so I'd guess that. | 00:00 |
kfox1111 | oh... | 00:01 |
kfox1111 | duh. | 00:01 |
kfox1111 | that job is the external openvswitch one. | 00:02 |
kfox1111 | so there are probably breaking issues there too. but won't see a pod for openvswitch. | 00:02 |
kfox1111 | http://logs.openstack.org/59/403959/10/experimental/gate-kolla-kubernetes-deploy-centos-binary-ceph-nv/76b6744/console.html looks to have just gotten stucdk. bad vm maybe. | 00:03 |
*** pbourke_ has joined #openstack-kolla | 00:03 | |
sbezverk | kfox1111: I sorted out a couple of things | 00:03 |
sbezverk | the main thing is image_tag | 00:03 |
kfox1111 | http://logs.openstack.org/59/403959/10/experimental/gate-kolla-kubernetes-deploy-ubuntu-binary-ceph-nv/3a4f6fc/console.html#_2016-11-30_22_31_22_087728 | 00:03 |
kfox1111 | heh. yeah. | 00:03 |
*** absubram has joined #openstack-kolla | 00:04 | |
sbezverk | I see you use image_tag for nuetron server | 00:04 |
sbezverk | but where do you define it? | 00:04 |
sbezverk | it does not seem to be all_values.yaml | 00:04 |
sbezverk | but it should be there since it is very generic name | 00:05 |
kfox1111 | hmm... let me look. | 00:05 |
kfox1111 | sbezverk: I'd say, its broken, and thats why the gates broken. :) | 00:07 |
kfox1111 | was debugging that myself. :) | 00:07 |
kfox1111 | so, yeah. :) | 00:07 |
harlowja | SamYaple was there a docker issue for the tag stuff u filed (or kew about?) | 00:07 |
harlowja | *knew | 00:07 |
kfox1111 | I'm also missinga haproxy tag. | 00:08 |
kfox1111 | kind of surprised helm renders it at all... I think this is a differeence between 2.0.0rc2 and 2.0.0 | 00:09 |
kfox1111 | I need to upgrade my box... | 00:09 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Adding openvswitch-ovsdb microservice https://review.openstack.org/403959 | 00:09 |
*** eaguilar has joined #openstack-kolla | 00:09 | |
sbezverk | kfox1111: it just add "nil" | 00:09 |
sbezverk | I mean helm instead of generating error | 00:09 |
kfox1111 | yeah. I was getting nils on my box and i didn't notice it, | 00:10 |
kfox1111 | but the gate's helm is giving me this weird error: http://logs.openstack.org/29/404929/5/experimental/gate-kolla-kubernetes-deploy-centos-binary-external-ovs-nv/8734904/console.html#_2016-11-30_23_25_59_591981 | 00:11 |
kfox1111 | still may be a differnet bug yet. | 00:11 |
sbezverk | ah I saw something similar when ident got screwed up somewhere | 00:11 |
kfox1111 | k. gona force upgrade to 2.0.0... thinking rc2 might be being more permissive... | 00:12 |
*** hieulq_ has joined #openstack-kolla | 00:12 | |
*** sayantan_ has quit IRC | 00:13 | |
*** sayantan_ has joined #openstack-kolla | 00:13 | |
kfox1111 | ever upgrade a tiller? | 00:14 |
kfox1111 | well, I can just minikube relaunch. nm. | 00:15 |
sbezverk | check out this command not sure though how support it ;-) kubectl edit deployment tiller-deploy | 00:17 |
sbezverk | then change image to point to version you want | 00:17 |
kfox1111 | ah. yeah. | 00:17 |
kfox1111 | thx. | 00:17 |
kfox1111 | I'll try that next time. | 00:17 |
sbezverk | np very usefull for upgrading kube cluster deployed by kubeadm :-) | 00:18 |
kfox1111 | ah. yeah. haven't tried that either yet. :) | 00:18 |
sbezverk | I went from 1.4.4 ro 1.4.6 with just a couple of changes | 00:19 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize neutron-server https://review.openstack.org/404929 | 00:19 |
kfox1111 | I keep a copy of all I deploy in git, so I can burn it down and put it back. | 00:19 |
kfox1111 | so I make that the definitive copy, not whats in k8s itself. | 00:19 |
kfox1111 | helm may change that pattern slightly though... not sure how that will work yet. | 00:20 |
sbezverk | kfox1111: I think we need two tags for images, one global applied for all images where more specific tag does not exist. otherwise we are forced for each microservice specify individual tag, thought? | 00:23 |
*** duonghq_ is now known as duonghq | 00:23 | |
kfox1111 | sbezverk: I think thats the pattern though. if your deploying from helm microservices packages, each has their= own standalone config. so have to override them all individually. | 00:24 |
kfox1111 | getting them all to do something common is what either operators or the service packages are for? | 00:24 |
kfox1111 | helm kind of throws out the idea of having a global config file, unless you are making a super package. | 00:26 |
sbezverk | kfox1111: so each microservice will have a mandatory parameter image_tag defined in all_values.yaml is that your idea? | 00:27 |
kfox1111 | well, not exactly sure what I'm thinking there... all the other ones have had an var name kind of specific to the package name. | 00:28 |
kfox1111 | so neutron-l3-agent package has a var called neutron_l3_agent_image_.... | 00:28 |
kfox1111 | but from a usability standpoint, that causes stuff like: | 00:28 |
sbezverk | kfox1111: I do not care how we call it let's just agree on some common approach | 00:28 |
kfox1111 | helm install neutron-l3-agent --set neutron_l3_agent_image_full=neutron-l3-agent..... | 00:29 |
kfox1111 | sbezverk: yeah. thats the point of the reviews. to start discussing some of this stuff. :/ | 00:29 |
*** yingjun has joined #openstack-kolla | 00:30 | |
*** yingjun has quit IRC | 00:30 | |
kfox1111 | do we make the primary conatainer image just image_full, as its a per microservices var, | 00:30 |
kfox1111 | and therefore garanteed to be unique. | 00:30 |
*** yingjun has joined #openstack-kolla | 00:30 | |
sbezverk | kfox1111: since we put code that can generate image name we need to make sure all variables defined and have some values | 00:30 |
kfox1111 | but slightly confusing when multiple contaienrs are involved, | 00:30 |
kfox1111 | or do we make it explicit always, and make the user have to look it up each time? | 00:30 |
kfox1111 | I could kind of go either way. :/ not sure which will be better long run. | 00:31 |
sbezverk | I ma more worried about image_built | 00:31 |
kfox1111 | sbezverk: to your question, https://hub.docker.com/r/kolla/centos-binary-openvswitch-vswitchd/tags/ | 00:31 |
kfox1111 | what about image_built? | 00:31 |
kfox1111 | looks like 3.0.0 doesn't exist for that container. | 00:32 |
sbezverk | I see if we have code that can potentially build full image name based on just vairables | 00:32 |
kfox1111 | yeah. | 00:32 |
sbezverk | then we need to make sure that all used variables have defaults | 00:32 |
kfox1111 | right. | 00:32 |
kfox1111 | thats what values is. | 00:32 |
kfox1111 | they realy should have named it "default_values.yaml" | 00:33 |
kfox1111 | its the defaults you get when launching, if a user doesn't override them, or a parent package doesn't override them. | 00:33 |
duonghq | can somebody give me current context? | 00:33 |
kfox1111 | heh. its all wrapped up in reviews. would be hard to give context without reading through them. :/ | 00:34 |
duonghq | kfox1111, you mean Helmize neutron-server? | 00:34 |
sbezverk | also in openvswitch-ovsdb PS | 00:34 |
duonghq | okay, this is what context I need | 00:34 |
sbezverk | we are trying to see how to tag images inside of templates | 00:35 |
sbezverk | based on default and in addition gives the way to override on per microservice bases | 00:35 |
kfox1111 | yeah. the way helm works is pretty different then what we did with kollakube. | 00:35 |
kfox1111 | so we have to redecide how to do things we had previously decided to do in a much different way. | 00:35 |
sbezverk | or no default tag at all and always go with a specific per microservice tag | 00:35 |
kfox1111 | and there's an uglyish 3rd option on top of it all I think. global override as the default, and package level overrides. | 00:36 |
kfox1111 | for example: | 00:36 |
kfox1111 | global.debug=False. | 00:37 |
kfox1111 | then in the templates, .Values.debug | default .Values.Global.debug | 00:37 |
kfox1111 | then a super package could set global.debug = true, and it woudl ripple down to sub packages entirely. | 00:37 |
duonghq | you guys mean common tag or default tag? | 00:38 |
kfox1111 | but still turn it back off for specific packages, with global.debug=true,neutron-server.debug=false | 00:38 |
kfox1111 | just speaking generally about variables for the moment. | 00:38 |
kfox1111 | there's a completly different issue running around in the back of my head with reguards to tags/versions. | 00:38 |
kfox1111 | there a version on a helm package. | 00:39 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Adding openvswitch-ovsdb microservice https://review.openstack.org/403959 | 00:39 |
kfox1111 | its a version suposed to uniquely identify the helm template inside it, and the subcharts, at minimum. | 00:39 |
kfox1111 | but if it does just that, | 00:39 |
kfox1111 | if you want to go from container 3.0.0-1 to 3.0.0-2 you would have to "upgrade" to the same chart, with different parameters. | 00:40 |
kfox1111 | not even sure helm will do that right. | 00:40 |
sbezverk | kfox1111: neither do I, I am still only at deploy stage ;-) | 00:41 |
kfox1111 | the other way to do it, is to bump the package release number every time the containers inside change versions and hardcocde the versions at build time, making them not easily changed, as thats not even really a good thing to change. | 00:41 |
kfox1111 | so, say you have a package, openvswitch 3.0.0-1. if either kolla-toolbox or openvswitch container gets updated, openvswitch helm pakcage gets rereleased as 3.0.0-2 pointing to the new fixed container. | 00:42 |
kfox1111 | then helm upgrade detects the newer version then whats on the system, and can do the right thing. | 00:42 |
kfox1111 | it means releasing a lot of packages though. a bump everytime a container changes as well as whne the chart or subcharts change. | 00:43 |
*** hieulq_ has quit IRC | 00:43 | |
duonghq | seems that it's painful when the helm's tree depth is high? | 00:43 |
kfox1111 | (which is why I was kind of pushing to see if we could automate that somewhat) | 00:43 |
*** jtriley has joined #openstack-kolla | 00:43 | |
kfox1111 | like, if we keep the versions all in one yaml, and push them to the Chart.yaml, it would be easier to detect and automate building new ones. | 00:44 |
kfox1111 | say, in a ci/cd job. | 00:44 |
duonghq | seem like micro-template engine will do the job | 00:44 |
kfox1111 | the ci/cd can detect a container needs updating, and then asks the build system to go build new affected helm packages. | 00:45 |
sbezverk | kfox1111: where ci/cd process detecting the change runs? | 00:45 |
kfox1111 | so, this is kind of the direction I've been tryign to take things, have a way where ci/cd could keep things built constantly and up to date, and crank out images. | 00:46 |
kfox1111 | sbezverk: some ci/cd system. could be the openstack infra (would be nice), | 00:46 |
kfox1111 | could be a local enterprises jenkins/gitlab system, | 00:46 |
kfox1111 | or maybe the ci/cd runs right in the k8s system the kolla system's using. | 00:46 |
kfox1111 | sharing that task. :) | 00:47 |
*** sdake_ has quit IRC | 00:48 | |
kfox1111 | I haven't really throught throgh all the ramifications and fleshed it all out, but I konw some big sites are doing this same kind of thing, and I think it will be pretty important going forward to support atleast the workflow, if we don't provide the solution. | 00:48 |
*** Marcellin__ has quit IRC | 00:48 | |
sbezverk | kfox1111: yeah I am not sure if ci/cd or integration is a part of kolla-kubernetes ;-) | 00:49 |
sbezverk | but think about it I guess it is important | 00:50 |
*** hfu has joined #openstack-kolla | 00:50 | |
kfox1111 | the problem is, without some kind of tool like that, it becomes much harder to do security volnerability remediation. :/ | 00:50 |
*** jtriley has quit IRC | 00:50 | |
kfox1111 | and if that doesn't happen, kolla based clouds become jumping off points for attacking others, and kolla gets a black eye. | 00:51 |
kfox1111 | same way iot device companies are getting beet up right now. | 00:51 |
*** tovin07_ has joined #openstack-kolla | 00:52 | |
kfox1111 | like you said, I'm not sure its belongs to kolla-kkubernetes to solve. its actually a bigger, docker/k8s issue. | 00:53 |
*** JRobinson__ is now known as JRobinson__afk | 00:53 | |
kfox1111 | but we need to be able to provide enough support somehow for something to make it possible. | 00:53 |
duonghq | kfox1111, can you define "enough"? | 00:55 |
kfox1111 | duonghq: not entirely sure what that looks like. peruprosfully a bit vague. | 00:55 |
kfox1111 | but, for example, | 00:55 |
*** hieulq_ has joined #openstack-kolla | 00:55 | |
kfox1111 | I've been adding an attribute to all services that should be completely safe to upgrade without users knowing it happened. | 00:56 |
duonghq | basically, we need something like version manager for package? (if I keep up correctly) | 00:56 |
kfox1111 | a ci/cd tool could look at those, look for updates on them, and automatically upgrade them. | 00:56 |
kfox1111 | duonghq: yeah, thats another way it coudl help. | 00:56 |
kfox1111 | so that if version numbers always increment with every change, | 00:56 |
kfox1111 | it would be easier to do a helm list, and a helm search ,and see if there is a new version available. | 00:57 |
kfox1111 | same way you do a 'yum upgrade' on your system, and if finds updated packages and puts them in place. | 00:57 |
kfox1111 | you don't nessisarily tell it what to updgrade individually. it goes and figures it out for you. | 00:57 |
duonghq | kfox1111, agreed | 00:57 |
*** fguillot has joined #openstack-kolla | 00:58 | |
duonghq | I do not have much experience in production, but does ci/cd is common in production environment? | 00:58 |
kfox1111 | I think we want to try and figure out a way to make it such that we don't prevent that kind of tool from working. | 00:58 |
kfox1111 | duonghq: its costly, as most sites don't share their stuff. :/ so it tends to be done when the SLA is high enough to justify the cost of standing one up. | 00:58 |
kfox1111 | but if opensource philosophy is adopted, and its contributed and worked on in the open, | 00:59 |
kfox1111 | it could be much cheeper for everyone to implement and benifit from. | 00:59 |
kfox1111 | same as with kolla-kubernetes itself. | 00:59 |
sbezverk | kfox1111: I think it is nice to have, but concern me is the time | 00:59 |
kfox1111 | everyone scrambled to build their own snowfflakes, in their own sites, | 01:00 |
kfox1111 | and then have to pay for every aspect of it. | 01:00 |
kfox1111 | kolla-kuberneetes is atempting to opensource and croudsource the development of it, making it cheeper ofr everyone. | 01:00 |
kfox1111 | sbezverk: yeah. this is far out stuff. not the next 6 months. | 01:00 |
kfox1111 | but, if we're a bit careful now in our decisions, it will make things easier down the road too. | 01:01 |
kfox1111 | so a tradeoff there. | 01:01 |
duonghq | so, I think whether people use ci/cd in production as greatly depend on their model, but if we want to use some kind of ci/cd, we need to do general, due to each company use difference ci/cd stack? | 01:01 |
sbezverk | kfox1111: I think if it is achieved within reasonable time, it would be huge win for everybody | 01:01 |
kfox1111 | duonghq: yeah. some jenkins, some gitlab, and there are a bunch of others. | 01:02 |
portdirect | v1k0d3n/mgiles - back with the living again | 01:02 |
kfox1111 | sbezverk: +1. it would be a killer feature. | 01:02 |
kfox1111 | sbezverk: (and I think mirantis just bought a company just to get their ci/cd tech for their openstack distro... ;) | 01:02 |
duonghq | so, which ci/cd do you want to PoC your idea? | 01:03 |
*** fguillot has quit IRC | 01:03 | |
kfox1111 | its far out that I'm not even thinking of a PoC. I havent' researched the alternatives much. | 01:03 |
duonghq | portdirect, you are Pete Birley? | 01:03 |
*** msimonin has quit IRC | 01:03 | |
duonghq | kfox1111, okay | 01:03 |
kfox1111 | I'm just at the point of thinking, what woudl be needed from the package build system and the microservices vars to enable the workflow. | 01:03 |
portdirect | duonghq: wish i could deny it - but yes :) | 01:04 |
sbezverk | kfox1111: back to tags, at least for now let's use something common and then revisit when you compose your ideas into something | 01:04 |
*** hieulq_ has quit IRC | 01:05 | |
kfox1111 | well, the easist thing is to stick to the <package_name>_image_... convention. I think its common to all the patches except neutron-server at the moment. | 01:05 |
duonghq | kfox1111, due to kolla's docker image is recommended to build locally and push to local registry, I think we should do the same for kolla-k8s? and the package build system is helm-related? | 01:05 |
kfox1111 | duonghq: two possible workflows there. | 01:06 |
sbezverk | kfox1111: ok, last thing do we introduce global image tag or keep at microservice specific level only? | 01:06 |
kfox1111 | that as you described is currently driven by a human. you could do the same by driving it by a ci/cd. | 01:06 |
kfox1111 | the seccond is, openstack could do that in the gate, and release prebuilt helm packages on a repo that pass the gate tests, and the prebuilt/updated containers to the hub. | 01:07 |
kfox1111 | then an end user would do something like: helm repo add https://helm.openstack.org | 01:07 |
kfox1111 | helm install kolla/neutron-server ... | 01:07 |
kfox1111 | or helm upgrade neutron-server | 01:08 |
*** hieulq_ has joined #openstack-kolla | 01:09 | |
kfox1111 | just like yum. | 01:11 |
kfox1111 | (or apt) | 01:11 |
kfox1111 | you don't usually think these days about building packages from source. you just go fetch some. | 01:11 |
sbezverk | kfox1111: sorry to interrupt you, need your input for gate. I see openvswitch-ovsdb got deployed but some other containers are crashing, could you check the logic | 01:11 |
kfox1111 | k. sec. | 01:11 |
sbezverk | http://logs.openstack.org/59/403959/12/experimental/gate-kolla-kubernetes-deploy-centos-binary-ceph-multi-nv/f0b0dbd/ | 01:11 |
kfox1111 | sbezverk: gota think a bit more about the global image tag. maybe we talk about it more tomorrow? sleeping on it a bit might help. | 01:11 |
duonghq | the 2nd option is using gate to trigger build and push process? | 01:11 |
*** zhurong has joined #openstack-kolla | 01:11 | |
duonghq | I do not know we have way to detect the ps is merged and run some script? | 01:11 |
sbezverk | kfox1111: sure thing, we have tons of things to do anyway | 01:11 |
*** hieulq_ has quit IRC | 01:11 | |
kfox1111 | sbezverk: http://logs.openstack.org/59/403959/12/experimental/gate-kolla-kubernetes-deploy-centos-binary-ceph-multi-nv/f0b0dbd/logs/pods/kolla-openvswitch-vswitchd-compute-3krpf.txt | 01:11 |
kfox1111 | there's the kubectl describe pod | 01:11 |
sbezverk | kfox1111: right but this one is gets lauched using old tool, so not modified by my PS | 01:12 |
kfox1111 | all running pods are here: http://logs.openstack.org/59/403959/12/experimental/gate-kolla-kubernetes-deploy-centos-binary-ceph-multi-nv/f0b0dbd/logs/pods/kolla-openvswitch-vswitchd-compute-3krpf.txt | 01:13 |
*** zhubingbing has joined #openstack-kolla | 01:13 | |
kfox1111 | rrr, bad paste. | 01:13 |
duonghq | kfox1111, I do not agree with you on "end user would do something like: helm repo add https://helm.openstack.org", it does not follow current Kolla (docker) guideline? | 01:13 |
kfox1111 | http://logs.openstack.org/59/403959/12/experimental/gate-kolla-kubernetes-deploy-centos-binary-ceph-multi-nv/f0b0dbd/logs/pods.yaml | 01:13 |
kfox1111 | so we can see what helm rendered there. | 01:13 |
kfox1111 | duonghq: there is no infrastructure in kolla today to keep packages up to date. so running off of out of date containers are counteridcated for good reason. | 01:14 |
kfox1111 | but. there is one use case at least where running out of date containers are still okish. | 01:14 |
kfox1111 | testing/kicking the tires. | 01:14 |
kfox1111 | which is why they still release prebuild containers to the docker hub. | 01:14 |
kfox1111 | the same can be done with helm packages even without having a mechanism to crank out up to date images continuously. | 01:15 |
duonghq | as I understand, currently we recommend user build and use local registry, so it's upto user to update their local registry | 01:15 |
kfox1111 | makes it much much easier to test deploy a kolla-kubernetes system. | 01:15 |
kfox1111 | right. the problem of watching for security updates is pushed apon the user, | 01:15 |
sbezverk | kfox1111: helm deployed only openvswitch-ovsdb and it is stable. other containers not deployed by helm were crashing | 01:15 |
kfox1111 | and hte user has no real tools to di it themselves. :/ | 01:15 |
*** hieulq_ has joined #openstack-kolla | 01:16 | |
duonghq | the prebuild containers are never built and updated usually | 01:16 |
kfox1111 | sbezverk: I don't think so. there are two instances of openvswitch here. I'm guessin the compute one is crashing. | 01:16 |
kfox1111 | sbezverk: see http://logs.openstack.org/59/403959/12/experimental/gate-kolla-kubernetes-deploy-centos-binary-ceph-multi-nv/f0b0dbd/logs/pods/kolla-openvswitch-vswitchd-compute-3krpf.txt its restarting and failing to init. | 01:17 |
*** tonanhngo has quit IRC | 01:17 | |
kfox1111 | wichi is probably causing neutron-openvswitch-agent grumpyness too. | 01:17 |
*** sayantan_ has quit IRC | 01:17 | |
portdirect | duonghq: juts looking through your comments on my docs, thanks man - they are really great - I'm currently going through the style guide and getting it setup so I can render them locally, but may have some q's - will you still be about in 45? | 01:18 |
sbezverk | kfox1111: openvswitch-vswitchd is NOT deployed by helm | 01:18 |
*** JRobinson__afk has quit IRC | 01:18 | |
kfox1111 | hmm... | 01:18 |
sbezverk | I suspect | 01:18 |
kfox1111 | just the db? | 01:18 |
sbezverk | yes | 01:18 |
kfox1111 | but vswitchd depends on the db? | 01:18 |
sbezverk | yes but it should be not relevant how they deployed right? | 01:18 |
*** sdake has joined #openstack-kolla | 01:19 | |
sbezverk | they are pods after all :-) | 01:19 |
duonghq | portdirect, I have meeting in about 10min, for 30min | 01:19 |
kfox1111 | sec... lets look at its logs then | 01:19 |
kfox1111 | http://logs.openstack.org/59/403959/12/experimental/gate-kolla-kubernetes-deploy-centos-binary-ceph-multi-nv/f0b0dbd/logs/openstack/kolla-openvswitch-ovsdb-compute-tp36f.txt | 01:19 |
kfox1111 | that looks suspicious. | 01:19 |
kfox1111 | the selector wrong probably? | 01:19 |
sbezverk | I think the problem is that two of them are rnning at the same time | 01:19 |
kfox1111 | yup. | 01:20 |
sbezverk | there must be a logic issue in my chage for ceph_workflow | 01:20 |
kfox1111 | http://logs.openstack.org/59/403959/12/experimental/gate-kolla-kubernetes-deploy-centos-binary-ceph-multi-nv/f0b0dbd/logs/pods.yaml | 01:20 |
sbezverk | since you developed can you check ceph_workflow in my ps? | 01:20 |
kfox1111 | grep for openvswitch-ovsdb-compute | 01:20 |
kfox1111 | and then nodeSelector | 01:20 |
kfox1111 | its showing kolla_controller: | 01:20 |
kfox1111 | so the helm install needs --set selector_key=kolla_compute | 01:21 |
kfox1111 | for that one instance. | 01:21 |
sbezverk | kfox1111: got it in the command line in ceph_workflow that is right? | 01:22 |
*** sayantan_ has joined #openstack-kolla | 01:23 | |
kfox1111 | sbezverk: commented in the review. | 01:24 |
sbezverk | kfox1111: thanks, it will take a bit of time to get used to new approach | 01:25 |
kfox1111 | yeah. it will take a bit of time to flesh out all the details too. I'm still trying to figure this stuff out too. :/ | 01:26 |
*** pbourke_ has quit IRC | 01:26 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Adding openvswitch-ovsdb microservice https://review.openstack.org/403959 | 01:26 |
*** tonanhngo has joined #openstack-kolla | 01:28 | |
kfox1111 | yeah. I think that one has a good chance of working. :) | 01:28 |
sbezverk | kfox1111: I hope so it is getting late on east coast | 01:29 |
*** pbourke_ has joined #openstack-kolla | 01:29 | |
kfox1111 | yeah. I need to head out in a few min too. | 01:30 |
Jeffrey4l_ | gate is broken. anyone can review this https://review.openstack.org/404844 which unblock the gate. | 01:30 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize neutron-server https://review.openstack.org/404929 | 01:32 |
*** tonanhngo has quit IRC | 01:32 | |
*** hieulq_ has quit IRC | 01:34 | |
*** kristian__ has quit IRC | 01:34 | |
*** kristian__ has joined #openstack-kolla | 01:35 | |
openstackgerrit | caoyuan proposed openstack/kolla: Consistent home directory creation for all the services https://review.openstack.org/390179 | 01:35 |
*** jtriley has joined #openstack-kolla | 01:38 | |
*** sdake_ has joined #openstack-kolla | 01:38 | |
*** pbourke_ has quit IRC | 01:39 | |
*** kristian__ has quit IRC | 01:39 | |
Jeffrey4l_ | sdake, please review https://review.openstack.org/404844 which blocks the gate. | 01:39 |
*** eaguilar has quit IRC | 01:40 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: Helmize openvswitch-agent and l3-agent https://review.openstack.org/404999 | 01:41 |
kfox1111 | k. gota head out. l8r all. | 01:41 |
sbezverk | kfox1111: ttyl | 01:41 |
*** sdake has quit IRC | 01:42 | |
*** portdirect has quit IRC | 01:43 | |
*** jtriley has quit IRC | 01:44 | |
*** adrian_otto has quit IRC | 01:45 | |
*** portdirect has joined #openstack-kolla | 01:51 | |
*** Pavo has quit IRC | 01:59 | |
duonghq | portdirect, around? | 02:02 |
portdirect | yep | 02:02 |
sbezverk | kfox1111: finally :-) all green. so gate can deploy helm charts!! | 02:02 |
duonghq | I start thinking we need another option for dev env setup or your guide need more details? I don't think vagrant have much issue like this | 02:03 |
*** Pavo has joined #openstack-kolla | 02:04 | |
portdirect | duonghq: dont parse the end of your line " I don't think vagrant have much issue like this", but open to suggestions - vagrant is by no means perfect solution (frankly I dont really like it - but it gets the job done) | 02:06 |
portdirect | duonghq: I'm goint to split the doc in my next ps into 2: vagrant setup & k8s/helm setup so that we can accomodate other platforms/envs | 02:08 |
duonghq | portdirect, thank you | 02:08 |
duonghq | I mean I don't think people can have trouble with your vagrant setup | 02:09 |
portdirect | duonghq, np - but I totally appricate it's not a perfect solution, for example (and dropping them in it) - I dont think that either kfox1111 or sbezverk will use it/or find it appropriate for their needs? | 02:11 |
*** JRobinson__afk has joined #openstack-kolla | 02:11 | |
portdirect | sbezverk: if you have any input on the about I'd appriacte it :) | 02:12 |
duonghq | me too, I'm using kube-deploy as Ryan suggested (before your guide is pushed to gerrit) | 02:12 |
*** JRobinson__afk is now known as JRobinson__ | 02:12 | |
portdirect | duonghq: thats fine :) the only thing I'd caution against is actually spending much long on this issue - as k8s deployment is already becoming a feild in its own right :/ | 02:14 |
duonghq | It toke me over 1 months for get one working k8s cluster :) | 02:15 |
duonghq | 2-3 months ago :) | 02:15 |
openstackgerrit | Li Yingjun proposed openstack/kolla-kubernetes: Fix different repo issue for documentation https://review.openstack.org/405054 | 02:16 |
portdirect | duonghq: I've also done a few bits and bobs round the edges of https://github.com/projectatomic/commissaire, which when ready will be a great solution for k8s/openshift cluster management - as it also looks after host and cluster upgrades, when both kolla-k8s and it are a bit further along I'll exlpore providing specific features there for kollas requirements | 02:18 |
duonghq | portdirect, good news | 02:18 |
portdirect | *expectantly waits* | 02:19 |
*** goldyfruit has joined #openstack-kolla | 02:19 | |
*** jmccarthy has joined #openstack-kolla | 02:19 | |
*** pbourke_ has joined #openstack-kolla | 02:21 | |
zhangshuai | @all, hello everybody, I have met one error in kolla-ansible. First, I pull all images over by running "kolla-build", it is all ok. and when I run "kolla-ansible pull", I received error msg like "Unknown error message: Tag 4.0.0 not found in repository docker.io/kolla/ubuntu-source-kolla-toolbox" | 02:29 |
*** JRobinson__ has quit IRC | 02:31 | |
*** jtriley has joined #openstack-kolla | 02:32 | |
*** sayantan_ has quit IRC | 02:36 | |
*** jtriley has quit IRC | 02:37 | |
harlowja | zhangshuai see https://github.com/docker/docker-py/pull/1324/ | 02:39 |
harlowja | ^ seems to be the cause | 02:39 |
*** sayantan_ has joined #openstack-kolla | 02:39 | |
*** bjolo_ has joined #openstack-kolla | 02:39 | |
*** JRobinson__ has joined #openstack-kolla | 02:40 | |
zhangshuai | ok, let me see, thank you | 02:40 |
zhangshuai | thank you @hariowja | 02:40 |
harlowja | np | 02:40 |
*** goldyfruit has quit IRC | 02:41 | |
*** goldyfruit has joined #openstack-kolla | 02:41 | |
srwilkers_ | duonghq, vagrant is just a means to provision something to deploy kubernetes to. you can use it to provision to openstack if you wish or even just run the halcyon-kubernetes playbooks without the vagrant aspect | 02:42 |
duonghq | srwilkers_, understood | 02:44 |
*** inc0 has joined #openstack-kolla | 02:48 | |
*** yuanying has quit IRC | 02:51 | |
*** yuanying has joined #openstack-kolla | 02:51 | |
*** goldyfruit has quit IRC | 02:52 | |
*** hieulq_ has joined #openstack-kolla | 02:55 | |
*** saneax-_-|AFK is now known as saneax | 02:56 | |
*** wwriverrat has joined #openstack-kolla | 02:59 | |
openstackgerrit | jianyi proposed openstack/kolla: Install senlin dashboard into horizon image https://review.openstack.org/405070 | 02:59 |
*** rkrum has joined #openstack-kolla | 03:00 | |
rkrum | hi | 03:01 |
rkrum | does anyone know who is looking after the code split of the deployment and orchestration? | 03:01 |
rkrum | i.e. https://github.com/openstack/kolla and https://github.com/openstack/kolla-ansible | 03:02 |
rkrum | some work I've done is missing: https://review.openstack.org/#/c/353261/ | 03:04 |
v1k0d3n | yo portdirect | 03:06 |
portdirect | yo! | 03:06 |
v1k0d3n | kfox1111 and sbezverk are you allowing for users to be able to install individual components and not rely on the global openstack helm deployment? | 03:07 |
portdirect | v1k0d3n: ^^ This is part of the spec for kolla-k8s so that should be possible | 03:09 |
v1k0d3n | portdirect: you're really familiar with our repos? | 03:09 |
portdirect | v1k0d3n: fairly | 03:10 |
portdirect | sup | 03:10 |
v1k0d3n | i can't keep up with both honestly. for us, our PoC is first class....but all this is pretty well flushed out already. | 03:10 |
v1k0d3n | want to just "take" that and have them reviewed? | 03:10 |
v1k0d3n | also...as i was telling the whole group at one point... | 03:10 |
v1k0d3n | make them so the helm templates can be turned over to kubernetes/helm. | 03:11 |
v1k0d3n | that's our goal. | 03:11 |
v1k0d3n | generic, user can add anything they want to the values, but more global support. | 03:11 |
v1k0d3n | so users can just install via "helm install stable/kolla-kubernetes" | 03:12 |
v1k0d3n | done | 03:12 |
v1k0d3n | if they aren't generic enough...kolla owns all work. not as wide of an audience. value is lost. | 03:12 |
v1k0d3n | also...a play like that... | 03:12 |
v1k0d3n | trumps all other solutions. | 03:12 |
v1k0d3n | it's the end all, be all.. | 03:12 |
v1k0d3n | beast mode | 03:13 |
srwilkers_ | big league! | 03:13 |
v1k0d3n | i've stated this before...but i feel meaning has been lost somehow/some way | 03:14 |
v1k0d3n | srwilkers_... :) | 03:14 |
v1k0d3n | funny dude ^ | 03:14 |
portdirect | v1k0d3n: take it you have seen this: https://github.com/openstack/kolla-kubernetes/blob/master/specs/kolla-kubernetes-arch.rst ? | 03:14 |
inc0 | portdirect, do you ever sleep? | 03:14 |
inc0 | like...ever? | 03:14 |
v1k0d3n | though i may have helped with this somewhat portdirect but not sure of final stages. | 03:15 |
v1k0d3n | there's been a lot of back and forth. | 03:15 |
portdirect | just woke up at 1am - been sick with the flu over the last week - so working from home on a mad scietish schedule... | 03:15 |
v1k0d3n | inc0: portdirect is a driven dude...driven dudes never sleep. | 03:16 |
v1k0d3n | (again...beast mode) | 03:16 |
inc0 | ahh, explains why I end work 5pm sharp | 03:16 |
v1k0d3n | damn sorry portdirect that sucks | 03:16 |
inc0 | v1k0d3n, hows poc going btw? | 03:16 |
portdirect | all good - it's only my body that hates me :) | 03:16 |
v1k0d3n | inc0: you miss some good convos when you're gone *sigh | 03:17 |
v1k0d3n | lol | 03:17 |
inc0 | I think helm stuff was is fairly defined by now | 03:17 |
inc0 | yeah, but not miss other stuff | 03:17 |
v1k0d3n | portdirect: a nice whiskey can cure that | 03:17 |
inc0 | ;) | 03:17 |
inc0 | like packing apartament :/ | 03:17 |
*** unicell has quit IRC | 03:17 | |
v1k0d3n | inc0: did you see my note about creating charts that can go into kubernetes | 03:17 |
inc0 | my second biggest love left me today ;( | 03:17 |
inc0 | (shipped my convertible to Oregon) | 03:18 |
v1k0d3n | i mentioned this goal a while back...but the message was lost in all of the back and forth. | 03:18 |
inc0 | v1k0d3n, well, I think helm install openstack is an utopia | 03:18 |
inc0 | frankly | 03:18 |
v1k0d3n | inc0: convertible? what kind? (and please don't say something like a chevy cavalier ) :) | 03:18 |
inc0 | mazda mx-5 miata 2016 | 03:19 |
v1k0d3n | inc0: that's what our mission is. ;) | 03:19 |
v1k0d3n | we wouldn't be doing this if there wasn't a utopia to aim for. | 03:19 |
v1k0d3n | (being serious about that) | 03:19 |
v1k0d3n | so first thing is...does the current helm mariadb work for kolla. | 03:19 |
v1k0d3n | yes / no? | 03:20 |
v1k0d3n | if not...what changes can be added. | 03:20 |
v1k0d3n | done | 03:20 |
v1k0d3n | next | 03:20 |
inc0 | thats actually something I'd love to see | 03:20 |
v1k0d3n | rabbitmq. what changes, and so on | 03:20 |
inc0 | remove infrastructure services from kolla-kubernetes | 03:20 |
v1k0d3n | that's what our PoC is for :) | 03:20 |
inc0 | would remove tons of complexity | 03:20 |
portdirect | +1 | 03:20 |
inc0 | hmm..funny enough I didn't think of it | 03:21 |
inc0 | I'm so deep in our kolla-ansible paradigm | 03:21 |
inc0 | I guess | 03:21 |
*** absubram has quit IRC | 03:21 | |
inc0 | nvm, +2 from me | 03:21 |
v1k0d3n | well...then things become CI | 03:21 |
portdirect | i can see some pushpack on that though | 03:21 |
v1k0d3n | so anyway...what i'm saying is...this should be the goal | 03:21 |
inc0 | why? and from whom? | 03:21 |
v1k0d3n | utopia | 03:21 |
openstackgerrit | jianyi proposed openstack/kolla-ansible: Enable senlin dashboard when enable_senlin is yes https://review.openstack.org/405078 | 03:21 |
v1k0d3n | this is what we're shooting for. | 03:21 |
v1k0d3n | so if kolla is on same page...great | 03:21 |
v1k0d3n | this is why i'd like to cross check both | 03:22 |
inc0 | I just want kolla-k8s released in ocata;) | 03:22 |
v1k0d3n | srwilkers_: is trying to help with this. | 03:22 |
duonghq | inc0, let do this | 03:22 |
inc0 | so I see clear way to do it | 03:22 |
inc0 | like - use helm mariadb | 03:22 |
portdirect | inc0: i suspect that some people will want to have full controll over their infra - and not trust upstreams/3rd party infra backing their deployment | 03:22 |
v1k0d3n | things are moving fast though. can't keep up with irc or else i would stop working :) | 03:22 |
v1k0d3n | lol | 03:22 |
inc0 | portdirect, well, so here's thing - if we assume mariadb external always | 03:22 |
portdirect | inc0: though I'm all for it - just playing devils advocate | 03:22 |
inc0 | (even if we deploy it) | 03:22 |
inc0 | you can go either way | 03:23 |
inc0 | modularity | 03:23 |
srwilkers_ | ^ | 03:23 |
v1k0d3n | portdirect: bingo! "i suspect that some people will want to have full controll over their infra" <<< ding ding ding. that's the only way we can do this. | 03:23 |
inc0 | I'm all for it, deploying just openstack is hard enough | 03:23 |
*** jtriley has joined #openstack-kolla | 03:24 | |
inc0 | kfox1111, sdake_ sbezverk reading this ^ | 03:24 |
v1k0d3n | so the tricky part is making a "kolla-kubernetes" master helm deployment... | 03:24 |
inc0 | what say you? how about we won't even write mariadb+rabbitmq support now? | 03:24 |
v1k0d3n | that uses all the others. | 03:24 |
portdirect | from my perspective ceph is another one on that list? | 03:24 |
v1k0d3n | BUT allows all the others to be individually deployed and completely customizable. | 03:24 |
inc0 | portdirect, ofc | 03:24 |
v1k0d3n | that;s the trick. but that's what we're doing now. | 03:24 |
portdirect | v1k0d3n: yeah your ceph (in aic) is nice | 03:25 |
v1k0d3n | so it does not end up like SAP....where you cannot throw away the top part of the deployment | 03:25 |
v1k0d3n | they are all going that way. | 03:25 |
v1k0d3n | look at the others. | 03:25 |
inc0 | v1k0d3n, they never really wanted modularity | 03:25 |
v1k0d3n | "they"? | 03:25 |
inc0 | they want their own solution | 03:25 |
inc0 | SAP | 03:25 |
v1k0d3n | oh | 03:25 |
inc0 | for their own needs | 03:25 |
inc0 | just sharing what they have, which is fair and fine | 03:26 |
v1k0d3n | well...i don't think they knew how to do that at first... | 03:26 |
v1k0d3n | but yes | 03:26 |
v1k0d3n | ours is a completely different approach | 03:26 |
inc0 | anyway, I'm all for minimizing complexity by removing work;) | 03:26 |
v1k0d3n | want everything to be modular, customizable, pluggable...but also have a global deployment, | 03:26 |
inc0 | if somebody will tell me that helm mariadb is well written that is | 03:27 |
inc0 | or if it isn't we can help it being well written | 03:27 |
v1k0d3n | well, we need a person who can reach out to helm and poke around with it. | 03:27 |
portdirect | inc0: I think we would need to help them | 03:27 |
v1k0d3n | all the other charts that require maria...just uses it as we're talking about now | 03:27 |
inc0 | I'd bet money on that, yes | 03:27 |
inc0 | but that's cool if they will be nice community | 03:27 |
portdirect | same with ceph - as they have a great set of kube templates - but no packaging | 03:27 |
portdirect | though v1k0d3n's crew have done that | 03:28 |
inc0 | well here's deal | 03:28 |
v1k0d3n | ^^ | 03:28 |
inc0 | I'm ok with us being people supporting mariadb helm package | 03:28 |
inc0 | but as helm community more than kolla | 03:29 |
inc0 | but yes v1k0d3n we need to merge our communities to some extent | 03:29 |
v1k0d3n | srwilkers_: is looking at this. | 03:29 |
yingjun | Hi, i’m trying to run kolla-kubernetes, when i running ‘kolla-kubernetes bootstrap mariadb’ i get error: | 03:29 |
yingjun | “jinja2.exceptions.UndefinedError: 'hostvars' is undefined” | 03:29 |
v1k0d3n | unfortunately he's getting beat to these reviews lol | 03:29 |
yingjun | am i missed to config something? | 03:30 |
v1k0d3n | we're trying to make it easier for him | 03:30 |
inc0 | yingjun, kolla-k8s is in total chaos now | 03:30 |
inc0 | be careful | 03:30 |
v1k0d3n | but for the most part...you can drop in what we have...tweak ever so slightly and should be good | 03:30 |
v1k0d3n | we're using kolla images | 03:30 |
yingjun | Oops.. | 03:30 |
inc0 | it's in very fast development now | 03:30 |
inc0 | and changing a lot | 03:30 |
*** tovin07 has quit IRC | 03:31 | |
inc0 | v1k0d3n, I do want you guys in Kolla community at the end of the day;) | 03:31 |
duonghq | yingjun, kolla-kubernetes is in re-constructing process, so the doc is totally outdated, | 03:31 |
inc0 | I'm sure we can figure out any differences we might find out having | 03:31 |
v1k0d3n | duonghq: is right yingjun...going to be much better with new approach. but refactoring now. | 03:31 |
portdirect | yingjun: if you followy the directions here: that should lest you kick the tyres of the last iteration: https://github.com/openstack/kolla-kubernetes/blob/master/doc/source/minikube-quickstart.rst | 03:32 |
yingjun | Yeah, found that, one can’t get kolla-kubernetes run correctly following the doc.. | 03:32 |
v1k0d3n | srwilkers_: you and inc0 have been working to merge in? | 03:32 |
v1k0d3n | portdirect: you wanna help srwilkers_ with this process? | 03:32 |
inc0 | I didn't, I'm lazy and I don't review or write code | 03:32 |
inc0 | just lecture other people | 03:32 |
srwilkers_ | inc0, ;) | 03:32 |
portdirect | v1k0d3n: sure | 03:32 |
v1k0d3n | we can keep things moving forward on our side...keeping the PoC moving (as we need it jan first), but this way we can stay in sync | 03:32 |
inc0 | that woudl be the plan | 03:33 |
inc0 | even if our standards will differ at the end, we want essentially same thing | 03:33 |
v1k0d3n | lol inc0...too busy in the cavalier convertible?? | 03:33 |
inc0 | and using same tools | 03:33 |
v1k0d3n | jk btw :) | 03:33 |
openstackgerrit | jianyi proposed openstack/kolla: code alphabetical order https://review.openstack.org/405082 | 03:33 |
srwilkers_ | i dont see how any human being fits in a miata by the way | 03:34 |
v1k0d3n | i'm just jealous | 03:34 |
srwilkers_ | theres zero chance id fit in one | 03:34 |
duonghq | guess that I should add a comment to kolla-k8s guide :) | 03:34 |
v1k0d3n | lol! | 03:34 |
inc0 | cavelier is so ultimately boring car | 03:34 |
v1k0d3n | srwilkers_: you may not fit into a miata...but inc0 isn't a big dude | 03:34 |
srwilkers_ | still better than a lebaron | 03:34 |
yingjun | i’d like to get involved to the kolla-k8s dev, where should i start from? | 03:34 |
v1k0d3n | inc0: absolutely! lol | 03:34 |
inc0 | v1k0d3n, new trim is a bit bigger than older | 03:34 |
inc0 | I'm 6'+ | 03:34 |
duonghq | yingjun, 1. setup dev env, 2. Read Ryan's spec | 03:34 |
v1k0d3n | dated a girl in college that had one...didn't last long. i couldn't ride in that thing. | 03:35 |
portdirect | duonghq: add comment re fast moving target atm? | 03:35 |
inc0 | 2015 - couldn't drive | 03:35 |
inc0 | 2016 - I'm lovin' it | 03:35 |
srwilkers_ | inc0, ill just leave this here for you: https://www.youtube.com/watch?v=kp1kuo6xkbE | 03:35 |
duonghq | portdirect, comment: normal user should not use kolla-k8s atm | 03:35 |
inc0 | but yeah, I can't grow any more | 03:35 |
*** tovin07 has joined #openstack-kolla | 03:35 | |
duonghq | portdirect, what do you think? | 03:35 |
v1k0d3n | (no offense to anyone who loves the chevy cavalier ... its a great car) | 03:35 |
yingjun | Thanks duonghq, any link for Ryan's spec? | 03:35 |
v1k0d3n | so inc0...you're talking to a major car dude here... | 03:35 |
duonghq | portdirect, it is late at your location, is not it? | 03:36 |
portdirect | duonghq: It almost early - I'm on a totally bacward scedule atm... | 03:36 |
v1k0d3n | https://farm8.staticflickr.com/7177/13859552003_c624ca1fa0_b.jpg | 03:37 |
v1k0d3n | that's me buddy :) | 03:37 |
duonghq | yingjun, in the source code, specs/kolla-kuberenets-arch.... | 03:37 |
srwilkers_ | yingjun, https://github.com/openstack/kolla-kubernetes/blob/master/specs/kolla-kubernetes-arch.rst | 03:37 |
inc0 | v1k0d3n, I'm not huge car dude myself, but when I be rich and have garage I'm buying ~2000 impreza STI and set rollcage in it | 03:37 |
yingjun | srwilkers_, thanks | 03:37 |
v1k0d3n | heck yeah inc0!!! | 03:37 |
v1k0d3n | do you autocross or track?! | 03:37 |
v1k0d3n | you have to with that comment | 03:37 |
inc0 | never tried, always wanted | 03:38 |
v1k0d3n | DO IT! | 03:38 |
v1k0d3n | i used to race...big time | 03:38 |
v1k0d3n | motorcycles and cars | 03:38 |
inc0 | I always loved rally | 03:38 |
inc0 | I'm eurpean mind you | 03:38 |
inc0 | so autocross would fit just fine | 03:38 |
v1k0d3n | i had a 2011 STI hatch and a 2007 STi sadan | 03:38 |
v1k0d3n | sedan | 03:38 |
inc0 | alternatively turpocharged older miata :# | 03:38 |
inc0 | :3 | 03:38 |
v1k0d3n | yup...euros love their cars | 03:39 |
inc0 | ahh, fellow subaru fan | 03:39 |
inc0 | there always will be subaru in my garage | 03:39 |
inc0 | now I have Crosstrek | 03:39 |
v1k0d3n | https://farm5.staticflickr.com/4095/4916713968_1ed7406f17_m.jpg | 03:39 |
v1k0d3n | that's me | 03:39 |
inc0 | and I'm going to drive through half of US with it next week;) | 03:39 |
v1k0d3n | https://farm6.staticflickr.com/5232/5910426942_640505a4a7_b.jpg | 03:39 |
v1k0d3n | that's me | 03:39 |
portdirect | nice parking. | 03:40 |
v1k0d3n | total tool i know | 03:40 |
duonghq | v1k0d3n, and portdirect in same tz? | 03:40 |
v1k0d3n | actually it was dusk in a business park...everyone was gone. | 03:40 |
v1k0d3n | duonghq: different | 03:41 |
v1k0d3n | i'm EST | 03:41 |
portdirect | and you took the bait :D | 03:41 |
v1k0d3n | -5 | 03:41 |
portdirect | im gmt | 03:41 |
v1k0d3n | so true | 03:41 |
duonghq | ah, ok | 03:41 |
v1k0d3n | duonghq: which TZ are you? | 03:41 |
srwilkers_ | i wont tell you what i drive -- its about as bad as a cavalier | 03:41 |
duonghq | v1k0d3n, I'm in UTC+7 | 03:42 |
v1k0d3n | ah. makes sense why morning was just a few hours ago :) | 03:42 |
duonghq | ya | 03:42 |
v1k0d3n | srwilkers_: don't need a cool car to be a cool dude | 03:42 |
v1k0d3n | :) | 03:42 |
v1k0d3n | i'm old | 03:42 |
v1k0d3n | i would do anything for youth :) | 03:42 |
inc0 | v1k0d3n, if you havent tried new miata, try it | 03:44 |
inc0 | this stick shift is magnificent | 03:44 |
inc0 | I think best manual transmission I've ever tried | 03:45 |
portdirect | srwilkers_: I'll have a look at the aic helm stuff and see what we can do with it - at the stage your at it would give us a real boost i think? you have keystone running? | 03:45 |
inc0 | and it handles like gocart | 03:45 |
*** yuanying has quit IRC | 03:46 | |
inc0 | ok, I'm off | 03:47 |
inc0 | farewell my dear friends, may night be soft as velvet | 03:47 |
portdirect | laters inc0 | 03:48 |
v1k0d3n | portdirect: yes, i think it will help a lot. keystone should be running, but we're taking our time with each component according to our mission....utopia. | 03:48 |
v1k0d3n | keystone is last in our current PoC...so it will be fully ready very soon. | 03:49 |
v1k0d3n | from there, things should be much easier (with exceptions to nova and neutron...obviously). | 03:49 |
portdirect | v1k0d3n: nova and neutron, ain't half as hard as people make out | 03:50 |
portdirect | designate - thats a real nightmare in k8s | 03:50 |
v1k0d3n | not to get working. but to make it completely independent, yet still use of the benefits of a global deploy yaml...AND add other things like gluon (which we're huge backers of, users of, and contributors to)...that's the challenge. | 03:51 |
v1k0d3n | portdirect eh....designate... | 03:52 |
v1k0d3n | new | 03:52 |
v1k0d3n | painful still | 03:52 |
*** inc0 has quit IRC | 03:52 | |
portdirect | v1k0d3n: fair points! are you using onos at all? | 03:53 |
v1k0d3n | portdirect: ha! that's bryan sullivan's pet. most definitely use onos. | 03:54 |
v1k0d3n | Bryan is near to our team | 03:54 |
v1k0d3n | we support him | 03:54 |
v1k0d3n | and that team | 03:55 |
portdirect | nice I've been keeping an eye on it - looks like really good stuff brewing over there | 03:55 |
v1k0d3n | glad to hear that | 03:56 |
v1k0d3n | sounds like you're tracking a lot of things we're involved with. | 03:56 |
portdirect | well it seems like your all over the interesting stuff :) | 03:57 |
v1k0d3n | we try. you know...we have this network to keep running. | 03:58 |
v1k0d3n | cell calls, tv...IoT now. making my head spin. | 03:59 |
*** sayantan_ has quit IRC | 03:59 | |
*** Pavo has quit IRC | 03:59 | |
v1k0d3n | a'ight guys. i'm out. tv time with the wife :) | 04:00 |
srwilkers_ | ++ | 04:00 |
v1k0d3n | srwilkers_ and portdirect see what works to get into review/garrit. i'll be around if you need anythng. | 04:00 |
v1k0d3n | always ready to help. | 04:00 |
srwilkers_ | v1k0d3n, sounds good | 04:01 |
v1k0d3n | and portdirect thanks for the help with halcyon. you have stuff going into that soon, i guess? | 04:01 |
portdirect | v1k0d3n: have a good 'un | 04:01 |
v1k0d3n | yeah man, hope you feel better soon portdirect. take some meds and rest up :) | 04:01 |
portdirect | yeah - I'll be ready for you in the morning | 04:01 |
*** hieulq_ has quit IRC | 04:01 | |
portdirect | *It'll | 04:02 |
v1k0d3n | cool later all | 04:02 |
*** Pavo has joined #openstack-kolla | 04:04 | |
openstackgerrit | zhubingbing proposed openstack/kolla: Install trove dashboard into horizon image https://review.openstack.org/405087 | 04:04 |
jascott1 | hey portdirect wanna rap about setup_dev.sh ? | 04:04 |
*** srwilkers_ is now known as srwilkers_zzzz | 04:04 | |
portdirect | jascott1: I could spit some bars | 04:05 |
jascott1 | :) | 04:05 |
jascott1 | prob a bad name for the script acutally | 04:05 |
jascott1 | did you have a start already? | 04:05 |
portdirect | what do you want from it? we can pick a name later | 04:06 |
jascott1 | complete the local dev setup | 04:06 |
jascott1 | I can run this and get everything I need to do a developers helm install <service> | 04:07 |
jascott1 | developer grade i guess I should say | 04:07 |
jascott1 | i have been looking at the setup_gate | 04:07 |
sbezverk | portdirect: hey man, are you almost done with vagrant? | 04:08 |
jascott1 | basically dissecting it into something that might work but just started | 04:08 |
jascott1 | otherwise helm install openstack is a lie! | 04:08 |
*** sayantan_ has joined #openstack-kolla | 04:08 | |
openstackgerrit | zhubingbing proposed openstack/kolla-ansible: Enable trove dashboard when enable_senlin is yes https://review.openstack.org/405090 | 04:09 |
jascott1 | or rather *just* requiring that have openstack is not the complete story :) | 04:09 |
portdirect | sbezverk: yeah, though I'll need an hour to wrap it up rather that chat on iRC :) | 04:09 |
*** hieulq_ has joined #openstack-kolla | 04:10 | |
sbezverk | portdirect: I was more curious where on your to do list keystone chart ;-) ?? | 04:10 |
jascott1 | oh sorry we can chat later if you are interested. you expressed some interest in the meeting so thought you might have some prior art | 04:10 |
portdirect | sbezverk: the changes th this stage will be iminila to whats on the docs ive got in pr | 04:10 |
openstackgerrit | zhubingbing proposed openstack/kolla: Install trove dashboard into horizon image https://review.openstack.org/405087 | 04:10 |
portdirect | jascott1: no this is important, lets do it now | 04:10 |
portdirect | sbezverk: so if you got going with that it should be fine? | 04:11 |
jascott1 | i really just have the desire am not currently loaded with opinions about its construction | 04:11 |
*** mdnadeem has joined #openstack-kolla | 04:11 | |
sbezverk | portdirect: it can wait probably dau but then we need it rolling, since it is at the center of everything | 04:12 |
portdirect | sbezverk: roger | 04:13 |
jascott1 | 1. haylcon vagrant 2. setup_kk8s_dev.sh 3. helm install openstack (or component thereof) | 04:14 |
portdirect | ok going through them: | 04:15 |
portdirect | 1 should be only covered in documentation - as everyone's dev machine will be starting from a diffeent place (unless you are up for making a script that sets up vagrant in a safe manner accross platforms) | 04:16 |
jascott1 | agree | 04:16 |
portdirect | 3 is where we need to be | 04:16 |
portdirect | so... | 04:16 |
portdirect | 2 | 04:16 |
jascott1 | oh i just want to create 2 | 04:16 |
jascott1 | but those could be quickstart steps to kk8s dev bliss | 04:16 |
portdirect | +1 | 04:17 |
*** hieulq_ has quit IRC | 04:17 | |
jascott1 | ya totally do not think we need a giant wrapper or anythin | 04:17 |
portdirect | ok - I think I'm in the best ostion to implement it - but would apprecate going ove the scaope of what it should do | 04:18 |
portdirect | pull repos, setup clients, pip install requiremnts | 04:18 |
portdirect | what else? | 04:18 |
jascott1 | yup | 04:18 |
jascott1 | well the gate has ceph scripts already :) | 04:18 |
jascott1 | even those steps you listed would be a start | 04:19 |
*** hieulq_ has joined #openstack-kolla | 04:19 | |
jascott1 | the docs are outdated or something cause I have struggled with the install and setup_gate was really helpful | 04:20 |
portdirect | ah ok - we dont yet have kolla-k8s ceph in helm but it should also setup a base env and exand as those servcices come online | 04:20 |
*** absubram has joined #openstack-kolla | 04:21 | |
jascott1 | we might not install ceph but refactor setup_ceph (if necesary) so that developer can use it until we land ceph | 04:21 |
*** duonghq has quit IRC | 04:21 | |
jascott1 | unless it will be very soon | 04:21 |
jascott1 | maybe we adapt the aic work | 04:21 |
portdirect | jascott1: just wanna double check, you Justin Scott? | 04:22 |
jascott1 | remains to be seen but the gate already works so seems foable | 04:22 |
jascott1 | I am sir! | 04:22 |
jascott1 | or madam | 04:22 |
jascott1 | :) | 04:22 |
portdirect | only at the weekends :P | 04:22 |
jascott1 | lol | 04:22 |
jascott1 | whereabouts are you located? sorry I may have asked before | 04:22 |
jascott1 | :D | 04:23 |
portdirect | I'd be up for adopting the AIC work - but that needs to be cleared by people other than me | 04:23 |
portdirect | Scotland | 04:23 |
jascott1 | woot! | 04:23 |
portdirect | and you? | 04:23 |
*** sayantan_ has quit IRC | 04:23 | |
jascott1 | Portland OR | 04:23 |
portdirect | nice | 04:23 |
jascott1 | well the Intel suburb of it anyway | 04:23 |
*** sayantan_ has joined #openstack-kolla | 04:23 | |
jascott1 | it really is | 04:24 |
jascott1 | nice | 04:24 |
jascott1 | like Linus said (who lives here also) ~"I dont ever go out but when I do its great" | 04:24 |
portdirect | anyway better get on with getting vargant up - dont want to start being a blocker - I'll also get a framwork for the setup_kolla_dev.sh in | 04:25 |
jascott1 | I will test and contribute! | 04:25 |
*** absubram_ has joined #openstack-kolla | 04:25 | |
portdirect | Now I know how to find you on gerrit you wont have any choice :) | 04:26 |
jascott1 | same nick | 04:26 |
*** absubram has quit IRC | 04:26 | |
*** absubram_ is now known as absubram | 04:26 | |
jascott1 | oh haha | 04:26 |
jascott1 | cool man thanks | 04:26 |
*** jgriffith is now known as jgriffith_away | 04:28 | |
*** JRobinson__ is now known as JRobinson__afk | 04:33 | |
*** rkrum has quit IRC | 04:35 | |
*** saneax is now known as saneax-_-|AFK | 04:36 | |
*** tonanhngo has joined #openstack-kolla | 04:37 | |
*** duonghq has joined #openstack-kolla | 04:39 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Hel-ming openvswitch-vswitchd https://review.openstack.org/405107 | 04:45 |
*** yuanying has joined #openstack-kolla | 04:48 | |
*** yuanying_ has joined #openstack-kolla | 04:55 | |
sayantan_ | Hi | 04:56 |
sayantan_ | I have been working on documentation. | 04:56 |
sayantan_ | I have a question. Is it advisable to upload the entire doc or should I upload it in different patch sets | 04:57 |
*** yuanying has quit IRC | 04:57 | |
*** yuanying has joined #openstack-kolla | 05:02 | |
*** bjolo_ has quit IRC | 05:03 | |
*** yuanying_ has quit IRC | 05:04 | |
*** yuanying_ has joined #openstack-kolla | 05:10 | |
*** JRobinson__afk is now known as JRobinson__ | 05:10 | |
*** yuanying has quit IRC | 05:11 | |
*** sayantan_ has quit IRC | 05:12 | |
openstackgerrit | Hui Kang proposed openstack/kolla: Pin requests to 2.12.1 https://review.openstack.org/405112 | 05:25 |
*** jrich523_ has joined #openstack-kolla | 05:35 | |
*** jrich523 has quit IRC | 05:38 | |
*** wwriverrat has quit IRC | 05:41 | |
*** msimonin has joined #openstack-kolla | 05:50 | |
*** duonghq has quit IRC | 05:59 | |
*** Pavo has quit IRC | 05:59 | |
*** duonghq has joined #openstack-kolla | 06:00 | |
*** Pavo has joined #openstack-kolla | 06:04 | |
*** ram___ has quit IRC | 06:04 | |
*** ram__ has joined #openstack-kolla | 06:05 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: TEST_NO_MERGE https://review.openstack.org/403978 | 06:07 |
*** kristian__ has joined #openstack-kolla | 06:08 | |
*** kristian__ has quit IRC | 06:12 | |
*** yuanying_ has quit IRC | 06:18 | |
ram__ | Hi. I am doing Kolla All-in-one deployment. I ran $sudo kolla-ansible pull then I was getting authentication issue for kolla-toolbox. How can I resolve this. How do I provide authentication . could anyone please help me in this. Issue Info : http://paste.openstack.org/show/591057/ | 06:21 |
Jeffrey4l_ | ram__, which docker registry are u using to pull? | 06:26 |
Jeffrey4l_ | seems you haven't permission to this registry FAILED! => {"changed": false, "failed": true, "msg": "Unknown error message: unauthorized: authentication required"} | 06:26 |
*** msimonin has quit IRC | 06:28 | |
*** msimonin has joined #openstack-kolla | 06:28 | |
*** msimonin has quit IRC | 06:28 | |
openstackgerrit | jianyi proposed openstack/kolla: modified code alphabetical order https://review.openstack.org/405121 | 06:29 |
*** unicell has joined #openstack-kolla | 06:30 | |
*** hieulq_ has quit IRC | 06:31 | |
*** hieulq_ has joined #openstack-kolla | 06:31 | |
*** jtriley has quit IRC | 06:33 | |
*** unicell has quit IRC | 06:34 | |
ram__ | Jeffrey4l_: Hi. I think I am not using any docker registry. In globals.yml also I didn't edited. In doc they gave like optional. Am I need to use registry to get authentication? | 06:35 |
ram__ | Jeffrey4l_ : I am new to this. Don't mind. Please tell me if I am doing anything wrong. thanks. | 06:36 |
Jeffrey4l_ | ram__, the kolla built images on your node already? | 06:36 |
ram__ | Jeffrey4l_: Yes. kolla build images. | 06:37 |
Jeffrey4l_ | ram__, run deploy with -vvvv parameter. it will tell us more info ;) | 06:37 |
ram__ | Jeffrey4l_: OK. | 06:37 |
*** senk has joined #openstack-kolla | 06:39 | |
*** kristian__ has joined #openstack-kolla | 06:43 | |
*** kristian__ has quit IRC | 06:44 | |
*** saneax-_-|AFK is now known as saneax | 06:45 | |
*** kristian__ has joined #openstack-kolla | 06:45 | |
*** kristia__ has joined #openstack-kolla | 06:47 | |
*** kristian__ has quit IRC | 06:50 | |
ram__ | Jeffrey41_: Hi. I ran $ sudo kolla-ansible pull -vvv. This time it crossed pulling kolla-toolbox step and it is trying to pull all other images. Thanks . | 06:55 |
ram__ | Jeffrey41_: Hi. Before running kolla-deploy ansible I want to enable magnum service in globals.yml to deploy that service also. Can I do that. It won't cause any problems right. | 06:56 |
*** kristia__ has quit IRC | 06:56 | |
*** kristian__ has joined #openstack-kolla | 06:57 | |
Jeffrey4l_ | ram__, yes. you can. | 06:57 |
*** kristian__ has quit IRC | 07:00 | |
SamYaple | hey Jeffrey4l_ | 07:01 |
Jeffrey4l_ | hi SamYaple ;) | 07:01 |
ram__ | Jeffrey41_: Even kuryr service also right? When I ran kolla-build , it build all docker images including magnum,kuryr. | 07:01 |
Jeffrey4l_ | ram__, are u talking about build or deploy? | 07:02 |
Jeffrey4l_ | in default when running kolla-build, it builds all images. | 07:02 |
ram__ | Jeffrey41_: sry for confusion. I am talking about deploy. | 07:03 |
*** msimonin has joined #openstack-kolla | 07:03 | |
ram__ | Jeffrey41_: In globals.yml : If enable magnum and kuryr services that is enough right. Remaining all openstack service automatically deploy after I ran kolla-deploy without enabling them right? | 07:05 |
Jeffrey4l_ | ram__, yep. some service is enable in default. you can check the globals.yml carefully. some service is using enable_*:yes | 07:06 |
*** zhurong__ has joined #openstack-kolla | 07:07 | |
ram__ | Jeffrey41_: Thanks. i will look at that carefully. | 07:09 |
Jeffrey4l_ | np ;) | 07:09 |
openstackgerrit | zhubingbing proposed openstack/kolla: Fix fluentd dockerfile * move repo to basefile https://review.openstack.org/405132 | 07:10 |
*** zhurong has quit IRC | 07:10 | |
openstackgerrit | zhubingbing proposed openstack/kolla: Fix fluentd dockerfile https://review.openstack.org/405132 | 07:12 |
*** hieulq_ has quit IRC | 07:13 | |
*** tonanhngo has quit IRC | 07:25 | |
*** tonanhngo has joined #openstack-kolla | 07:26 | |
*** williamcaban has quit IRC | 07:28 | |
*** kristian__ has joined #openstack-kolla | 07:31 | |
*** JRobinson__ has quit IRC | 07:32 | |
*** kristia__ has joined #openstack-kolla | 07:32 | |
*** mkoderer has joined #openstack-kolla | 07:35 | |
*** kristian__ has quit IRC | 07:35 | |
openstackgerrit | zhubingbing proposed openstack/kolla-ansible: Move watcher precheck into its own role https://review.openstack.org/404468 | 07:36 |
*** hieulq_ has joined #openstack-kolla | 07:36 | |
jascott1 | review.openstack.org is down | 07:39 |
jascott1 | for maintenance. must've missed the memo | 07:40 |
jascott1 | its back. woah those guys are fast | 07:41 |
*** unicell has joined #openstack-kolla | 07:42 | |
*** kristia__ has quit IRC | 07:48 | |
*** kristian__ has joined #openstack-kolla | 07:49 | |
*** kristian__ has quit IRC | 07:53 | |
*** Serlex has joined #openstack-kolla | 07:55 | |
magicboiz | morning | 07:56 |
*** jtriley has joined #openstack-kolla | 07:56 | |
openstackgerrit | jianyi proposed openstack/kolla: Install senlin dashboard into horizon image https://review.openstack.org/405070 | 07:57 |
magicboiz | Can I ask someone about some heat-api problems?? Basically, I make REQ to heat-api asking for specific stacks/resources through the publicURL, but heat-api replies with incorrect publicURL location... | 07:57 |
*** unicell has quit IRC | 07:59 | |
*** Pavo has quit IRC | 07:59 | |
*** Pavo has joined #openstack-kolla | 08:04 | |
*** jtriley has quit IRC | 08:05 | |
*** slagle has quit IRC | 08:05 | |
*** psanchez has quit IRC | 08:06 | |
*** stevebaker has quit IRC | 08:07 | |
*** matrohon has joined #openstack-kolla | 08:08 | |
openstackgerrit | jianyi proposed openstack/kolla: Install senlin dashboard into horizon image https://review.openstack.org/405070 | 08:09 |
*** kristian__ has joined #openstack-kolla | 08:09 | |
*** kristian__ has quit IRC | 08:14 | |
openstackgerrit | Javier Castillo Alcíbar proposed openstack/kolla-ansible: Fix ceilometer bootstrap https://review.openstack.org/398190 | 08:20 |
*** jascott1 has quit IRC | 08:24 | |
*** jascott1 has joined #openstack-kolla | 08:25 | |
openstackgerrit | Javier Castillo Alcíbar proposed openstack/kolla-ansible: Make sure heka-ceilometer.toml manages correctly apache and openstack ceilometer logs https://review.openstack.org/400716 | 08:26 |
*** wznoinsk has quit IRC | 08:26 | |
*** Jeffrey4l_ has quit IRC | 08:26 | |
*** yuanying has joined #openstack-kolla | 08:27 | |
*** williamcaban has joined #openstack-kolla | 08:28 | |
*** pprokop has quit IRC | 08:28 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: TEST_NO_MERGE https://review.openstack.org/403978 | 08:29 |
*** shardy has joined #openstack-kolla | 08:29 | |
*** pprokop has joined #openstack-kolla | 08:29 | |
*** kristian__ has joined #openstack-kolla | 08:30 | |
*** Jeffrey4l_ has joined #openstack-kolla | 08:30 | |
*** wznoinsk has joined #openstack-kolla | 08:30 | |
*** Mech422 has quit IRC | 08:32 | |
*** Mech422 has joined #openstack-kolla | 08:33 | |
openstackgerrit | jianyi proposed openstack/kolla: Install senlin dashboard into horizon image https://review.openstack.org/405070 | 08:35 |
*** Guest66666 has quit IRC | 08:36 | |
*** egonzalez90 has joined #openstack-kolla | 08:36 | |
*** Guest66666 has joined #openstack-kolla | 08:38 | |
*** matrohon has quit IRC | 08:39 | |
*** kristian__ has quit IRC | 08:43 | |
ram__ | HI. i am getting an issue while running kolla-deploy. Issue detail: http://paste.openstack.org/show/591067/. I was getting the issue while initializing kolla-toolbox container | 08:43 |
*** yatin has joined #openstack-kolla | 08:44 | |
egonzalez90 | ram__: what docker logs kolla_toolbox says | 08:49 |
ram__ | egonzalez90:"Error response from daemon: Container e78ca4ec68429615df2964575d06ac425f9c26cc4c14f2523f1b2d856cd6705b is restarting, wait until the container is running" | 08:50 |
ram__ | egonzalez90: Hi. Koola log showing like that. | 08:51 |
*** hieulq_ has quit IRC | 08:51 | |
*** rmart04 has joined #openstack-kolla | 08:52 | |
*** kristian__ has joined #openstack-kolla | 08:52 | |
*** kristia__ has joined #openstack-kolla | 08:53 | |
ram__ | egonzalez90: Hi. I removed Container e78ca4ec68429615df2964575d06ac425f9c26cc4c14f2523f1b2d856cd6705b and again run $sudo kolla-deploy ansible. This it was not showing the previous issue but it was failing at "Initializing the kolla-toolbox container only. | 08:54 |
*** kristi___ has joined #openstack-kolla | 08:55 | |
*** kristian__ has quit IRC | 08:55 | |
ram__ | egonzalez90:This time I am getting the following issue :http://paste.openstack.org/show/591068/ | 08:57 |
*** kristian__ has joined #openstack-kolla | 08:57 | |
*** kristia__ has quit IRC | 08:58 | |
*** williamcaban has quit IRC | 08:59 | |
*** kristi___ has quit IRC | 09:00 | |
*** msimonin has quit IRC | 09:01 | |
*** klindgren has quit IRC | 09:03 | |
egonzalez90 | ram__: whats your ansible version? | 09:04 |
ram__ | egonzalez90: ansible 2.2.0.0 | 09:05 |
jascott1 | ram__ can you run that docker exec -t kolla_toolbox /usr/bin/ansible --version ? | 09:07 |
ram__ | jascott1: I ran. It was again giving Error response from daemon: Container 028ec90cd874a76a9ceb6bae4fc1033b47e4926ea601094f9e6ed70947383e5b is restarting, wait until the container is running | 09:08 |
egonzalez90 | docker logs kolla_toolbox to see whats happening in the container | 09:09 |
jascott1 | if you need to inspect the container, change the command in the resource yaml to /sbin/init and the container can start. but do check logs first | 09:10 |
*** hogepodge has quit IRC | 09:11 | |
jascott1 | oh sorry got k8s on the brain | 09:11 |
*** hogepodge has joined #openstack-kolla | 09:12 | |
*** matrohon has joined #openstack-kolla | 09:12 | |
jascott1 | can do docker run -ti kolla_toolbox /sbin/init | 09:13 |
jascott1 | oops | 09:13 |
jascott1 | run -tid | 09:13 |
jascott1 | then docker exec -ti kolla_toolbox /bin/bash | 09:13 |
jascott1 | if you want to look around in the container | 09:14 |
*** mewald has joined #openstack-kolla | 09:14 | |
ram__ | jascott1: Ok | 09:14 |
*** msimonin has joined #openstack-kolla | 09:16 | |
ram__ | egonzalez90: $sudo docker logs kolla_toolbox http://paste.openstack.org/show/591075/ | 09:18 |
egonzalez90 | selinux enabled? | 09:19 |
openstackgerrit | Duong Ha-Quang proposed openstack/kolla-ansible: Implement Keystone zero-downtime upgrade https://review.openstack.org/405215 | 09:20 |
*** hieulq_ has joined #openstack-kolla | 09:21 | |
*** igordcard has quit IRC | 09:21 | |
*** igordcard has joined #openstack-kolla | 09:23 | |
duonghq | egonzalez90, ping | 09:24 |
egonzalez90 | duonghq: pong | 09:24 |
duonghq | egonzalez90, can you get it merged? https://review.openstack.org/#/c/404092/ | 09:25 |
duonghq | the policy.json one | 09:25 |
ram__ | jascott1: sudo docker run -tid kolla_toolbox /sbin/init shows as http://paste.openstack.org/show/591077/ | 09:25 |
egonzalez90 | ram__: add the docker tag | 09:25 |
jascott1 | docker images | 09:25 |
jascott1 | to list images | 09:26 |
egonzalez90 | kolla_toolbox:3.0.2 or whatever version | 09:26 |
ram__ | egonzalez90: selinux is in disabled mode | 09:26 |
openstackgerrit | Duong Ha-Quang proposed openstack/kolla: Implement Keystone zero-downtime upgrade https://review.openstack.org/388544 | 09:27 |
jascott1 | gotta bounce. good luck! | 09:27 |
ram__ | egonzalez90 : kolla/ubuntu-source-kolla-toolbox 3.0.1 18a965c0e7c5 4 weeks ago 760.7 MB. docker tag is 3.0.1 | 09:28 |
openstackgerrit | zhubingbing proposed openstack/kolla-ansible: Enable trove dashboard when enable_trove is yes https://review.openstack.org/405090 | 09:28 |
egonzalez90 | sudo docker run -tid kolla_toolbox:3.0.1 /sbin/init | 09:29 |
*** yingjun has quit IRC | 09:29 | |
egonzalez90 | ram__: as far I can find, is this ulimit issue https://github.com/docker/docker/issues/11740 | 09:29 |
*** jev has joined #openstack-kolla | 09:30 | |
ram__ | egonzalez90: I ran $ sudo docker run -tid kolla_toolbox:3.0.1 /sbin/init. Unable to find image 'kolla_toolbox:3.0.1' locally. trying to pull from repo. but unable to find that image. | 09:30 |
egonzalez90 | ram__: tried to face the same issue but cannot reproduce it | 09:30 |
egonzalez90 | ram__: sorry missed repository kolla/ubuntu-source-kolla-toolbox:3.0.1 | 09:31 |
*** hogepodge has quit IRC | 09:31 | |
berendt | morning everybody | 09:32 |
*** hogepodge has joined #openstack-kolla | 09:32 | |
egonzalez90 | morning berendt | 09:32 |
ram__ | egonzalez90: What I need to do in this case to resolve the issue. | 09:32 |
*** mariso has joined #openstack-kolla | 09:32 | |
openstackgerrit | zhubingbing proposed openstack/kolla: Fix fluentd dockerfile https://review.openstack.org/405132 | 09:32 |
duonghq | thank you egonzalez90 | 09:34 |
egonzalez90 | np | 09:34 |
*** mewald has quit IRC | 09:34 | |
ram__ | egonzalez90: how do I get that missing repository kolla/ubuntu-source-kolla-toolbox:3.0.1 ? | 09:35 |
openstackgerrit | Merged openstack/kolla-ansible: Fix placement of policy.json https://review.openstack.org/404092 | 09:37 |
*** yuanying_ has joined #openstack-kolla | 09:38 | |
*** yuanying has quit IRC | 09:41 | |
*** hfu has quit IRC | 09:44 | |
ram__ | egonzalez90: Hi. Are you there? | 09:51 |
mariso | Is it possible with current sources from Git (kolla, kolla-kubernetes, kube-deploy) deploy kolla-kubernetes? | 09:53 |
*** openstack has joined #openstack-kolla | 10:03 | |
ram__ | 1 | 10:03 |
*** harbor has quit IRC | 10:04 | |
*** Pavo has joined #openstack-kolla | 10:04 | |
*** jtriley has quit IRC | 10:06 | |
*** matrohon has quit IRC | 10:11 | |
egonzalez90 | ram__: yes | 10:15 |
*** Satya_ has joined #openstack-kolla | 10:20 | |
Jeffrey4l_ | egonzalez90, please review https://review.openstack.org/404844 ;) | 10:20 |
*** matrohon has joined #openstack-kolla | 10:23 | |
egonzalez90 | Jeffrey4l_: is that what broke our gates or different issue? | 10:24 |
Jeffrey4l_ | egonzalez90, it is a upstream issue which block our gate. | 10:25 |
Jeffrey4l_ | this is a workaround | 10:25 |
egonzalez90 | Jeffrey4l_: thanks | 10:25 |
Jeffrey4l_ | maybe i need add some NOTE for this? | 10:25 |
*** tonanhngo has quit IRC | 10:26 | |
egonzalez90 | umm, yes so this will be reverted in a future | 10:26 |
*** kristian__ has quit IRC | 10:27 | |
egonzalez90 | Jeffrey4l_: have you ever seen this error from ram__ ? http://paste.openstack.org/show/591075/ | 10:27 |
Jeffrey4l_ | hrm, i got another idea. during install ansible, we need restrict the package version from requirements fproject. | 10:27 |
Jeffrey4l_ | egonzalez90, yep. Saw this before. | 10:28 |
Jeffrey4l_ | which host OS are u using? and docker storage driver ram__ | 10:28 |
Jeffrey4l_ | seem like this issue https://github.com/docker/docker/issues/5899 egonzalez90 ram__ | 10:29 |
Jeffrey4l_ | ram__, kernel version? | 10:30 |
*** tonanhngo has joined #openstack-kolla | 10:32 | |
Serlex | hi, there is no limitation to how many internal networks I can attach to an active instance in kolla? | 10:34 |
*** tonanhngo has quit IRC | 10:34 | |
Jeffrey4l_ | Serlex, no. it is not related to kolla. | 10:35 |
ram__ | Jeffrey41_: Hi. I am using Ubuntu 14.04. kernal version: Linux kolla 4.2.0-42-generic #49~14.04.1-Ubuntu SMP Wed Jun 29 20:22:11 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux | 10:36 |
*** kristian__ has joined #openstack-kolla | 10:36 | |
*** zhubingbing_ has joined #openstack-kolla | 10:37 | |
Jeffrey4l_ | hrm, sound another issue. it is report the issue i post above is fixed on kernel 3.15. | 10:37 |
*** kristian__ has quit IRC | 10:38 | |
Jeffrey4l_ | ram__, post your docker info? | 10:38 |
*** zhubingbing has quit IRC | 10:38 | |
*** kristian__ has joined #openstack-kolla | 10:38 | |
*** cmceniry has quit IRC | 10:39 | |
*** smekel_ has quit IRC | 10:39 | |
*** tonyb has quit IRC | 10:39 | |
ram__ | Jeffrey41_: my docker info http://paste.openstack.org/show/591097/ | 10:39 |
Jeffrey4l_ | ram__, sorry no idea. it should be bug in docker ;( | 10:41 |
Jeffrey4l_ | ram__, try ubuntu 16.04 or centos if u can ;) | 10:41 |
Serlex | ok thanks Jeffrey4l_...im struggling | 10:41 |
Jeffrey4l_ | Serlex, try to find the error message from log file. | 10:42 |
ram__ | Jeffrey4l_: OK. Thanks. | 10:42 |
*** kristian__ has quit IRC | 10:43 | |
*** cmceniry has joined #openstack-kolla | 10:43 | |
*** smekel_ has joined #openstack-kolla | 10:43 | |
*** tonyb has joined #openstack-kolla | 10:43 | |
*** eaguilar has joined #openstack-kolla | 10:44 | |
Serlex | "InterfaceAttachFailed: Failed to attach network adapter device to" either I'm doing something I shouldn't or running a broken setup or it just doesn't work | 10:45 |
*** tonanhngo has joined #openstack-kolla | 10:45 | |
*** tonanhngo has quit IRC | 10:47 | |
*** jascott1 has quit IRC | 10:51 | |
*** yingjun has joined #openstack-kolla | 10:52 | |
*** matrohon has quit IRC | 10:54 | |
*** ipsecguy has quit IRC | 10:54 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Install ansible with OpenStack requirements constrains https://review.openstack.org/405273 | 10:54 |
*** ipsecguy has joined #openstack-kolla | 10:54 | |
*** williamcaban has joined #openstack-kolla | 10:56 | |
*** tonanhngo has joined #openstack-kolla | 10:58 | |
*** yingjun has quit IRC | 11:00 | |
*** tonanhngo has quit IRC | 11:00 | |
openstackgerrit | zhubingbing proposed openstack/kolla-ansible: Fix aodh role https://review.openstack.org/405275 | 11:00 |
*** yingjun has joined #openstack-kolla | 11:00 | |
*** williamcaban has quit IRC | 11:01 | |
*** zhubingbing_ has quit IRC | 11:01 | |
*** jtriley has joined #openstack-kolla | 11:02 | |
*** yingjun has quit IRC | 11:04 | |
*** matrohon has joined #openstack-kolla | 11:06 | |
*** tonanhngo has joined #openstack-kolla | 11:06 | |
*** jtriley has quit IRC | 11:07 | |
*** tonanhngo has quit IRC | 11:07 | |
mariso | Which versions of kolla, kolla-kubernetes, docker, docker-py should be used to deploy kolla with kubernetes? | 11:14 |
*** msimonin has quit IRC | 11:26 | |
*** sean-k-mooney has joined #openstack-kolla | 11:28 | |
*** kristian__ has joined #openstack-kolla | 11:29 | |
*** msimonin has joined #openstack-kolla | 11:30 | |
mliima | morning guys | 11:33 |
*** athomas has joined #openstack-kolla | 11:34 | |
*** eaguilar has quit IRC | 11:36 | |
mliima | need some act here https://review.openstack.org/#/c/403689/ | 11:42 |
*** yingjun has joined #openstack-kolla | 11:51 | |
*** kristian__ has quit IRC | 11:51 | |
*** kristian__ has joined #openstack-kolla | 11:52 | |
*** yingjun has quit IRC | 11:54 | |
*** yingjun has joined #openstack-kolla | 11:54 | |
*** igordcard has quit IRC | 11:55 | |
*** jtriley has joined #openstack-kolla | 11:56 | |
*** williamcaban has joined #openstack-kolla | 11:58 | |
*** yingjun has quit IRC | 11:59 | |
*** Pavo has quit IRC | 11:59 | |
*** awidders_ has quit IRC | 11:59 | |
*** jtriley has quit IRC | 12:01 | |
*** williamcaban has quit IRC | 12:02 | |
*** Pavo has joined #openstack-kolla | 12:04 | |
*** tonanhngo has joined #openstack-kolla | 12:04 | |
*** tonanhngo has quit IRC | 12:05 | |
*** srwilkers_zzzz has quit IRC | 12:08 | |
*** slagle has joined #openstack-kolla | 12:10 | |
*** kristia__ has joined #openstack-kolla | 12:12 | |
*** kristia__ has quit IRC | 12:14 | |
*** kristia__ has joined #openstack-kolla | 12:14 | |
*** kristian__ has quit IRC | 12:16 | |
*** mewald has joined #openstack-kolla | 12:17 | |
*** tonanhngo has joined #openstack-kolla | 12:18 | |
*** kristia__ has quit IRC | 12:18 | |
*** tonanhngo has quit IRC | 12:18 | |
*** hieulq_ has quit IRC | 12:22 | |
*** klindgren has joined #openstack-kolla | 12:27 | |
*** athomas has quit IRC | 12:28 | |
*** strigazi_AFK is now known as strigazi | 12:34 | |
*** tonanhngo has joined #openstack-kolla | 12:37 | |
*** hieulq has joined #openstack-kolla | 12:38 | |
*** tonanhngo has quit IRC | 12:38 | |
pbourke_ | Jeffrey4l_: you tried magnum-ui in horizon yet by any chance? | 12:45 |
Jeffrey4l_ | pbourke_, sorry, not tried. | 12:46 |
pbourke_ | Jeffrey4l_: np, I have a patch for it to submit but I think the plugin itself might be broken | 12:46 |
Jeffrey4l_ | the plugin mech? or just magnum dashboard? | 12:46 |
Jeffrey4l_ | pbourke_, ^^ | 12:47 |
pbourke_ | Jeffrey4l_: the magnum dashboard | 12:47 |
pbourke_ | Jeffrey4l_: just shows up empty | 12:47 |
*** kristian__ has joined #openstack-kolla | 12:48 | |
Jeffrey4l_ | pbourke_, patch link? let me check. | 12:48 |
pbourke_ | Jeffrey4l_: sure that would be great, one sec | 12:48 |
Jeffrey4l_ | egonzalez90, pbourke_ please review this https://review.openstack.org/405273 which unblock the gate. | 12:49 |
mliima | hey Jeffrey4l_, morning | 12:50 |
Jeffrey4l_ | sup mliima ;) | 12:50 |
mliima | can you review it https://review.openstack.org/#/c/403689/ | 12:50 |
*** kristian__ has quit IRC | 12:50 | |
Jeffrey4l_ | np | 12:50 |
*** kristian__ has joined #openstack-kolla | 12:50 | |
*** jtriley has joined #openstack-kolla | 12:50 | |
openstackgerrit | Paul Bourke (pbourke) proposed openstack/kolla: Install magnum dashboard into the horizon image https://review.openstack.org/405350 | 12:51 |
Jeffrey4l_ | mliima, done | 12:51 |
mliima | thank Jeffrey4l_ | 12:51 |
*** kristian__ has quit IRC | 12:52 | |
*** kristian__ has joined #openstack-kolla | 12:52 | |
openstackgerrit | Paul Bourke (pbourke) proposed openstack/kolla-ansible: Enable magnum dashboard when enable_magnum is yes https://review.openstack.org/405351 | 12:52 |
pbourke_ | Jeffrey4l_: those are the two patches ^ | 12:52 |
Jeffrey4l_ | roger. | 12:53 |
*** hieulq has quit IRC | 12:53 | |
openstackgerrit | Merged openstack/kolla-ansible: Install ansible with OpenStack requirements constrains https://review.openstack.org/405273 | 12:55 |
Jeffrey4l_ | pbourke_, seems OK. did u checked the horizon container. whether the files under enabled/ is copied to openstack_dashbaord/local/enabled? | 12:56 |
*** msimonin has quit IRC | 12:56 | |
pbourke_ | Jeffrey4l_: yes they're all there. the entry shows in the side panel but the page is blank. I think it may be the same as https://bugs.launchpad.net/magnum-ui/+bug/1606382 | 12:57 |
openstack | Launchpad bug 1606382 in Magnum UI "Magnum UI: Blank page on horizon -> Orchestration" [Undecided,New] - Assigned to Bradley Jones (bradjones) | 12:57 |
*** kristian__ has quit IRC | 12:57 | |
Jeffrey4l_ | hrm. i will try to deploy a env with patch. | 12:57 |
*** zhubingbing has joined #openstack-kolla | 12:57 | |
*** tonanhngo has joined #openstack-kolla | 12:57 | |
*** unicell has joined #openstack-kolla | 12:58 | |
*** tonanhngo has quit IRC | 12:59 | |
*** williamcaban has joined #openstack-kolla | 12:59 | |
*** kristian__ has joined #openstack-kolla | 13:00 | |
*** jtriley has quit IRC | 13:01 | |
*** hvlad has joined #openstack-kolla | 13:02 | |
*** hvlad has quit IRC | 13:03 | |
*** unicell has quit IRC | 13:03 | |
*** williamcaban has quit IRC | 13:04 | |
*** srwilkers_ has joined #openstack-kolla | 13:06 | |
srwilkers_ | good morning everyone | 13:07 |
*** kristian__ has quit IRC | 13:08 | |
openstackgerrit | Paul Bourke (pbourke) proposed openstack/kolla-ansible: Add missing pull task for neutron-vpnaas-agent https://review.openstack.org/401240 | 13:08 |
*** eaguilar has joined #openstack-kolla | 13:09 | |
*** tonanhngo has joined #openstack-kolla | 13:11 | |
*** tonanhngo has quit IRC | 13:12 | |
*** msimonin has joined #openstack-kolla | 13:15 | |
*** Satya_ has quit IRC | 13:17 | |
*** mgiles has joined #openstack-kolla | 13:21 | |
*** mgiles has quit IRC | 13:23 | |
*** mgiles has joined #openstack-kolla | 13:24 | |
*** tonanhngo has joined #openstack-kolla | 13:24 | |
*** prameswar has joined #openstack-kolla | 13:24 | |
*** eaguilar has quit IRC | 13:24 | |
*** jrist has quit IRC | 13:25 | |
*** gfidente has joined #openstack-kolla | 13:25 | |
*** gfidente has quit IRC | 13:25 | |
*** gfidente has joined #openstack-kolla | 13:25 | |
*** mdnadeem has quit IRC | 13:26 | |
*** tonanhngo has quit IRC | 13:27 | |
openstackgerrit | Merged openstack/kolla: Add manila-ui in horizon image https://review.openstack.org/403689 | 13:27 |
*** mdnadeem has joined #openstack-kolla | 13:29 | |
Jeffrey4l_ | morning ;) | 13:32 |
Jeffrey4l_ | pbourke_, testing you magnum ui patch, it works in my env ;) | 13:33 |
Jeffrey4l_ | ( even though i do not deploy magnum service, just enabled ui ) | 13:33 |
*** lamt has joined #openstack-kolla | 13:34 | |
*** rhallisey has joined #openstack-kolla | 13:36 | |
*** zhubingbing has quit IRC | 13:40 | |
*** Serlex has quit IRC | 13:43 | |
*** yingjun has joined #openstack-kolla | 13:43 | |
*** yingjun has quit IRC | 13:46 | |
*** yingjun has joined #openstack-kolla | 13:46 | |
*** yingjun has quit IRC | 13:52 | |
*** yingjun has joined #openstack-kolla | 13:52 | |
pbourke_ | Jeffrey4l_: o_O | 13:55 |
pbourke_ | wonder what could be different | 13:56 |
*** yingjun has quit IRC | 13:56 | |
*** mdnadeem has quit IRC | 13:58 | |
openstackgerrit | Merged openstack/kolla-ansible: Added support for a stop playbook in kolla https://review.openstack.org/399289 | 13:59 |
*** Pavo has quit IRC | 13:59 | |
*** Pavo has joined #openstack-kolla | 13:59 | |
*** eaguilar has joined #openstack-kolla | 14:00 | |
*** lamt has quit IRC | 14:01 | |
openstackgerrit | caoyuan proposed openstack/kolla-ansible: Remove the "nova-network" from the ansible https://review.openstack.org/404200 | 14:02 |
*** athomas has joined #openstack-kolla | 14:02 | |
*** jev has quit IRC | 14:05 | |
jmccarthy | Ok I've got a weird sort of legacy upgrade issue relating to neutron/nova and nova.conf.j2 template. Nowadays in the [neutron] block of this nova template username = {{ neutron_keystone_user }} is used (this fits ok no probs) https://github.com/openstack/kolla-ansible/blob/master/ansible/roles/nova/templates/nova.conf.j2#L128 | 14:05 |
jmccarthy | However if I'm upgrading from an *old setup that has this username hardcoded to a different user in the template itself, then instances seem to have trouble getting to their metadata post upgrade. Of course I'd rather get on board with the new name. I'm not seeing what it is that needs doing (manually or otherwise) in the db? to migrate/facilitate this upgrade ? | 14:05 |
*** lamt has joined #openstack-kolla | 14:07 | |
jmccarthy | Any ideas ? :) | 14:07 |
*** zhubingbing has joined #openstack-kolla | 14:09 | |
openstackgerrit | caoyuan proposed openstack/kolla-ansible: Remove the "nova-network" from the ansible https://review.openstack.org/404200 | 14:09 |
*** athomas has quit IRC | 14:10 | |
pbourke_ | jmccarthy: what was the old value? | 14:12 |
*** yingjun has joined #openstack-kolla | 14:12 | |
jmccarthy | "nova" | 14:14 |
jmccarthy | @pbourke_ nova | 14:15 |
pbourke_ | jmccarthy: so just set those vars to nova before upgrading? | 14:15 |
*** fguillot has joined #openstack-kolla | 14:15 | |
jmccarthy | Not sure I want to set neutron_keystone_user to nova - since this property is used elsewhere | 14:16 |
jmccarthy | (also I think I tried that but it broke the upgrade somewhere else) | 14:16 |
Jeffrey4l_ | pbourke_, i used checkout code. so it is the same. | 14:18 |
jmccarthy | @pbourke_ I might be able to give that another go later on to confirm | 14:18 |
*** jrist has joined #openstack-kolla | 14:21 | |
*** goldyfruit has joined #openstack-kolla | 14:21 | |
*** jtriley has joined #openstack-kolla | 14:22 | |
*** yingjun_ has joined #openstack-kolla | 14:23 | |
*** athomas has joined #openstack-kolla | 14:24 | |
*** yingjun has quit IRC | 14:26 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Enable trove dashboard when enable_trove is yes https://review.openstack.org/405090 | 14:27 |
openstackgerrit | caoyuan proposed openstack/kolla: Remove the "nova-network" container file https://review.openstack.org/404099 | 14:31 |
*** yingjun has joined #openstack-kolla | 14:32 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: NO_MERGE: TEST all tempest test https://review.openstack.org/402122 | 14:33 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: TEST_NO_MERGE https://review.openstack.org/403978 | 14:33 |
*** yingjun_ has quit IRC | 14:35 | |
*** yingjun has quit IRC | 14:37 | |
*** yingjun has joined #openstack-kolla | 14:37 | |
*** msimonin has quit IRC | 14:38 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Enable tempest test in gate https://review.openstack.org/402122 | 14:38 |
*** jtriley has quit IRC | 14:39 | |
*** yingjun_ has joined #openstack-kolla | 14:40 | |
*** goldyfruit has quit IRC | 14:41 | |
*** goldyfruit has joined #openstack-kolla | 14:41 | |
*** yingjun has quit IRC | 14:41 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Hel-ming Rabbitmq https://review.openstack.org/405422 | 14:42 |
sdake_ | Jeffrey4l_ yo bro around? | 14:43 |
Jeffrey4l_ | sdake_, yep | 14:43 |
sdake_ | Jeffrey4l_ if you are, I wanted to give you a thumbs up for running the meeting yesterday - first time ever! | 14:43 |
sdake_ | Jeffrey4l_ you did a great job | 14:43 |
Jeffrey4l_ | sdake_, thanks ;) more thing need to learn | 14:43 |
sdake_ | Jeffrey4l_ meeting was a bit hectic | 14:43 |
sdake_ | Jeffrey4l_ thats not on you thats on me :) | 14:44 |
sdake_ | had multiple things going on | 14:44 |
sdake_ | apologies for the disruption | 14:44 |
v1k0d3n | mornin' ya'll from the south! | 14:44 |
Jeffrey4l_ | yes. need take some time to prepare the meeting . | 14:44 |
sdake_ | sup v1k0d3n | 14:44 |
v1k0d3n | sdake_: sup | 14:44 |
sdake_ | chillen | 14:44 |
sdake_ | got most of my backlog sorted out | 14:45 |
sdake_ | *finally* | 14:45 |
sdake_ | so can actually get to work | 14:45 |
sdake_ | i still owe sbezverk a review and i'll be all caught up :) | 14:45 |
sdake_ | v1k0d3n who specifically is from the south? | 14:46 |
v1k0d3n | me | 14:47 |
sdake_ | didn't catch the southern accent | 14:47 |
v1k0d3n | "from" the south | 14:47 |
v1k0d3n | land of biscuits and gravy | 14:47 |
v1k0d3n | i hide it :) | 14:47 |
sdake_ | i see that :) | 14:47 |
v1k0d3n | unless i have whiskey it stays hidden :) | 14:48 |
sdake_ | when my wife travels to the south - she comes back with an accent | 14:48 |
sdake_ | like if she is there for a week | 14:48 |
sdake_ | arizona has no accent | 14:48 |
v1k0d3n | it's addictive. | 14:48 |
sdake_ | yup it is | 14:48 |
v1k0d3n | there's no doubt :) | 14:48 |
*** duonghq has quit IRC | 14:48 | |
*** duonghq_ has joined #openstack-kolla | 14:48 | |
duonghq_ | evening guys | 14:48 |
sdake_ | i think there rea a few other states without accent as well | 14:48 |
sdake_ | sup duonghq_ | 14:48 |
sdake_ | oh backlog incomplete | 14:49 |
v1k0d3n | yeah, one would think that AZ would have an accent of some sort...but you're right...not really at all. | 14:49 |
sdake_ | damn i almost declared victory on my backlog | 14:49 |
*** duonghq has joined #openstack-kolla | 14:49 | |
*** duonghq has quit IRC | 14:49 | |
*** duonghq_ is now known as duonghq | 14:49 | |
sdake_ | GTD failed me there | 14:49 |
*** duonghq_ has joined #openstack-kolla | 14:50 | |
sdake_ | or rather i failed at GTD :) | 14:50 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Enable cloudkitty ui when enable_cloudkitty is yes https://review.openstack.org/403520 | 14:50 |
duonghq | has GTD failed you before? | 14:50 |
sdake_ | i have failed at keeping gtd updated yes | 14:51 |
sdake_ | its one way to lead a stress free life | 14:51 |
sdake_ | one of many techniques | 14:51 |
duonghq | sure, so you will continue with GTD or not? | 14:52 |
sdake_ | of course | 14:52 |
sdake_ | error on my part - not problem with gtd | 14:52 |
*** DTadrzak has joined #openstack-kolla | 14:52 | |
sdake_ | duonghq are you familiar with GTD? | 14:52 |
*** eaguilar has quit IRC | 14:53 | |
sdake_ | one of my mentors turned me on to it about 8-10 years ago | 14:53 |
sdake_ | before it I was a mess | 14:53 |
v1k0d3n | portdirect: hey man | 14:53 |
duonghq | not yet, I found it is hard to keep up myself, maybe I need for life experience | 14:53 |
v1k0d3n | updated PR's and submodule | 14:53 |
sdake_ | duonghq ya it requires upkeep | 14:53 |
sdake_ | duonghq and practice | 14:54 |
duonghq | sure | 14:54 |
v1k0d3n | very close to CI as well. have the vagrant part working well, just need to get the connection/reports going from our external jenkins to the repo for users (so folks can correct any issues with PR's). | 14:54 |
sdake_ | v1k0d3n nice ci :) | 14:54 |
v1k0d3n | everyone...i think halcyon is a go now for kolla-kube! | 14:54 |
v1k0d3n | check out the readme. portdirect made things crazy simple.... | 14:54 |
sdake_ | v1k0d3n doe that mean portdirect doesn't have the fork anymore? | 14:55 |
v1k0d3n | so developers should be able to get in, do work, get out. easy. | 14:55 |
sdake_ | is readme in gerrit or somewhere else? | 14:55 |
duonghq | did we have consensus on helm version tagging? | 14:55 |
v1k0d3n | sdake_: no, i don't think so from what i've seen in the PR's | 14:55 |
sdake_ | duonghq i think we are punting on that for the moment | 14:55 |
sdake_ | v1k0d3n cool | 14:55 |
v1k0d3n | sdake_: maybe try it out? | 14:56 |
sdake_ | v1k0d3n i am running an older version | 14:56 |
v1k0d3n | give us any feedback? | 14:56 |
sdake_ | v1k0d3n yesterday's version | 14:56 |
*** inc0 has joined #openstack-kolla | 14:56 | |
sdake_ | i can spin up todays version if you like | 14:56 |
sdake_ | sup inc0 | 14:56 |
v1k0d3n | sdake_: this has been added :) https://github.com/att-comdev/halcyon-vagrant-kubernetes#configuration-helper | 14:56 |
inc0 | good morning | 14:57 |
v1k0d3n | inc0: good morning | 14:57 |
*** gfidente has quit IRC | 14:57 | |
*** eaguilar has joined #openstack-kolla | 14:57 | |
*** saneax is now known as saneax-_-|AFK | 14:57 | |
inc0 | sdake_, kfox1111 sbezverk how about we don't write mariadb at all in kolla-k8s? | 14:57 |
v1k0d3n | sdake_: that would be awesome | 14:58 |
v1k0d3n | keep that old one around...just in case :) | 14:58 |
v1k0d3n | but i think it should be god | 14:58 |
v1k0d3n | haha or GOOD not god | 14:58 |
sdake_ | v1k0d3n nice - i assume your ok with kolla-kubernetes making its own docs to point to those repos | 14:58 |
sdake_ | inc0 mariadb is critical path | 14:58 |
v1k0d3n | absolutely | 14:58 |
*** msimonin has joined #openstack-kolla | 14:58 | |
sdake_ | v1k0d3n cool | 14:58 |
inc0 | sdake_, to get it somehow up | 14:58 |
inc0 | somehow up can mean multitude of things | 14:58 |
v1k0d3n | think someone is working on this. either srwilkers_ or portdirect | 14:58 |
inc0 | including kolla-ansible deploy -t mariadb | 14:59 |
inc0 | and I for one would think this actually would be quite elegant | 14:59 |
sdake_ | inc0 we have a standard api model for communicating with helm packages - deploying mariadb from the incubator wont work | 14:59 |
*** Marcellin__ has joined #openstack-kolla | 14:59 | |
inc0 | sdake_, if our standard model is not compatible with helm incubator standard model | 15:00 |
sdake_ | v1k0d3n portdirect as i recall | 15:00 |
inc0 | I'd say we need to get back to drawing board | 15:00 |
v1k0d3n | ok. just checking. | 15:00 |
v1k0d3n | inc0: drawing board? | 15:00 |
*** unicell has joined #openstack-kolla | 15:00 | |
inc0 | what I'm saying | 15:00 |
v1k0d3n | per our convo last night? | 15:00 |
inc0 | no about that: "sdake_> inc0 we have a standard api model for communicating with helm packages - deploying mariadb from the incubator wont work" | 15:01 |
*** williamcaban has joined #openstack-kolla | 15:01 | |
inc0 | does that mean that our standard model is super different from helm incubator model? | 15:01 |
srwilkers_ | good morning sdake_ v1k0d3n inc0 | 15:01 |
sdake_ | inc0 kfox1111 is the best person to talk to about the api | 15:01 |
sdake_ | sup srwilkers_ | 15:01 |
srwilkers_ | sdake_, yes portdirect is working on that | 15:01 |
v1k0d3n | sup srwilkers_ | 15:01 |
mewald | Hi guys, have a question: haproxy config looks wrong to me. We set X-Forwarded-Proto if HAproxy manages SSL termination (if { ssl_fc}) but we remove it no matter what. I think it should only be removed also if haproxy manages SSL, or not? (https://github.com/openstack/kolla-ansible/blob/master/ansible/roles/haproxy/templates/haproxy.cfg.j2#L66) | 15:02 |
inc0 | mewald, yeah I caught that too, seemed wrong to me | 15:04 |
*** unicell has quit IRC | 15:04 | |
*** williamcaban has quit IRC | 15:05 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Hel-ming Rabbitmq https://review.openstack.org/405422 | 15:05 |
portdirect | morning | 15:06 |
*** jtriley has joined #openstack-kolla | 15:07 | |
*** tonanhngo has joined #openstack-kolla | 15:07 | |
portdirect | srwilkers_/v1k0d3n/sdake: just about to get the readme update - the att-comdev repos are good to go for kolla-dev | 15:07 |
*** tonanhngo has quit IRC | 15:08 | |
mewald | inc0: Also I noticed that Nova api returns wrong references. For example when you list servers you get an href to the flavor of each instance. The URL to the flavor is http (not https) I am not sure where that comes from but it might be related | 15:08 |
inc0 | mewald, well nova doesn't have idea of public endpoint | 15:09 |
inc0 | I'd say it's bug in nova | 15:09 |
mewald | well, when I curl the Keystone API get the listing of available APIs and it does say "https" and not "http". Therefore, keystone knows about the public endpoint (I assume from x-forwarded-proto), so either nova doesn't care about it, or the header is not set | 15:10 |
*** gfidente has joined #openstack-kolla | 15:11 | |
*** jtriley has quit IRC | 15:12 | |
*** absubram has quit IRC | 15:12 | |
v1k0d3n | portdirect: awesome | 15:12 |
v1k0d3n | we noticed the same issues with helm 2.0.0 and 1.5 kube the other day | 15:13 |
v1k0d3n | started down the path of mentioning this in the helm slack channel....but nobody seemed to know what i was talking about. | 15:13 |
v1k0d3n | there's obviously an issue there. | 15:13 |
sdake_ | portdirect ok will wait for doc udpate to validate the env | 15:14 |
portdirect | v1k0d3n: chime in here if you can: https://github.com/kubernetes/helm/issues/1589 | 15:15 |
*** mewald has quit IRC | 15:15 | |
v1k0d3n | portdirect: saw that. one sec. | 15:15 |
v1k0d3n | portdirect: done | 15:18 |
v1k0d3n | i spoke with matt about this for a little while over slack | 15:18 |
portdirect | v1k0d3n: great cheers | 15:19 |
*** lrensing has joined #openstack-kolla | 15:20 | |
*** tonanhngo has joined #openstack-kolla | 15:21 | |
*** stevemar has joined #openstack-kolla | 15:21 | |
*** tonanhngo has quit IRC | 15:22 | |
openstackgerrit | Duong Ha-Quang proposed openstack/kolla-ansible: Specify 'become' for only neccesary tasks (default roles) https://review.openstack.org/398684 | 15:23 |
*** dave-mccowan has joined #openstack-kolla | 15:26 | |
duonghq | goodnight guys | 15:28 |
*** duonghq has quit IRC | 15:28 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Hel-ming Rabbitmq https://review.openstack.org/405422 | 15:29 |
*** stevemar_znc has joined #openstack-kolla | 15:29 | |
srwilkers_ | goodnight DuncanT | 15:31 |
srwilkers_ | oops | 15:31 |
srwilkers_ | duonghq_ | 15:31 |
*** jgriffith_away is now known as jgriffith | 15:32 | |
openstackgerrit | caoyuan proposed openstack/kolla-ansible: Remove the deprecated command and the useless "if" condition https://review.openstack.org/404896 | 15:35 |
*** mdnadeem has joined #openstack-kolla | 15:35 | |
*** senk has quit IRC | 15:38 | |
*** jtriley has joined #openstack-kolla | 15:39 | |
*** jheroux has joined #openstack-kolla | 15:40 | |
openstackgerrit | zhubingbing proposed openstack/kolla-ansible: Fix aodh role https://review.openstack.org/405275 | 15:40 |
*** tonanhngo has joined #openstack-kolla | 15:40 | |
*** tonanhngo has quit IRC | 15:43 | |
openstackgerrit | zhubingbing proposed openstack/kolla: Install trove dashboard into horizon image https://review.openstack.org/405087 | 15:45 |
*** athomas has quit IRC | 15:46 | |
*** dims has quit IRC | 15:46 | |
openstackgerrit | Pete Birley proposed openstack/kolla-kubernetes: Kolla-Kubernetes Vagrant development environment documentation https://review.openstack.org/404493 | 15:46 |
*** adrian_otto has joined #openstack-kolla | 15:47 | |
openstackgerrit | zhubingbing proposed openstack/kolla-ansible: Enable trove dashboard when enable_trove is yes https://review.openstack.org/405090 | 15:49 |
inc0 | kfox1111, around? | 15:49 |
*** zhubingbing has quit IRC | 15:50 | |
*** yingjun_ has quit IRC | 15:51 | |
*** yingjun has joined #openstack-kolla | 15:54 | |
*** msimonin has quit IRC | 15:56 | |
*** Pavo has quit IRC | 15:59 | |
inc0 | SamYaple, sooo..have you noticed that gates were working yesteday? | 16:00 |
*** dims has joined #openstack-kolla | 16:00 | |
*** adrian_otto has quit IRC | 16:01 | |
*** jrich523_ is now known as jrich523 | 16:02 | |
*** dave-mccowan has quit IRC | 16:02 | |
*** msimonin has joined #openstack-kolla | 16:03 | |
sdake_ | sup dims | 16:03 |
sdake_ | thanks for the understanding and +1 on the pruning email | 16:04 |
dims | sdake_ : howdy | 16:04 |
*** Pavo has joined #openstack-kolla | 16:04 | |
sdake_ | that was difficult for me personally | 16:04 |
*** Satya_ has joined #openstack-kolla | 16:04 | |
dims | i should have done it before :) | 16:04 |
sdake_ | dims its ok - it was my job to finish | 16:04 |
sdake_ | and its done | 16:04 |
kfox1111 | morning | 16:05 |
sdake_ | sup kfox1111 | 16:05 |
kfox1111 | sdake_: hi | 16:05 |
kfox1111 | inc0: here now. | 16:05 |
sdake_ | so basically kfox1111 i haven't gotten to the review queue this mroning | 16:05 |
sdake_ | kfox1111 will get to it shortly | 16:05 |
inc0 | kfox1111, sooo...what would stop us from using mariadb from helm? | 16:06 |
inc0 | helm incubator that is | 16:06 |
inc0 | or rabbitmq | 16:06 |
Satya_ | While using master branch kolla module, I am getting following error "msg": "Unknown error message: Get http://: http: no Host in request URL" | 16:06 |
sbezverk | kfox1111: ovsdb was all green, I think we need to merge it to test properly your new common code. | 16:06 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Add qemu-img into glance api image https://review.openstack.org/405468 | 16:06 |
Satya_ | it seems the target controller machine is not able to pull the image from deployment server | 16:07 |
kfox1111 | inc0: I don't see any reason that shouldn't work. I think we should support external rabbit/mariadb anyway. | 16:07 |
inc0 | Satya_, interesting, docker is acting wierdly | 16:07 |
kfox1111 | sbezverk: cool. will give it one more look. | 16:07 |
Satya_ | i tried to pull the image manually it worked | 16:07 |
kfox1111 | sbezverk: could you try a rebased version on the other patch though? then we woudln't merge a copy of _common_lib just to remove it one patch later. | 16:07 |
inc0 | kfox1111, so follow up question - we will do all external at some point | 16:08 |
inc0 | and not even support mariadb other than helm incubator | 16:08 |
Satya_ | after that if i start deployment it is giving following error NoneType object has no attribute __getitem__ | 16:08 |
inc0 | and work with helm community to make sure helmy maria is happy maria | 16:08 |
Satya_ | it seems some version mismatch of package, did any one observed this issue with master branch code | 16:09 |
inc0 | Satya_, let me guess, you have registry on one node and spawned couple fresh vms? | 16:09 |
Satya_ | my registry in on deployment node where I am running the ansible-playbook | 16:09 |
kfox1111 | inc0: could be? I don't think it matters too much to me. thats probably a long ways out though. | 16:09 |
inc0 | would be elegant imho | 16:10 |
inc0 | and lower complexity of project tremendously | 16:10 |
*** sayantan_ has joined #openstack-kolla | 16:11 | |
*** sayantan_ has quit IRC | 16:11 | |
kfox1111 | inc0: yeah, maybe. if they can come up with an operator for mariadb (or make k8s itself stateful sets flexible enough to support galera proper) then that would be awesome. | 16:11 |
inc0 | well we can still help | 16:11 |
*** sayantan_ has joined #openstack-kolla | 16:11 | |
kfox1111 | or bake that into helm somehow.. (kind of scary) | 16:11 |
inc0 | but there is good chance we'll get broader community around there | 16:11 |
inc0 | for this one particular case | 16:11 |
Satya_ | inc0 any suggestions for my issue | 16:12 |
*** absubram has joined #openstack-kolla | 16:12 | |
inc0 | Satya_, soo...try to build images in one of non-registry nodes | 16:12 |
inc0 | and tell me if it works | 16:12 |
*** stevemar_znc is now known as topol_ | 16:12 | |
kfox1111 | inc0: yeah. if the k8s community builds around a community mariadb / rabbit, that woudl be nice. | 16:13 |
Satya_ | Ok I will try it thanks | 16:13 |
kfox1111 | long term, that would be good for all parties involved. | 16:13 |
inc0 | v1k0d3n, ^ I think we can help on that front right? | 16:14 |
*** eaguilar has quit IRC | 16:14 | |
kfox1111 | in the mean time though, I think we should continue to work with our own stuff, and then work on either upstreaming it, or merging features of ours until the upstream community one works for us. | 16:14 |
inc0 | kfox1111, agree, just to get us up and running | 16:14 |
inc0 | but not spend too much time on running stuff like galera proper here, and ensure we can use galera proper chart when it's available | 16:14 |
*** dave-mccowan has joined #openstack-kolla | 16:16 | |
*** Bico_Fino has joined #openstack-kolla | 16:16 | |
*** wwriverrat has joined #openstack-kolla | 16:16 | |
SamYaple | inc0: im not sure | 16:17 |
SamYaple | inc0: any news on the issue though? | 16:17 |
inc0 | only thats blowing minds | 16:17 |
inc0 | but issue Satya_ had appears in my env to | 16:18 |
inc0 | too | 16:18 |
inc0 | I have working registry and broken compute nodes | 16:18 |
inc0 | and deployment no bueno | 16:18 |
*** mdnadeem has quit IRC | 16:19 | |
*** topol_ is now known as topol | 16:20 | |
inc0 | Jeffrey4l_, around? | 16:20 |
Jeffrey4l_ | inc0, yep. | 16:20 |
inc0 | I have favor to ask | 16:21 |
inc0 | remember our talk about submitting registry to tarballs.o.o? | 16:21 |
inc0 | https://review.openstack.org/#/c/404435/ <- this is this change | 16:21 |
inc0 | kinda | 16:21 |
sdake_ | inc0++ | 16:21 |
kfox1111 | inc0: I need galera soon I think though. so I may contribute it anyway. | 16:21 |
kfox1111 | then we can work on getting that upstream. | 16:22 |
inc0 | kfox1111, cool, just let's keep in mind that we want to keep doors open | 16:22 |
kfox1111 | inc0: absolutelly. :) | 16:22 |
inc0 | at some point we might want to move to upstream helmy impl | 16:22 |
sdake_ | another inc0++ :) | 16:22 |
Jeffrey4l_ | inc0, yep. | 16:22 |
kfox1111 | inc0: I'd like to see trove involved in that endevor too. I could totally see them helping a lot with it. | 16:22 |
inc0 | Jeffrey4l_, I might need help with this change in case I won't manage to get it up before I leave | 16:23 |
inc0 | (next week) | 16:23 |
inc0 | mind taking over this work? | 16:23 |
sdake_ | inc0 i'd offer to help but puppet is not a thing i know :( | 16:23 |
kfox1111 | if there was a mariadb-gallera operator, trove api could wrap around it. | 16:23 |
Jeffrey4l_ | inc0, we need create new jobs for this. rather than modified the current job. | 16:23 |
inc0 | sdake_, all the infra pieces are in place already | 16:23 |
sdake_ | inc0 sweet | 16:23 |
kfox1111 | so trove create-db ends up being wired into a kubectl create -f mariadbcluster.yaml | 16:23 |
sdake_ | inc0 did something i never could do :) | 16:24 |
Jeffrey4l_ | inc0, need add a jobs into post pipeline http://zuul.openstack.org/ | 16:24 |
Jeffrey4l_ | i.e. push the registry to tarball after change is merged. | 16:25 |
sbezverk | kfox1111: rebasing on something which has not been tested at the gate does not sound attractive and I hate rebasing :-( if we merge template then we can test new common right at the gate with template which is 100% working.. | 16:25 |
*** rmart04 has quit IRC | 16:25 | |
kfox1111 | sbezverk: I spent a lot of time yesterday rebasing. :/ so somewhat sympethetic. but, if you don't rebase the one patch, I'll have to rebase 3. :/ | 16:26 |
kfox1111 | and you can still test the gate against the rebase. | 16:27 |
Jeffrey4l_ | in your patch, you just add the packed tarball file to jenkins logs collector. hrm, it is useless and is not accessible by other patch. inc0 | 16:27 |
sdake_ | fsa deadline march 31st | 16:28 |
* sdake_ groans | 16:28 | |
inc0 | no, no, saving is in tarballs is done | 16:28 |
sbezverk | kfox1111: so do I :-) because vswitchd and rabbit sit on top of ovsdb, but I will give it a try.. | 16:28 |
inc0 | just do recheck experimental | 16:28 |
kfox1111 | sbezverk: oh. i didn't konw you had more patches in review too. | 16:28 |
inc0 | and it should save every .tar.gz in $WORKSPACE/images to tarballs | 16:29 |
sbezverk | kfox1111: I will need then sdake help to this as I have never tried rebasing parent PS with bunch of childs depending on a parent | 16:30 |
sdake_ | sbezverk i can help with that today at some point | 16:30 |
* Jeffrey4l_ is checking the project-config | 16:30 | |
kfox1111 | sbezverk: ok. then don't worry about it. I'll rebase one more time my set. | 16:30 |
kfox1111 | the procedure is still fresh in my head. | 16:30 |
*** adrian_otto has joined #openstack-kolla | 16:31 | |
sbezverk | kfox1111: thanks if you could share steps for future would be great. | 16:31 |
kfox1111 | can we drop the built sources out of your review real quick so we don't have to pull them back out? | 16:31 |
kfox1111 | the gate prebuilds all the time, so will still work. | 16:31 |
sbezverk | do you mean common part? | 16:31 |
kfox1111 | yeah. comon lib and the built vars | 16:32 |
Jeffrey4l_ | inc0, seems so. | 16:32 |
sbezverk | kfox1111: ok will do it now | 16:32 |
inc0 | if it works, all the pieces are in our hands now | 16:32 |
inc0 | we need to pack it up, save to tarballs and in other gates download tarball and unpack | 16:33 |
inc0 | https://review.openstack.org/#/c/404405/3 <- this review is key | 16:33 |
kfox1111 | sbezverk: cool. rest looks good. we can merge then. :) | 16:33 |
inc0 | we unpack tarball to /tmp/kolla_registry | 16:33 |
Jeffrey4l_ | yep. so need launch a docker registry service first. then use tar tool. | 16:33 |
inc0 | and we should have populated registry | 16:33 |
sbezverk | kfox1111: done | 16:34 |
Jeffrey4l_ | inc0, i think i got the whole story. | 16:34 |
Jeffrey4l_ | so what i need to do? inc0 | 16:35 |
inc0 | yeah, if you could take care of it when I'm gone, I'd appreciate it | 16:35 |
Jeffrey4l_ | ok. np. | 16:35 |
kfox1111 | workflowed. | 16:35 |
*** msimonin has quit IRC | 16:35 | |
inc0 | many thanks, this will allow us to have registry-like thing in our gates, big improvement | 16:35 |
Jeffrey4l_ | inc0, have u ever thought push tag/stable tar.gz file? | 16:35 |
Jeffrey4l_ | inc0, it will be also helpful for hub.docker.com , imo. | 16:36 |
Jeffrey4l_ | just push that tar.gz file. :) | 16:36 |
inc0 | not sure how useful that would be | 16:36 |
inc0 | we're hacking around the system here | 16:36 |
inc0 | having proper registry would still be better | 16:37 |
Jeffrey4l_ | inc0, no. we can tread tarball.o.o as a new release tool. | 16:37 |
Jeffrey4l_ | other guys can get latest or stable or certain tag of kolla image fro that site. | 16:37 |
inc0 | nah, we should use docker.io | 16:38 |
Jeffrey4l_ | rather than from docker.io. | 16:38 |
Jeffrey4l_ | docker.io hasn't master images. | 16:38 |
Jeffrey4l_ | lots of guy ( in china, network is bad, you know ) want to try the master code. ;) | 16:38 |
Jeffrey4l_ | ok. edge case. | 16:38 |
inc0 | I wouldn't call China "edge";) | 16:39 |
inc0 | but it might be useful to hack around this registry too for them I guess | 16:40 |
inc0 | well, let's think about it later | 16:40 |
Jeffrey4l_ | yep. | 16:40 |
inc0 | for now let's get our deploy jobs voting | 16:40 |
Jeffrey4l_ | yes. | 16:40 |
Jeffrey4l_ | i will work on that PS. | 16:40 |
sbezverk | kfox1111: I guess when your new common code gets merged I will need to add _common to ovsdb so it would be in git repo | 16:41 |
inc0 | thanks | 16:41 |
Jeffrey4l_ | np :) | 16:41 |
inc0 | I'll get back to this strange docker behavior bag | 16:41 |
inc0 | bug | 16:41 |
sdake_ | sbezverk sounds like you dont need help with a rebase on top of someone elses patch, if you do shout out | 16:41 |
kfox1111 | sbezverk: do you think its onerious to need to do a prebuild after git checking out, but before doing helm installs? | 16:41 |
kfox1111 | it would be cleaner I think. | 16:42 |
kfox1111 | then we don't have to check in built sources. | 16:42 |
Jeffrey4l_ | inc0, bug? | 16:42 |
inc0 | hard to say what's happening | 16:42 |
inc0 | so I spawned new vm yesterday for deployment | 16:42 |
inc0 | ubuntu xenial | 16:42 |
inc0 | fresh master installation of kolla | 16:42 |
inc0 | when I build images there, they're not getting tagged | 16:43 |
sbezverk | kfox1111: agree it is safer, specifically for the case we have now with ovsdb :-) | 16:43 |
inc0 | same happends to SamYaple and it seems Satya_ | 16:43 |
Jeffrey4l_ | what's mean by `not getting tagged`? | 16:43 |
sbezverk | inc0: I hit the same issue as SamYaple too | 16:43 |
kfox1111 | sbezverk: ok. lets leave the built stuff out for now then, and see how things go. we can always reevaluate as we go. | 16:43 |
inc0 | image is built, but it doesn't get its name after build | 16:44 |
Jeffrey4l_ | hmm. | 16:44 |
Satya_ | Hi Inc0 . what did SamYaple do to fix that issue | 16:44 |
Jeffrey4l_ | you hit requests bug i think ;) | 16:44 |
inc0 | we don't have idea | 16:44 |
*** adrian_otto has quit IRC | 16:44 | |
*** DTadrzak has quit IRC | 16:44 | |
Jeffrey4l_ | check the requests version. | 16:44 |
inc0 | Jeffrey4l_, oh? you know something? | 16:44 |
*** strigazi is now known as strigazi_AFK | 16:44 | |
Jeffrey4l_ | yep. | 16:44 |
Jeffrey4l_ | just fixed a issue in the gate. | 16:44 |
inc0 | 2.12.2 | 16:45 |
Jeffrey4l_ | requests 2.12.2 breaks the docker-py. | 16:45 |
inc0 | oh, how fun | 16:45 |
*** egonzalez90 has quit IRC | 16:45 | |
inc0 | we need to change requirements asap | 16:45 |
inc0 | which version works? | 16:45 |
Jeffrey4l_ | requirements is OK. | 16:45 |
*** portdirect is now known as portdirect_inmee | 16:46 | |
Jeffrey4l_ | inc0, check this https://github.com/openstack/kolla-ansible/commit/27ee3973a7e95db3f3f6c43c0d514f332b7d1602 | 16:46 |
*** portdirect_inmee is now known as portdirect_meetn | 16:46 | |
Jeffrey4l_ | inc0, sbezverk upstream bug https://github.com/docker/docker-py/issues/1321 | 16:46 |
*** dave-mccowan has quit IRC | 16:46 | |
*** bmace has quit IRC | 16:46 | |
Jeffrey4l_ | you can check your requests version first, i think it should be the same. | 16:46 |
*** bmace has joined #openstack-kolla | 16:47 | |
Satya_ | does it work with latest master branch code | 16:47 |
inc0 | SamYaple, ^ | 16:47 |
*** msimonin has joined #openstack-kolla | 16:47 | |
Jeffrey4l_ | Satya_, yep. but this patch just fixed the master. | 16:47 |
Jeffrey4l_ | other branch still broken ;( | 16:47 |
Jeffrey4l_ | so gate is green now :) | 16:48 |
*** yingjun has quit IRC | 16:48 | |
*** yingjun has joined #openstack-kolla | 16:48 | |
Satya_ | if I make the changes mentioned in "bug https://github.com/docker/docker-py/issues/1321" in my code then it should be working in my environment? | 16:49 |
Jeffrey4l_ | Satya_, what kinda changes? | 16:50 |
inc0 | Satya_, just do pip install --upgrade requests==2.11.1 on all nodes | 16:50 |
Jeffrey4l_ | inc0, ++ | 16:50 |
inc0 | it seems to be working | 16:50 |
inc0 | just tested | 16:50 |
Jeffrey4l_ | just do not use requests==2.12.2 , other version is OK. | 16:51 |
inc0 | Jeffrey4l_, if someone does pip install kolla, it's gonna install 2.11.2 now? | 16:51 |
Jeffrey4l_ | upstream upstream bug id https://github.com/kennethreitz/requests/issues/3734 | 16:51 |
*** adrian_otto has joined #openstack-kolla | 16:52 | |
Satya_ | Ok thanks inc0 and Jeffrey41 I will try that | 16:52 |
Jeffrey4l_ | inc0, i think so. | 16:52 |
inc0 | we need to fix that | 16:52 |
Jeffrey4l_ | normally, the best practice is using upper-constrains.txt file. | 16:52 |
Jeffrey4l_ | it is well tested. | 16:52 |
inc0 | let's use it then | 16:52 |
inc0 | we have broken master nw | 16:53 |
inc0 | now | 16:53 |
*** yingjun has quit IRC | 16:53 | |
Jeffrey4l_ | we can do nothing at this point. | 16:53 |
Jeffrey4l_ | it is located in another repo. | 16:53 |
inc0 | yes we can | 16:53 |
Jeffrey4l_ | how? | 16:53 |
openstackgerrit | sayantani proposed openstack/kolla: Add new guide for single-node deployment https://review.openstack.org/404993 | 16:53 |
inc0 | hold on | 16:53 |
Jeffrey4l_ | kolla's tox.ini file is using upper-con* | 16:54 |
Jeffrey4l_ | install_command = pip install -c{env:UPPER_CONSTRAINTS_FILE:https://git.openstack.org/cgit/openstack/requirements/plain/upper-constraints.txt} {opts} {packages} | 16:54 |
inc0 | our setup.py has to | 16:55 |
kfox1111 | oh... its the tunnel interface on the helm job... right... | 16:56 |
openstackgerrit | Merged openstack/kolla-kubernetes: Adding openvswitch-ovsdb microservice https://review.openstack.org/403959 | 16:56 |
Jeffrey4l_ | imo, this should be maintained by requirements.txt | 16:56 |
*** dave-mccowan has joined #openstack-kolla | 16:56 | |
*** unicell has joined #openstack-kolla | 16:56 | |
Jeffrey4l_ | btw, i hate to install package direct from pypi ;( . unstable. | 16:56 |
openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla: Fix master https://review.openstack.org/405525 | 16:56 |
Jeffrey4l_ | inc0, you will get a shoot https://github.com/openstack/requirements/blob/master/global-requirements.txt#L241 lol | 16:57 |
kfox1111 | yeah... this will change things a bit... setup_config won't be where config's applied. it will ahppen at package launch time.... | 16:58 |
inc0 | no, that still allows us to have 2.11.1 for example | 16:58 |
inc0 | which works | 16:58 |
*** msimonin has quit IRC | 16:58 | |
Jeffrey4l_ | inc0, it works but requirements check will be red. | 16:58 |
Jeffrey4l_ | ever requirements.txt file need follow global-requirements.txt, but you break it in your patch. | 16:59 |
inc0 | then let's propose this to global requirements | 16:59 |
kfox1111 | sbezverk: goign to rebase the common one now. | 17:00 |
Jeffrey4l_ | inc0, that will work. | 17:00 |
Jeffrey4l_ | requests>=2.10.0,!=2.12.2 | 17:00 |
SamYaple | Jeffrey4l_: inc0 yep all good here | 17:01 |
inc0 | 2.12.3 is on pypi already | 17:01 |
inc0 | let's see | 17:01 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helm common package https://review.openstack.org/404808 | 17:02 |
Jeffrey4l_ | cool | 17:02 |
*** williamcaban has joined #openstack-kolla | 17:02 | |
*** matrohon has quit IRC | 17:02 | |
inc0 | seems like it's working | 17:02 |
inc0 | so master seems to be unblocked | 17:02 |
kfox1111 | sbezverk: k. that one should be good to go. | 17:02 |
inc0 | uff | 17:02 |
Jeffrey4l_ | anyway, in gate ( or any other place ) install python packages with requirements/upper-constrains.txt will be more stable. | 17:03 |
*** igordcard has joined #openstack-kolla | 17:03 | |
sbezverk | kfox1111: the common one? | 17:03 |
kfox1111 | sbezverk: yeah | 17:03 |
SamYaple | inc0: yea thats what i meant | 17:04 |
SamYaple | 2.12.3 is working | 17:04 |
SamYaple | i havent tested 2.11 | 17:04 |
Jeffrey4l_ | i am always using 2.11.1 | 17:05 |
sbezverk | kfox1111: should we wait for check experimental results? | 17:05 |
inc0 | that's ok | 17:05 |
inc0 | newest version is working | 17:05 |
inc0 | pip install kolla works, thats what I care about | 17:06 |
inc0 | https://review.openstack.org/#/c/405534/1/global-requirements.txt | 17:06 |
SamYaple | Jeffrey4l_: agreed on upper-constraints | 17:07 |
*** williamcaban has quit IRC | 17:07 | |
SamYaple | glad i managed to start working on this literally seconds after a bad packages was pushed | 17:07 |
SamYaple | yay me | 17:07 |
*** Bico_Fino has quit IRC | 17:07 | |
*** Bico_Fino has joined #openstack-kolla | 17:09 | |
Jeffrey4l_ | fyi enabled tempest in gate https://review.openstack.org/402122 http://logs.openstack.org/22/402122/9/check/gate-kolla-ansible-dsvm-deploy-centos-binary-centos-7-nv/8d16727/logs/results.html | 17:09 |
Jeffrey4l_ | lol | 17:09 |
openstackgerrit | sayantani proposed openstack/kolla: Add new guide for single-node deployment https://review.openstack.org/404993 | 17:11 |
kfox1111 | sbezverk: hmmm.... | 17:11 |
kfox1111 | yeah. then it would test your patch against it. | 17:11 |
*** TxGirlGeek has joined #openstack-kolla | 17:12 | |
*** sdake_ has quit IRC | 17:16 | |
*** TxGirlGe_ has joined #openstack-kolla | 17:16 | |
SamYaple | so where are we at in this cycle people? is there time for me to get monasca merged into kolla for O? | 17:16 |
Jeffrey4l_ | SamYaple, I think so. | 17:17 |
*** TxGirlGeek has quit IRC | 17:17 | |
SamYaple | at the very least the docker _images_ would make me happy | 17:17 |
pbourke_ | SamYaple: you cant build? | 17:20 |
pbourke_ | currently? | 17:20 |
SamYaple | pbourke_: yesterday | 17:21 |
SamYaple | and you werent here to help | 17:21 |
SamYaple | its fine now | 17:21 |
SamYaple | just go | 17:21 |
SamYaple | pbourke_: it was an issue with requests as Jeffrey4l_ found. all good now | 17:21 |
pbourke_ | I hit that issue | 17:22 |
SamYaple | well its fixed, no thanks to you! | 17:22 |
pbourke_ | spent about an hour debugging, figured out I should be using requests and moved on | 17:22 |
pbourke_ | *using requirements | 17:22 |
*** senk has joined #openstack-kolla | 17:22 | |
pbourke_ | its all thanks to me | 17:22 |
SamYaple | haha well this was my first time building kolla in a few months so i spent a few hours in IRC trying to find out if anyone had built master at all | 17:23 |
SamYaple | last time it "worked" for me was newton b3 i think | 17:23 |
SamYaple | so i had to point of reference for troubleshooting | 17:23 |
pbourke_ | SamYaple: if you get horizon back up I can run my irc bouncer again and provide 24/7 support | 17:24 |
SamYaple | pbourke_: im wokring on it. | 17:25 |
SamYaple | pbourke_: if it makes you feel better, openstack is working just fine in my house | 17:26 |
pbourke_ | but you need images to do that, chicken and egg ;) | 17:26 |
pbourke_ | python packaging is horrible though | 17:30 |
*** jascott1 has joined #openstack-kolla | 17:30 | |
pbourke_ | the dependencies to run the tests for openstack/requirements are hilariously hard to install | 17:31 |
mliima | guys, need some act here https://review.openstack.org/#/c/396626/ inc0 pbourke_ | 17:34 |
mliima | the gate tests are ok after recheck | 17:34 |
*** diogogmt has joined #openstack-kolla | 17:34 | |
kfox1111 | sbezverk: https://review.openstack.org/#/c/404808 looks ready | 17:38 |
kfox1111 | tests passed | 17:38 |
kfox1111 | nm. you saw it. :) | 17:38 |
openstackgerrit | Merged openstack/kolla-kubernetes: Helm common package https://review.openstack.org/404808 | 17:39 |
*** sayantan_ has quit IRC | 17:43 | |
*** senk has quit IRC | 17:44 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize neutron-server https://review.openstack.org/404929 | 17:46 |
sbezverk | kfox1111: after rebasing on recently merged common, I see this: http://paste.openstack.org/show/591168/ | 17:46 |
*** tonanhngo has joined #openstack-kolla | 17:47 | |
*** unicell has quit IRC | 17:47 | |
*** ayoung has quit IRC | 17:48 | |
jascott1 | kfox1111 sbezverk we are shipping PVs? | 17:48 |
*** tonanhngo has quit IRC | 17:48 | |
*** tonanhngo has joined #openstack-kolla | 17:49 | |
sbezverk | jascott1: I think we should, at least rabbitmq has both pv and pvc | 17:52 |
jascott1 | its against the helm recommendations but I dont have strong feelings about it | 17:52 |
sbezverk | jascott1: can you point me to this recommendation? WOuld like to undersatnd reasoning behind it | 17:55 |
qwang | jascott1: why is helm against pv/pvc? | 17:55 |
jascott1 | not against pvc just pv | 17:56 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize openvswitch-agent and l3-agent https://review.openstack.org/404999 | 17:56 |
jascott1 | PV are the actual volumes, and we cant know what storage systems a user has | 17:56 |
kfox1111 | sbezverk: looking... | 17:56 |
kfox1111 | jascott1: I think it doesn't hurt to ship the templates for the pv/pvc's if they are in their own packages. can be skipped then. | 17:56 |
jascott1 | kfox1111 makes sense | 17:57 |
kfox1111 | sbezverk: /root/.helm/repository/local missing? weird. I thought 'helm init' did that for you. | 17:57 |
kfox1111 | do you have an older ~/.helm dir? | 17:58 |
jascott1 | qwang sorry it was k8s docs not helm docs http://kubernetes.io/docs/user-guide/persistent-volumes/ | 17:58 |
jascott1 | "Do not include PersistentVolume objects in the config, since the user instantiating the config may not have permission to create PersistentVolumes." | 17:58 |
jascott1 | but as kfox1111 said, if they are optional i dont see a problem | 17:58 |
kfox1111 | got a meeting to run to. be back in a bit. | 17:58 |
qwang | jascott1: thanks. will look into it | 17:59 |
kfox1111 | jascott1: the pv's shoudl be created by k8s admins, not regular users. and helm's docs are more oriented towards regular users. | 17:59 |
*** Pavo has quit IRC | 17:59 | |
*** athomas has joined #openstack-kolla | 17:59 | |
jascott1 | makes sense | 17:59 |
*** Pavo has joined #openstack-kolla | 17:59 | |
openstackgerrit | Merged openstack/kolla: Remove the "nova-network" container file https://review.openstack.org/404099 | 18:00 |
kfox1111 | jascott1: since k8s 1.5 supports pv less ceph creation, I think we might start switching to just that once 1.5 is released and require it. | 18:00 |
*** eaguilar has joined #openstack-kolla | 18:00 | |
jascott1 | woot! | 18:00 |
sbezverk | jascott1: so we will ship them as options whether they are used or not will be up to the operator | 18:00 |
openstackgerrit | Merged openstack/kolla: Move manila-share service to network node https://review.openstack.org/396626 | 18:00 |
Pavo | evening | 18:01 |
Pavo | got a question anyone have an issue using kolla-ansible with docker on target nodes after they did a yum -y update and docker engine got updated to 1.12? | 18:01 |
*** williamcaban has joined #openstack-kolla | 18:03 | |
*** senk has joined #openstack-kolla | 18:04 | |
Bico_Fino | Anyone using ironic-inspector ? | 18:04 |
Bico_Fino | Trying to deploy an instance and got this on ironic-inspector | 18:06 |
Bico_Fino | InstanceDeployFailure: Couldn't determine the UUID of the root partition or the disk identifier after deploying node | 18:06 |
*** williamcaban has quit IRC | 18:07 | |
openstackgerrit | Paul Bourke (pbourke) proposed openstack/kolla-ansible: Add missing pull task for neutron-vpnaas-agent https://review.openstack.org/401240 | 18:15 |
klindgren | So coming back around to the consistent userid/groupid mapping. | 18:15 |
klindgren | Someone said that they would not have been happy with adding the mappings into globals.yml, and would rather see them just done in code. But if I wanted to overide those mappings into something else. Would it be acceptable to default to a uid/gid but if one is found in globals.yml to use that instead? | 18:16 |
*** senk has quit IRC | 18:17 | |
jascott1 | klindren thats what I was thinking | 18:18 |
*** krtaylor has quit IRC | 18:20 | |
*** klindgren_ has joined #openstack-kolla | 18:22 | |
*** prameswar has quit IRC | 18:22 | |
*** klindgren has quit IRC | 18:23 | |
*** saneax-_-|AFK is now known as saneax | 18:23 | |
openstackgerrit | Merged openstack/kolla-ansible: Remove the "nova-network" from the ansible https://review.openstack.org/404200 | 18:25 |
openstackgerrit | Merged openstack/kolla-ansible: Remove the deprecated command and the useless "if" condition https://review.openstack.org/404896 | 18:25 |
*** klindgren_ is now known as klindgren | 18:25 | |
klindgren | sorry - internet died on me. | 18:26 |
klindgren | jascott1, how would you expect to se that overide happing in globals.yml? | 18:26 |
klindgren | <project>_userid, <project>_groupid | 18:30 |
klindgren | ? | 18:30 |
jascott1 | idk im not kolla expert. looking at the docker templates now | 18:31 |
klindgren | or would a dictionary be better? userid_mapping={'nova':1000, 'glance': 1001} | 18:32 |
jascott1 | the stated problem is taht uid/gid arent consistent which leads to problems. we could start just by using the same for all services | 18:32 |
klindgren | people are against that | 18:32 |
klindgren | where containers need to interact between each other | 18:33 |
klindgren | such as nova-compute with libvirt | 18:33 |
*** mkoderer has quit IRC | 18:33 | |
jascott1 | does kola create a libvirt user? | 18:33 |
klindgren | I think the libvirt package creates the libvirt user | 18:33 |
jascott1 | ok I will defer to the kolla experts :) | 18:33 |
*** vhosakot has joined #openstack-kolla | 18:34 | |
*** shardy has quit IRC | 18:34 | |
*** gfidente has quit IRC | 18:34 | |
jascott1 | ppl might hate this but it might be better to make a separate file for these mappings since we will need a bunch | 18:34 |
klindgren | yea that was my thought and Jeffrey4l_ was not for that idea. Just trying to flush out how people expect this to happen. SO that I can get a spec up. | 18:34 |
klindgren | Since we need to be able to set a consistent userid gfor at a minimum glance. Due to our use of nfs stores in our older deployments | 18:35 |
klindgren | and each container changing the the user/group of all the files on the nfs stores is not a great thing. | 18:35 |
jascott1 | I say post how you think it should work and comments will ensue :) | 18:36 |
sbezverk | folks anybody tried to push PS recently, it looks like gerrit is down | 18:36 |
*** tonanhngo has quit IRC | 18:37 | |
*** msimonin has joined #openstack-kolla | 18:39 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: PS restores two previously removed files https://review.openstack.org/405569 | 18:40 |
*** unicell has joined #openstack-kolla | 18:41 | |
*** unicell1 has joined #openstack-kolla | 18:42 | |
*** unicell has quit IRC | 18:42 | |
jmccarthy | Any nova wizards about ? :) I'm trying to work out about in nova.conf how/what these credentials are used: | 18:46 |
jmccarthy | [neutron] | 18:46 |
jmccarthy | ... | 18:46 |
jmccarthy | auth_type = password | 18:46 |
jmccarthy | username = {{ neutron_keystone_user }} | 18:46 |
jmccarthy | password = {{ neutron_keystone_password }} | 18:46 |
jascott1 | jmccarthy i believe that controls how nova authenticates to keystone | 18:49 |
jascott1 | :) | 18:49 |
*** hyang has joined #openstack-kolla | 18:49 | |
jmccarthy | :) Hehe yea I have a sort of legacy upgrade issue where it's been set to some other username | 18:50 |
v1k0d3n | yo portdirect_meetn | 18:50 |
v1k0d3n | don't mean to bother...but when you're back around... have a Q for you. | 18:50 |
jmccarthy | I'm trying to figure out how I should be switching over so that I can use the {{ neutron_keystone_user }} name, not the old one | 18:51 |
jascott1 | i believe you can change the user/pass on keystone | 18:54 |
jascott1 | wouldnt that fix it? | 18:54 |
jmccarthy | How did we come up with this format ? I can't seem to find openstack docs that list username/password like this ? | 18:55 |
jmccarthy | @jascott1 Yea I think I have tried that but it still didn't seem to work, unless I am just doing something wrong | 18:55 |
*** msimonin has quit IRC | 18:57 | |
klindgren | jmccarthy, I am pretty sure that as long as that user is in the admin "services" tenant keystone it doesn't actually matter? | 18:58 |
*** dave-mccowan has quit IRC | 18:59 | |
dmsimard | o/ hello Kolla, FYI I just tagged a new ARA release, I'll watch your gate in case anything breaks but it shouldn't be the case! Details are here: https://dmsimard.com/drafts/ara-0.10-the-biggest-release-yet-is-out.html | 18:59 |
jmccarthy | klindgren I sort of thought that too, but it does seem to matter, I can only get upgrade working when I have it set to old name - how do I get it to list that ? | 19:00 |
*** williamcaban has joined #openstack-kolla | 19:04 | |
*** mewald has joined #openstack-kolla | 19:04 | |
klindgren | I will have to defer to someone else - I am new to kolla/kolla-ansible as well. Though since its a jinja template. You could modify the template to use {{ nova_keystone_user }}? assuming thats the user you want? | 19:05 |
Bico_Fino | Can I install byfrost and communicate with one Kolla install? | 19:05 |
*** sayantan_ has joined #openstack-kolla | 19:06 | |
hyang | Hi all, I’m trying to use Kolla for all-in-one deployment of OpenStack on CentOS 7 in virtualbox. I’m following the QuickStart doc and when it comes to set up “network_interface” variable in globals.yml, I’m confused about which network interface should I use here. | 19:06 |
hyang | My vm has 1 NAT interface by default, but using that will get error “Hostname has to resolve to IP address of api_interface” in kolla pre-checks. I also tried to use more network interfaces of NAT, Host-only, Bridged, but still got the same error. | 19:07 |
hyang | Could someone point out which network interface type is the correct choice? Appreciate for any help. | 19:07 |
klindgren | But for ease going forward, most likely it would be easier to figure out/fix the problem inside keystone, and/or debug why that doesn't work. | 19:08 |
jmccarthy | klindgren Yea I can adjust template alright, or change property, but I still only get successful upgrade if I use old name from original deploy (and not the newer 'neutron' default I would prefer) - I'm just talking out loud here, I'm not all that familiar with keystone | 19:08 |
*** williamcaban has quit IRC | 19:08 | |
jascott1 | sbezverk should I wait for your last PS to merge before trying to add helmized memcached? | 19:08 |
klindgren | jmccarthy, suggest to look at: http://docs.openstack.org/juno/install-guide/install/apt/content/neutron-controller-node.html | 19:09 |
klindgren | and verify that the output of some of the role/tenant commands to make sure the user is correctly added and an admin | 19:10 |
*** tyrola has joined #openstack-kolla | 19:12 | |
jmccarthy | Ok cool, yea trying to look into this sort of thing now - thanks ! | 19:15 |
sean-k-mooney | mliima: hi, so i think i know why my manila issues are happening. | 19:15 |
mliima | really? | 19:16 |
sean-k-mooney | mliima: i am using the ovs conntrack security group diriver and i think it is blocking the comunnitation between the manila_share agent and the vm | 19:16 |
mliima | hmm... | 19:16 |
sean-k-mooney | arp requests and responces are working but icmp and ssh are being blocked | 19:17 |
*** vhosakot has quit IRC | 19:17 | |
sbezverk | jascott1: I think you can go ahead with memcached | 19:17 |
mliima | maybe we should add something in our guide sean-k-mooney | 19:18 |
jascott1 | ok things are changing quickly so feels like frogger trying to dodge conflicts | 19:18 |
*** vhosakot has joined #openstack-kolla | 19:18 | |
sean-k-mooney | the irritating thing is that i can ssh fine from vms to it just not from the tap devices that are created on the host for the agent to connect too. | 19:18 |
sbezverk | common part got merged, the rest has no impact on memcached | 19:18 |
sayantan_ | Hi. I have a quick question. How do I wrap a line to have 79 characters? | 19:19 |
jascott1 | sbezverk should I make it dependent on 404808 (helm common package)? | 19:19 |
sean-k-mooney | mliima: i would have to do a little diggin but im guessing there is a bug in neutron or manila and the prots used by the agent are not in the correct secutiy group | 19:20 |
mliima | but now, are you able to create a share sean-k-mooney ? | 19:20 |
jascott1 | oh it doesnt matter cause its already merged right? sorry Im newish to gerrit | 19:20 |
sbezverk | jascott1: it got merged, so just clone the master, it should be there | 19:20 |
sean-k-mooney | mliima: no i have not reverted to iptables | 19:20 |
jascott1 | sbezverk thank you | 19:20 |
sbezverk | jascott1: np | 19:20 |
sean-k-mooney | mliima: if i did i probaly could | 19:20 |
mliima | that's weird, I created a share yesterday | 19:20 |
sean-k-mooney | mliima: with the ovs conntrack securty group dirver enabeld? | 19:21 |
sean-k-mooney | on stable/newton | 19:21 |
mliima | how can i check it quickly sean-k-mooney ? | 19:22 |
*** classicsnail has quit IRC | 19:22 | |
mliima | I have nothing ready to check now :/ | 19:22 |
*** vhosakot has quit IRC | 19:22 | |
sean-k-mooney | mliima: you need to add one line to the ml2_conf.ini and run a reconfigure. but i dont expect to go and check it out. if you have time that woudl be cool but this was just and fyi | 19:24 |
mliima | i'll check it tomorrow | 19:24 |
*** srwilkers_ has quit IRC | 19:24 | |
mliima | ok? | 19:24 |
*** Bico_Fino has quit IRC | 19:24 | |
sean-k-mooney | mliima: add [securitygroup] | 19:24 |
sean-k-mooney | firewall_driver = openvswitch | 19:24 |
*** sdake has joined #openstack-kolla | 19:25 | |
mliima | firewall_driver = neutron.agent.linux.iptables_firewall.OVSHybridIptablesFirewallDriver | 19:25 |
mliima | sean-k-mooney, ? | 19:25 |
sean-k-mooney | mliima: yes that is the default iptables firewall. i am useing the alternitve openvswitch firewall dirver that uses openflow + conntrack instead of iptables | 19:26 |
*** matrohon has joined #openstack-kolla | 19:26 | |
sean-k-mooney | mliima: iptables is really slow in comparision | 19:26 |
mliima | hmm... | 19:27 |
*** eaguilar has quit IRC | 19:27 | |
mliima | that's the our problem sean-k-mooney | 19:27 |
mliima | i'm using neutron.agent.linux.iptables_firewall.OVSHybridIptablesFirewallDriver | 19:27 |
mliima | https://github.com/openstack/kolla/blob/stable/newton/ansible/roles/neutron/templates/ml2_conf.ini.j2#L42 | 19:28 |
sean-k-mooney | mliima: basically i can see the drop rule in the ovs br-int that is dropping all the traffic and i know that rule was installed by the openflow firewall | 19:28 |
mliima | but tommorow i check it | 19:28 |
*** david-lyle_ has joined #openstack-kolla | 19:28 | |
mliima | and fill a bug to solve it | 19:28 |
sean-k-mooney | mliima: cool. you should be able to swap between those two driver freely but the openvswitch driver is very new and i dont think it is handeling ports that are not attached to a vm correctly | 19:29 |
mliima | okay | 19:30 |
mliima | you can ask to manila community guys | 19:30 |
mliima | #openstack-manila | 19:30 |
mliima | about it | 19:31 |
mliima | :) | 19:31 |
*** david-lyle has quit IRC | 19:31 | |
mliima | i have to go now | 19:31 |
*** mliima has quit IRC | 19:31 | |
*** srwilkers_ has joined #openstack-kolla | 19:31 | |
openstackgerrit | satroutr proposed openstack/kolla-ansible: RabbitMQ cluster should work on api interface https://review.openstack.org/404944 | 19:32 |
sean-k-mooney | mliima: ill bring this up in the neutron channel too because i think this is there issue | 19:32 |
*** openstackgerrit has quit IRC | 19:32 | |
*** openstackgerrit has joined #openstack-kolla | 19:33 | |
openstackgerrit | Mathias Ewald proposed openstack/kolla-ansible: Fix X-Forward-Proto Header Deletion https://review.openstack.org/405589 | 19:36 |
*** classicsnail has joined #openstack-kolla | 19:36 | |
*** openstackgerrit has quit IRC | 19:36 | |
*** tyrola has quit IRC | 19:37 | |
qwang | Is there anyone working on helm chart for horizon right now? I want to pick up this one | 19:39 |
*** mewald has quit IRC | 19:40 | |
*** krtaylor has joined #openstack-kolla | 19:40 | |
*** saneax is now known as saneax-_-|AFK | 19:40 | |
*** eaguilar has joined #openstack-kolla | 19:40 | |
*** dave-mccowan has joined #openstack-kolla | 19:41 | |
jascott1 | qwang https://blueprints.launchpad.net/kolla-kubernetes/+spec/helm-packaging | 19:42 |
jascott1 | edit that list and add your name to horizon if want :) | 19:42 |
kfox1111 | back | 19:43 |
jascott1 | kfox1111 I created a deployment for memcache thats a full yaml but I also created a version using you new common tpl... I dont know alot about memcached use cases with openstack, do you think the common format will be fine? | 19:46 |
kfox1111 | which common template? common lib? | 19:46 |
jascott1 | like this https://review.openstack.org/#/c/404929/7/helm/microservice/neutron-server/templates/neutron_server.yaml | 19:47 |
kfox1111 | I've been very weary of memcached in general. :/ | 19:47 |
*** openstackgerrit has joined #openstack-kolla | 19:47 | |
openstackgerrit | Pete Birley proposed openstack/kolla-kubernetes: Kolla-Kubernetes Vagrant development environment documentation https://review.openstack.org/404493 | 19:47 |
*** msimonin has joined #openstack-kolla | 19:47 | |
kfox1111 | there's a gotcha in its scaling that most people don't pay too much attention to. | 19:47 |
kfox1111 | although, k8s makes it proably irelevent. | 19:47 |
kfox1111 | welll... still probably has... | 19:48 |
kfox1111 | split brained systems. | 19:48 |
jascott1 | ah | 19:48 |
kfox1111 | ok. so. common deployment.. hmm... | 19:49 |
inc0 | I usually don't assume memcache to be HA anyway | 19:49 |
inc0 | and don't deploy it as such | 19:49 |
jascott1 | ya my lack of openstack exp really hurts me here cause I dont really know all the ways memcached is used | 19:50 |
kfox1111 | since it has no stateful tracking, and while it has state, its explicitly expected that the state can vanish. | 19:50 |
jascott1 | i made it a deployment, should it just be a pod then? | 19:50 |
kfox1111 | so a k8s deployment style object with replicas=1 is probably a good fit. | 19:50 |
kfox1111 | haproxy in front is overkill though. | 19:50 |
jascott1 | well it does have persistence. sounds like openstack doesnt use it tho? | 19:50 |
kfox1111 | memcached has persistance? | 19:50 |
jascott1 | ya | 19:51 |
inc0 | jascott1, not really, it's mostly used for token cache | 19:51 |
jascott1 | i dont think anyone use it | 19:51 |
inc0 | why would you | 19:51 |
jascott1 | cool | 19:51 |
jascott1 | right | 19:51 |
jascott1 | :) | 19:51 |
inc0 | redis is SO MUCH BETTER in this space | 19:51 |
kfox1111 | jascott1: it use the persistance for a hot cache? | 19:51 |
jascott1 | oh hell I was thinking about redis sorry | 19:51 |
kfox1111 | ah. yeah. thats what I was thinking. :) | 19:52 |
jascott1 | i blame the cold I am apparently contracting | 19:52 |
inc0 | yeah, different story | 19:52 |
kfox1111 | jascott1: yeah. its going around. I've had it for about 4 days now. :/ | 19:52 |
kfox1111 | I want to try out redis at some point, but never got around to it yet. | 19:52 |
jascott1 | its fast | 19:52 |
kfox1111 | inc0: so, I think common_deployment is probably overkill for memcached. would recommend making a deployment based base template and link to common_lib though. | 19:53 |
jascott1 | inc0 for compute kit where is memcached used? | 19:53 |
kfox1111 | jascott1: it can be used in just about all the api services. | 19:54 |
inc0 | kfox1111, redis is awesome. I wrote mechanism to deal with BGP one day on it, loading of 500k routes took 20s on small vm | 19:54 |
kfox1111 | inc0: nice. :) | 19:54 |
inc0 | embedded lua is also insane | 19:54 |
jascott1 | TTL feature is great for tokens | 19:54 |
jascott1 | inc0 staff not canceled? thought that was to be the case | 19:55 |
sdake | inc0 are you talking about heka? | 19:55 |
jascott1 | kfox1111 I have a memcached deployment not based on neutron server as well so I will jsut roll with that | 19:56 |
inc0 | sdake, no, redis | 19:56 |
sdake | yikes | 19:56 |
sean-k-mooney | jascott1 i belive memcached is only used in the controler by the api services | 19:56 |
sdake | didn't know that used lua | 19:56 |
inc0 | it has embedded lua interpreter | 19:56 |
inc0 | really insanely fast | 19:56 |
sdake | is it used for anything? | 19:56 |
inc0 | you write lua script to deal with redis data and you have extremally efficient data store | 19:57 |
openstackgerrit | Pete Birley proposed openstack/kolla-kubernetes: Kolla-Kubernetes Vagrant development environment documentation https://review.openstack.org/404493 | 19:57 |
sean-k-mooney | inc0: is the lua interperated uses as part db querys | 19:57 |
sdake | interesting | 19:57 |
sdake | haven't tried out redis | 19:57 |
inc0 | sean-k-mooney, kinda | 19:57 |
sdake | i've started it before but never actually used it ;) | 19:57 |
inc0 | you literally put lua code as one of records and do EXEC;) | 19:58 |
sean-k-mooney | inc0: oh i could do such bad things with that | 19:58 |
kfox1111 | inc0: what could go wrong with having a turing complete language in your db query language.... ;) | 19:58 |
kfox1111 | insanly powerful... | 19:58 |
kfox1111 | but with great power... :) | 19:58 |
jascott1 | it also has sets and sorted lists and other k/v structures | 19:58 |
inc0 | well SQL is turing complete | 19:58 |
sdake | kfox1111 we got languages for your languages, YO DAWG | 19:59 |
jascott1 | haha | 19:59 |
kfox1111 | inc0: strict sql isn't. | 19:59 |
*** Pavo has quit IRC | 19:59 | |
sean-k-mooney | incO: sql cant load system libs written in c and the exec arbitry function calls | 19:59 |
kfox1111 | no explicit loops and such. | 20:00 |
inc0 | well lua in redis can't as awell | 20:00 |
sean-k-mooney | inc0: if you have embeded you lua interperter correctly it cant either but lua itsefl can | 20:00 |
kfox1111 | though once you get into the embeded trigger languages, then you would be better off with lua or some standard language rather then the db specific languages. | 20:00 |
inc0 | you have stdlib plus some stuff for redis | 20:00 |
inc0 | and that's it | 20:00 |
kfox1111 | since you are turing complete at that point. | 20:00 |
sean-k-mooney | anyway reddis is used everywhere | 20:01 |
sean-k-mooney | im sure they have looked at the securtiy | 20:01 |
*** david-lyle_ is now known as david-lyle | 20:01 | |
inc0 | anyway, if you ever find yourself needing something super fast to deal with even complex logic around k/v stor, try redis+lua | 20:01 |
kfox1111 | itsa tradoff, as everything. security is not a bool. | 20:01 |
*** stevebaker has joined #openstack-kolla | 20:02 | |
inc0 | jascott1, confirmed - no staff | 20:02 |
jascott1 | k thanks | 20:02 |
openstackgerrit | Pete Birley proposed openstack/kolla-kubernetes: Kolla-Kubernetes Vagrant development environment documentation https://review.openstack.org/404493 | 20:03 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize neutron-server https://review.openstack.org/404929 | 20:04 |
sean-k-mooney | so what kicked of the reddis conversation by the way | 20:04 |
*** vhosakot has joined #openstack-kolla | 20:04 | |
*** Pavo has joined #openstack-kolla | 20:04 | |
jascott1 | i got confused about memcached and started attributing redis features to it | 20:04 |
inc0 | I'm sure there is medical name for it | 20:05 |
jascott1 | cause I use redis and not memcache in prev life | 20:05 |
sean-k-mooney | jascott1: ah ok | 20:05 |
*** williamcaban has joined #openstack-kolla | 20:05 | |
jascott1 | coldbrain | 20:05 |
openstackgerrit | Pete Birley proposed openstack/kolla-kubernetes: Kolla-Kubernetes Vagrant development environment documentation https://review.openstack.org/404493 | 20:05 |
inc0 | I was thinking more in lines computer-systemic synesthesia | 20:06 |
sean-k-mooney | well memcached is not presetend though so they have slightly different target usecase but both are k/v stores at the end of the day | 20:06 |
jascott1 | sounds like a fun VR game | 20:06 |
jascott1 | Pre order "Computer-Systemic Synesthesia"! | 20:07 |
jascott1 | inc0 I got kids psvr afterall | 20:07 |
jascott1 | for xmas | 20:07 |
jascott1 | not a vive but better than nothing ;) | 20:08 |
openstackgerrit | Pete Birley proposed openstack/kolla-kubernetes: Kolla-Kubernetes Vagrant development environment documentation https://review.openstack.org/404493 | 20:08 |
jascott1 | "for the kids" | 20:08 |
inc0 | I hope you won't need ps pro to run it effectivelly | 20:08 |
sean-k-mooney | what you really want is the comfort/build quality of the rift with the usablity of vive and price of psvr | 20:08 |
inc0 | I'm ok with comfort of vive | 20:09 |
jascott1 | what ive read so far, the differences are only in a few titles right now most of the comparisons I can barely tell any improvement | 20:09 |
inc0 | jascott1, in vive you can walk around | 20:09 |
inc0 | in oculus you can't | 20:09 |
kfox1111 | jascott1: played with the psvr yet? been meaning to give it a shot. | 20:09 |
*** williamcaban has quit IRC | 20:09 | |
jascott1 | I did at gamestop also vive | 20:10 |
kfox1111 | jascott1: played with the oculus and vive. | 20:10 |
kfox1111 | vive's amazing. :) | 20:10 |
jascott1 | vive is best | 20:10 |
sean-k-mooney | inc0: well you can but they said they were originally targeting sitting with the rift | 20:10 |
openstackgerrit | Pete Birley proposed openstack/kolla-kubernetes: Kolla-Kubernetes Vagrant development environment documentation https://review.openstack.org/404493 | 20:10 |
jascott1 | havent played the released occulus tho | 20:10 |
kfox1111 | onlything stopping me from pulling the trigger on it, is the resolution. if they can upgrade it to a 4k display fro the same price, I'm sold. :) | 20:10 |
kfox1111 | heard cheep 4k screens are becoming available early next year. | 20:10 |
sean-k-mooney | inc0: so next openstack kolla demo. vive experience for kolla k8s so you can interact with your cloud in vr :) | 20:11 |
kfox1111 | hehe | 20:11 |
jascott1 | haha someone made that for k8s already | 20:11 |
inc0 | docker exec -it vr bash | 20:11 |
inc0 | now I'm *INSIDE* container | 20:11 |
kfox1111 | oooo :) | 20:11 |
*** tyrola has joined #openstack-kolla | 20:11 | |
kfox1111 | aahhhh :) | 20:11 |
inc0 | jascott1, say what? link plz | 20:12 |
inc0 | I'll test it when I get home | 20:12 |
inc0 | yeah resolution is a bit iffy | 20:12 |
inc0 | but on the bright note you get used to it pretty fast | 20:12 |
jascott1 | https://www.youtube.com/watch?v=Qb1siOeTIrk | 20:13 |
inc0 | and usually your sensors are overloaded anyway | 20:13 |
inc0 | my god this is brillian | 20:13 |
inc0 | t | 20:13 |
jascott1 | a good start | 20:13 |
sean-k-mooney | well if you had a 4k vive and a rack of servers to power it you would not need monitors anymore as the resolution would be good enough to read fine text | 20:13 |
jascott1 | he used mozilla's aframe | 20:13 |
jascott1 | i wouldve used unity | 20:14 |
inc0 | jascott1, let's write it | 20:14 |
inc0 | I mean it | 20:14 |
*** portdirect__ has joined #openstack-kolla | 20:14 | |
jascott1 | sounds fun to me | 20:14 |
inc0 | let's write a deployment tool that uses k8s and vive | 20:14 |
srwilkers_ | can we? | 20:14 |
kfox1111 | inc0: heh. | 20:14 |
srwilkers_ | please? | 20:14 |
jascott1 | i could see servers emitting colored particle systems to represent something | 20:14 |
srwilkers_ | :p | 20:14 |
kfox1111 | I'd rather see it used for monitoring. | 20:14 |
jascott1 | like errors or connectino problems | 20:14 |
inc0 | kfox1111, one word - tmux | 20:15 |
kfox1111 | lets you see the nodes. then grab one and stretch it to see all the services on it, | 20:15 |
kfox1111 | etc. :) | 20:15 |
inc0 | "hold on, I had this running tcpdump somewhere around my left foot" | 20:15 |
jascott1 | ala lawnmower man | 20:15 |
kfox1111 | then touch a service and it shows you the logs. | 20:15 |
kfox1111 | :) | 20:15 |
jascott1 | haha | 20:15 |
kfox1111 | can touch scroll the logs liike a tablet.... :) | 20:15 |
kfox1111 | etc. | 20:16 |
*** tyrola_ has joined #openstack-kolla | 20:16 | |
sean-k-mooney | hehe i dont think this was the original plan for a kolla ui but hay i have no issues with it lol | 20:16 |
jascott1 | could have hundreds of containers like a city and see whih ones are 'on fire' | 20:16 |
kfox1111 | jascott1: heh. ever see 'doom ps' ? :) | 20:16 |
jascott1 | no whats that? | 20:16 |
kfox1111 | http://psdoom.sourceforge.net/ | 20:17 |
kfox1111 | but make it vr. :) | 20:17 |
openstackgerrit | Pete Birley proposed openstack/kolla-kubernetes: Kolla-Kubernetes Vagrant development environment documentation https://review.openstack.org/404493 | 20:17 |
kfox1111 | "what are you doing? you can't play doom at work" | 20:17 |
inc0 | killall them | 20:18 |
kfox1111 | "I'm admining the systems..." :) | 20:18 |
sean-k-mooney | kfox1111: well there is mindcraft and second life intergration with openstack already. gamify all the things lol | 20:18 |
kfox1111 | hehe. cool. | 20:18 |
jascott1 | its novel to us but i am convinced we will use the interfaces someday... of course they will be much better I would think. looking at you magic leap https://www.wired.com/2016/04/magic-leap-vr/ | 20:21 |
*** goldyfruit has quit IRC | 20:22 | |
jascott1 | ^ these are the ones they say will change everything | 20:22 |
kfox1111 | yeah. been keeping an eye on them too. :) | 20:23 |
portdirect__ | v1k0d3n: ping | 20:23 |
portdirect__ | "<inc0> let's write a deployment tool that uses k8s and vive" <<-- this is less hard than you would think | 20:25 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize openvswitch-agent and l3-agent https://review.openstack.org/404999 | 20:26 |
kfox1111 | portdirect__: you need a visual pipeline programming language in vr. then plug in the helm stuff as building blocks. | 20:27 |
kfox1111 | then you just plug each thing together :) | 20:27 |
srwilkers_ | portdirect__, this is looking good man :) | 20:29 |
portdirect__ | kfox1111: exactly :) Pointless but fun - it would take about 2 days to peice together from the k8s toplology map thats up on github | 20:30 |
portdirect__ | srwilkers_: cheers dude, just want to get it to bed asap now - so I can get on with building services using the work that kfox and sbezverk have done :) | 20:31 |
*** vhosakot has quit IRC | 20:32 | |
*** vhosakot has joined #openstack-kolla | 20:32 | |
srwilkers_ | portdirect__, absolutely | 20:33 |
srwilkers_ | portdirect__, i mentioned the blueprint for that guide in my most recent review as well. i changed you to the assignee, so if you add the implements blueprint tag, itll link back to the blueprint in launchpad :) | 20:33 |
openstackgerrit | Pete Birley proposed openstack/kolla-kubernetes: Kolla-Kubernetes Vagrant development environment documentation https://review.openstack.org/404493 | 20:33 |
srwilkers_ | and youll get credit for it :) | 20:33 |
*** matrohon has quit IRC | 20:34 | |
portdirect__ | srwilkers_: sweet dude :) Also means I get to continue dos'ing irc with openstackgerritbot | 20:35 |
srwilkers_ | portdirect__, >:) | 20:35 |
*** sdake_ has joined #openstack-kolla | 20:37 | |
srwilkers_ | sup sdake | 20:37 |
openstackgerrit | Pete Birley proposed openstack/kolla-kubernetes: Kolla-Kubernetes Vagrant development environment documentation https://review.openstack.org/404493 | 20:38 |
*** unicell1 has quit IRC | 20:38 | |
*** goldyfruit has joined #openstack-kolla | 20:40 | |
*** sdake has quit IRC | 20:40 | |
*** openstack has joined #openstack-kolla | 20:46 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize neutron-server https://review.openstack.org/404929 | 20:46 |
openstackgerrit | Pete Birley proposed openstack/kolla-kubernetes: Kolla-Kubernetes Vagrant development environment documentation https://review.openstack.org/404493 | 20:47 |
*** mariso has quit IRC | 20:47 | |
inc0 | portdirect_ I found typo mwhahahah! | 20:47 |
inc0 | mwhahahahhahahaha << evil >> | 20:47 |
*** adrian_otto has quit IRC | 20:48 | |
srwilkers_ | haha | 20:49 |
portdirect__ | inc0: srwilkers made me do it | 20:49 |
srwilkers_ | tis true | 20:49 |
inc0 | but you had suffered for it | 20:49 |
inc0 | kudos to srwilkers | 20:49 |
portdirect__ | :) | 20:50 |
inc0 | new metallica album - great | 20:50 |
srwilkers_ | inc0, im torn on it | 20:51 |
srwilkers_ | inc0, james sounds flat | 20:51 |
inc0 | it sounds to me like death magnetic | 20:51 |
inc0 | which I liked a lot | 20:51 |
*** tyrola_ has quit IRC | 20:51 | |
srwilkers_ | that one wasnt bad | 20:51 |
srwilkers_ | ill probably catch flak for it, but i really liked saint anger | 20:52 |
inc0 | ahh, ok, I'm not big st anger fan;) | 20:52 |
openstackgerrit | Merged openstack/kolla-ansible: clean up kolla related files https://review.openstack.org/404330 | 20:52 |
srwilkers_ | not many people are :p | 20:52 |
inc0 | I'm more ride the lightning guy than battery | 20:52 |
srwilkers_ | yes, RTL is my favorite | 20:53 |
inc0 | +1 to that | 20:53 |
kfox1111 | inc0: interesting. I'll have to give it a shot. :) | 20:53 |
*** JRobinson__ has joined #openstack-kolla | 20:54 | |
inc0 | ok, enough with this girly metallica, let's play something tad more nordic;) | 20:55 |
*** vhosakot has joined #openstack-kolla | 20:55 | |
*** vhosakot has quit IRC | 20:55 | |
kfox1111 | :) | 20:55 |
inc0 | https://www.youtube.com/watch?v=pwU22xcDiJ4 | 20:57 |
sbezverk | kfox1111: need your ack on this one: https://review.openstack.org/405569 | 21:00 |
*** signed8bit has joined #openstack-kolla | 21:02 | |
kfox1111 | sbezverk: I thought we decided we were going to just do a precreate after checkout rather then check it in? | 21:04 |
*** Pavo has quit IRC | 21:04 | |
*** vhosakot has joined #openstack-kolla | 21:05 | |
*** Pavo has joined #openstack-kolla | 21:06 | |
*** Satya_ has quit IRC | 21:06 | |
sbezverk | kfox1111: if somebody wants to deploy manually and check out kolla-kubernetes, then missing these files will prevent chart from being deployed | 21:06 |
kfox1111 | sbezverk: sure. but in the document that tells folks how to use the repo, could just document "git checkout repo, run tools/helm_prebuild.py; cd helm/microservices; helm install ..... | 21:08 |
sbezverk | kfox1111: I am afraid it will take 1. time to get used to run helm_prebuild even before trying to deploy | 21:08 |
*** lrensing has quit IRC | 21:09 | |
sbezverk | also to be consistent with other microsercices | 21:09 |
sbezverk | which will include these files | 21:10 |
kfox1111 | sbezverk: there isn't much consistency here I think. the incubator for example, | 21:11 |
kfox1111 | depends on binary built packages in a seperate repo I think. | 21:11 |
kfox1111 | I think the build of build of a repo with multiple helm packages is kind of uncharted teritory. :/ | 21:12 |
kfox1111 | I rally want a build all capability. | 21:12 |
openstackgerrit | sayantani proposed openstack/kolla: Add new guide for single-node deployment https://review.openstack.org/404993 | 21:12 |
kfox1111 | for example, | 21:13 |
kfox1111 | github.com/kubernetes/charts | 21:13 |
kfox1111 | you can't build it by itself. | 21:13 |
kfox1111 | see: | 21:13 |
kfox1111 | https://github.com/kubernetes/charts/blob/master/stable/owncloud/requirements.yaml | 21:13 |
kfox1111 | it points to a prebuilt binary helm package for mariadb. | 21:14 |
kfox1111 | so, you can't just tweak the mariadb chart and rebuild everything. | 21:14 |
kfox1111 | :/ | 21:14 |
*** eaguilar has quit IRC | 21:16 | |
*** fguillot has quit IRC | 21:17 | |
Pavo | anyone know anything about using custom configs with kolla-ansible? | 21:18 |
sbezverk | kfox1111: I was not aware that we are going to re-use template coming with helm for mariadb | 21:18 |
kfox1111 | sbezverk: not quite the point I'm tryhing to make. | 21:18 |
*** eaguilar has joined #openstack-kolla | 21:18 | |
kfox1111 | just saying, other places, for example the github.com/kubernetes/charts do things differently, and not nessisarily better. so we should be careful doing things like others just because they do it a particular way. its not nessisarily better. | 21:19 |
kfox1111 | if a user wants something super convenient, being able to just checkout and build, maybe the best use case for that really is to privde them the packages prebuilt. | 21:20 |
openstackgerrit | sayantani proposed openstack/kolla: Add new guide for single-node deployment https://review.openstack.org/404993 | 21:21 |
*** imcsk8 has quit IRC | 21:21 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Hel-ming openvswitch-vswitchd https://review.openstack.org/405107 | 21:22 |
*** hyang has quit IRC | 21:22 | |
*** Pavo has quit IRC | 21:23 | |
sbezverk | kfox1111: here is the problem I have with this approach. if I do not have these files I create them with helm_prebuild, it creates the files, but then git detects new files and if I need to send new commit it tries to add these files. so I need to remove them explicitely | 21:24 |
*** vhosakot has quit IRC | 21:24 | |
kfox1111 | sbezverk: hmm... I use adifferent git commit workflow, so I dont notice that.... maybe its a .gitignore thing? | 21:26 |
*** mewald has joined #openstack-kolla | 21:28 | |
*** inc0 has quit IRC | 21:28 | |
kfox1111 | maybe we add '**/charts' and 'values.yaml' to .gitignore? | 21:29 |
*** vhosakot has joined #openstack-kolla | 21:30 | |
sbezverk | kfox1111: that could helm, cause it is very annoying | 21:30 |
sbezverk | oops helm/help | 21:30 |
*** adrian_otto has joined #openstack-kolla | 21:30 | |
srwilkers_ | i dont think thats a terrible idea kfox1111 | 21:31 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Hel-ming Rabbitmq https://review.openstack.org/405422 | 21:31 |
kfox1111 | k. give that a shot and see if it helps. if so, then lets merge it asap. | 21:31 |
kfox1111 | I don't use that workflow, so I'm not sure how to test it. | 21:31 |
kfox1111 | I think yo ucan just add those two lines into kolla-kubernetes/.gitignore | 21:31 |
*** TxGirlGe_ has quit IRC | 21:33 | |
*** TxGirlGeek has joined #openstack-kolla | 21:33 | |
*** Jeffrey4l_ has quit IRC | 21:34 | |
openstackgerrit | Kris Lindgren proposed openstack/kolla: Spec to enable consistent use of user id's and group id's https://review.openstack.org/405647 | 21:36 |
*** Jeffrey4l_ has joined #openstack-kolla | 21:37 | |
sbezverk | kfox1111: will do that soon | 21:40 |
*** rhallisey has quit IRC | 21:41 | |
openstackgerrit | Merged openstack/kolla: Install trove dashboard into horizon image https://review.openstack.org/405087 | 21:41 |
*** JRobinson__ is now known as JRobinson_afk | 21:42 | |
*** portdirect__ has quit IRC | 21:44 | |
*** diogogmt has quit IRC | 21:45 | |
kfox1111 | sbezverk: https://review.openstack.org/#/c/404999/ finally checked out. | 21:45 |
*** mgiles has quit IRC | 21:46 | |
kfox1111 | sbezverk: k. thanks. | 21:46 |
*** vhosakot_ has joined #openstack-kolla | 21:47 | |
srwilkers_ | gunna head home -- have a few things to do, then have a few questions for you kfox1111 if you'll have a little bit of bandwidth either later this evening or tomorrow morning :) | 21:48 |
kfox1111 | srwilkers_: sounds good. | 21:48 |
*** jrist has quit IRC | 21:49 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Adds filtering to .gitignore for values.yaml and **/charts https://review.openstack.org/405661 | 21:49 |
*** vhosakot has quit IRC | 21:50 | |
sbezverk | kfox1111: tested .gitignore works well, I just hope there will not be any corner cases where we do need to change values.yaml or something in ./charts | 21:50 |
*** JRobinson_afk is now known as JRobinson_ | 21:51 | |
kfox1111 | sbezverk: cool. | 21:51 |
kfox1111 | heh. yeah. there are always corner cases. :/ thats why I always git add everything explicitly, git diff --cached before commits and only git commit from staging. | 21:52 |
kfox1111 | more work, but less risky. always tradeoffs there. | 21:52 |
openstackgerrit | Kris Lindgren proposed openstack/kolla: Spec to enable consistent use of user id's and group id's https://review.openstack.org/405647 | 21:52 |
*** srwilkers_ has quit IRC | 21:53 | |
kfox1111 | seeing some weirdness out of helm at times.... | 21:54 |
mewald | Guys, we have an empty /etc/cinder/rootwrap.d/ for cinder-backup. This currently breaks my plans to add nfs driver for cinder backup as it tries to run "mount" but fails | 21:54 |
kfox1111 | http://logs.openstack.org/29/404929/10/experimental/gate-kolla-kubernetes-deploy-centos-binary-ceph-multi-nv/e12aba9/console.html#_2016-12-01_21_47_05_669946 | 21:54 |
openstackgerrit | Merged openstack/kolla-kubernetes: Adds filtering to .gitignore for values.yaml and **/charts https://review.openstack.org/405661 | 21:55 |
*** krtaylor has quit IRC | 21:56 | |
sbezverk | kfox1111: the same issue att guys reported to helm people | 21:56 |
sbezverk | kfox1111: but I never hit it in baremetal env. | 21:57 |
kfox1111 | sbezverk: oh, they saw that too? interesting. | 21:58 |
kfox1111 | sbezverk: yeah. this is the first I've seen it too. and not always. just sometimes. | 21:58 |
kfox1111 | and only a few times today. not before. | 21:58 |
kfox1111 | how long ago did the att folks spot it? | 21:59 |
sbezverk | kfox1111: not sure but I saw some discussions of helm slack yesterday | 21:59 |
kfox1111 | ah. | 21:59 |
sbezverk | need to ping v1k0d3n | 22:00 |
sbezverk | I think he can really confirm if it is the same or not | 22:00 |
sbezverk | kfox1111: common defines are not merged yet? | 22:01 |
kfox1111 | sbezverk: should be. | 22:01 |
sbezverk | kfox1111: the one where you define entrypoint image name right? | 22:02 |
kfox1111 | yeah. that should be merged. | 22:02 |
kfox1111 | need a rebase maybe? | 22:02 |
*** sayantan_ has quit IRC | 22:03 | |
*** openstackgerrit has quit IRC | 22:03 | |
sbezverk | kfox1111: no I just forgot to include it :-( | 22:04 |
*** sayantan_ has joined #openstack-kolla | 22:05 | |
*** sayantan_ has quit IRC | 22:06 | |
*** sayantan_ has joined #openstack-kolla | 22:06 | |
*** williamcaban has joined #openstack-kolla | 22:06 | |
*** unicell has joined #openstack-kolla | 22:07 | |
*** goldyfruit has quit IRC | 22:08 | |
sbezverk | kfox1111: I do not think entrypoint image was pushed to docker hub with 3.0.1 tag | 22:08 |
*** unicell has quit IRC | 22:08 | |
sbezverk | kfox1111: can we push it there? | 22:08 |
*** unicell has joined #openstack-kolla | 22:09 | |
mewald | https://bugs.launchpad.net/kolla/+bug/1646630 | 22:10 |
openstack | mewald: Error: Could not gather data from Launchpad for bug #1646630 (https://launchpad.net/bugs/1646630). The error has been logged | 22:10 |
*** williamcaban has quit IRC | 22:11 | |
kfox1111 | sbezverk: would have to ask inc0. I think we should though. | 22:13 |
sbezverk | I think so too, otherwise we will have a problem at the gate | 22:16 |
*** openstackgerrit has joined #openstack-kolla | 22:16 | |
sbezverk | but I guess inc0 is already on PTO, so will check with sdake | 22:16 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize neutron-server https://review.openstack.org/404929 | 22:16 |
kfox1111 | sounds good. | 22:16 |
*** vhosakot_ has quit IRC | 22:17 | |
kfox1111 | sbezverk: it sounds like helm search doesn't and package downloading are client actions. | 22:17 |
kfox1111 | and helm init returns before things are started. | 22:17 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Hel-ming openvswitch-vswitchd https://review.openstack.org/405107 | 22:17 |
kfox1111 | so there is a race there. | 22:17 |
kfox1111 | I added a new barior to see if it will block until its ready. should fix the issue I think. | 22:18 |
*** portdirect_meetn is now known as portdirect | 22:19 | |
*** mewald has quit IRC | 22:21 | |
*** jtriley has quit IRC | 22:21 | |
sbezverk | kfox1111: let's hope so | 22:22 |
*** jheroux has quit IRC | 22:23 | |
*** jascott1 has quit IRC | 22:29 | |
*** DaveTurner has joined #openstack-kolla | 22:30 | |
*** jascott1 has joined #openstack-kolla | 22:30 | |
openstackgerrit | sayantani proposed openstack/kolla: Corrected Punctuation in quickstart.rst https://review.openstack.org/404907 | 22:31 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize neutron-server https://review.openstack.org/404929 | 22:32 |
*** eaguilar has quit IRC | 22:33 | |
*** krtaylor has joined #openstack-kolla | 22:42 | |
portdirect | sbezverk: hel-ming? http://static.giantbomb.com/uploads/original/0/1987/853982-ming.jpg | 22:47 |
*** signed8bit is now known as signed8bit_Zzz | 22:47 | |
sbezverk | portdirect: lol | 22:52 |
*** adrian_otto has quit IRC | 22:55 | |
kfox1111 | portdirect: http://www.johnmsteele.com/wp-content/uploads/2013/06/AD&D/Settings/GammaWrld/GammaWorld/Warhamster.jpg | 22:55 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize neutron-server https://review.openstack.org/404929 | 22:57 |
*** athomas has quit IRC | 22:57 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Hel-ming openvswitch-vswitchd https://review.openstack.org/405107 | 23:00 |
sbezverk | where do you guys find these pictures ;-) | 23:02 |
kfox1111 | sbezverk: hehe. | 23:03 |
portdirect | the internet sbezverk, its a magical place poplulated by cats and angry people shouting at each other. | 23:04 |
sbezverk | kfox1111: about mariadb, I thought sdake was going to work on the chart for it, but you are saying we might/should use the one that comes with helm? | 23:04 |
sbezverk | portdirect: :-) I know what internet is, it runs partialy on our routers ;-) | 23:05 |
kfox1111 | sbezverk: sdake, yeah. | 23:06 |
*** signed8bit_Zzz is now known as signed8bit | 23:06 | |
kfox1111 | I was saying, we shouldn't block getting helm's one from being used. | 23:06 |
kfox1111 | and long term, maybe switch to it by default. | 23:06 |
sbezverk | kfox1111: got it, so we still need to develop internal.. | 23:07 |
kfox1111 | yeah. for now. | 23:07 |
* portdirect 'nods in agreement' | 23:08 | |
sbezverk | kfox1111: how difficult would be to make the gate to run on a nonzuul infra? | 23:10 |
kfox1111 | sbezverk: setup_gate.sh is pretty neutral. | 23:11 |
kfox1111 | I think the only thing may be the iptables stuff. and that could be easily conditionaled around. | 23:11 |
*** msimonin has quit IRC | 23:12 | |
*** ayoung has joined #openstack-kolla | 23:12 | |
sbezverk | kfox1111: it just I have tons of compute power, I thought to use it to get results faster | 23:12 |
sbezverk | and when everything is working then push it to the official gate | 23:13 |
kfox1111 | sbezverk: yeah. would be cool to try. | 23:13 |
kfox1111 | I modularized it to try and run most of it in minikube too a while back, so it should mostly work. just havent tried it yet. | 23:14 |
*** dave-mccowan has quit IRC | 23:15 | |
sbezverk | so all I would need is to automatically provision VM with the gate bits | 23:15 |
*** vhosakot has joined #openstack-kolla | 23:19 | |
*** sayantan_ has quit IRC | 23:19 | |
*** sayantan_ has joined #openstack-kolla | 23:20 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize neutron-server https://review.openstack.org/404929 | 23:21 |
*** signed8bit is now known as signed8bit_Zzz | 23:22 | |
*** absubram has quit IRC | 23:22 | |
kfox1111 | should only need to spawn a vm with a kolla-kubernetes checkout. setup gate script should do the rest. | 23:22 |
kfox1111 | if yo uwhave an openstack setup, you could do it with a heat template with a cloud init script. | 23:23 |
kfox1111 | or jenkins or something. | 23:23 |
*** unicell has quit IRC | 23:23 | |
openstackgerrit | Justin Scott proposed openstack/kolla-kubernetes: WIP Helm-ize memcached https://review.openstack.org/405677 | 23:24 |
*** sayantan_ has quit IRC | 23:27 | |
jascott1 | kfox1111 i committed my 'use the common template' version but I will delete it and keep the other deployment | 23:27 |
kfox1111 | k. sounds good. | 23:29 |
*** dave-mccowan has joined #openstack-kolla | 23:31 | |
jascott1 | thanks! | 23:32 |
sbezverk | kfox1111: I would use virsh, it is much easier | 23:32 |
jascott1 | noob question here, forgot my gerrit-ninja training. how do I mark experimental? | 23:35 |
jascott1 | or do i need ot? | 23:36 |
jascott1 | ot/to | 23:36 |
sbezverk | jascott1: check experimental | 23:36 |
*** unicell has joined #openstack-kolla | 23:36 | |
*** dave-mccowan has quit IRC | 23:37 | |
*** unicell1 has joined #openstack-kolla | 23:38 | |
*** unicell has quit IRC | 23:38 | |
sbezverk | jascott1: if you do it, it will not do much unless you change the workflow to use helm memcached chart | 23:38 |
sbezverk | instead of using the old way | 23:39 |
jascott1 | any reference on its use (experimental)? im kinda fuzzy on it | 23:40 |
jascott1 | and dont know how to set it | 23:40 |
kfox1111 | it would be really nice to cache known good bits for gate usage. :/ | 23:40 |
jascott1 | docker! | 23:41 |
*** vhosakot has quit IRC | 23:41 | |
kfox1111 | I think the hub may be having issues. | 23:41 |
*** signed8bit_Zzz is now known as signed8bit | 23:41 | |
*** signed8bit has quit IRC | 23:42 | |
sbezverk | jascott1: if you check this PS https://review.openstack.org/#/c/403959/14 check the change in cepf_workflow.sh to get an idea | 23:43 |
kfox1111 | sbezverk: Error: image kolla/centos-binary-kubernetes-entrypoint not found | 23:43 |
kfox1111 | I didn't catch it in the review. did you enable entrypoint? | 23:44 |
sbezverk | kfox1111: it is not in the docker hub, that is why it is failing, I suspect as soon as we psuhed it should either work or fail for a different reason | 23:45 |
jascott1 | kfox1111 for my clarity is the idea for user to be able to do 'helm install memcached --set resourceName=memcached-nova and then repeat for other services? | 23:45 |
kfox1111 | jascott1: helm install memcached --name memcached-nova | 23:46 |
jascott1 | oh ok thanks | 23:46 |
kfox1111 | there's a magic variable I think in helm to pull that instance name in the template. | 23:46 |
kfox1111 | don't know what it is off hand though. | 23:46 |
sbezverk | so we are going with memcached per service? | 23:47 |
kfox1111 | sbezverk: I'm kind of surprised the gate didn't reject it outright. | 23:47 |
sbezverk | kfox1111: why would it reject? it only fails to pull the image for init container | 23:48 |
kfox1111 | ... very cnfused.. | 23:48 |
kfox1111 | http://logs.openstack.org/07/405107/4/experimental/gate-kolla-kubernetes-deploy-ubuntu-binary-ceph-nv/dede055/logs/pods/kolla-openvswitch-vswitchd-network-3k6cw.txt | 23:49 |
*** klindgren_ has joined #openstack-kolla | 23:49 | |
kfox1111 | oh. maybe that was yours? | 23:49 |
sbezverk | kfox1111: it is perfectly normal behavior, I saw it many time when the image does not exist at the registry | 23:50 |
kfox1111 | https://review.openstack.org/#/c/404929/ is ready | 23:50 |
sbezverk | yes it is mine :-) | 23:50 |
sbezverk | kfox1111: why you go with 3 replicas, I think I already asked it but do not remember if I saw your reply | 23:51 |
kfox1111 | sbezverk: oh. didn't see the question. | 23:51 |
kfox1111 | sbezverk: should be enough to make rolling upgrades to always work smoothly. and not enough I think that it will cause a problem on single node I think. | 23:52 |
*** klindgren has quit IRC | 23:52 | |
sbezverk | kfox1111: if you request 3 replicas, how kube reacts when it does not have enought compute nodes to run them, do they scheduled to run on the same node? | 23:55 |
sbezverk | another question, you use "helm serve&" but I thought we use tiller running on the cluster? could you clarify why? | 23:58 |
kfox1111 | sbezverk: it will just run 3 of them unless theres a reason not to. | 23:58 |
kfox1111 | for example, if you expose them with hostPort in the pod, there's no way to do that except 1 per host. | 23:59 |
kfox1111 | so to scale out the system, you can add a second node, then delete one of the pods. | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!