#google-containers

/

      • meniv joined the channel
      • jseaidoun has quit
      • bobra joined the channel
      • [kubot] joined the channel
      • NOTICE: [kubernetes] thockin closed pull request #13345: Don't take the proxy mutex in the traffic path (master...proxy-contention) http://git.io/vGswe
      • [kubot] has left the channel
      • agupta has quit
      • sdake_ joined the channel
      • odinho has quit
      • jalateras joined the channel
      • qma17 joined the channel
      • markturansky has quit
      • sdake joined the channel
      • sdake_ has quit
      • ssandke joined the channel
      • qma17 has quit
      • rbatiz has quit
      • [kubot] joined the channel
      • NOTICE: [kubernetes] yujuhong closed pull request #13397: kubelet: define the housekeeping period (master...housekeeping) http://git.io/vGERv
      • [kubot] has left the channel
      • craig_ has quit
      • rossk has quit
      • briangrant has quit
      • gtaylor
        I'm seeing a ton of error spam in kubelet since my docker engine isn't running (because of the -b=cbr0 which doesn't exist yet). Seems like a chicken before the egg type scenario, since I can't run etcd, meaning I can't run the API server from hyperkube.
      • Should I create but not configure cbr0, start docker, then etcd, then api server (thus causing cbr0 to be configured by apiserver+kubelet?)
      • Or do I ignore the errors, fire up etcd within kubelet, then fire up apiserver within kubelet and it'll all magically work?
      • KubeBot joined the channel
      • KubeBot has quit
      • nightling
        does the out-of-the-box k8s cluster on GCE runs on google's debian?
      • gtaylor
        nightling: yeah, Wheezy 7.8
      • nightling
        was hoping for coreos :)
      • odinho joined the channel
      • gtaylor
        container-vm-v20150806 is the name of the image I last played with on GKE.
      • nightling
        out of curiosity whats the current stable kernel version there?
      • gtaylor
        Linux gke-test-cluster-d600919f-node-47f0 3.16.0-0.bpo.4-amd64 #1 SMP Debian 3.16.7-ckt11-1+deb8u2~bpo70+1 (2015-07-22) x86_64 GNU/Linux
      • nightling
        hmm, so no overlay yet
      • gtaylor
        not really needed within GKE
      • mehola joined the channel
      • PacketCollision
        what does the "kubernetes" binary actually do? Because of the name, it's almost impossible to Google
      • michael_mbp has quit
      • doxon
        beeps_: Thanks. It was overlapping labels.
      • michael_mbp joined the channel
      • jseaidoun joined the channel
      • sdake_ joined the channel
      • JohanSJA joined the channel
      • pwhiteside joined the channel
      • [kubot] joined the channel
      • NOTICE: [kubernetes] nikhiljindal opened pull request #13419: Adding support for deployments to client library (master...deploymentController) http://git.io/vGz9Y
      • [kubot] has left the channel
      • resno joined the channel
      • sdake__ joined the channel
      • mehola has quit
      • sdake has quit
      • nightling
        gtaylor, i'm signing up for that trial soon-ish to test it out myself
      • i don't think there is anything we use that wouldn't be compatible with 3.16
      • sdake_ has quit
      • sdake__ is now known as sdake
      • b_mo has quit
      • pwhiteside has left the channel
      • Hazelesque joined the channel
      • Hazelesque joined the channel
      • alivigni joined the channel
      • [kubot] joined the channel
      • NOTICE: [kubernetes] JanetKuo opened pull request #13420: Change deprecated 'kubectl stop' to 'delete' (master...doc-rolling) http://git.io/vGzFU
      • [kubot] has left the channel
      • alivigni has quit
      • Perceptes
        am I understanding correctly that on first run, kube-apiserver will create a TLS cert and key for kubelets, and these files must be distributed to all the nodes for their kubelets to authenticate with the apiserver?
      • vilobhmm1 has quit
      • roque has quit
      • ebala has quit
      • ebala joined the channel
      • erkules_ joined the channel
      • meniv has quit
      • jseaidoun has quit
      • jseaidoun joined the channel
      • [kubot] joined the channel
      • NOTICE: [kubernetes] jdef opened pull request #13421: WIP/MESOS: Add --contain-pod-resources flag to determine how k8s-mesos will contain pod resources (master...optout_pod_resource_constraints) http://git.io/vGzxE
      • [kubot] has left the channel
      • briangrant joined the channel
      • beeradb joined the channel
      • bobra joined the channel
      • haomaiwang joined the channel
      • beeradb has quit
      • matt993_ has quit
      • markturansky joined the channel
      • meniv joined the channel
      • bmeng joined the channel
      • Zach joined the channel
      • Zach
        zachary12345
      • longguang has quit
      • Hello
      • Zach is now known as Guest82097
      • Guest82097
        Would anybody be able to help me debug my issue with the HTTP load balancers? Specifically, that my NodePort instances are always returning unhealthy
      • jayunit100 joined the channel
      • markturansky has quit
      • Guest82097 has quit
      • jseaidoun has quit
      • mitton has quit
      • mitton joined the channel
      • wojtek-t has quit
      • mitton has quit
      • vilobhmm1 joined the channel
      • zhenghc joined the channel
      • matt993_ joined the channel
      • vilobhmm1 has left the channel
      • beeps_
        how did you setup the loadbalancer?
      • b_mo joined the channel
      • markturansky joined the channel
      • bobra1 joined the channel
      • mitton joined the channel
      • mitton has quit
      • markturansky has quit
      • ghodss has quit
      • jalateras joined the channel
      • jseaidoun joined the channel
      • ghodss joined the channel
      • `hidexvi joined the channel
      • [kubot] joined the channel
      • NOTICE: [kubernetes] hurf closed pull request #13379: Use unique service name for expose when --name is not specified (master...expose_name) http://git.io/vG45v
      • [kubot] has left the channel
      • mitton joined the channel
      • ceposta has quit
      • mitton has quit
      • joshm1 has quit
      • wanghua joined the channel
      • humble_ joined the channel
      • `hidexvi has quit
      • wanghua has quit
      • JohanSJA joined the channel
      • markturansky joined the channel
      • Casan
        how are people solving the dynamic pod address issue for https-http forwarding with nginx as load balancer for a node.js application pod gorup? is it solved, and if so what are the known patterns? this is really halting my adoption.
      • achanda joined the channel
      • using something like confd, etcd, the kubernetes api with go.. for maintaining the nginx.conf serverblocks automatically seems to be what I've discovered as the plausible solutions people are working on, but wouldn't the best be if the nginx-pods had its conf files on a shared persistent disks, and could send the request to a service which already monitors the load and can keep distributing the request?
      • jwoods joined the channel
      • I'm referring to something like this http://anandmanisankar.com/posts/docker-contain... which although is for docker, describes the problem scenario.
      • Esya has quit
      • markturansky has quit
      • humble__ joined the channel
      • [kubot] joined the channel
      • NOTICE: [kubernetes] feihujiang opened pull request #13422: Support setting env vars in kubectl run (master...supportSettingEnvInKubectlRun) http://git.io/vGgOT
      • [kubot] has left the channel
      • zhenghc has quit
      • jayunit100 joined the channel
      • humble_ has quit
      • gongysh joined the channel
      • rbatiz joined the channel
      • [kubot] joined the channel
      • NOTICE: [kubernetes] brendandburns closed pull request #13413: Release 1.0.5 (release-1.0...release-1.0) http://git.io/vGzEz
      • [kubot] has left the channel
      • pmorie has quit