1.3.6 is slightly different but the result should be the same - 
https://github.com/kubernetes/kubernetes/blob/v1.3.6/cluster/gce/gci/configure-helper.sh#L322

The logic in the code seems contrary to the kubenet docs - 'Kubenet plugin: 
implements basic cbr0 using the bridge and host-local CNI plugins 
<http://kubernetes.io/docs/admin/network-plugins/>' i.e. 'cbr0 exists when 
cni is being used' but perhaps I'm just misinterpreting something.


On Tuesday, September 20, 2016 at 3:06:33 AM UTC+9, Amey Deshpande wrote:

> From what I can tell, the use of kubenet and cbr0 is mutually exclusive: 
> https://github.com/kubernetes/kubernetes/blob/master/cluster/gce/gci/configure-helper.sh#L372
>
> When cbr0 is being used, Docker daemon is restarted with "--bridge=cbr0" 
> on commandline to use the correct bridge device.  However, when cni is 
> being used, I do not see any evidence of deleting docker0 in the code.  
> I'll let K8s/GKE networking experts comment on it.
>
> On Mon, Sep 19, 2016 at 9:52 AM, Wil Reichert <wil.re...@gmail.com 
> <javascript:>> wrote:
>
>> I've got a GKE cluster that was create with a command something like
>>
>> # gcloud container clusters create "cluster" --machine-type 
>>> "n1-standard-1" --image-type=GCI --num-nodes "3" --network "default"
>>
>>
>> The resulting vm image os-release file is
>>
>> BUILD_ID=8530.71.0
>>> NAME="Google Container-VM Image"
>>> GOOGLE_CRASH_ID=Lakitu
>>> VERSION_ID=53
>>> BUG_REPORT_URL=https://crbug.com/new
>>> PRETTY_NAME="Google Container-VM Image"
>>> VERSION=53
>>> GOOGLE_METRICS_PRODUCT_ID=26
>>> HOME_URL="https://cloud.google.com/compute/docs/containers/vm-image/";
>>> ID=gci
>>
>>
>> GKE is configured with kubenet & a cbr0 bridge. Docker bridge networking 
>> is setup to use the docker0 bridge. Since the docker0 bridge does not exist 
>> any attempt by docker to use this network fails like:
>>
>> # docker run --rm -i -t busybox sh
>>
>> docker: Error response from daemon: failed to create endpoint 
>>> pedantic_lalande on network bridge: adding interface vethdc5e518 to bridge 
>>> docker0 failed: could not find bridge docker0: route ip+net: no such 
>>> network interface.
>>
>>
>> Adding a --net=Host to the above run statement works fine. The primary 
>> problem is docker does not accept the --net parameter for builds. The 
>> relevant Docker issue <https://github.com/docker/docker/issues/10324> 
>> has been open for a year & a half, a PR 
>> <https://github.com/docker/docker/pull/20987> fix for 9 months.
>>
>> My primary use case is running Docker builds on Jenkins slave pods which 
>> is a complete no go with the above configuration.
>>
>> Wil
>>
>> -- 
>> You received this message because you are subscribed to the Google Groups 
>> "Kubernetes user discussion and Q&A" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to kubernetes-use...@googlegroups.com <javascript:>.
>> To post to this group, send email to kubernet...@googlegroups.com 
>> <javascript:>.
>> Visit this group at https://groups.google.com/group/kubernetes-users.
>> For more options, visit https://groups.google.com/d/optout.
>>
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"Kubernetes user discussion and Q&A" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to kubernetes-users+unsubscr...@googlegroups.com.
To post to this group, send email to kubernetes-users@googlegroups.com.
Visit this group at https://groups.google.com/group/kubernetes-users.
For more options, visit https://groups.google.com/d/optout.
  • [kubernetes-us... Wil Reichert
    • Re: [kube... 'Amey Deshpande' via Kubernetes user discussion and Q&A
      • Re: [... Wil Reichert
        • R... 'Amey Deshpande' via Kubernetes user discussion and Q&A

Reply via email to