Unable to communicate between containers on different hosts

[Craig Reeves]

Hi all,

I have spun-up a sample platform on vsphere infrastructure which looks like this:

docker-manager01 - has the latest version of the Rancher server running. This vm is CentOS 6.6.

docker-node01 - this is a RancherOS node, with version 0.3.1 (latest) version of RancherOS running.

• IP Address 172.17.42.1 
• CPU 2x2.67 GHz
• Memory 1.96 GiB
• Storage 18.6 GiB
• Kernel 3.19.3-rancher
• Docker 1.6.0, build 4749651

docker-node02 - same as node01.

• IP Address 172.17.42.1 
• CPU 4x2.67 GHz
• Memory 3.87 GiB
• Storage 7.4 GiB
• Kernel 3.19.3-rancher
• Docker 1.6.0, build 4749651

These docker nodes and the manager all belong to a single, common physical network. I have created a service with the following components:

2x "official" nginx container - both are on node01

2x loadbalancer - one loadbalancer is on node01, the other is on node02.

If i visit the port exposed to the common physical network that belongs to the loadbalancer on node01, I can see the nginx test page. If I visit the loadbalancer on node02, however, I get a standard haproxy 'no servers are available to handle the request' page. This looks to me like some kind of networking issue between the hosts, but I am not clued up enough with Docker to figure it out. Something that does strike me as odd, though, is that both hosts (node01 and node02) have the same IP address - 172.17.42.1

Would you be able to point me in the right direction - I am sure there is some config property I have not seen. 

My /var/lib/rancher/conf/cloud-config-rancher.yml file looks like this:

network:

  dns:

    domain: internal.example.com

    nameservers:

    - 10.200.4.1

  interfaces:

    eth*:

      dhcp: true

Many thanks

Craig

[Craig Reeves]

Hi,

I have just rebuilt my environment in the same way as below, but this time with the newest version of rancher - 0.24 (much slicker, by the way!). I am experiencing the same networking behaviour. I'm not using any special configuration at all for the nginx container (maybe that is where I am going wrong?) but the load balancer has the same behaviour.

Thanks

Craig

[Denise Schannon]

I was trying to reproduce and have been unable to.

Are you able to ping between the 2 containers on different hosts? Click on the “Execute shell” in the drop down of one of the containers. Ping one of the container IPs on the other host. 

--
You received this message because you are subscribed to the Google Groups "Rancher" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rancherio+...@googlegroups.com.
To post to this group, send email to ranc...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/rancherio/0760f9bd-2c1c-4b2c-9d0c-cf740a3e749c%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.