Docker Swarm Network Issues


We have a Docker swarm running on 9 nodes (RHEL 7.4 VMs), 3 masters and 6 workers. We are facing networking issues and trying to identify the root cause.

  1. Port collisions - Multiple services assigned same port during creation - Ports are assigned randomly > 30000
  2. LB randomly failing to route request to the host running the service/container - timeouts

Bounding services off the node appears to fix the problem in some cases.

Our docker VMs are part of DRS vmotion (vmware). I would like to know if there are any specific recommendations on using dedicated hosts vs VMs (already seen performance improvement related posts, looking for something related our pain points).

Version: 1.13.1
LB: Apache
Container Services : Java/Tomcat