skip to Main Content

We have setup three docket containers

  1. PHP
  2. nginx
  3. FluentD

Docker is running on bridge network mode, we have find huge TCP retransmission (18%) from Docker.

Docker Version is: Docker version 18.09.0, build 4d60db4

Anyone facing similar issues or any work around ?

enter image description here

2

Answers


  1. Same workaround mentioned by “solvease”(Feb 5 at 8:58) > use ‘–network host’ or ‘network_mode: “host”‘ for docker-compose definition.

    Login or Signup to reply.
  2. We have had a similar issue when instance network MTU was 1450 due to VXLAN usage but docker by default creates bridged network with MTU 1500.

    Configuring /etc/docker/daemon.json to be like:

    {
      "mtu": 1400
    }
    

    And restarting docker fixed it for us.
    Usage of host network also works but it limits docket usage a lot.
    For docker-compose use network and define MTU as well, e.g.

    version: '2'
    services:
      redis:
        restart: always
        image: 'redis:alpine'
        command: redis-server
        networks:
          - backend
    
    networks:
        backend:
          driver: bridge
          driver_opts:
            com.docker.network.driver.mtu: 1400
    

    This behaviour describe here:
    https://mlohr.com/docker-mtu/ and here
    https://sylwit.medium.com/how-we-spent-a-full-day-figuring-out-a-mtu-issue-with-docker-4d81fdfe2caf

    Login or Signup to reply.
Please signup or login to give your own answer.
Back To Top
Search