PCI passthrough strategy in Docker or oVirt

We have to deploy a test system where a Docker container or a VM (oVirt 3.5) shares up to 4x 10GB network cards with other containers/VMs.

So far we are using just oVirt for this purpose but we would like to shift to a Dockerized system to save some resources on the machines.

  • docker compose, vagrant and insecure Repository
  • kubernetes installation and kube-dns: open /run/flannel/subnet.env: no such file or directory
  • Docker - Mount Directory From Container to Host
  • php-fpm process hight memory usage, and the host jammed
  • Howto modify an docker image that was created from an existing one
  • How to use the “Remote Systems” view in Eclipse to explore a Docker container file system?
  • Does anybody have some experience or suggestion?

  • docker-compose does not start postgres image
  • ssh pseudo-tty causes PTY allocation request failed on channel 0
  • Accessing a Docker MySQL server from the host machine, without needing the container's IP address
  • Warning: Module 'curl' already loading in Unknown on line 0 (Fix this from a custom.php.ini file)
  • How to set Docker-compose arguments in console?
  • celery pdb running inside docker container with changing listening port
  • One Solution collect form web for “PCI passthrough strategy in Docker or oVirt”

    Docker containers are really just processes; it can run them each in a separate network namespace (the default) or let them use the host’s network directly (--net=host).

    If running in a separate network namespace then they won’t have any access to the host’s network cards; in the default config (--net=bridge) they are NAT networked via a Linux bridge, so if that matches your requirements, you’re away.

    Link to Docker docs on networking

    Docker will be the best open platform for developers and sysadmins to build, ship, and run distributed applications.