Linux Plumbers Conf 2010

One Line Summary

Abstract

In the course of live migration of an OpenVZ Linux container, we have encountered a chicken-and-egg class problem.

To restore an NFS mount one should restore the container’s network first. To restore network one should restore the container’s processes first. To restore processes one should restore all the opened files first. In case opened files are on NFS … Houston, we have a problem!

After researching possible solutions, we decided to break this vicious circle at the NFS mount restoration stage. In order to do that, we use a special purpose file system called delayfs. It helps to postpone bringing up NFS mount and files to a later time when the networking is fully functional.

Tags

kernel, NFS, network, openvz, checkpointing, live migration

Speakers

• Pavel Emelyanov

  OpenVZ
  • Website: http://openvz.org/

  Biography

  Pavel Emelyanov is a core OpenVZ and PVC kernel developer working in the project for the last six years. His current role is the kernel team leader. Pavel is also the leading force behind a few mainstream kernel features, such as the network and the PID namespaces. He holds a PhD in Mathematics from the Moscow Institute of Physics and Technology.

• Kir Kolyshkin

  OpenVZ
  • Website: http://openvz.org/
  • Blog: http://blog.openvz.org/

  Biography

  Kirill Kolyshkin was named leader and project manager for the OpenVZ project in 2005 to further the adoption of containers virtualization for Linux. He spearheads the overall development and manages all key architecture, updates and feature upgrades for OpenVZ. Kolyshkin has more than 10 years of Linux experience and has long been an active open source advocate. He is a frequent speaker about virtualization technology and his 15-years career experience includes positions in information technology at Deutsche Bank and telecommunications company, Severtelecom. He holds a degree in Computer Science from the Ukhta State Technical University.