mesos-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Benjamin Hindman" <>
Subject Re: Review Request 24754: Introduced "states" for Docker containers to transition between.
Date Fri, 15 Aug 2014 23:43:28 GMT

This is an automatically generated e-mail. To reply, visit:

(Updated Aug. 15, 2014, 11:43 p.m.)

Review request for mesos, Jie Yu and Timothy Chen.

Repository: mesos-git


The DockerContainerier needs to be able to properly clean up Docker
containers, regardless of when they are destroyed. For example, if a
container gets destroyed while we are fetching, we need to not keep
running the fetch, nor should we try and start the Docker
container. For this reason, we've split out the states into:

In particular, we made 'PULLING' be it's own state so that we could
easily destroy and cleanup when a user initiated pulling a really big
image but we timeout due to the executor registration timeout. Since
we curently have no way to discard a Docker::run, we needed to
explicitely do the pull (which is the part that takes the longest) so
that we can assume we won't have to wait very long for Docker::run to

Diffs (updated)

  src/docker/docker.hpp 3270c910d8b8d87aaf838f8494ee90ed93322a81 
  src/docker/docker.cpp 8f04babb9cd06ff5b2a39033664326d7d44cd6c6 
  src/slave/containerizer/docker.cpp d5292b609a9348e36d1327c9719f347bba84efb2 
  src/slave/slave.cpp 1eaab04137835d348bac881854d8c81a5f5dd280 
  src/tests/docker_containerizer_tests.cpp e0fd62f83387635f503817ced7a592cc3ae6e775 



make check


Benjamin Hindman

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message