Instead of asking systemd-nspawn to write its "pid" file, rkt can write
the parent pid "ppid". Then, update getPID() to get the pid file from
/proc/$ppid/task/$ppid/children. In this way, we don't need to patch
systemd-nspawn.
getPID() is able to read both the "pid" file and the "ppid" file, in
order to work with alternative stage1s which choose to implement one or
the other method.
coreos/rkt#910
Before this change, when I run `sudo rkt trust` it fails. I hope
we can also install rkt under usr/local/bin by default. So our
users can have a better experience when following our rkt basic
tutorial.
This issue was introduced by a recent script rewrote (#652)
Add a script to copy busybox into stage1 for debugging purposes. It
takes a container UUID and outputs the nsenter command needed to enter
stage1's namespaces.
When using overlay filesystem, a new mount namespace will be created for
stage1. This script enters this namespace to copy busybox.
Previously the vagrant configuration would grab the last release
of Rocket to build. Now Rocket will be built on whatever is in the
checked out git repository (/vagrant inside the VM)
This will allow testing of changes to Rocket using the "vagrant
provision" command.
It will also automatically use the latest rocket version.
Assumes ./build script handles cleanup.
Go and APPC versions still manually configured in provisioning scripts
This commit adds the necessary Vagrantfile and
associated provisioning scripts to stand-up a working
rocket installation using the Vagrant tooling.
see vagrantup.com for more details about Vagrant
Environment details:
* Ubuntu trusty 64bit (Official Image)
* golang version 1.4.1
* Rocket version 0.3.2
* APPC version 0.3.0
files directly copied from commit 03592ce of:
https://github.com/brookerj11211/vagrant-rocket
Alban Crequy